**Meet the editors**

Joan Bech is Associate Professor at the Department of Astronomy and Meteorology, University of Barcelona, Spain where he obtained his Physics degree and PhD. He participated in different research projects dealing with atmospheric physics, environmental quality, and meteorological remote sensing before joining the Meteorological Service of Catalonia where he worked 14

years in charge of remote sensing systems and applications, including a network of 4 C-band Doppler radars and a total lightning detection system. He was national delegate in several EU COST Projects related to weather radar applications and has participated in a number of projects related to anomalous propagation assessment, and in severe weather case studies including hailfalls, microbursts and tornadoes, as well as heavy rainfall events.

Jorge Chau is a Senior Researcher at the Radio Observatorio de Jicamarca, Peru where he has served as Director since 2001. His research has focused on radar studies of the equatorial neutral atmosphere and ionosphere. He is interested in the development of radar techniques to improve the atmospheric/ionospheric measurements as well as to improve the understanding of the atmos-

pheric/ionospheric physics at low latitudes and/or trospheric/stratospheric altitudes. Topics of ongoing research include the development of radio imaging techniques for 2D and 3D applications, incoherent scatter radar measurements at D and E region altitudes under the equatorial electrojet region, and meteor-head echo studies to determine the sources of the sporadic meteor population as well as other parameters of aeronomical and astronomical importance.

Contents

**Preface IX** 

**Part 1 Doppler Radar and** 

Dusan S. Zrnic

Chapter 1 **Doppler Radar for** 

Chapter 4 **Nowcasting 97** 

**Weather Surveillance 1** 

Chapter 2 **Automated Processing of Doppler** 

Paul Joe, Sandy Dance,

Paul James, Peter Lang,

P.W. Li, Hon-Yin Yeung,

Chapter 3 **Aviation Applications of Doppler** 

Clive Pierce, Alan Seed,

Ernesto Caetano,

**Weather Radar 159** 

Chapter 6 **Measuring Snow with** 

Elena Saltikoff

Chapter 5 **Use of Radar Precipitation Estimates** 

P.W. Chan and Pengfei Zhang

**USA Weather Surveillance 3** 

**Radar Data for Severe Weather Warnings 33** 

**Radars in the Alerting of Windshear and Turbulence 75** 

Valliappa Lakshmanan, Dirk Heizenreder,

Osamu Suzuki, Keiji Doi and Jianhua Dai

**Part 2 Precipitation Estimation and Nowcasting 95** 

Sue Ballard, David Simonin and Zhihong Li

Baldemar Méndez-Antonio and Víctor Magaña

**in Urban Areas: A Case Study of Mexico City 143** 

Thomas Hengstebeck, Yerong Feng,

### Contents

	- **Part 2 Precipitation Estimation and Nowcasting 95**

X Contents


Contents VII

**Part 6 Other Advanced Doppler Radar Applications 407** 

**Doppler Radars: A Review and Examples** 

Chapter 18 **Doppler Radar Tracking Using Moments 447** 

**from a Transportable Pulse Radar in L-Band 409** 

Mohammad Hossein Gholizadeh and Hamidreza Amindavar

Chapter 17 **Volcanological Applications of** 

Franck Donnadieu

#### **Part 6 Other Advanced Doppler Radar Applications 407**

Chapter 17 **Volcanological Applications of Doppler Radars: A Review and Examples from a Transportable Pulse Radar in L-Band 409**  Franck Donnadieu

VI Contents

Chapter 7 **A Network of Portable,** 

**Part 3 Tropospheric Wind and** 

Olivier Bousquet

Shingo Shimizu

Chris G. Collier

**Part 4 Weather Radar Quality** 

Chapter 12 **Quality Control Algorithms** 

Chapter 13 **Effects of Anomalous Propagation** 

**Part 5 Advanced Techniques for** 

D.L. Hysell and J.L. Chau

Erhan Kudeki and Marco Milla

Chapter 15 **Aperture Synthesis Radar** 

Chapter 16 **Incoherent Scatter Radar –**

Lars Norin and Günther Haase

**Probing the Ionosphere 355** 

Masayuki K. Yamamoto

Chapter 11 **Synergy Between Doppler Radar** 

Chapter 8 **Retrieving High Resolution 3-D** 

**Low-Cost, X-Band Radars 175**

**Turbulence Observations 203**

Marco Allegretti and Giovanni Perona

Marco Gabella, Riccardo Notarpietro, Silvano Bertoldo,

**Wind Vector Fields from Operational Radar Networks 205** 

**Three-Dimensional Wind Field Within Thunderstorms 231**

**and Lidar for Atmospheric Boundary Layer Research 271** 

Andrea Prato, Claudio Lucianaz, Oscar Rorato,

Chapter 9 **Multiple Doppler Radar Analysis for Retrieving the** 

Chapter 10 **New Observations by Wind Profiling Radars 247** 

**Control and Related Applications 287** 

Chapter 14 **Doppler Weather Radars and Wind Turbines 333** 

**Applied on Weather Radar Reflectivity Data 289** Jan Szturc, Katarzyna Ośródka and Anna Jurczyk

**Conditions on Weather Radar Observations 307**  Joan Bech, Adolfo Magaldi, Bernat Codina and Jeroni Lorente

**Imaging for Upper Atmospheric Research 357**

**Spectral Signal Model and Ionospheric Applications 377** 

Chapter 18 **Doppler Radar Tracking Using Moments 447**  Mohammad Hossein Gholizadeh and Hamidreza Amindavar

Preface

Over the last decades Doppler radar systems have been instrumental to improve our understanding and monitoring capabilities of phenomena and processes taking place in the low, middle, and upper atmosphere. Weather radars, wind profilers, and incoherent and coherent scatter radars implementing Doppler techniques are now used routinely both in research and operational applications by scientists and practitioners. This book brings together a collection of essays by international leading authors devoted to different applications of ground based Doppler radars. The target audiences are graduate students looking for an introduction to the field or professionals intending to refresh or update their knowledge. The book is organized in

The first section deals with the use of Doppler radar in weather surveillance and is made up by three chapters. The first one gives a brief introduction to Doppler radar fundamentals and an overview of weather radar surveillance in the USA. The second one offers an updated description of operational processing systems used in severe weather monitoring, and the third chapter is focused on aviation applications of Doppler radars.

The second section, devoted to precipitation estimation and very short range forecasting, or *nowcasting*, has four different chapters. In the first one a description of historical and current development of *nowcasting* techniques is given in detail, while the second chapter describes a specific implementation of rainfall estimates in Mexico. The third chapter includes a description of snowfall estimates and related applications in cold climates, and a fourth chapter details an innovative network of portable radars

Tropospheric wind and turbulence observations are the topic of the third section. In the first three chapters of this section readers will find several new methodologies developed in Japan and France to retrieve low level wind fields, along with an innovative approach to retrieve wind estimates with wind profilers. The section contains as well a chapter devoted to the complementary use of weather radar and

The fourth section covers three chapters related to quality control of weather radar data and related topics. The first chapter offers a comprehensive methodology applied

eighteen chapters grouped into six different sections.

designed to improve precipitation estimates.

lidar data to probe the atmospheric boundary layer.

### Preface

Over the last decades Doppler radar systems have been instrumental to improve our understanding and monitoring capabilities of phenomena and processes taking place in the low, middle, and upper atmosphere. Weather radars, wind profilers, and incoherent and coherent scatter radars implementing Doppler techniques are now used routinely both in research and operational applications by scientists and practitioners. This book brings together a collection of essays by international leading authors devoted to different applications of ground based Doppler radars. The target audiences are graduate students looking for an introduction to the field or professionals intending to refresh or update their knowledge. The book is organized in eighteen chapters grouped into six different sections.

The first section deals with the use of Doppler radar in weather surveillance and is made up by three chapters. The first one gives a brief introduction to Doppler radar fundamentals and an overview of weather radar surveillance in the USA. The second one offers an updated description of operational processing systems used in severe weather monitoring, and the third chapter is focused on aviation applications of Doppler radars.

The second section, devoted to precipitation estimation and very short range forecasting, or *nowcasting*, has four different chapters. In the first one a description of historical and current development of *nowcasting* techniques is given in detail, while the second chapter describes a specific implementation of rainfall estimates in Mexico. The third chapter includes a description of snowfall estimates and related applications in cold climates, and a fourth chapter details an innovative network of portable radars designed to improve precipitation estimates.

Tropospheric wind and turbulence observations are the topic of the third section. In the first three chapters of this section readers will find several new methodologies developed in Japan and France to retrieve low level wind fields, along with an innovative approach to retrieve wind estimates with wind profilers. The section contains as well a chapter devoted to the complementary use of weather radar and lidar data to probe the atmospheric boundary layer.

The fourth section covers three chapters related to quality control of weather radar data and related topics. The first chapter offers a comprehensive methodology applied

#### XIV Preface

in Poland to control radar data used for quantitative applications, and the second one describes the effects of anomalous propagation conditions on radar observations. Finally a study developed in Sweden examines the effects of wind turbines on weather radars.

The study of the ionosphere by Doppler radars with different techniques is covered in the fifth section, made up of two chapters. The first one covers the incoherent scatter technique, including recent improvements needed when the technique is applied to beams pointing perpendicular to the magnetic field. The second chapter presents advances in imaging aperture synthesis techniques applied to coherent echoes from ionospheric irregularities. Such techniques have been borrowed, improved, and adapted from the radio astronomy community.

The sixth section includes two chapters presenting other advanced techniques of Doppler radars. The first one depicts the use of Doppler radar for volcanological applications, and the second one describes a tracking technique based on the use of moments.

We finally would like to show our gratitude to all the contributing authors of this book for their eagerness and enthusiastic cooperation during the preparation and review of the chapters. We are particularly indebted to Dr Dusan Zrnic (National Severe Storms Laboratory, USA) and Professor David Hysell (Cornell University, USA) for their valuable suggestions and help with some chapter reviews. Last but not least we wish to thank our Intech editorial manager, Ms Marina Jozipovic, for her permanent assistance and professionalism. They all made this book possible and we sincerely hope that the readers will benefit from it.

> **Joan Bech**  Department of Astronomy and Meteorology, University of Barcelona, Spain

**Jorge Chau**  The Jicamarca Radio Observatory, Institute of Geophysics of Peru, Peru

## **Part 1**

### **Doppler Radar and Weather Surveillance**

**1** 

*USA* 

Dusan S. Zrnic

*NOAA, National Severe Storms Laboratory* 

**Doppler Radar for USA Weather Surveillance** 

Weather radar had its beginnings at the end of Word War II when it was noticed that storms clutter radar displays meant to reveal enemy aircraft. Thus radar meteorology was born. Until the sixties only the return power from weather tracers was measured which offered the first glimpses into precipitation structure hidden inside clouds. Possibilities opened up to recognize hail storms, regions of tornadoes (i.e., hook echoes), the melting zone in stratiform precipitation, and even determine precipitation rates at the ground, albeit with

Technology innovations and discoveries made in government laboratories and universities were quickly adopted by the National Weather Service (NWS). Thus in 1957 the Miami Hurricane Forecast Center commissioned the first modern weather radar (WSR-57) the type subsequently installed across the continental United States. The radar operated in the 10 cm band of wavelengths and had beamwidth of about 2o. In 1974 more radars were added: the WSR-74S operating in the band of 10 cm wavelengths and WSR-

Development of Doppler radars followed, providing impressive experience to remotely observe internal motions in convective storms and infer precipitation amounts. Thus scientists quickly discovered tell tale signatures of kinematic phenomena (rotation, storm

After demonstrable successes with this technology the NWS commissioned a network of Doppler radars (WSR-88D=Weather Surveillance Radars, year 1988, Doppler), the last of which was installed in 1997. Much had happened since that time and the current status

The nineties saw an accelerated development of information technology so much so that, upon installation of the last radar, computing and signal processing capabilities available to the public were about an order of magnitude superior to the ones on the radar. And scientific advancements were still coming in strong implying great improvements for operations if an upgrade in processing power were to be made. This is precisely what the NWS did by continuing infusion of the new technology into the system. Two significant upgrades have been made. The first involved replacement of the computer with distributed workstations (on the Ethernet in about 2002) for executing algorithms for precipitation estimation, tornado detection, storm tracking, and other. The second upgrade (in 2005)

pertinent to Doppler measurements and future trends are discussed herein.

**1. Introduction** 

considerable uncertainty.

74C in the 5 cm band.

outflows, divergence) in the fields of radial velocities.

### **Doppler Radar for USA Weather Surveillance**

Dusan S. Zrnic *NOAA, National Severe Storms Laboratory USA* 

#### **1. Introduction**

Weather radar had its beginnings at the end of Word War II when it was noticed that storms clutter radar displays meant to reveal enemy aircraft. Thus radar meteorology was born. Until the sixties only the return power from weather tracers was measured which offered the first glimpses into precipitation structure hidden inside clouds. Possibilities opened up to recognize hail storms, regions of tornadoes (i.e., hook echoes), the melting zone in stratiform precipitation, and even determine precipitation rates at the ground, albeit with considerable uncertainty.

Technology innovations and discoveries made in government laboratories and universities were quickly adopted by the National Weather Service (NWS). Thus in 1957 the Miami Hurricane Forecast Center commissioned the first modern weather radar (WSR-57) the type subsequently installed across the continental United States. The radar operated in the 10 cm band of wavelengths and had beamwidth of about 2o. In 1974 more radars were added: the WSR-74S operating in the band of 10 cm wavelengths and WSR-74C in the 5 cm band.

Development of Doppler radars followed, providing impressive experience to remotely observe internal motions in convective storms and infer precipitation amounts. Thus scientists quickly discovered tell tale signatures of kinematic phenomena (rotation, storm outflows, divergence) in the fields of radial velocities.

After demonstrable successes with this technology the NWS commissioned a network of Doppler radars (WSR-88D=Weather Surveillance Radars, year 1988, Doppler), the last of which was installed in 1997. Much had happened since that time and the current status pertinent to Doppler measurements and future trends are discussed herein.

The nineties saw an accelerated development of information technology so much so that, upon installation of the last radar, computing and signal processing capabilities available to the public were about an order of magnitude superior to the ones on the radar. And scientific advancements were still coming in strong implying great improvements for operations if an upgrade in processing power were to be made. This is precisely what the NWS did by continuing infusion of the new technology into the system. Two significant upgrades have been made. The first involved replacement of the computer with distributed workstations (on the Ethernet in about 2002) for executing algorithms for precipitation estimation, tornado detection, storm tracking, and other. The second upgrade (in 2005)

Doppler Radar for USA Weather Surveillance 5

Table 1 lists the radar parameters with which the surveillance mission is supported. Discussions of the reasons behind choices in volume coverage and other radar attributes of

the WSR-88D network, with principal emphasis on Doppler measurements, follows.

SNR > 0 dB, for Z= - 8 dBZ at r=50 km (exceeded by ~5 dB)

1 dB; SNR>10 dB; v = 4 m s-1 1 m s-1; SNR> 8 dB; v = 4 m s-1 1 m s-1; SNR>10 dB; v = 4 m s-1

Surveillance range is limited to about 460 km because storms beyond this range are usually below the horizon. Without beam blockage, the horizon's altitude at 460 km is 12.5 km; thus only the tops of strong convective storms are intercepted. Quantitative measurements of precipitation are required for storms at ranges less than 230 km. Nevertheless, in the region beyond 230 km, storm cells can be identified and their tracks established. Even at the range of about 230 km, the lowest altitude that the radar can observe under normal propagation conditions is about 3 km. Extrapolation of rainfall measurements from this height to the ground is subject to large errors, especially if the beam is above the melting layer and is

Surveillance time is determined by the time of growth of hazardous phenomena as well as the need for timely warnings. Five minutes for a repeat time is sufficient for detecting and confirming features with lifetime of about 15 min or more. Typical mesocyclone life time is 90 minutes (Burgess et al., 1982). Ordinary storms last tens of minutes but microbursts from these storms can produce dangerous shear in but a few minutes. Similarly tornadoes can rapidly develop from mesocyclones. For such fast evolving hazards a revisit time of less than a minute is desirable but not achievable if the whole three dimensional volume has to be covered. The principal driver to decrease the surveillance time is prompt detection of the tornadoes so that timely warning of their presence can be issued. Presently, the lead time for tornado warnings (i.e., the time that a warning is issued to the time the tornado does

Δ*r*  1 km; 0 < r 230 km; Δ*r*  2 km; r 460 km

460 km < 5 min hemispherical

Δ*r* = 250 m

**2.1 Considerations and requirements for storm surveillance** 

Requirement Values

Angular resolution 1o

Table 1. Requirements for weather radar observations.

detecting scatter from snow or melting ice particles.

damage) is about 12 minutes (see Section 5).

Surveillance: Range Time

For reflectivity For velocity

Estimate accuracy: Reflectivity Velocity

Spectrum width

**2.1.1 Range** 

**2.1.2 Time** 

Volumetric coverage

Range sampling interval:

brought in fully programmable signal processor and replaced the analogue receiver with the digital receiver. In 2009 the NWS started the process of converting the radars to dual polarization which should be accomplished by mid 2013.

The number of radars used continuously for operations is 159 and there are two additional radars for other use. One is for supporting changes in the network brought by infusion of new science or caused by deficiencies in existing components (designated KCRI in Norman, OK). The evolution involves both hardware and software and the update in the former are typically made annually. The other (designated KOUN in Norman, OK, USA) is for research and development. Therefore its configuration is more flexible allowing experimental changes in both hardware and software.

Conference articles and presentation about the WSR-88D and its data abound and there are few descriptions of its basic hardware. Very recent improvements are summarized by Saxion & Ice (2011) and a look into the future is presented in Ice & Saxion (2011). Yet only few journal articles describing the system have been published. The one by Heiss et al. (1990) presents hardware details from the manufacturer's point of view. The paper by Crum et al. (1993) describes data and archiving and the one by Crum & Alberty (1993) contain valuable information about algorithms. The whole No. 2 issue of Weather and Forecasting (1998), Vol. 13 is devoted to applications of the WSR-88D with a good part discussing products that use Doppler information. A look at the network with the view into the future is summarized by Serafin & Wilson (2000).

As twenty years since deployment of the last WSR-88D is approaching there are concerns about future upgrades and replacements. High on the list is the Multifunction Phased Array Radar (MPAR). At its core is a phased array antenna wherein beam position and shape are electronically controlled allowing rapid and adaptable scans. Thus, observations of weather (Zrnic et al., 2007) and tracking/detecting aircraft for traffic management and security purposes is proposed (Weber et al., 2007). Another futuristic concept is exemplified in proposed networks for Cooperative Adaptive Sensing of the Atmosphere (CASA) consisting of low power 3 cm wavelength phased array radars (McLaughlin et al., 2009).

Very few books on weather radar have been written and most include Doppler measurements. Here I list some published within the last 20 years. The one by Doviak & Zrnic (2006) primarily concentrates on Doppler aspects and contains information about the WSR-88D. The book by Bringi & Chandrasekar (2001) emphasizes polarization diversity and has sections relevant to Doppler. Role of Doppler radar in aviation weather surveillance is emphasized in the book by Mahapatra (1999). The compendium of chapters written by specialists and edited by Meishner (2004) concentrates on precipitation measurements but has chapters on Doppler principles as well as application to severe weather detection. Radar for meteorologists (Rinehart, 2010) is equally suited for engineers, technicians, and students who will enjoy its easy writing style and informative content.

#### **2. Basic radar**

The surveillance range, time, and volumetric coverage are routed in practical considerations of basic radar capabilities and the size and lifetimes of meteorological phenomena the radar is supposed to observe. This is considered next.

Doppler Radar Observations – 4 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

brought in fully programmable signal processor and replaced the analogue receiver with the digital receiver. In 2009 the NWS started the process of converting the radars to dual

The number of radars used continuously for operations is 159 and there are two additional radars for other use. One is for supporting changes in the network brought by infusion of new science or caused by deficiencies in existing components (designated KCRI in Norman, OK). The evolution involves both hardware and software and the update in the former are typically made annually. The other (designated KOUN in Norman, OK, USA) is for research and development. Therefore its configuration is more flexible allowing experimental

Conference articles and presentation about the WSR-88D and its data abound and there are few descriptions of its basic hardware. Very recent improvements are summarized by Saxion & Ice (2011) and a look into the future is presented in Ice & Saxion (2011). Yet only few journal articles describing the system have been published. The one by Heiss et al. (1990) presents hardware details from the manufacturer's point of view. The paper by Crum et al. (1993) describes data and archiving and the one by Crum & Alberty (1993) contain valuable information about algorithms. The whole No. 2 issue of Weather and Forecasting (1998), Vol. 13 is devoted to applications of the WSR-88D with a good part discussing products that use Doppler information. A look at the network with the view into the future

As twenty years since deployment of the last WSR-88D is approaching there are concerns about future upgrades and replacements. High on the list is the Multifunction Phased Array Radar (MPAR). At its core is a phased array antenna wherein beam position and shape are electronically controlled allowing rapid and adaptable scans. Thus, observations of weather (Zrnic et al., 2007) and tracking/detecting aircraft for traffic management and security purposes is proposed (Weber et al., 2007). Another futuristic concept is exemplified in proposed networks for Cooperative Adaptive Sensing of the Atmosphere (CASA) consisting

Very few books on weather radar have been written and most include Doppler measurements. Here I list some published within the last 20 years. The one by Doviak & Zrnic (2006) primarily concentrates on Doppler aspects and contains information about the WSR-88D. The book by Bringi & Chandrasekar (2001) emphasizes polarization diversity and has sections relevant to Doppler. Role of Doppler radar in aviation weather surveillance is emphasized in the book by Mahapatra (1999). The compendium of chapters written by specialists and edited by Meishner (2004) concentrates on precipitation measurements but has chapters on Doppler principles as well as application to severe weather detection. Radar for meteorologists (Rinehart, 2010) is equally suited for engineers, technicians, and students

The surveillance range, time, and volumetric coverage are routed in practical considerations of basic radar capabilities and the size and lifetimes of meteorological phenomena the radar

of low power 3 cm wavelength phased array radars (McLaughlin et al., 2009).

who will enjoy its easy writing style and informative content.

is supposed to observe. This is considered next.

**2. Basic radar** 

polarization which should be accomplished by mid 2013.

changes in both hardware and software.

is summarized by Serafin & Wilson (2000).

#### **2.1 Considerations and requirements for storm surveillance**

Table 1 lists the radar parameters with which the surveillance mission is supported. Discussions of the reasons behind choices in volume coverage and other radar attributes of the WSR-88D network, with principal emphasis on Doppler measurements, follows.


Table 1. Requirements for weather radar observations.

#### **2.1.1 Range**

Surveillance range is limited to about 460 km because storms beyond this range are usually below the horizon. Without beam blockage, the horizon's altitude at 460 km is 12.5 km; thus only the tops of strong convective storms are intercepted. Quantitative measurements of precipitation are required for storms at ranges less than 230 km. Nevertheless, in the region beyond 230 km, storm cells can be identified and their tracks established. Even at the range of about 230 km, the lowest altitude that the radar can observe under normal propagation conditions is about 3 km. Extrapolation of rainfall measurements from this height to the ground is subject to large errors, especially if the beam is above the melting layer and is detecting scatter from snow or melting ice particles.

#### **2.1.2 Time**

Surveillance time is determined by the time of growth of hazardous phenomena as well as the need for timely warnings. Five minutes for a repeat time is sufficient for detecting and confirming features with lifetime of about 15 min or more. Typical mesocyclone life time is 90 minutes (Burgess et al., 1982). Ordinary storms last tens of minutes but microbursts from these storms can produce dangerous shear in but a few minutes. Similarly tornadoes can rapidly develop from mesocyclones. For such fast evolving hazards a revisit time of less than a minute is desirable but not achievable if the whole three dimensional volume has to be covered. The principal driver to decrease the surveillance time is prompt detection of the tornadoes so that timely warning of their presence can be issued. Presently, the lead time for tornado warnings (i.e., the time that a warning is issued to the time the tornado does damage) is about 12 minutes (see Section 5).

Doppler Radar for USA Weather Surveillance 7

this increases the error of the Doppler velocity estimates by the square root of two, the improved angular resolution can increase the range, by about 50% (Brown et al., 2002 and 2005), to which mesocyclones and violent tornadoes can be detected. Therefore in the recently introduced scanning patterns, the data (i.e., spectral moments) are provided at 0.5o

The essence of the hardware (Fig. 1) is what radar operators see on the console. To the left of the data link (R,V,W,D) is the radar data acquisition (RDA) part consisting of the transmitter, antenna, microwave circuits, receiver, and signal processor. These components are located at radar site and data is transmitted to the local forecast office (LFO) where Radar Product Generation (RPG, Fig. 1) takes place. Operators at the LFO control (the block Control in Fig. 1) the radar and observe/analyze displays of data fields. At a glance of a console they can see the operating status of the radar and data flow. In the RPG the data is transformed into meteorologically meaningful information (Products in Fig. 1) by

increments in azimuth (Section 3.5).

algorithms executed on Ethernet cluster of workstation.

Fig. 1. Block diagram of the WSR-88D seen on the console of operators.

**2.2 Radar operation** 

#### **2.1.3 Volumetric coverage**

The volume scan patterns currently available on the WSR-88D have maximum elevations up to 20° and many are accomplished in about 5 minutes. Meteorologists have expressed a desire to extend the coverage to higher elevations to reduce the cone of silence. It is fair to state that the 30o elevation might be a practical upper limit for the WSR-88D. Top elevations higher than 20o have not been justified by strong meteorological reasons.

#### **2.1.4 Signal to noise ratio**

The SNR listed in Table 1 provides the specified accuracy of velocity and spectrum width measurements to the range of 230 km for both rain and snowfall rates of about 0.3 mm of liquid water depth per hour. That is, at a range of 230 km the SNR is larger than 10 dB thus the accuracy of Doppler measurements to shorter ranges is independent of noise and solely a function of number of samples and Doppler spectrum width.

#### **2.1.5 Spatial resolution**

The angular resolution is principally determined by the need to resolve meteorological phenomena such as tornados and mesocyclones to ranges of about 230 km, and the practical limitations imposed by antenna size at wavelength of 0.1 m. Even though beamwidth of 1o provides relatively high resolution, the spatial resolution at 230 km is 4 km. Because the beam of the WSR-88D is scanning azimuthally, the effective angular resolution in the azimuthal direction is somewhat larger (Doviak & Zrnic, 2006, Section 7.8); typically, about 40% at the 3 RPM scan rates of the WSR-88D. This exceeds many mesocyclone diameters, and thus these important weather phenomena, precursors of many tornadoes, can be missed. Tornadoes have even smaller diameters and therefore can not be resolved at the 230 km range.

The range resolution is indirectly influenced by the angular resolution; there is marginal gain in having range resolution finer than the angular one. For example better range resolution can provide additional shear segments and therefore improve detection of vortices at larger distance. The range resolution for reflectivity is coarser for two reasons: (1) reflectivity is principally used to measure rainfall rates over watersheds which are much larger than mesocyclones and (2) reflectivity samples at a resolution of 250 m are averaged in range (Doviak & Zrnic, 2006, Section 6.3.2) to achieve the required accuracy of 1 dB.

#### **2.1.6 Precision of measurements**

The specified 1 dB precision of reflectivity measurements (Table 1) provides about a 15% relative error of stratiform rain rate (Doviak & Zrnic, 2006, eq 8.22a). This has been accepted by the meteorological community. The specified precisions of velocity and spectrum width estimates are those derived from observations of mesocyclones with research radars. The 8 dB SNR is roughly that level beyond which the precision of velocity and spectrum width estimates do not improve significantly (Doviak & Zrnic, 2006, Sections 6.4, 6.5). But, it is possible that lower precisions can be tolerated and benefits can be derived therefrom. For example, it has been proposed (Wood et al., 2001) that velocity estimates be made with less samples (e.g., by a factor of two) in order to improve the azimuthal resolution. Although this increases the error of the Doppler velocity estimates by the square root of two, the improved angular resolution can increase the range, by about 50% (Brown et al., 2002 and 2005), to which mesocyclones and violent tornadoes can be detected. Therefore in the recently introduced scanning patterns, the data (i.e., spectral moments) are provided at 0.5o increments in azimuth (Section 3.5).

#### **2.2 Radar operation**

Doppler Radar Observations – 6 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

The volume scan patterns currently available on the WSR-88D have maximum elevations up to 20° and many are accomplished in about 5 minutes. Meteorologists have expressed a desire to extend the coverage to higher elevations to reduce the cone of silence. It is fair to state that the 30o elevation might be a practical upper limit for the WSR-88D. Top elevations

The SNR listed in Table 1 provides the specified accuracy of velocity and spectrum width measurements to the range of 230 km for both rain and snowfall rates of about 0.3 mm of liquid water depth per hour. That is, at a range of 230 km the SNR is larger than 10 dB thus the accuracy of Doppler measurements to shorter ranges is independent of noise and solely

The angular resolution is principally determined by the need to resolve meteorological phenomena such as tornados and mesocyclones to ranges of about 230 km, and the practical limitations imposed by antenna size at wavelength of 0.1 m. Even though beamwidth of 1o provides relatively high resolution, the spatial resolution at 230 km is 4 km. Because the beam of the WSR-88D is scanning azimuthally, the effective angular resolution in the azimuthal direction is somewhat larger (Doviak & Zrnic, 2006, Section 7.8); typically, about 40% at the 3 RPM scan rates of the WSR-88D. This exceeds many mesocyclone diameters, and thus these important weather phenomena, precursors of many tornadoes, can be missed. Tornadoes have even smaller diameters and therefore can not be resolved at the 230

The range resolution is indirectly influenced by the angular resolution; there is marginal gain in having range resolution finer than the angular one. For example better range resolution can provide additional shear segments and therefore improve detection of vortices at larger distance. The range resolution for reflectivity is coarser for two reasons: (1) reflectivity is principally used to measure rainfall rates over watersheds which are much larger than mesocyclones and (2) reflectivity samples at a resolution of 250 m are averaged in range (Doviak & Zrnic, 2006, Section 6.3.2) to achieve the required accuracy of 1 dB.

The specified 1 dB precision of reflectivity measurements (Table 1) provides about a 15% relative error of stratiform rain rate (Doviak & Zrnic, 2006, eq 8.22a). This has been accepted by the meteorological community. The specified precisions of velocity and spectrum width estimates are those derived from observations of mesocyclones with research radars. The 8 dB SNR is roughly that level beyond which the precision of velocity and spectrum width estimates do not improve significantly (Doviak & Zrnic, 2006, Sections 6.4, 6.5). But, it is possible that lower precisions can be tolerated and benefits can be derived therefrom. For example, it has been proposed (Wood et al., 2001) that velocity estimates be made with less samples (e.g., by a factor of two) in order to improve the azimuthal resolution. Although

higher than 20o have not been justified by strong meteorological reasons.

a function of number of samples and Doppler spectrum width.

**2.1.3 Volumetric coverage** 

**2.1.4 Signal to noise ratio** 

**2.1.5 Spatial resolution** 

**2.1.6 Precision of measurements** 

km range.

The essence of the hardware (Fig. 1) is what radar operators see on the console. To the left of the data link (R,V,W,D) is the radar data acquisition (RDA) part consisting of the transmitter, antenna, microwave circuits, receiver, and signal processor. These components are located at radar site and data is transmitted to the local forecast office (LFO) where Radar Product Generation (RPG, Fig. 1) takes place. Operators at the LFO control (the block Control in Fig. 1) the radar and observe/analyze displays of data fields. At a glance of a console they can see the operating status of the radar and data flow. In the RPG the data is transformed into meteorologically meaningful information (Products in Fig. 1) by algorithms executed on Ethernet cluster of workstation.

Fig. 1. Block diagram of the WSR-88D seen on the console of operators.

Doppler Radar for USA Weather Surveillance 9

pattern function (one way voltage), *η* is the free space impedance (120*π* Ω), *c* speed of light, *f* radar frequency, and *ψ<sup>t</sup>* arbitrary phase at the antenna. *U*(*t-r/c*) designates the pulse function

The effective beam cross section and pulse width define the intrinsic radar resolution volume but processing by the receiver increases it in range. Scatterers (hydrometeors such as rain, hail, snow and also insects, birds etc.,) within the resolution volume contribute to the backscattered electric field which upon reception by the antenna is transformed into a microwave signal. The signal is converted to an intermediate frequency *f*if then passed through anti-alias filter (nominal passband ~ 14 MHz), digitized (as per Table 2), and down

At intermediate frequency the signal coming from a continuum of scatterers can be

scatterers and *ω<sup>d</sup>* is the instantaneous Doppler shift caused by their motions toward (positive shift) and/or away (negative shift) from the radar. To determine the mean sense of motion (sign of Doppler shift) the intermediate frequency is removed and the signal is decomposed into its sinusoidal and cosinusoidal components, the inphase I and quadrature phase Q parts. These carry information about the number and sizes of scatterers as well as their motion. Samples of I and Q components are taken at consecutive delays with respect to the transmitted pulse. The delays are proportional to the range within the cloud from which the transmitted pulse is reflected. Samples from the same range locations (delays) are combined to obtain estimates of the spectral moments: reflectivity factor *Z*, mean Doppler velocity *v,* and spectrum width *σ*v (Doviak & Zrnic, 2006). The Doppler velocity *v* is related to the

 *f*d =2*v/λ*, (2)

Radars display (and store) equivalent reflectivity factor (often denoted with *Ze*) which is computed from the power and other parameters in the radar equation (Doviak & Zrnic 2006) assuming the scatterers have refractive index of liquid water. For small (compared to wavelength) spherical scatterers, *Z*e expressed as function of the distribution of sizes *N*(*D*),

max

*D*

0

Top left part in Fig 2 illustrates the continuum of returns (either I or Q), after each transmitted pulse from 1,…to *M*. Thus *M* samples at a fixed range delay (double vertical line) are operated on in various ways to produce estimates. There are as many estimates

6

*Z N D D dD <sup>e</sup>* (3)

() .

*<sup>d</sup>* where the amplitude *A*(*t*) fluctuates due to contribution by

  is the antenna

where *Pa* is the power radiated by the antenna, *r* is the distance, *f* (,)

such that it is 1 if its argument is between 0 and *τ* (the pulse width).

converted to audio frequencies (base band) for further processing.

frequency shift *f*d and wavelength λ via the Doppler equation

**2.2.2 Processing path from signals to algorithms** 

**2.2.1 Radar signal and Doppler shift** 

represented as if ( )cos( ) *At t t*

and so is the spectrum width.

equals

 

The radar is fully coherent pulsed Doppler and pertinent parameters are listed in Table 2 (see also Doviak & Zrnic, 2006 page 47). Each radar is assigned a fixed frequency in the band (Table 2), hence some values like the beamwidth and unambiguous velocities (not listed) depend on the exact frequency.


Table 2. Radar characteristics.

The data coming out of the RDA consist of housekeeping (time, pointing direction of the antenna, status, operating mode, and fields of reflectivity factor, mean radial velocity, and spread of velocities (designated as R, V, W in the console, Fig. 1), collectively called spectral moments. A wideband communication link is used to exchange base data and radar status/control between RDA and RPG. Depending on distance this link is by direct wire (up to 120 m), microwave line-of-site (to 38 km), or telephone company T1 line (unlimited).

Pulse of high peak power and narrow width (Table 2) generated at the output of the power amplifier is guided to the antenna. It is radiated in form of electromagnetic (EM) field confined within the narrow (1o) antenna beam. The propagating EM field interacts with intervening scatterers (precipitation, biological, and other). Part of the field is reflected forming a continuous stream at the antenna where it is intercepted and transformed for further processing by the receiver. Concise mathematical expression for the magnitude of the electric field at a distance *r* from the radar is

$$E = \left[\frac{P\_a \eta}{\pi}\right]^{1/2} \frac{f(\theta, \phi)}{2r} \cos\left[2\pi f \left(t - \frac{r}{c}\right) + \varphi\_t\right] \mathcal{U}(t - r/c),\tag{1}$$

Doppler Radar Observations – 8 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

The radar is fully coherent pulsed Doppler and pertinent parameters are listed in Table 2 (see also Doviak & Zrnic, 2006 page 47). Each radar is assigned a fixed frequency in the band (Table 2), hence some values like the beamwidth and unambiguous velocities (not listed)

1°

2.7 to 3 GHz

44.5 to 45.5 dB

1.57 μs and 4.57 μs

6 MHz (3 dB bandwidth)

322, 446, 644, 857, 1014, 1095, 1181, 1282 466, 336, 233, 175, 148, 137, 127, 117

94 dB at 1.57 μs pulse and 99 dB at 4.57 μs


Output samples spaced at 250 m/500m

750 kW

57.6 MHz 14 bits 71.9 MHz

1o or 0.5o

The data coming out of the RDA consist of housekeeping (time, pointing direction of the antenna, status, operating mode, and fields of reflectivity factor, mean radial velocity, and spread of velocities (designated as R, V, W in the console, Fig. 1), collectively called spectral moments. A wideband communication link is used to exchange base data and radar status/control between RDA and RPG. Depending on distance this link is by direct wire (up to 120 m), microwave line-of-site (to 38 km), or telephone company T1 line

Pulse of high peak power and narrow width (Table 2) generated at the output of the power amplifier is guided to the antenna. It is radiated in form of electromagnetic (EM) field confined within the narrow (1o) antenna beam. The propagating EM field interacts with intervening scatterers (precipitation, biological, and other). Part of the field is reflected forming a continuous stream at the antenna where it is intercepted and transformed for further processing by the receiver. Concise mathematical expression for the magnitude of

> 1/2 (,) cos 2 ( / ), 2 *<sup>a</sup> <sup>t</sup> <sup>P</sup> <sup>f</sup> <sup>r</sup> <sup>E</sup> f t Ut r c r c*

 

(1)

0.002

depend on the exact frequency.

PRFs (Hz, 5 sets of 8, variation ~3% )

IF Digital matched, short/long pulse

the electric field at a distance *r* from the radar is

 

Unambiguous range (km)

 Intermediate frequency (IF) A/D converter at IF Sampling rate Noise figure

Filter bandwidth or type: Front end analogue

Radial spacing in azimuth

Table 2. Radar characteristics.

(unlimited).

Frequency Beamwidth Antenna gain

Transmitter: Pulse power Pulse width Rf duty cycle

Receiver linear: Dynamic range where *Pa* is the power radiated by the antenna, *r* is the distance, *f* (,) is the antenna pattern function (one way voltage), *η* is the free space impedance (120*π* Ω), *c* speed of light, *f* radar frequency, and *ψ<sup>t</sup>* arbitrary phase at the antenna. *U*(*t-r/c*) designates the pulse function such that it is 1 if its argument is between 0 and *τ* (the pulse width).

#### **2.2.1 Radar signal and Doppler shift**

The effective beam cross section and pulse width define the intrinsic radar resolution volume but processing by the receiver increases it in range. Scatterers (hydrometeors such as rain, hail, snow and also insects, birds etc.,) within the resolution volume contribute to the backscattered electric field which upon reception by the antenna is transformed into a microwave signal. The signal is converted to an intermediate frequency *f*if then passed through anti-alias filter (nominal passband ~ 14 MHz), digitized (as per Table 2), and down converted to audio frequencies (base band) for further processing.

At intermediate frequency the signal coming from a continuum of scatterers can be represented as if ( )cos( ) *At t t <sup>d</sup>* where the amplitude *A*(*t*) fluctuates due to contribution by scatterers and *ω<sup>d</sup>* is the instantaneous Doppler shift caused by their motions toward (positive shift) and/or away (negative shift) from the radar. To determine the mean sense of motion (sign of Doppler shift) the intermediate frequency is removed and the signal is decomposed into its sinusoidal and cosinusoidal components, the inphase I and quadrature phase Q parts. These carry information about the number and sizes of scatterers as well as their motion. Samples of I and Q components are taken at consecutive delays with respect to the transmitted pulse. The delays are proportional to the range within the cloud from which the transmitted pulse is reflected. Samples from the same range locations (delays) are combined to obtain estimates of the spectral moments: reflectivity factor *Z*, mean Doppler velocity *v,* and spectrum width *σ*v (Doviak & Zrnic, 2006). The Doppler velocity *v* is related to the frequency shift *f*d and wavelength λ via the Doppler equation

$$f\_{\rm d} = 2\text{v}/\Omega\_{\rm \epsilon} \tag{2}$$

and so is the spectrum width.

Radars display (and store) equivalent reflectivity factor (often denoted with *Ze*) which is computed from the power and other parameters in the radar equation (Doviak & Zrnic 2006) assuming the scatterers have refractive index of liquid water. For small (compared to wavelength) spherical scatterers, *Z*e expressed as function of the distribution of sizes *N*(*D*), equals

$$Z\_e = \int\_0^{D\_{\text{max}}} N(D) D^6 dD. \tag{3}$$

#### **2.2.2 Processing path from signals to algorithms**

Top left part in Fig 2 illustrates the continuum of returns (either I or Q), after each transmitted pulse from 1,…to *M*. Thus *M* samples at a fixed range delay (double vertical line) are operated on in various ways to produce estimates. There are as many estimates

Doppler Radar for USA Weather Surveillance 11

estimated from the change in phase of the returned signal (Doviak & Zrnic 2006). Thus the WSR-88D is a phase sampling and measuring instrument. The change in phase of the return from one pulse to the next 2*πf*d*T*s is proportional to the Doppler velocity *v* as

If the phase change caused by precipitation is outside the – *π* to *π* interval it cannot be easily distinguished from the change within this interval. These limits define the unambiguous frequency *f*a = 1/(2*T*s) and through the Doppler relation (2) the

 *v*a=*λ/*(4*T*s). (4) Scatterers do cause a Doppler shift within the pulse as it is propagating and reflecting, but this shift is very small and can not be measured reliably as the following argument demonstrates. Consider the *τ =* 1.57 μs pulse width (WSR-88D) and scatterers moving at 10 m s-1 (36 km h-1). The corresponding Doppler frequency shift is 200 Hz (at 10 cm wavelength) and it produces a phase difference of 0.11o (2*πf*d*τ*) between the beginning and end of the pulse return. This tiny difference can not be measured with sufficient accuracy to

To mitigate the ambiguity problem the WSR-88D has some options one of which is special phase coding and processing. The result is seen in Fig. 2 where the pink ring at 137 km indicates the unambiguous range for velocity measurements (see discussion in section 3.2.3); it represents censored data because the ground clutter from nearby range and weather signals from the second trip range are comparable in power and can not be reliably

Operators of the WSR-88D have at their disposal preprogrammed volume coverage patterns (VCP – see example in Fig. 2). These are consecutive scans starting from elevation of 0.5o and incrementing until a top elevation is reached. Most algorithms require a full volume scan to generate a product. The one in Fig. 2 (bottom left) reconstructs a vertical profile of Doppler velocities along a radial; the radar is located to the right and green colors indicate velocities toward the radar in 5 m s-1 increments starting with 0 (gray color). Cylindrical protrusion below 5 km in the middle with some velocities toward the radar (red color) is indicative of a

The block diagram (Fig. 3) of the WSR-88D radar is typical for pulsed Doppler radars. Essential components are the Frequency and Timing generator, the transmitter and the receiver. Radar and antenna controls are omitted from the figure. Intermediate frequency (if) on the radars is 57.6 MHz, and the local oscillator (lo) frequency is adjustable to cover the range between 2.7 and 3 GHz (the operating band, see Table 2). The power amplifier is a klystron. The transmit/receive switch is comprised of a circulator and additional devices to protect the receiver from the transmitted high power pulse. The low noise amplifier (LNA) has a noise figure ~ 0.8 dB and the receiver bandwidth is 6 MHz up to the input of the digital receiver. The digital receiver is a proprietary product of SIGMET Co (now Vaisala)

indicated in (2).

unambiguous velocity as

yield useful estimate.

separated.

tornado.

**3. Signal processing and display** 

and its essence is described next.

along range time as there are samples. That is, sample spacing is typically equal to pulse duration and therefore consecutive samples are almost independent. Closer sampling (i.e., oversampling) has some advantages (Section 4.2).

Radials of spectral moments are transmitted to the RPG (a radial of velocities is in the top right part of Fig. 2). Spectral moments are displayed at Weather Forecast Offices, are recorded, and are also processed by algorithms to automatically identify hazardous weather features, estimate amounts of precipitation, and to be used in numerical models among other applications. Example displayed in Fig. 2 (right bottom) is the field of Doppler velocities obtained by the WSR-88D in Dove, North Carolina during the Hurricane Irene on Aug 28th, 2011 at 2:29 UTC. The end range on the display is 230 km which is also the range up to which quantitative measurements are currently being made. Extension to 300 km is planned.

Fig. 2. Information path from time series to output of algorithms.

The radar is sufficiently sensitive to detect precipitation at much larger ranges where the beamwidth and observations high above ground mar quantitative interpretation of impending weather on the ground. At the elevation of 0.5o, the radar makes two scans: one with the longest PRT (3.1 ms) for estimating reflectivities unambiguously up to 465 km in range, the other with one of the short PRTs to estimate unambiguously velocity over a sufficiently large span. The ambiguities in range and velocity are inherent to pulsed Doppler radars. Reflections from scatterers spaced by the unambiguous range (*r*<sup>a</sup> = *cT*s/2 where *T*s is pulse repetition time) appear at the same delay with respect to the reference time (determined by the last of two transmitted pulse). Obvious increase in range can be made by increasing *T*s. And this is fine for measurements of reflectivity but would harm measurements of velocity. At the 10 cm wavelength Doppler velocities are Doppler Radar Observations – 10 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

along range time as there are samples. That is, sample spacing is typically equal to pulse duration and therefore consecutive samples are almost independent. Closer sampling (i.e.,

Radials of spectral moments are transmitted to the RPG (a radial of velocities is in the top right part of Fig. 2). Spectral moments are displayed at Weather Forecast Offices, are recorded, and are also processed by algorithms to automatically identify hazardous weather features, estimate amounts of precipitation, and to be used in numerical models among other applications. Example displayed in Fig. 2 (right bottom) is the field of Doppler velocities obtained by the WSR-88D in Dove, North Carolina during the Hurricane Irene on Aug 28th, 2011 at 2:29 UTC. The end range on the display is 230 km which is also the range up to which

quantitative measurements are currently being made. Extension to 300 km is planned.

Fig. 2. Information path from time series to output of algorithms.

The radar is sufficiently sensitive to detect precipitation at much larger ranges where the beamwidth and observations high above ground mar quantitative interpretation of impending weather on the ground. At the elevation of 0.5o, the radar makes two scans: one with the longest PRT (3.1 ms) for estimating reflectivities unambiguously up to 465 km in range, the other with one of the short PRTs to estimate unambiguously velocity over a sufficiently large span. The ambiguities in range and velocity are inherent to pulsed Doppler radars. Reflections from scatterers spaced by the unambiguous range (*r*<sup>a</sup> = *cT*s/2 where *T*s is pulse repetition time) appear at the same delay with respect to the reference time (determined by the last of two transmitted pulse). Obvious increase in range can be made by increasing *T*s. And this is fine for measurements of reflectivity but would harm measurements of velocity. At the 10 cm wavelength Doppler velocities are

oversampling) has some advantages (Section 4.2).

estimated from the change in phase of the returned signal (Doviak & Zrnic 2006). Thus the WSR-88D is a phase sampling and measuring instrument. The change in phase of the return from one pulse to the next 2*πf*d*T*s is proportional to the Doppler velocity *v* as indicated in (2).

If the phase change caused by precipitation is outside the – *π* to *π* interval it cannot be easily distinguished from the change within this interval. These limits define the unambiguous frequency *f*a = 1/(2*T*s) and through the Doppler relation (2) the unambiguous velocity as

$$v\_a = \lambda/(4\,T\_s).\tag{4}$$

Scatterers do cause a Doppler shift within the pulse as it is propagating and reflecting, but this shift is very small and can not be measured reliably as the following argument demonstrates. Consider the *τ =* 1.57 μs pulse width (WSR-88D) and scatterers moving at 10 m s-1 (36 km h-1). The corresponding Doppler frequency shift is 200 Hz (at 10 cm wavelength) and it produces a phase difference of 0.11o (2*πf*d*τ*) between the beginning and end of the pulse return. This tiny difference can not be measured with sufficient accuracy to yield useful estimate.

To mitigate the ambiguity problem the WSR-88D has some options one of which is special phase coding and processing. The result is seen in Fig. 2 where the pink ring at 137 km indicates the unambiguous range for velocity measurements (see discussion in section 3.2.3); it represents censored data because the ground clutter from nearby range and weather signals from the second trip range are comparable in power and can not be reliably separated.

Operators of the WSR-88D have at their disposal preprogrammed volume coverage patterns (VCP – see example in Fig. 2). These are consecutive scans starting from elevation of 0.5o and incrementing until a top elevation is reached. Most algorithms require a full volume scan to generate a product. The one in Fig. 2 (bottom left) reconstructs a vertical profile of Doppler velocities along a radial; the radar is located to the right and green colors indicate velocities toward the radar in 5 m s-1 increments starting with 0 (gray color). Cylindrical protrusion below 5 km in the middle with some velocities toward the radar (red color) is indicative of a tornado.

#### **3. Signal processing and display**

The block diagram (Fig. 3) of the WSR-88D radar is typical for pulsed Doppler radars. Essential components are the Frequency and Timing generator, the transmitter and the receiver. Radar and antenna controls are omitted from the figure. Intermediate frequency (if) on the radars is 57.6 MHz, and the local oscillator (lo) frequency is adjustable to cover the range between 2.7 and 3 GHz (the operating band, see Table 2). The power amplifier is a klystron. The transmit/receive switch is comprised of a circulator and additional devices to protect the receiver from the transmitted high power pulse. The low noise amplifier (LNA) has a noise figure ~ 0.8 dB and the receiver bandwidth is 6 MHz up to the input of the digital receiver. The digital receiver is a proprietary product of SIGMET Co (now Vaisala) and its essence is described next.

Doppler Radar for USA Weather Surveillance 13

Fig. 4. Conceptual diagram of the digital receiver and down converter indicating the essential operations. The dashed vertical line shows where digital processing begins.

Fig. 5. Conceptual timing diagram of processes in the digital receiver. The return signal is assumed to be sinusoidal pulse such as would be produced by a single point scatterer.

At the lowest two (sometime three) elevations (< 1.6o) two consecutive scans at each elevation are made. For surveillance and reflectivity measurement the longest PRT is used so that the unambiguous range is ~460 km. It is followed by one or more of the higher PRTs for measurement of Doppler velocity and spectrum width whereby the unambiguous velocity interval is larger than ~ 20 m s-1. Thus Doppler estimates can be ambiguous and overlaid in range. To determine the location of the Doppler estimates, powers along the radial at the same azimuth but in the surveillance scans are examined. The echoes from ranges spaced by *n*PRT*c*/2 of the Doppler scan, where *n* is 1, 2, 3, 4, can be overlaid in the Doppler scan; the echo for *n*=1 is said to come from the first trip because it corresponds to the round trip shorter than the separation between consecutive pulses. Powers from

**3.2.1 Lowest elevation scans** 

Fig. 3. Block diagram of the receiver (without signal processing part) and the transmitter.

#### **3.1 Digital receiver**

The analogue signal if ( )cos( ) *At t t <sup>d</sup>* is sampled at a rate of 71.9 MHz producing a stream (time *ti*) of 14 bit numbers. These are multiplied (Fig. 4) with sin(*ω*if*ti*) and cos(*ω*if*ti*) and digitally filtered to obtain the base band I and Q components (at times *tk*). Although the nominal short pulse duration is 1.57 μs same as sample spacing in range, 155 samples spaced at ~ 13.8 ns over 2.15 μs interval are used for multiplication and filtering (in the long pulse mode the number of samples is 470 over a 6.53 μs interval). The digital low pass filter is adjusted to match the shape of the transmitted long or short pulse. Matching is achieved by passing the attenuated transmitted pulse ("burst") through the receiver and taking the discrete Fourier transform of the output. The inverse of this transform gives the coefficients of the matched impulse response filter. Amplitude and phase of the "burst" is sampled upon each transmission to monitor power, compensate for phase instabilities, and use in phase codes for mitigating range ambiguities. The timing diagram (Fig. 5) illustrates the relations between transmitted sequence, digital oscillator samples, the sampled sequence from a point scatterer and its I and Q values (after the matched filter).

#### **3.2 Transmitted sequences and volume scans**

Several volume coverage patterns are available. With the exception of one all utilize the short pulse. The exception has a uniform sequence of long pulses at the longest PRT for observations in clear air or snow where weak reflections are from insects, birds, ice and/or refractive index fluctuations. For storm observations the volume coverage patterns have three distinct modes depending on the elevation.

Doppler Radar Observations – 12 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Fig. 3. Block diagram of the receiver (without signal processing part) and the transmitter.

stream (time *ti*) of 14 bit numbers. These are multiplied (Fig. 4) with sin(*ω*if*ti*) and cos(*ω*if*ti*) and digitally filtered to obtain the base band I and Q components (at times *tk*). Although the nominal short pulse duration is 1.57 μs same as sample spacing in range, 155 samples spaced at ~ 13.8 ns over 2.15 μs interval are used for multiplication and filtering (in the long pulse mode the number of samples is 470 over a 6.53 μs interval). The digital low pass filter is adjusted to match the shape of the transmitted long or short pulse. Matching is achieved by passing the attenuated transmitted pulse ("burst") through the receiver and taking the discrete Fourier transform of the output. The inverse of this transform gives the coefficients of the matched impulse response filter. Amplitude and phase of the "burst" is sampled upon each transmission to monitor power, compensate for phase instabilities, and use in phase codes for mitigating range ambiguities. The timing diagram (Fig. 5) illustrates the relations between transmitted sequence, digital oscillator samples, the sampled sequence from a point scatterer and its I and Q values (after the matched

Several volume coverage patterns are available. With the exception of one all utilize the short pulse. The exception has a uniform sequence of long pulses at the longest PRT for observations in clear air or snow where weak reflections are from insects, birds, ice and/or refractive index fluctuations. For storm observations the volume coverage patterns have

*<sup>d</sup>* is sampled at a rate of 71.9 MHz producing a

**3.1 Digital receiver** 

filter).

The analogue signal if ( )cos( ) *At t t*

**3.2 Transmitted sequences and volume scans** 

three distinct modes depending on the elevation.

 

Fig. 4. Conceptual diagram of the digital receiver and down converter indicating the essential operations. The dashed vertical line shows where digital processing begins.

Fig. 5. Conceptual timing diagram of processes in the digital receiver. The return signal is assumed to be sinusoidal pulse such as would be produced by a single point scatterer.

#### **3.2.1 Lowest elevation scans**

At the lowest two (sometime three) elevations (< 1.6o) two consecutive scans at each elevation are made. For surveillance and reflectivity measurement the longest PRT is used so that the unambiguous range is ~460 km. It is followed by one or more of the higher PRTs for measurement of Doppler velocity and spectrum width whereby the unambiguous velocity interval is larger than ~ 20 m s-1. Thus Doppler estimates can be ambiguous and overlaid in range. To determine the location of the Doppler estimates, powers along the radial at the same azimuth but in the surveillance scans are examined. The echoes from ranges spaced by *n*PRT*c*/2 of the Doppler scan, where *n* is 1, 2, 3, 4, can be overlaid in the Doppler scan; the echo for *n*=1 is said to come from the first trip because it corresponds to the round trip shorter than the separation between consecutive pulses. Powers from

Doppler Radar for USA Weather Surveillance 15

desired phase and actual phase might not be exactly equal, the transmitted phase is sampled

Fig. 6. Transmitted pulse sequence (vertical color lines) and the corresponding received powers (wiggly curves). The phases of the transmitted pulses are indicated and indexed from -1 to 3. The location of overlay at one fixed range is indicated by two black vertical lines. The phases of the received signals from the first and second trip are indicated as well as the phase of the second trip signal after subtracting (correcting) the phase of the first trip.

In case of overlaid echoes the phase coding allows separation of the contributions by the first and second trip signals. This is accomplished by first cohering (correcting) the phases of the stronger echo, then filtering it out. For example if first trip is cohered the second trip signal spectrum (complex with magnitudes and phases) is split into eight replicas over the unambiguous interval. Then frequency domain filtering of the first (strong) trip signal with a notch centered on its spectrum and having a width of ¾ unambiguous interval leaves two spectral replicas of the second trip signal spectrum. From these replicas it is possible to reconstruct the second trip spectrum and compute spectral moments. It turns out that cohering for the first trip signal induces 4 spectral replicas in the third trip signal and again eight replicas

into the fourth trip signal; the fifth trip signal has two replicas and can not be recovered.

Determination of the ranges where overlaid echoes might be is made using powers from the surveillance scan (long PRT) which precedes the Doppler scan (phase coded short PRT). The overlay trip number and powers are needed to make proper cohering-recohering order and notch filter application. In case ground clutter is present Blackman window is applied to time series data and clutter is taken out with a special frequency domain filter (Sec 3.3). If there is no clutter contamination but overlaid echoes are present the von Hann window is chosen. An example of Doppler velocity fields obtained with the SZ(8/64) phase code is in Fig. 7 (left side). The same field obtained by processing and censoring with no phase coding is also plotted (right side); note the large pink area in the second trip region indicative of non recoverable velocities. Small pink areas in the first trip region (SE of radar) signify that

and used in processing to precisely cohere the signal from the desired trip.

locations spaced by *n*PRT*c*/2 are compared to determine the correct range of the Doppler estimates and presence of overlaid echoes. If one of the overlaid powers is larger than user specified threshold (typically 5 dB) the corresponding Doppler spectral moment is assigned to the correct range whereas the values at location of the other overlaid echoes are censored. If the powers are within 5 dB, the variables at all locations where the overlay is possible are censored. Because the Doppler spectral moments are computed and recorded only to the distance of at most twice the unambiguous range the censoring is also done to that distance.

There is a special VCP (Zittel et al., 2008) with three scans at same elevation on five consecutive lowest elevations whereby velocities from three PRFs (No. 4, 6, and 8 in table 2) are combined to increase the *va* and display it up to the distance of 175 km.

#### **3.2.2 Scans at mid and high elevations**

At elevations between 1.6o and 7o a "batch" sequence is transmitted. It is a dual PRF in which the first few (3 to 12) pulses are at the lowest PRF and the rest (between about 25 and 90) are transmitted at one of the four highest PRFs (shortest PRTs, Table 2). The lowest PRF pulses are for surveillance, reflectivity measurements, and censoring and assignment of range to Doppler spectral moments; just the same as in the lowest scans. To improve accuracy of the reflectivity estimates powers from the Doppler sequence (high PRF) are included in the averaging provided there is no contamination by overlaid echoes. Beyond 7o elevation uniform PRTs are transmitted because the tops of storms at locations where overlay can occur are below the radar beam.

#### **3.2.3 Phase coding**

To mitigate range overlay some volume scanning patterns at the lowest elevations (<2o) have transmitted sequences encoded with the SZ(8/64) phase code (Sachidananda & Zrnic, 1999). The concept is depicted in Fig. 6 and explained in the caption. The prescribed phases *Ψk* (i.e., switching phases) are applied to the transmitted pulses. Formally this is represented by multiplication of the sequence with the switching code *ak* = exp(j*Ψk*). The first trip return signal is made coherent by multiplying it with the conjugate \* *<sup>k</sup> a* = exp(-j*Ψk*). With this multiplication the 2nd trip signal is phase modulated by the code \* *k kk* <sup>1</sup> *c aa* . The 2nd trip can be made coherent by multiplying the incoming sequence with *ak-1*\*, in which case the 1st trip signal is modulated by the code *ck*\*. The code, *ak* is designed such that the modulation code *ck* has a phase shift given by <sup>2</sup> <sup>1</sup> 8 /64 *kk k k* . The special property of this code is that its autocorrelation is unity for lags in multiples of 8 (lags 8*n*; *n*=0,1,2,...), and is zero for all other lags. Therefore the power spectrum has only 8 non-zero coefficients separated by *M*/8 coefficients. The SZ(8/64) switching code is given by

$$a\_k = \exp[-j\sum\_{m=0}^k \left(\pi m^2 \;/\; 8\right)]; \quad k = 0, 1, 2...63. \tag{5}$$

It has periodicity of 32 hence the number of samples *M* must be an integer multiple of 32. From (5) it is obvious that the phase sequence consists of a binary sub multiple of 360o hence it is generated without round-off errors using standard binary phase shifters. Because the Doppler Radar Observations – 14 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

locations spaced by *n*PRT*c*/2 are compared to determine the correct range of the Doppler estimates and presence of overlaid echoes. If one of the overlaid powers is larger than user specified threshold (typically 5 dB) the corresponding Doppler spectral moment is assigned to the correct range whereas the values at location of the other overlaid echoes are censored. If the powers are within 5 dB, the variables at all locations where the overlay is possible are censored. Because the Doppler spectral moments are computed and recorded only to the distance of at most twice the unambiguous range the censoring is also done to that distance. There is a special VCP (Zittel et al., 2008) with three scans at same elevation on five consecutive lowest elevations whereby velocities from three PRFs (No. 4, 6, and 8 in table 2)

At elevations between 1.6o and 7o a "batch" sequence is transmitted. It is a dual PRF in which the first few (3 to 12) pulses are at the lowest PRF and the rest (between about 25 and 90) are transmitted at one of the four highest PRFs (shortest PRTs, Table 2). The lowest PRF pulses are for surveillance, reflectivity measurements, and censoring and assignment of range to Doppler spectral moments; just the same as in the lowest scans. To improve accuracy of the reflectivity estimates powers from the Doppler sequence (high PRF) are included in the averaging provided there is no contamination by overlaid echoes. Beyond 7o elevation uniform PRTs are transmitted because the tops of storms at locations where

To mitigate range overlay some volume scanning patterns at the lowest elevations (<2o) have transmitted sequences encoded with the SZ(8/64) phase code (Sachidananda & Zrnic, 1999). The concept is depicted in Fig. 6 and explained in the caption. The prescribed phases *Ψk* (i.e., switching phases) are applied to the transmitted pulses. Formally this is represented by multiplication of the sequence with the switching code *ak* = exp(j*Ψk*). The first trip return

be made coherent by multiplying the incoming sequence with *ak-1*\*, in which case the 1st trip signal is modulated by the code *ck*\*. The code, *ak* is designed such that the modulation code *ck*

its autocorrelation is unity for lags in multiples of 8 (lags 8*n*; *n*=0,1,2,...), and is zero for all other lags. Therefore the power spectrum has only 8 non-zero coefficients separated by *M*/8

2

It has periodicity of 32 hence the number of samples *M* must be an integer multiple of 32. From (5) it is obvious that the phase sequence consists of a binary sub multiple of 360o hence it is generated without round-off errors using standard binary phase shifters. Because the

exp[ ( /8)]; 0,1,2...63.

*k* . The special property of this code is that

(5)

  *<sup>k</sup> a* = exp(-j*Ψk*). With this

*k kk* <sup>1</sup> *c aa* . The 2nd trip can

signal is made coherent by multiplying it with the conjugate \*

multiplication the 2nd trip signal is phase modulated by the code \*

<sup>1</sup> 8 /64 *kk k*

0

*k*

*m a jm k* 

are combined to increase the *va* and display it up to the distance of 175 km.

**3.2.2 Scans at mid and high elevations** 

overlay can occur are below the radar beam.

has a phase shift given by <sup>2</sup>

coefficients. The SZ(8/64) switching code is given by

*k*

**3.2.3 Phase coding** 

desired phase and actual phase might not be exactly equal, the transmitted phase is sampled and used in processing to precisely cohere the signal from the desired trip.

Fig. 6. Transmitted pulse sequence (vertical color lines) and the corresponding received powers (wiggly curves). The phases of the transmitted pulses are indicated and indexed from -1 to 3. The location of overlay at one fixed range is indicated by two black vertical lines. The phases of the received signals from the first and second trip are indicated as well as the phase of the second trip signal after subtracting (correcting) the phase of the first trip.

In case of overlaid echoes the phase coding allows separation of the contributions by the first and second trip signals. This is accomplished by first cohering (correcting) the phases of the stronger echo, then filtering it out. For example if first trip is cohered the second trip signal spectrum (complex with magnitudes and phases) is split into eight replicas over the unambiguous interval. Then frequency domain filtering of the first (strong) trip signal with a notch centered on its spectrum and having a width of ¾ unambiguous interval leaves two spectral replicas of the second trip signal spectrum. From these replicas it is possible to reconstruct the second trip spectrum and compute spectral moments. It turns out that cohering for the first trip signal induces 4 spectral replicas in the third trip signal and again eight replicas into the fourth trip signal; the fifth trip signal has two replicas and can not be recovered.

Determination of the ranges where overlaid echoes might be is made using powers from the surveillance scan (long PRT) which precedes the Doppler scan (phase coded short PRT). The overlay trip number and powers are needed to make proper cohering-recohering order and notch filter application. In case ground clutter is present Blackman window is applied to time series data and clutter is taken out with a special frequency domain filter (Sec 3.3). If there is no clutter contamination but overlaid echoes are present the von Hann window is chosen. An example of Doppler velocity fields obtained with the SZ(8/64) phase code is in Fig. 7 (left side). The same field obtained by processing and censoring with no phase coding is also plotted (right side); note the large pink area in the second trip region indicative of non recoverable velocities. Small pink areas in the first trip region (SE of radar) signify that

Doppler Radar for USA Weather Surveillance 17

call Clutter Phase Alignment (CPA) defined as *CPA=|*sum*Vk*|/sum|*Vk|*, where *Vk* is the complex voltage (I + j Q) from a fixed clutter location at consecutive times (spaced by the PRT) indicated by time index *k* and the sum is over the total number of pulses in the dwell time. Local standard deviation (termed texture) of reflectivity factor *Zi* in range (*i* index indicates adjacent values in range) and changes in sign of the differences *Zi+1* - *Zi* are also used; the frequency of change in reflectivity gradient along range is obtained from this difference and it defines the spin variable. The *CPA,* texture, and spin are combined in a fuzzy classification scheme to identify locations where clutter filter should be applied.

Fig. 8. Doppler spectrum of simulated weather signal (red) and clutter (blue). Interpolated (filtered) Gaussian part and estimated noise level are shown. The *va*=32 m s-1. (Figure as in

The GMAP filter and censoring (Free & Patel, 2007) is applied to surveillance and Doppler scans. In the "batch" mode the number of samples is insufficient for spectral processing hence the average voltage (i.e., DC) from the samples spaced by the long PRT is removed. The system also employs strong point clutter (typically caused by aircraft) removal along radials. It is done on each spectral moment independently by comparing the sample power with two adjacent values either side of it. If the value is outside prescribed criteria it is

In computations of *Z* and *σ*v receiver noise powers are subtracted from the returned powers. Thus, the receiver noise power is estimated at the end of each volume scan at high elevation angle. The noise depends on the elevation angle because contributions from ground radiation and air constituents are larger if the beam is closer to the ground. To account for

The reflectivity factor is obtained by summing the pulse powers, subtracting the noise power, and using the radar equation (Doviak & Zrnic, 2006). At the lowest few elevations *Z* is computed from the long PRT (surveillance scan). At mid elevations ("batch mode") the

the increase the noise is extrapolated to lower elevations using empirical relations.

Torres et al., 2004c).

replaced by interpolation of neighboring values.

**3.4 Computation of spectral moments** 

overlaid powers of first and second trip signals are within 10 dB and hence velocities can not be confidently recovered. There is a narrow pink ring of censored data in the image where phase code is applied. The beginning range of the ring is at the start of the second trip (175 km) and is caused by automatic receiver shut down during transmission followed by the strong first trip ground clutter overwhelming the weaker second trip signal.

Fig. 7. Fields of Doppler velocities. Left: obtained from phased coded sequence. Right: obtained from non coded sequences. Elevation is 0.5o, the unambiguous velocity *va*= 23.7 m s-1 and range *ra* = 175 km. Data were collected on 10/08/2002 with the research WSR-88D (KOUN). The color bar indicates velocities in m s-1. (Figure from Torres et al., 2004c).

#### **3.3 Ground clutter filter**

The ground clutter filter implemented on the network is a frequency domain filter with interpolation over the removed clutter spectral coefficients. The filter called Gaussian Model Adaptive Processing (GMAP) has been developed by Siggia and Passarelli (2004). Its first premise is: clutter has a Gaussian shape power spectrum with width linearly related to the antenna rotation rate; hence the width can be computed. The second is the signal spectrum has also Gaussian shape and has width larger than clutter's. The Blackman window is applied followed by Fourier transform. Receiver noise is externally provided to the filter and used to establish the spectral noise level which helps determine how many spectral coefficients either side of zero to remove (Fig. 8, blue peak is from ground clutter). The removed coefficients are replaced (iteratively) with a Gaussian curve obtained from Doppler moments and the spectrum of the weather signal (dotted curve) is restored. Then the inverse discrete Fourier transform is performed to obtain the autocorrelation at lag 1. The argument of the autocorrelation is linearly related to the mean Doppler velocity (see section 3.4).

Several options exist to decide where to filter clutter. One relies on the clutter map to locate azimuths and ranges. It is also possible but undesirable to apply clutter filter everywhere. The operators can select regions between azimuths and ranges where to turn the filter on. Recently an adaptive algorithm called Clutter Mitigation Decision has been implemented (Hubbert et al., 2009). It uses coherency of the clutter signal exemplified in what the authors Doppler Radar Observations – 16 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

overlaid powers of first and second trip signals are within 10 dB and hence velocities can not be confidently recovered. There is a narrow pink ring of censored data in the image where phase code is applied. The beginning range of the ring is at the start of the second trip (175 km) and is caused by automatic receiver shut down during transmission followed by the

strong first trip ground clutter overwhelming the weaker second trip signal.

Fig. 7. Fields of Doppler velocities. Left: obtained from phased coded sequence. Right: obtained from non coded sequences. Elevation is 0.5o, the unambiguous velocity *va*= 23.7 m s-1 and range *ra* = 175 km. Data were collected on 10/08/2002 with the research WSR-88D (KOUN). The color bar indicates velocities in m s-1. (Figure from Torres et al., 2004c).

The ground clutter filter implemented on the network is a frequency domain filter with interpolation over the removed clutter spectral coefficients. The filter called Gaussian Model Adaptive Processing (GMAP) has been developed by Siggia and Passarelli (2004). Its first premise is: clutter has a Gaussian shape power spectrum with width linearly related to the antenna rotation rate; hence the width can be computed. The second is the signal spectrum has also Gaussian shape and has width larger than clutter's. The Blackman window is applied followed by Fourier transform. Receiver noise is externally provided to the filter and used to establish the spectral noise level which helps determine how many spectral coefficients either side of zero to remove (Fig. 8, blue peak is from ground clutter). The removed coefficients are replaced (iteratively) with a Gaussian curve obtained from Doppler moments and the spectrum of the weather signal (dotted curve) is restored. Then the inverse discrete Fourier transform is performed to obtain the autocorrelation at lag 1. The argument of the autocorrelation is linearly related to the mean Doppler velocity (see section 3.4).

Several options exist to decide where to filter clutter. One relies on the clutter map to locate azimuths and ranges. It is also possible but undesirable to apply clutter filter everywhere. The operators can select regions between azimuths and ranges where to turn the filter on. Recently an adaptive algorithm called Clutter Mitigation Decision has been implemented (Hubbert et al., 2009). It uses coherency of the clutter signal exemplified in what the authors

**3.3 Ground clutter filter** 

call Clutter Phase Alignment (CPA) defined as *CPA=|*sum*Vk*|/sum|*Vk|*, where *Vk* is the complex voltage (I + j Q) from a fixed clutter location at consecutive times (spaced by the PRT) indicated by time index *k* and the sum is over the total number of pulses in the dwell time. Local standard deviation (termed texture) of reflectivity factor *Zi* in range (*i* index indicates adjacent values in range) and changes in sign of the differences *Zi+1* - *Zi* are also used; the frequency of change in reflectivity gradient along range is obtained from this difference and it defines the spin variable. The *CPA,* texture, and spin are combined in a fuzzy classification scheme to identify locations where clutter filter should be applied.

Fig. 8. Doppler spectrum of simulated weather signal (red) and clutter (blue). Interpolated (filtered) Gaussian part and estimated noise level are shown. The *va*=32 m s-1. (Figure as in Torres et al., 2004c).

The GMAP filter and censoring (Free & Patel, 2007) is applied to surveillance and Doppler scans. In the "batch" mode the number of samples is insufficient for spectral processing hence the average voltage (i.e., DC) from the samples spaced by the long PRT is removed.

The system also employs strong point clutter (typically caused by aircraft) removal along radials. It is done on each spectral moment independently by comparing the sample power with two adjacent values either side of it. If the value is outside prescribed criteria it is replaced by interpolation of neighboring values.

#### **3.4 Computation of spectral moments**

In computations of *Z* and *σ*v receiver noise powers are subtracted from the returned powers. Thus, the receiver noise power is estimated at the end of each volume scan at high elevation angle. The noise depends on the elevation angle because contributions from ground radiation and air constituents are larger if the beam is closer to the ground. To account for the increase the noise is extrapolated to lower elevations using empirical relations.

The reflectivity factor is obtained by summing the pulse powers, subtracting the noise power, and using the radar equation (Doviak & Zrnic, 2006). At the lowest few elevations *Z* is computed from the long PRT (surveillance scan). At mid elevations ("batch mode") the

Doppler Radar for USA Weather Surveillance 19

window and so are data from the adjacent azimuth centered 0.5o off from the previous. This produces more radials of data (spaced by 0.5o as opposed to 1o) increasing resolution to facilitate recognition of small phenomena such as tornado vortices (Brown et al., 2002, and 2005). The contrast between the routine and enhanced resolution of a tornado vortex signature is evident in the example in Fig. 9. The reflectivity field (top figures in dBZ as indicated by the color bars) displays a "hook echo" associated with low level circulation.

The velocity field (bottom in Fig. 9) displays three circular features ("balls") in its center: the lighter green and red adjacent to it in azimuth indicate cyclonic circulation (mesocyclone). Its diameter is about four km and it is estimated from the distance between maximum inbound (green) and outbound (red) velocities. The sharp discontinuity in the center (light green ~ -30 m s-1 to > 30 m s-1) is the tornado vortex signature (TVS). The transition between the red "ball" and the green one farther in range marks the zero radial velocity suggesting converging flow (i.e., red and green velocities pushing air toward each other) near ground. Bottom right: same as in the left but the resolution in azimuth is enhanced to 0.5o. The TVS is

Fig. 9. Top Left: Z, resolution 1 km x 1o. Right: resolution 250 m x 0.5o. Bottom Left: V field, resolution 250 m x 1o. Right: resolution 250 m x 0.5o. X, Y sizes are 25 by 20 km; radar is at x=

4 km and y = -25 km with respect to each image left corner. (Courtesy, S. Torres).

The crisp pattern (top right) is the result of the enhanced resolution.

better defined and so are other small scale features.

reflectivity is computed from both the long and short PRTs if no overlay is indicated; otherwise only samples from the surveillance scan (long PRT) are used.

Computation of Doppler variables starts with the discrete Fourier transform. In absence of clutter, time series data is equally weighted (uniform window) and the power spectrum estimate (at some range location) is

$$\hat{S}(k) = \left| \frac{1}{M} \sum\_{m=0}^{M-1} V(m) e^{-j\frac{2\pi mk}{M}} \right|^2, \quad k = 0, 1, \ldots, M - 1 \tag{6}$$

The discrete inverse Fourier transform applied to (6) produces the value of circular autocorrelation function at lag 1 (i.e., *Ts*) which contains one erroneous term, namely the product of first and last member of the time series (Torres et al., 2007). This term is subtracted so that the autocorrelation at lag one (i.e., *Ts*) becomes

$$
\hat{R}(1) = \sum\_{m=0}^{M-1} \hat{S}(k) e^{\int \frac{2\pi k}{M}} - \frac{1}{M} V^\*(M-1)V(0),\tag{7}
$$

and the mean velocity estimate comes out to be (Doviak & Zrnic, 2006. eq 6.19)

$$
\hat{\psi} = -(\frac{\mathcal{X}}{4\pi T\_s}) \arg[\hat{R}(1)].\tag{8}
$$

The spectrum width for most VCPs is estimated by combining the lag one autocorrelation

and the signal power <sup>1</sup> <sup>2</sup> 0 ˆ ( ) *M s n m P Vm P* , from which the noise power *Pn* is subtracted, as follows (Doviak & Zrnic 2006, eq 6.27)

$$
\hat{\sigma}\_v = \frac{\lambda}{2\sqrt{2}\pi T\_s} \left| \ln \left( \frac{\hat{P}\_s}{\left| \hat{R}(1) \right|} \right) \right|^{1/2} \,\mathrm{.}\tag{9}
$$

But, if the logarithm term is negative ˆ *<sup>v</sup>* is set to zero. In case of phase coding and presence of overlaid echoes equation (9) is used for the weaker signal in the surveillance scan (long PRT). The spectrum width of the strong signal is computed for the Doppler scan using the ratio ˆ ˆ *R R* (1) / (2) as in Doviak & Zrnic (2006, eq. 6.32), because it is not biased by presence of the weak signal.

#### **3.5 Oversampling in azimuth (overlapping radials)**

Until recent upgrades all VCPs had spacing of radials at 1o azimuth and reflectivities were averaged and recorded at 1 km range intervals but velocities retained inherent spacing of 250 m (Table 1). Newly added VCPs employ a strategy whereby at the lowest two elevations time series data from overlapping (in azimuth) beams are processed to produce spectral moments. Thus data obtained over one degree azimuth are weighted with the von Hann Doppler Radar Observations – 18 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

reflectivity is computed from both the long and short PRTs if no overlay is indicated;

Computation of Doppler variables starts with the discrete Fourier transform. In absence of clutter, time series data is equally weighted (uniform window) and the power spectrum

<sup>1</sup> ˆ( ) ( ) , 0,1,..., 1

The discrete inverse Fourier transform applied to (6) produces the value of circular autocorrelation function at lag 1 (i.e., *Ts*) which contains one erroneous term, namely the product of first and last member of the time series (Torres et al., 2007). This term is

(6)

(7)

(8)

*<sup>v</sup>* is set to zero. In case of phase coding and presence

. (9)

, from which the noise power *Pn* is subtracted, as

1/2

*P*

 

<sup>2</sup> <sup>2</sup> <sup>1</sup>

*Sk Vme k M*

<sup>2</sup> <sup>1</sup> \*

*R Ske V M V*

<sup>1</sup> <sup>ˆ</sup> <sup>ˆ</sup> (1) ( ) ( 1) (0),

<sup>ˆ</sup> <sup>ˆ</sup> ( )arg[ (1)]. <sup>4</sup> *<sup>s</sup> v R T* 

The spectrum width for most VCPs is estimated by combining the lag one autocorrelation

<sup>ˆ</sup> <sup>ˆ</sup> ln 2 2 ˆ(1) *<sup>s</sup> <sup>v</sup> s*

*T R*

of overlaid echoes equation (9) is used for the weaker signal in the surveillance scan (long PRT). The spectrum width of the strong signal is computed for the Doppler scan using the

Until recent upgrades all VCPs had spacing of radials at 1o azimuth and reflectivities were averaged and recorded at 1 km range intervals but velocities retained inherent spacing of 250 m (Table 1). Newly added VCPs employ a strategy whereby at the lowest two elevations time series data from overlapping (in azimuth) beams are processed to produce spectral moments. Thus data obtained over one degree azimuth are weighted with the von Hann

*R R* (1) / (2) as in Doviak & Zrnic (2006, eq. 6.32), because it is not biased by presence of

*M*

*<sup>M</sup> mk <sup>j</sup> <sup>M</sup>*

otherwise only samples from the surveillance scan (long PRT) are used.

0

0

*m*

<sup>1</sup> <sup>2</sup>

*s n*

*P Vm P* 

0 ˆ ( ) *M*

*m*

**3.5 Oversampling in azimuth (overlapping radials)** 

follows (Doviak & Zrnic 2006, eq 6.27)

But, if the logarithm term is negative ˆ

*<sup>M</sup> <sup>k</sup> <sup>j</sup> <sup>M</sup>*

and the mean velocity estimate comes out to be (Doviak & Zrnic, 2006. eq 6.19)

*m*

*M*

subtracted so that the autocorrelation at lag one (i.e., *Ts*) becomes

estimate (at some range location) is

and the signal power

ratio ˆ ˆ

the weak signal.

window and so are data from the adjacent azimuth centered 0.5o off from the previous. This produces more radials of data (spaced by 0.5o as opposed to 1o) increasing resolution to facilitate recognition of small phenomena such as tornado vortices (Brown et al., 2002, and 2005). The contrast between the routine and enhanced resolution of a tornado vortex signature is evident in the example in Fig. 9. The reflectivity field (top figures in dBZ as indicated by the color bars) displays a "hook echo" associated with low level circulation. The crisp pattern (top right) is the result of the enhanced resolution.

The velocity field (bottom in Fig. 9) displays three circular features ("balls") in its center: the lighter green and red adjacent to it in azimuth indicate cyclonic circulation (mesocyclone). Its diameter is about four km and it is estimated from the distance between maximum inbound (green) and outbound (red) velocities. The sharp discontinuity in the center (light green ~ -30 m s-1 to > 30 m s-1) is the tornado vortex signature (TVS). The transition between the red "ball" and the green one farther in range marks the zero radial velocity suggesting converging flow (i.e., red and green velocities pushing air toward each other) near ground. Bottom right: same as in the left but the resolution in azimuth is enhanced to 0.5o. The TVS is better defined and so are other small scale features.

Fig. 9. Top Left: Z, resolution 1 km x 1o. Right: resolution 250 m x 0.5o. Bottom Left: V field, resolution 250 m x 1o. Right: resolution 250 m x 0.5o. X, Y sizes are 25 by 20 km; radar is at x= 4 km and y = -25 km with respect to each image left corner. (Courtesy, S. Torres).

Doppler Radar for USA Weather Surveillance 21

the research radar. Also, the large pink area of overlaid echoes has almost disappeared in the measurement made utilizing the staggered PRT. The small circle closest to the radar origin indicates overlaid echo contaminating the first trip velocities of the operational radar.

Fig. 11. Velocity fields of a storm system. Left: field obtained with the operational WSR-88D radar in Oklahoma City on April 06, 2003, elevation 2.5 deg, batch mode with unambiguous range of 148 km and velocity of 25.4 m s-1. Pink regions locate censored velocities which can not be reliably recovered due to overlaid first and second trip echoes. Right: same as on the left but obtained with the research WSR-88D (KOUN) utilizing staggered PRT. This radar is about 20 km south from the operational radar. The color bar indicates velocities (m s-1), red

Oversampling here indicates spacing of I, Q samples smaller than the pulse duration. Operations on few of these range consecutive "oversamples" can reduce error in estimates and/or data acquisition time (Torres & Zrnic, 2003). Simplest of operations is averaging in range of oversampled spectral moments. Somewhat more involved is the whitening transformation in which the signal vector **v =** [V(*m*,0), V(*m*,1),... V(*m*, *l*),....V(*m*, *L*)] consisting of *L* oversampled correlated complex voltages is transformed into a set of *L* orthogonal voltages (Torres & Zrnic, 2003). The time index *m* refers to the usual sample time and *l* to the oversampled range time. The transformation takes the form **x** = **H**-1 **v** with **H** related to the normalized correlation matrix **C** of **v** via **C**=**H\*HT**. The correlation matrix can be precomputed (or measured e.g., Ivic et al., 2003) because it depends solely on the envelope of the transmitted pulse and the baseband equivalent receiver filter shape for a uniform Z. The *L* transformed samples are independent and averaging of spectral moments obtained from each (in absence of noise) yields smaller error of estimates. Whitening is effective at large SNRs but fails otherwise. To achieve *L* independent samples the receiver filter bandwidth needs to be increased *L* times over the matched filter bandwidth and this enhances the noise by the same factor. In addition the whitening transformation also increases the noise hence

away from and green toward the radar. (Figure adapted from Torres et al., 2003).

**4.2 Oversampling techniques** 

#### **4. Near term enhancements**

Currently a significant transformation of the radars is ongoing; it is addition of dual polarization (Zrnic et al., 2008). By mid 2013 all radars on the network should have this capability. Although Doppler capability is not a prerequisite for dual polarization, the coherency of transmit-receive signals within one PRT is for differential phase measurement. Dual polarization offers ample possibilities for application of spectral analysis to polarimetric signals and these are being explored (e.g., to discriminate between insects and birds, Bachman & Zrnic, 2007; to suppress ground clutter, Unal, 2009; or to achieve adaptive clutter and noise suppression, Moisseev & Chandrasekar, 2009).

Three improvements approved for soon inclusion on the network are pending. These are staggered PRT, processing of range oversampled signals, and adaptive recognition and filtering of ground clutter. Brief description follows.

#### **4.1 Staggered PRT**

It is planned for mitigating range velocity ambiguities at mid elevation angles with possible use at the lower elevations. The scheme consists of alternating interval between transmitted pulses (Fig. 10) and estimating arguments of two autocorrelations at the two lags, arg[*R*(*T*1)] and arg[*R*(*T2*)]. The velocities estimated from these arguments have a different unambiguous interval (each inversely proportional to the corresponding separation *T*i, i=1 or 2) as can be deduced from eq. (8). Therefore the difference of the velocities uniquely tags the proper unambiguous interval for either PRT so that correct dealiasing can be achieved (Torres et al., 2004a) up to larger *v*a than possible with only one of these PRTs . For the example in Fig. 10, *v*a = 3*v*a2 = 2*v*a1. Consider *T*1=1 ms *T*2=1.5 ms which produces *v*a = 50 m s-1 (unambiguous interval is -50 to 50 m s-1) and unambiguous range of at least 150 km.

Fig. 10. Staggered PRT. The stagger ratio *T*1/*T*2 = 2/3. The continuous curve depicts the return from precipitation extending up to c*T*2/2 but not further (from Torres et al., 2009 and adapted from Sachidananda & Zrnic, 2003).

Power estimates in range sections I, II, and III (Fig.10) are computed separately for the short PRT and the long PRT to check if data censoring is needed. Comparison of powers in the two PRT intervals indicates if there is overlay and how severe it is so that appropriate censoring can be applied. In Fig. 11 contrasted are two fields of velocities obtained with two radars (spaced about 20 km apart). The left field comes from the operational WSR-88D in Oklahoma City and was obtained with the "batch mode" and parameters as indicated. On the right is the same storm complex but obtained with staggered PRT on the research WSR-88D radar in Norman OK some 20 km SSW from Oklahoma City. Highlighted in yellow circles are regions where significant aliasing occurs on the operational radar (exemplified by abrupt discontinuities in the field, change from red to green) but are absent in the field from Doppler Radar Observations – 20 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Currently a significant transformation of the radars is ongoing; it is addition of dual polarization (Zrnic et al., 2008). By mid 2013 all radars on the network should have this capability. Although Doppler capability is not a prerequisite for dual polarization, the coherency of transmit-receive signals within one PRT is for differential phase measurement. Dual polarization offers ample possibilities for application of spectral analysis to polarimetric signals and these are being explored (e.g., to discriminate between insects and birds, Bachman & Zrnic, 2007; to suppress ground clutter, Unal, 2009; or to achieve adaptive

Three improvements approved for soon inclusion on the network are pending. These are staggered PRT, processing of range oversampled signals, and adaptive recognition and

It is planned for mitigating range velocity ambiguities at mid elevation angles with possible use at the lower elevations. The scheme consists of alternating interval between transmitted pulses (Fig. 10) and estimating arguments of two autocorrelations at the two lags, arg[*R*(*T*1)] and arg[*R*(*T2*)]. The velocities estimated from these arguments have a different unambiguous interval (each inversely proportional to the corresponding separation *T*i, i=1 or 2) as can be deduced from eq. (8). Therefore the difference of the velocities uniquely tags the proper unambiguous interval for either PRT so that correct dealiasing can be achieved (Torres et al., 2004a) up to larger *v*a than possible with only one of these PRTs . For the example in Fig. 10, *v*a = 3*v*a2 = 2*v*a1. Consider *T*1=1 ms *T*2=1.5 ms which produces *v*a = 50 m s-1

(unambiguous interval is -50 to 50 m s-1) and unambiguous range of at least 150 km.

Fig. 10. Staggered PRT. The stagger ratio *T*1/*T*2 = 2/3. The continuous curve depicts the return from precipitation extending up to c*T*2/2 but not further (from Torres et al., 2009 and

Power estimates in range sections I, II, and III (Fig.10) are computed separately for the short PRT and the long PRT to check if data censoring is needed. Comparison of powers in the two PRT intervals indicates if there is overlay and how severe it is so that appropriate censoring can be applied. In Fig. 11 contrasted are two fields of velocities obtained with two radars (spaced about 20 km apart). The left field comes from the operational WSR-88D in Oklahoma City and was obtained with the "batch mode" and parameters as indicated. On the right is the same storm complex but obtained with staggered PRT on the research WSR-88D radar in Norman OK some 20 km SSW from Oklahoma City. Highlighted in yellow circles are regions where significant aliasing occurs on the operational radar (exemplified by abrupt discontinuities in the field, change from red to green) but are absent in the field from

clutter and noise suppression, Moisseev & Chandrasekar, 2009).

filtering of ground clutter. Brief description follows.

adapted from Sachidananda & Zrnic, 2003).

**4. Near term enhancements** 

**4.1 Staggered PRT** 

the research radar. Also, the large pink area of overlaid echoes has almost disappeared in the measurement made utilizing the staggered PRT. The small circle closest to the radar origin indicates overlaid echo contaminating the first trip velocities of the operational radar.

Fig. 11. Velocity fields of a storm system. Left: field obtained with the operational WSR-88D radar in Oklahoma City on April 06, 2003, elevation 2.5 deg, batch mode with unambiguous range of 148 km and velocity of 25.4 m s-1. Pink regions locate censored velocities which can not be reliably recovered due to overlaid first and second trip echoes. Right: same as on the left but obtained with the research WSR-88D (KOUN) utilizing staggered PRT. This radar is about 20 km south from the operational radar. The color bar indicates velocities (m s-1), red away from and green toward the radar. (Figure adapted from Torres et al., 2003).

#### **4.2 Oversampling techniques**

Oversampling here indicates spacing of I, Q samples smaller than the pulse duration. Operations on few of these range consecutive "oversamples" can reduce error in estimates and/or data acquisition time (Torres & Zrnic, 2003). Simplest of operations is averaging in range of oversampled spectral moments. Somewhat more involved is the whitening transformation in which the signal vector **v =** [V(*m*,0), V(*m*,1),... V(*m*, *l*),....V(*m*, *L*)] consisting of *L* oversampled correlated complex voltages is transformed into a set of *L* orthogonal voltages (Torres & Zrnic, 2003). The time index *m* refers to the usual sample time and *l* to the oversampled range time. The transformation takes the form **x** = **H**-1 **v** with **H** related to the normalized correlation matrix **C** of **v** via **C**=**H\*HT**. The correlation matrix can be precomputed (or measured e.g., Ivic et al., 2003) because it depends solely on the envelope of the transmitted pulse and the baseband equivalent receiver filter shape for a uniform Z. The *L* transformed samples are independent and averaging of spectral moments obtained from each (in absence of noise) yields smaller error of estimates. Whitening is effective at large SNRs but fails otherwise. To achieve *L* independent samples the receiver filter bandwidth needs to be increased *L* times over the matched filter bandwidth and this enhances the noise by the same factor. In addition the whitening transformation also increases the noise hence

Doppler Radar for USA Weather Surveillance 23

Fig. 12. Fields of reflectivity and velocity from a severe storm obtained on 2 Apr 2010 10:54 UTC, with the phased array radar (NWRT) in Norman, OK. Top two panels resulted for pseudowhitening applied to *L =* 4 samples of time series data; the number of samples *M* per radial was 12 for Z and 26 for *v*. Data in the lower panels have been obtained by processing as on the WSR-88D (16 for Z and 64 for *v*). The curved discontinuity in the velocity field delineates outflow boundary (gust front) generated by this storm. The peak reflectivity values of ~ 65 dBZ are likely caused by hail. (Adapted from Curtis & Torres, 2011).

the net SNR reduction is proportional to *L*2. Practical *L* is about 3 to 6, so the decrease is not catastrophic considering that weather SNRs are mostly larger than 20 dB. Another issue concerning whitening is the shape of the range weighting function compared to the matched filter. The two weighting functions have the same range extent but the one from whitening has rectangular shape smearing slightly its range resolution.

Increasing the number of independent samples when it is advantageous and gradually reverting to the matched filter has also been proposed (Torres et al. 2004b) and implemented (Curtis & Torres, 2011) on the National Weather Radar Testbed (NWRT), a phased array radar antenna powered by a WSR-88D transmitter (Zrnic et al., 2007). The processing is called adaptive pseudowhitening. It requires initial estimates of SNR and spectrum width.

Vivid example contrasting adaptive pseudowhitening to standard processing illustrates the much smoother fields obtained with the former (see Fig.12, and caption). The gradient of Doppler velocities (indicated with an arrow) is at the interface of the storms outflow and the environmental flow. This type of discontinuity is the key feature detected by algorithms for locating gust fronts and quantifying wind shear across the boundary; such information is extremely useful for air traffic management and safety at airports.

In contrast to whitening techniques pulse compression does not degrade the SNR (Doviak and Zrnic, 2006) but is not considered due to excessive bandwidth and current hardware constraints. A very simple alternative to speed volume coverage at lowest elevations (where tornadoes are observed) is a VCP with adaptive top elevation angle based on radar measurements (Chrisman et al., 2009). It will soon be added to the VCPs on the network.

#### **4.3 Clutter detection and filtering**

A novel way to recognize and filter ground clutter is planned. Its acronym CLEAN-AP stands for clutter environment analysis using adaptive processing (Warde & Torres, 2009). The essence of the technique is spectral analysis (decomposition) of the autocorrelation at lag 1 and use of its phase at and near zero Doppler shift. The conventional estimate

$$
\hat{R}\_b(1) = \frac{1}{M^2} \sum\_{0}^{M-1} \left| Z(k) \right|^2 e^{j2\pi k/M} \,\prime \,\tag{10}
$$

where *Z*(*k*) is the discrete Fourier transform of the returned signal, is biased (indicated by subscript b) and can be unbiased as in (7). Another way to avoid the bias is by computing two Fourier transform as proposed by (Warde & Torres 2009). One, *Z*0(*k*) is the complex spectrum of *d*(*m*)*V*(*m*), *d*(*m*)=window function, and the other *Z*1(*k*) is the spectrum of *d*(*m*)*V*(*m*+1) from the sequence shifted in time by one unit (*Ts*). Then the unbiased estimate is

$$\hat{R}(1) = \sum\_{k=0}^{M-1} Z\_0^\*(k) Z\_1(k) / \, M^2. \tag{11}$$

Individual terms \* 1 01 *S k Z kZ k* () () () constitute the spectral density (over Doppler index *k*) of the lag 1 autocorrelation function. Thus the autocorrelation spectral density is estimated in CLEAN-AP from the cross spectrum.

Doppler Radar Observations – 22 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

the net SNR reduction is proportional to *L*2. Practical *L* is about 3 to 6, so the decrease is not catastrophic considering that weather SNRs are mostly larger than 20 dB. Another issue concerning whitening is the shape of the range weighting function compared to the matched filter. The two weighting functions have the same range extent but the one from whitening

Increasing the number of independent samples when it is advantageous and gradually reverting to the matched filter has also been proposed (Torres et al. 2004b) and implemented (Curtis & Torres, 2011) on the National Weather Radar Testbed (NWRT), a phased array radar antenna powered by a WSR-88D transmitter (Zrnic et al., 2007). The processing is called adaptive pseudowhitening. It requires initial estimates of SNR and spectrum width. Vivid example contrasting adaptive pseudowhitening to standard processing illustrates the much smoother fields obtained with the former (see Fig.12, and caption). The gradient of Doppler velocities (indicated with an arrow) is at the interface of the storms outflow and the environmental flow. This type of discontinuity is the key feature detected by algorithms for locating gust fronts and quantifying wind shear across the boundary; such information is

In contrast to whitening techniques pulse compression does not degrade the SNR (Doviak and Zrnic, 2006) but is not considered due to excessive bandwidth and current hardware constraints. A very simple alternative to speed volume coverage at lowest elevations (where tornadoes are observed) is a VCP with adaptive top elevation angle based on radar measurements (Chrisman et al., 2009). It will soon be added to the VCPs on the network.

A novel way to recognize and filter ground clutter is planned. Its acronym CLEAN-AP stands for clutter environment analysis using adaptive processing (Warde & Torres, 2009). The essence of the technique is spectral analysis (decomposition) of the autocorrelation at

where *Z*(*k*) is the discrete Fourier transform of the returned signal, is biased (indicated by subscript b) and can be unbiased as in (7). Another way to avoid the bias is by computing two Fourier transform as proposed by (Warde & Torres 2009). One, *Z*0(*k*) is the complex spectrum of *d*(*m*)*V*(*m*), *d*(*m*)=window function, and the other *Z*1(*k*) is the spectrum of *d*(*m*)*V*(*m*+1) from the sequence shifted in time by one unit (*Ts*). Then the unbiased estimate is

ˆ(1) ( ) ( ) / .

the lag 1 autocorrelation function. Thus the autocorrelation spectral density is estimated in

*R Z kZ k M* 

<sup>1</sup> <sup>2</sup> 2 /

\* 2 0 1

1 01 *S k Z kZ k* () () () constitute the spectral density (over Doppler index *k*) of

, (10)

(11)

lag 1 and use of its phase at and near zero Doppler shift. The conventional estimate

2 0 <sup>1</sup> <sup>ˆ</sup> (1) ( ) *M <sup>j</sup> k M Rb Zk e*

1

*M*

*k*

0

*M*

has rectangular shape smearing slightly its range resolution.

extremely useful for air traffic management and safety at airports.

**4.3 Clutter detection and filtering** 

Individual terms \*

CLEAN-AP from the cross spectrum.

Fig. 12. Fields of reflectivity and velocity from a severe storm obtained on 2 Apr 2010 10:54 UTC, with the phased array radar (NWRT) in Norman, OK. Top two panels resulted for pseudowhitening applied to *L =* 4 samples of time series data; the number of samples *M* per radial was 12 for Z and 26 for *v*. Data in the lower panels have been obtained by processing as on the WSR-88D (16 for Z and 64 for *v*). The curved discontinuity in the velocity field delineates outflow boundary (gust front) generated by this storm. The peak reflectivity values of ~ 65 dBZ are likely caused by hail. (Adapted from Curtis & Torres, 2011).

Doppler Radar for USA Weather Surveillance 25

Fig. 14. Fields of reflectivities (top) and velocities (bottom) with no filter (right) and after application of the CLEAN-AP. Close to the radar the strong reflectivities in the top right panel encircled in red (red indicates > 50 dBZ) are caused by ground clutter which also biases the velocities toward zero (lower right panel). CLEAN-AP eliminates most of the clutter in both fields (left panels). To the NE within the yellow circle there are areas of near zero velocities (lower panels gray areas are velocities within ±5 m s-1). These appear unaffected by the filter. The data were collected with the agile beam phased array radar

The spectrum width estimator (9) is deficient at narrow widths where significant bias occurs. This shortcoming will be overcome with the Hybrid estimator which chooses an appropriate equation depending on a rough initial estimate of *σ<sup>v</sup>* (Meymaris et al., 2009). Initial estimate of the spectrum width is made using thee estimators) (9), ˆ ˆ *R R* (1) / (2) as in (Doviak & Zrnic, 2006 eq. 6.32) and an estimator based on ˆˆ ˆ *RR R* (1), (2), and (3) . Criteria applied to the results produce three categories of widths, large, medium, and small. Then (9) is used as estimate for the large category, ˆ ˆ *R R* (1) / (2) for the medium and ˆ ˆ *R R* (1) / (3) for the

Mesocylone refers to a rotational part of storm with the diameter of maximum wind typically between 3 and 10 km. It is depicted with a couplet of Doppler velocity features (see Fig. 9). Storms having mesocyclones can produce devastating tornadoes (Fig. 9 exhibits a tornado vortex signature associated with the mesocylone), strong winds, and hail. Thus, much effort has been devoted to detecting and quantifying these phenomena (No. 2 issue of Weather and Forecasting, 1998). One of the motivating reasons for installing Doppler radars

(NWRT) in Norman, OK. (Figure adapted from Warde & Torres, 2009).

**4.4 Hybrid spectrum width estimator** 

**5. Observations of phenomena** 

small.

Fig. 13. Autocorrelation spectral density (ASD) of a radar return, top: magnitude and bottom: phase. Clutter is well defined with its peak at zero and flat phase (red). Based on this phase five coefficients are replaced with interpolated values resulting in 14.4 dB of suppression (defined as the ratio of total S+C power to remaining power). Interpolated powers are indicated by the dotted line; dash line represents linear phase; *va=* 27 m s-1. Data obtained with the phased array radar (NWRT). (Figure courtesy of Sebastian Torres).

The choice of window function *d*(*m*) is very important because its sidelobes limit the amount of power that can be filtered. The clutter power is computed from the sum of *V*(*m*) to obtain the clutter to noise ratio (CNR). Then the CNR is compared with the peak to first sidelobe level (PSw) ratio of four windows (w=rectangular, von Hann, Blackman, and Blackman-Nuttall) and the window whose PSw exceeds the CNR by the smallest amount is chosen. That way the leakage of the clutter signal away from zero will be below the noise level, while the notch width will be smaller than the one for the other windows satisfying the condition PSw>CNR.

Data windows spread the phase of clutter's *S*1(*k*) either side of zero (*k*=0) Doppler (Fig. 13). Recognition of the flat phase identifies clutter's presence. Doppler index at which the phase begins to depart from zero (according to a set of criteria) defines the clutter filter width. In the mean the autocorrelation spectral density of noise has linear phase as seen in Fig.13 but semi coherent signals have flattened phases in the vicinity of their mean Doppler shifts. Panels in Fig. 14 demonstrate qualitatively performance of this clutter mitigation technique and the caption highlights results.

Doppler Radar Observations – 24 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Fig. 13. Autocorrelation spectral density (ASD) of a radar return, top: magnitude and bottom: phase. Clutter is well defined with its peak at zero and flat phase (red). Based on this phase five coefficients are replaced with interpolated values resulting in 14.4 dB of suppression (defined as the ratio of total S+C power to remaining power). Interpolated powers are indicated by the dotted line; dash line represents linear phase; *va=* 27 m s-1. Data obtained with the phased array radar (NWRT). (Figure courtesy of Sebastian Torres).

The choice of window function *d*(*m*) is very important because its sidelobes limit the amount of power that can be filtered. The clutter power is computed from the sum of *V*(*m*) to obtain the clutter to noise ratio (CNR). Then the CNR is compared with the peak to first sidelobe level (PSw) ratio of four windows (w=rectangular, von Hann, Blackman, and Blackman-Nuttall) and the window whose PSw exceeds the CNR by the smallest amount is chosen. That way the leakage of the clutter signal away from zero will be below the noise level, while the notch width will be smaller than the one for the other windows satisfying the

Data windows spread the phase of clutter's *S*1(*k*) either side of zero (*k*=0) Doppler (Fig. 13). Recognition of the flat phase identifies clutter's presence. Doppler index at which the phase begins to depart from zero (according to a set of criteria) defines the clutter filter width. In the mean the autocorrelation spectral density of noise has linear phase as seen in Fig.13 but semi coherent signals have flattened phases in the vicinity of their mean Doppler shifts. Panels in Fig. 14 demonstrate qualitatively performance of this clutter mitigation technique

condition PSw>CNR.

and the caption highlights results.

Fig. 14. Fields of reflectivities (top) and velocities (bottom) with no filter (right) and after application of the CLEAN-AP. Close to the radar the strong reflectivities in the top right panel encircled in red (red indicates > 50 dBZ) are caused by ground clutter which also biases the velocities toward zero (lower right panel). CLEAN-AP eliminates most of the clutter in both fields (left panels). To the NE within the yellow circle there are areas of near zero velocities (lower panels gray areas are velocities within ±5 m s-1). These appear unaffected by the filter. The data were collected with the agile beam phased array radar (NWRT) in Norman, OK. (Figure adapted from Warde & Torres, 2009).

#### **4.4 Hybrid spectrum width estimator**

The spectrum width estimator (9) is deficient at narrow widths where significant bias occurs. This shortcoming will be overcome with the Hybrid estimator which chooses an appropriate equation depending on a rough initial estimate of *σ<sup>v</sup>* (Meymaris et al., 2009). Initial estimate of the spectrum width is made using thee estimators) (9), ˆ ˆ *R R* (1) / (2) as in (Doviak & Zrnic, 2006 eq. 6.32) and an estimator based on ˆˆ ˆ *RR R* (1), (2), and (3) . Criteria applied to the results produce three categories of widths, large, medium, and small. Then (9) is used as estimate for the large category, ˆ ˆ *R R* (1) / (2) for the medium and ˆ ˆ *R R* (1) / (3) for the small.

#### **5. Observations of phenomena**

Mesocylone refers to a rotational part of storm with the diameter of maximum wind typically between 3 and 10 km. It is depicted with a couplet of Doppler velocity features (see Fig. 9). Storms having mesocyclones can produce devastating tornadoes (Fig. 9 exhibits a tornado vortex signature associated with the mesocylone), strong winds, and hail. Thus, much effort has been devoted to detecting and quantifying these phenomena (No. 2 issue of Weather and Forecasting, 1998). One of the motivating reasons for installing Doppler radars

Doppler Radar for USA Weather Surveillance 27

generated by subsequent storm. From the vertical cross section of the velocities it is evident that the positive velocity perturbation (toward the radar) ends at about 4000 ft, above which the ambient flow (green color) resumes. The velocities measured by the radar can quantify the structure of the perturbation, tell the thickness and wavelength. Propagation speed can

Fig. 16. Vertical cross sections of reflectivity field (left) and Doppler velocity field through a microburst reconstructed from conical scans (up to 19.5o elevation) of the WSR-88D radar in Phoenix Az on Aug 15, 1995. Height is in kft and distances are in nautical miles. The radar is located to the right of each cross section (at about 26 nautical miles). The top color bar depicts velocity categories in non linear increments with red away from the radar: light red

= 0-5 kts, dark red 5 to 10, next 10-20; green indicates toward the radar in categories symmetric to red. The bottom bar refers to reflectivities starting at 0 dBZ in steps of 5 dBZ

Fig. 17. Doppler velocities at 0.5o elevation and superposed vertical cross sections of the velocities obtained with Oklahoma City radar on Aug 10, 2011. Red color indicates motion away and green toward the radar located ESE of the bottom right corner. Height lines are in

(white category indicates values larger than 65 dBZ).

kft above ground level.

be estimated by tracking the wave position in space and time.

in the USA was the potential to detect mesocyclones and tornadoes. The investment in this technology paid off as demonstrated by the graph in Fig. 15. Trend of improvement is seen on all three performance indicators with the steepest rise in the years the Doppler radar network (NEXRAD) was being installed. This is logical: as the new tool was spreading across the country more forecasters were beginning to use it. Improvement continues few years past the completion of the network likely because it took time to train all forecasters and gain experience with the Doppler radar. The data indicates a plateau from about 2002 until present suggesting maturity of the technology with little room left for significant advancements. Further progress might come from combining radar data with short term numerical weather prediction models and/or introduction of rapidly scanning agile beam phase array radars (Zrnic et al., 2007 and Weber et al., 2007).

Fig. 15. Probability of detection, false alarms and lead time in tornado warnings issued by the National Weather Service as function of year. (Figure courtesy Don Burgess).

Doppler velocities are potent indicators of diverging (converging) flows such as observed in strong outflows from collapsing storms. These "microbursts" have been implicated in several aircraft accidents motivating deployment of terminal Doppler weather radars (TDWR) at forty seven airports in the USA (Mahapatra, 1999, sec 7.4). Vertical profiles of reflectivity and Doppler velocity in Fig. 16 indicate a pulsing microburst; the intense reflectivity core (red below 5 kft) near ground is the first precipitation shaft and the elongated portion above is the following shaft. On the velocity display the yellow arrows indicate direction of motion. Clear divergence near ground and at the top of the storm (in the anvil) is visible and so is the convergence over the deep mid storm layer (5 to 14 kft). The horizontal change in wind speed near ground of ~ 20 kts at this stage is not strong to pose treat to aviation (35 kts is considered significant for light aircraft).

An atmospheric undular bore (Fig. 17) was observed with the WSR-88D near Oklahoma City. This phenomena is a propagating step disturbance in air properties (temperature, pressure, velocity) followed by oscillation. Spaced by about 10 km the waves propagate in a surface-based stable layer. The layer came from storm outflow and the bore might have been Doppler Radar Observations – 26 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

in the USA was the potential to detect mesocyclones and tornadoes. The investment in this technology paid off as demonstrated by the graph in Fig. 15. Trend of improvement is seen on all three performance indicators with the steepest rise in the years the Doppler radar network (NEXRAD) was being installed. This is logical: as the new tool was spreading across the country more forecasters were beginning to use it. Improvement continues few years past the completion of the network likely because it took time to train all forecasters and gain experience with the Doppler radar. The data indicates a plateau from about 2002 until present suggesting maturity of the technology with little room left for significant advancements. Further progress might come from combining radar data with short term numerical weather prediction models and/or introduction of rapidly scanning agile beam

Fig. 15. Probability of detection, false alarms and lead time in tornado warnings issued by

Doppler velocities are potent indicators of diverging (converging) flows such as observed in strong outflows from collapsing storms. These "microbursts" have been implicated in several aircraft accidents motivating deployment of terminal Doppler weather radars (TDWR) at forty seven airports in the USA (Mahapatra, 1999, sec 7.4). Vertical profiles of reflectivity and Doppler velocity in Fig. 16 indicate a pulsing microburst; the intense reflectivity core (red below 5 kft) near ground is the first precipitation shaft and the elongated portion above is the following shaft. On the velocity display the yellow arrows indicate direction of motion. Clear divergence near ground and at the top of the storm (in the anvil) is visible and so is the convergence over the deep mid storm layer (5 to 14 kft). The horizontal change in wind speed near ground of ~ 20 kts at this stage is not strong to pose

An atmospheric undular bore (Fig. 17) was observed with the WSR-88D near Oklahoma City. This phenomena is a propagating step disturbance in air properties (temperature, pressure, velocity) followed by oscillation. Spaced by about 10 km the waves propagate in a surface-based stable layer. The layer came from storm outflow and the bore might have been

the National Weather Service as function of year. (Figure courtesy Don Burgess).

treat to aviation (35 kts is considered significant for light aircraft).

phase array radars (Zrnic et al., 2007 and Weber et al., 2007).

generated by subsequent storm. From the vertical cross section of the velocities it is evident that the positive velocity perturbation (toward the radar) ends at about 4000 ft, above which the ambient flow (green color) resumes. The velocities measured by the radar can quantify the structure of the perturbation, tell the thickness and wavelength. Propagation speed can be estimated by tracking the wave position in space and time.

Fig. 16. Vertical cross sections of reflectivity field (left) and Doppler velocity field through a microburst reconstructed from conical scans (up to 19.5o elevation) of the WSR-88D radar in Phoenix Az on Aug 15, 1995. Height is in kft and distances are in nautical miles. The radar is located to the right of each cross section (at about 26 nautical miles). The top color bar depicts velocity categories in non linear increments with red away from the radar: light red = 0-5 kts, dark red 5 to 10, next 10-20; green indicates toward the radar in categories symmetric to red. The bottom bar refers to reflectivities starting at 0 dBZ in steps of 5 dBZ (white category indicates values larger than 65 dBZ).

Fig. 17. Doppler velocities at 0.5o elevation and superposed vertical cross sections of the velocities obtained with Oklahoma City radar on Aug 10, 2011. Red color indicates motion away and green toward the radar located ESE of the bottom right corner. Height lines are in kft above ground level.

Doppler Radar for USA Weather Surveillance 29

leaving roost early in the morning. The critters are diverging away from the roost in search of food. Close to the radar the continuous field of velocities is principally from reflections off insects filling a good part of the boundary layer (this is deduced from polarimetric

Fig. 19. Field of velocities obtained from the radar at Moorhead City, NC, on July 27, 2011 at 5:08 in the morning. The color bar indicates categories in kts; red away from the radar and

The WSR-88D network has been indispensable for issuing warnings of precipitation and wind related hazards in the USA. And its real time display of storm locations has become one of most popular and common applications on cellular phones. Its role in quantitative precipitation estimation is matching that of rain gages. So, what is beyond these achievements for the WSR-88D? Dual polarization upgrade combined with Doppler capability is the panacea a radar with the dish antenna on a rotating pedestal can achieve. Promising possibilities are: polarimetric confirmation of tornado touchdown at places where Doppler velocities indicate rotation; improvement of ground clutter filtering; polarimetric spectral analysis for extracting/separating features within radar resolution volume; significant improvement in data interpretation; inclusion of wind and precipitation type/amount in numerical prediction models; and other. Clearly the evolutionary trend continues and will do so at a decelerating pace until a plateau is reached. Complementary shorter wavelength (3 cm and 5 cm) surveillance radars are being considered for closing gaps or providing extra coverage at opportune places. (The TDWRs 5 cm wavelength radar data has been supplied to the NWS for several years). Explored are networks of tightly

coordinated 3 cm wavelength radars for surveillance close to the ground.

signatures, but not shown here).

green is toward. Elevation is 0.5o.

**6. Epilogue** 

Doppler radar is valued for measuring winds in hurricanes and detecting tornadoes that can be imbedded in the bands. Combined with polarimetric capability, its utility greatly increases because of improved quantitative measurement of rainfall. Observation of hurricane Irene which swept the US East coast at the end of August 2011 is the case in point. Rotation speed of over 110 km h-1 is apparent in Fig. 18 where the color categories are too coarse to estimate the maximum values. The cyan color captures well Irene's rotational winds because they are aligned with radials. Color categories are coarse precluding precise estimation of velocities but recorded values are quantized to 0.5 m s-1. Although the unambiguous velocity is ~ 28 m s-1 values more negative than -30 m s-1 are displayed. These and other outside the unambiguous interval have been correctly dealiased by imposing spatial continuity to the field.

Fig. 18. Left: Rain rate in Hurricane Irene, obtained with a polarimetric algorithm using differential phase and reflectivity factor (surveillance scan with unambiguous range of ~ 465 km). Right: Velocity field obtained with the SZ(2) phase code (Doppler scan with unambiguous range of ~135 km and velocity ~ 28 m s-1). Elevation is 0.5o, time 12:26 UTC, on Aug 27, 2011. The range circles are spaced 50 km apart. Color categories for rain rate are in mm h-1 and for velocity in m s-1. (Figure courtesy of Pengfei Zhang).

The rain rate field depicts Irene's bands some containing values larger than 100 mm h-1. These are instantaneous measurements and over time accumulations caused significant flooding which brought 43 deaths and ~ 20 billion \$ damage to the NE coast of the USA. The obviously large spatial extent of Irene amply justifies use of surveillance scan for maximum storm coverage and Doppler scan for wind hazard detection.

Atmospheric biota is routinely observed with the WSR-88D network (Rinehart, 2010). Examples are insects, birds, and bats. Many insects are passive wind tracers providing a way to estimate winds in the planetary boundary layer (extending up to 2 km above ground).

Biota can be tracked for ecological or other purposes. The radar can also provide location of bird migrating paths, roosts, and other congregating places; this could be important for aircraft safety. The three donut shaped features in Fig. 19 represent Doppler speeds of birds Doppler Radar Observations – 28 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Doppler radar is valued for measuring winds in hurricanes and detecting tornadoes that can be imbedded in the bands. Combined with polarimetric capability, its utility greatly increases because of improved quantitative measurement of rainfall. Observation of hurricane Irene which swept the US East coast at the end of August 2011 is the case in point. Rotation speed of over 110 km h-1 is apparent in Fig. 18 where the color categories are too coarse to estimate the maximum values. The cyan color captures well Irene's rotational winds because they are aligned with radials. Color categories are coarse precluding precise estimation of velocities but recorded values are quantized to 0.5 m s-1. Although the unambiguous velocity is ~ 28 m s-1 values more negative than -30 m s-1 are displayed. These and other outside the unambiguous interval have been correctly dealiased by imposing

Fig. 18. Left: Rain rate in Hurricane Irene, obtained with a polarimetric algorithm using differential phase and reflectivity factor (surveillance scan with unambiguous range of ~ 465

unambiguous range of ~135 km and velocity ~ 28 m s-1). Elevation is 0.5o, time 12:26 UTC, on Aug 27, 2011. The range circles are spaced 50 km apart. Color categories for rain rate are

The rain rate field depicts Irene's bands some containing values larger than 100 mm h-1. These are instantaneous measurements and over time accumulations caused significant flooding which brought 43 deaths and ~ 20 billion \$ damage to the NE coast of the USA. The obviously large spatial extent of Irene amply justifies use of surveillance scan for maximum

Atmospheric biota is routinely observed with the WSR-88D network (Rinehart, 2010). Examples are insects, birds, and bats. Many insects are passive wind tracers providing a way to estimate winds in the planetary boundary layer (extending up to 2 km above

Biota can be tracked for ecological or other purposes. The radar can also provide location of bird migrating paths, roosts, and other congregating places; this could be important for aircraft safety. The three donut shaped features in Fig. 19 represent Doppler speeds of birds

km). Right: Velocity field obtained with the SZ(2) phase code (Doppler scan with

in mm h-1 and for velocity in m s-1. (Figure courtesy of Pengfei Zhang).

storm coverage and Doppler scan for wind hazard detection.

ground).

spatial continuity to the field.

leaving roost early in the morning. The critters are diverging away from the roost in search of food. Close to the radar the continuous field of velocities is principally from reflections off insects filling a good part of the boundary layer (this is deduced from polarimetric signatures, but not shown here).

Fig. 19. Field of velocities obtained from the radar at Moorhead City, NC, on July 27, 2011 at 5:08 in the morning. The color bar indicates categories in kts; red away from the radar and green is toward. Elevation is 0.5o.

#### **6. Epilogue**

The WSR-88D network has been indispensable for issuing warnings of precipitation and wind related hazards in the USA. And its real time display of storm locations has become one of most popular and common applications on cellular phones. Its role in quantitative precipitation estimation is matching that of rain gages. So, what is beyond these achievements for the WSR-88D? Dual polarization upgrade combined with Doppler capability is the panacea a radar with the dish antenna on a rotating pedestal can achieve. Promising possibilities are: polarimetric confirmation of tornado touchdown at places where Doppler velocities indicate rotation; improvement of ground clutter filtering; polarimetric spectral analysis for extracting/separating features within radar resolution volume; significant improvement in data interpretation; inclusion of wind and precipitation type/amount in numerical prediction models; and other. Clearly the evolutionary trend continues and will do so at a decelerating pace until a plateau is reached. Complementary shorter wavelength (3 cm and 5 cm) surveillance radars are being considered for closing gaps or providing extra coverage at opportune places. (The TDWRs 5 cm wavelength radar data has been supplied to the NWS for several years). Explored are networks of tightly coordinated 3 cm wavelength radars for surveillance close to the ground.

Doppler Radar for USA Weather Surveillance 31

Hubbert, J.C., M. Dixon, & S.M. Ellis (2009). Weather radar ground clutter. Part II) Real- time identification and filtering. *Jour. Atmosph. Oceanic. Tech*. Vol. *26,* pp. 1181-1197. Ice, L.R., & D.S. Saxion (2011). Enhancing the foundational data from the WSR-88D) Part II,

Mahapatra, P. (1999). *Aviation weather surveillance systems*. Published by IEE and AIAA,

Meischner, P. (2004). *Weather radar, principles and advanced applications.* Springer-Verlag,

Meymaris, G., J.K. Williams, & J.C. Hubbert (2009). Performance of a proposed hybrid

McLaughlin, D., & Coauthors (2009). Short-wavelength technology and the potential for

Moisseev, D.N., & V. Chandrasekar (2009). Polarimetric spectral filter for adaptive clutter

Rinehart , R.E. (2010). *Radar for meteorologists.* Fifth edition. Rinehart Publications, Nevada,

Sachidananda, M., & D.S. Zrnic (2003). Unambiguous Range Extension by Overlay

Sachidananda, M., & D.S. Zrnic (1999). Systematic phase codes for resolving range overlaid

Saxion, D, S., & R. L. Ice (2011). Enhancing the foundational data from the WSR- 88D) Part I, a history of success. *35th Conference on Radar Meteorology,* AMS, Pittsburgh, PA. Serafin, R.J. & J.W. Wilson (2000). Operational weather radar in the United States Progress

Siggia, A.D., & R. E. Passarelli, Jr. (2004). Gaussian model adaptive processing (GMAP) for

Torres, M.S., C.D. Curtis, D.S. Zrnic, & M. Jain (2007). Analysis of new Nexrad spectrum width estimator. *33rd Inter. Conf. on Radar Meteorology,* AMS, Cairns, Australia. Torres, M.S., Y.F. Dubel, & D.S. Zrnic (2004a). Design, implementation, and demonstration

Torres, M.S., C.D. Curtis, & J.R. Cruz (2004b). Pseudowhitening of weather radar signals to

Torres S., Sachidananda, M, & D. Zrnic (2004c). Signal Design and Processing Techniques

available from http://publications.nssl.noaa.gov/wsr88d\_reports/.

and opportunity. *Bull. Amer. Meteor. Soc.*, Vol. 81, pp. 501-518.

*(2004)*, pp. 67-73. Visby, Island of Gotland, Sweden.

ratios. *IEEE Trans. Geosc. Remote Sens.* Vol. 42, pp. 941-949.

and noise suppression. *J. Atmos. Oceanic Technol.*, Vol. 26, 215-228.

spectrum width estimator for the NEXRAD ORDA. *25th Int. Conf. on IIPS.* AMS,

distributed networks of small radar systems. *Bull. Amer. Meteor. Soc.,* Vol. 90, pp.

Resolution in Staggered PRT Technique. *J. Atmos. Oceanic Technol.*, Vol. 20, pp. 673-

signals in a Doppler weather radar. *J. Atmos. Oceanic Technol.*, Vol. 16, pp. 1351-

improved ground clutter cancellation and moment calculation. *Proceedings of ERAD* 

of a staggered PRT algorithm for the WSR-88D. *J. Atmos. Oceanic Technol.*, 21, 1389-

improve spectral moment and polarimetric variable estimates at low signal-to-noise

for WSR-88D Ambiguity Resolution) Phase coding and staggered PRT, implementation, data collection, and processing. NOAA/NSSL Report, Part 8,

*Technol.*, Vol. 20, pp. 1449-1459.

Berlin, Germany.

Phoenix, AZ, USA.

1797–1817.

MO, USA.

684.

1363.

1399.

printed by Short Run Press, Ltd, Exeter, UK.

the future, *35th Conference on Radar Meteorology*, AMS, Pittsburgh, PA, USA. Ivic, R.I., D.S. Zrnic, & S. M. Torres (2003). Whitening in range to improve weather radar spectral moment estimates. Part II) Experimental evaluation. *J. Atmos. Oceanic* 

One emerging technology is rapid scan agile beam phased array radar. This might be the ultimate radar providing it exceeds all the capabilities on the current network at faster scan rates. If in addition it proves to fulfill security and aviation needs (tracking of airplanes, missiles) it could revolutionize the current radar paradigm.

### **7. Acknowledgment**

The author is grateful to Rich Ice, Darcy Saxion, Alan Free, and Dave Zittel for advice and valuable information about the WSR-88D. Sebastian Torres provided several figures and comments concerning technical aspects and designed signal processing for the MPAR. Dave Warde contributed figures and details about ground clutter and some VCPs. Collaboration with Dick Doviak is reflected in the requirements section. Allen Zahrai was in charge of engineering developments on MPAR and KOUN; Doug Forsyth lead the MPAR team in outstanding support and development of that platform.

#### **8. References**


Doppler Radar Observations – 30 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

One emerging technology is rapid scan agile beam phased array radar. This might be the ultimate radar providing it exceeds all the capabilities on the current network at faster scan rates. If in addition it proves to fulfill security and aviation needs (tracking of airplanes,

The author is grateful to Rich Ice, Darcy Saxion, Alan Free, and Dave Zittel for advice and valuable information about the WSR-88D. Sebastian Torres provided several figures and comments concerning technical aspects and designed signal processing for the MPAR. Dave Warde contributed figures and details about ground clutter and some VCPs. Collaboration with Dick Doviak is reflected in the requirements section. Allen Zahrai was in charge of engineering developments on MPAR and KOUN; Doug Forsyth lead the MPAR team in

Bachmann, S., & D. Zrnic (2007). Spectral Density of Polarimetric Variables Separating

Bringi, V. N., & V. Chandrasekar (2001). *Polarimetric Doppler Weather Radar*. Cambridge

Brown, R. A., V. T. Wood, & D. Sirmans (2002). Improved tornado detection using simulated

Brown, R. A., B. A. Flickinger, E. Forren, D. M. Schultz, D. Sirmans, P. L. Spencer, V. T.

Burgess, D. W., V. T. Wood, & R. A. Brown (1982). Mesocyclone evolution statistics, *Severe* 

Chrisman, J.N. (2009). Automated volume scan evaluation and termination (AVSET). *34th* 

Crum, T.D., & R.L. Alberty (1993). The WSR-88D & the WSR-88D operational support

Crum, T.D., R.L. Alberty, & D.W. Burgess (1993). Recording, archiving, and using WSR-88D

Curtis, D.C., & S. M. Torres (2011). Adaptive range oversampling to achieve faster scanning

Doviak, R.J., & D. S. Zrnic (2006). *Doppler radar and weather observations*. Second edition,

Free, A.D., & N.K. Patel (2007). Clutter censoring theory and application for the WSR-88D. *32nd Conference on Radar Meteorology,* AMS, Albuquerque, NM, USA. Heiss, W.H, D.L McGrew, & D. Sirmans (1990). Next generation weather radar (WSR- 88D).

and the national weather radar testbed phased array radar. *J. Atmos. Oceanic* 

*Storm Conf. Proc*., pp. 422-424, AMS, Boston, MA, USA.

*Conference on Radar Meteorology*, AMS, Williamsburg, VA, USA.

facility." *B. American Meteorological Society,* Vol. 74, pp. 1669-1687.

data." *B. American Meteorological Society,* Vol*.* 74, pp. 645-653.

Biological Scatterers in the VAD Display. *J. Atmos. Oceanic Technol.*, Vol. 24,

and actual WSR-88D data with enhanced resolution. *J. Atmos. Oceanic Technol*., Vol.

Wood, & C. L. Ziegler (2005). Improved detection of severe storms using experimental fine-resolution WSR-88D measurements. *Weather and Forecasting*, Vol.

missiles) it could revolutionize the current radar paradigm.

outstanding support and development of that platform.

University Press, Cambridge, UK.

**7. Acknowledgment** 

**8. References** 

pp.1186–1198.

19, pp. 1759-1771.

*Technol*., in press.

reprinted by Dover, Mineola, NY, USA.

*Microwave J.,* Vol. 33, pp. 79-98.

20, 3-14.


**2** 

*1Environment Canada 2Bureau of Meteorology,* 

*4Deutcher Wetterdienst,* 

*6Hong Kong Observatory, 7Japan Meteorological Agency,* 

> *1Canada 2Australia 3USA 4Germany 7Japan 5,6,8China*

**Automated Processing of Doppler Radar Data** 

Paul Joe1, Sandy Dance2, Valliappa Lakshmanan3, Dirk Heizenreder 4, Paul James4, Peter Lang4, Thomas Hengstebeck4, Yerong Feng5, P.W. Li6,

Radar is the only operational tool that provides observations of severe weather producing thunderstorms on a fine enough temporal or spatial resolution (minutes and kilometers) that enables warnings of severe weather. It can provide a three- dimensional view about every five to ten minutes at a spatial resolution of the order of 1 km or less. The development and evolution of intense convective precipitation is closely linked to thunderstorms and so understanding of the microphysics and dynamics of precipitation is needed to understand the evolution of thunderstorms as diabatic and precipitation

The characteristics and proportion of severe weather is climatologically or geographically dependent. For example, the highest incidence of tornadoes is in the central U.S. whereas the tallest thunderstorms are found in Argentina (Zipser et al, 2006). Warning services developed at National Hydrological and Meteorological Services (NHMS) often originate because of a particular damaging severe weather event and ensuing expectations of the public. Office organization, resources and expertise are critical considerations in the use of radar for the preparation of severe weather warnings. Warnings also imply a level of legal liability requiring the authority of an operational National Hydrological Meteorological

processes modify and create hazardous rain, hail, wind and lightning.

**1. Introduction** 

Hon-Yin Yeung6, Osamu Suzuki7, Keiji Doi7 and Jianhua Dai8

*5Guandong Meteorological Bureau, China Meteorological Agency,* 

*8Shanghai Meteorological Bureau, China Meteorological Agency,* 

**for Severe Weather Warnings** 

*3CIMMS/OU/National Severe Storms Laboratory,* 


### **Automated Processing of Doppler Radar Data for Severe Weather Warnings**

Paul Joe1, Sandy Dance2, Valliappa Lakshmanan3, Dirk Heizenreder 4, Paul James4, Peter Lang4, Thomas Hengstebeck4, Yerong Feng5, P.W. Li6, Hon-Yin Yeung6, Osamu Suzuki7, Keiji Doi7 and Jianhua Dai8 *1Environment Canada 2Bureau of Meteorology, 3CIMMS/OU/National Severe Storms Laboratory, 4Deutcher Wetterdienst, 5Guandong Meteorological Bureau, China Meteorological Agency, 6Hong Kong Observatory, 7Japan Meteorological Agency, 8Shanghai Meteorological Bureau, China Meteorological Agency, 1Canada 2Australia 3USA 4Germany 7Japan 5,6,8China* 

#### **1. Introduction**

Doppler Radar Observations – 32 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Torres S., D. Zrnic, & Y. Dubel (2003). Signal Design and Processing Techniques for WSR-

Torres, S.M., & D.S. Zrnic (2003). Whitening in range to improve weather radar spectral

Unal, C. (2009). Spectral polarimetric radar clutter suppression to enhance atmospheric

Warde, A. D., & S. M. Torres (2009). Automatic detection and removal of ground clutter

Weber, M., J.Y.N. Cho, J.S. Flavin, J. M. Herd, W. Benner, & G. Torok (2007). The next

Wood, V.T, R. A. Brown, & D. Sirmans (2001). Technique for improving detection of WSR-

Zittel, W.D. , D. Saxion, R. Rhoton, & D.C. Crauder (2008). Combined WSR-88D technique to

Zrnic, D. S., J. F. Kimpel, D. F. Forsyth, A. Shapiro, G. Crain, R. Ferek, J. Heimmer, W.

weather observations. *Bull. Amer. Meteorol. Soc.*, Vol. 88, pp. 1753-1766. Zrnic, D., S. V. M. Melnikov, & I. Ivic, 2008: Processing to obtain polarimetric variables on

the ORDA (final version) NOAA/NSSL Report available from

http://publications.nssl.noaa.gov/wsr88d\_reports/.

collection, and processing. NOAA/NSSL Report, Part 7, available from

http://publications.nssl.noaa.gov/wsr88d\_reports/.

echoes. *J. Atmos. Oceanic Technol*., Vol. 26, pp. 1781-1797.

Vol. 20, pp. 1443-1448.

Williamsburg, VA, USA.

Vol. 88, pp. 1739-1751.

*Forecasting.* Vol. 16, pp. 177-184.

*Conference,* AMS. New Orleans.

88D Ambiguity Resolution) Phase coding and staggered PRT, implementation, data

moment estimates. Part I) Formulation and simulation. *J. Atmos. Oceanic Technol.,* 

contamination on weather radars. *34th Conference on Radar Meteorology*, AMS,

generation multi-mission U.S. surveillance radar network. *Bull. Amer. Meteorol. Soc.*,

88D mesocyclone signatures by increasing angular sampling. *Weather and* 

reduce range aliasing using phase coding and multiple Doppler scans. *24th IIPS* 

Benner, T. J. McNellis, & R. J. Vogt (2007). Agile beam phased array radar for

Radar is the only operational tool that provides observations of severe weather producing thunderstorms on a fine enough temporal or spatial resolution (minutes and kilometers) that enables warnings of severe weather. It can provide a three- dimensional view about every five to ten minutes at a spatial resolution of the order of 1 km or less. The development and evolution of intense convective precipitation is closely linked to thunderstorms and so understanding of the microphysics and dynamics of precipitation is needed to understand the evolution of thunderstorms as diabatic and precipitation processes modify and create hazardous rain, hail, wind and lightning.

The characteristics and proportion of severe weather is climatologically or geographically dependent. For example, the highest incidence of tornadoes is in the central U.S. whereas the tallest thunderstorms are found in Argentina (Zipser et al, 2006). Warning services developed at National Hydrological and Meteorological Services (NHMS) often originate because of a particular damaging severe weather event and ensuing expectations of the public. Office organization, resources and expertise are critical considerations in the use of radar for the preparation of severe weather warnings. Warnings also imply a level of legal liability requiring the authority of an operational National Hydrological Meteorological

Automated Processing of Doppler Radar Data for Severe Weather Warnings 35

Severe weather is defined here as heavy rains, hail, strong winds including tornadoes and lightning. In the production of warnings, thresholds need to be defined. The thresholds are necessarily locally defined by climatology, local infrastructure and familiarity will dictate what is extreme. Table 1-4 show the warning criteria for Canada circa 1995. Canada is a very big country covering many different weather climatologies and therefore is illustrative of the variation of the severe weather thresholds (see also Galway, 1989). For example, Newfoundland on the east coast of Canada is a very windy location and hence strong winds are a common occurrence and the people have adapted to their environment and therefore it has the highest wind threshold in Canada. Each service needs to define these for them selves.

Fig. 1. The envisioned warning process from outlook to tornado watch. This is typical of the process that is used in most countries providing severe weather warning services. Getting the message out to and understood by the public is very important aspect of the utility of the warning service. Superimposing the warning on television, internet, mobile devices and

directed messaging are critical to have the message heard.

**2.1 Severe weather definition** 

Service. All available data and timely access is critical and requires substantial infrastructure, ongoing support and maintenance. Besides meteorological data, eye witness observations and reports are also essential element in the issuance of tornado warnings (Doswell et al, 1999; Moller, 1978).

This contribution will discuss operational or operational prototypical radar processing, visualization systems for the production of convective severe weather warnings. The focus will be on the severe weather identification algorithms, the underlying philosophy for its usage, the level of expertise required, decision-making and the preparation of the warning. Radar is also used for the precipitation estimation and its application for flash flood warnings. This is discussed elsewhere (Wilson and Brandes, 1979). Only a few countries have convective thunderstorm warning services and the target audience for this contribution are those countries or NHMS' considering developing such a service. The intent is to provide a broad overview and global survey of radar processing systems for the provision of severe weather warning services. There is a considerable literature in convective weather forecasting and warning, this contribution can only explore a few aspects of this topic (Doswell, 1982: Doswell, 1985; Johns and Doswell, 1992; Wilson et al, 1998).

The forecasting and the warning of severe weather are very briefly described. Then, the underlying technique for the identification of severe thunderstorms using radar is presented. This forms the basis for the radar algorithms that identify the severe storm features. The basic components of the system are then described. Some details and unique innovations are incorporated in the global survey of operational or near operational use. This is concluded by a summary.

#### **2. Forecasting/Nowcasting/Severe weather warnings**

Severe weather predictions are divided into severe weather watches and severe weather warnings. In the preceding days, thunderstorm outlooks may be issued. Watches are predictions of the potential of severe weather. They are strategic in nature and fairly coarse in spatial and temporal resolution. They are often issued on a schedule or in conjunction with the public forecast. The expected behaviour is that the public would be aware of the possibility of severe weather and to listen for future updates. Warnings are predictions of the occurrence or imminent occurrence (with high certainty) of severe weather. They are tactical and more specific in location and time. They are also specific in weather element. They are a call to action and to protect one's property and one's self. They are issued and updated as necessary. Fig. 1 shows an overview of the process from the Japanese Meteorological Agency.

Weather advisories are issued if the weather is a concern but not hazardous. Specific types of warning, such as tornado or hail warnings may then be issued and generally after the more generic severe thunderstorm warning is issued.

The key difference is that the watch is a forecast or very short range forecast service as strategic in nature whereas the warning is a nowcast (based on existing data, precise in time, location and weather element) and tactical in nature.

#### **2.1 Severe weather definition**

Doppler Radar Observations – 34 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Service. All available data and timely access is critical and requires substantial infrastructure, ongoing support and maintenance. Besides meteorological data, eye witness observations and reports are also essential element in the issuance of tornado warnings

This contribution will discuss operational or operational prototypical radar processing, visualization systems for the production of convective severe weather warnings. The focus will be on the severe weather identification algorithms, the underlying philosophy for its usage, the level of expertise required, decision-making and the preparation of the warning. Radar is also used for the precipitation estimation and its application for flash flood warnings. This is discussed elsewhere (Wilson and Brandes, 1979). Only a few countries have convective thunderstorm warning services and the target audience for this contribution are those countries or NHMS' considering developing such a service. The intent is to provide a broad overview and global survey of radar processing systems for the provision of severe weather warning services. There is a considerable literature in convective weather forecasting and warning, this contribution can only explore a few aspects of this topic (Doswell, 1982: Doswell, 1985; Johns and Doswell, 1992; Wilson et al,

The forecasting and the warning of severe weather are very briefly described. Then, the underlying technique for the identification of severe thunderstorms using radar is presented. This forms the basis for the radar algorithms that identify the severe storm features. The basic components of the system are then described. Some details and unique innovations are incorporated in the global survey of operational or near operational use.

Severe weather predictions are divided into severe weather watches and severe weather warnings. In the preceding days, thunderstorm outlooks may be issued. Watches are predictions of the potential of severe weather. They are strategic in nature and fairly coarse in spatial and temporal resolution. They are often issued on a schedule or in conjunction with the public forecast. The expected behaviour is that the public would be aware of the possibility of severe weather and to listen for future updates. Warnings are predictions of the occurrence or imminent occurrence (with high certainty) of severe weather. They are tactical and more specific in location and time. They are also specific in weather element. They are a call to action and to protect one's property and one's self. They are issued and updated as necessary. Fig. 1 shows an overview of the process from the Japanese

Weather advisories are issued if the weather is a concern but not hazardous. Specific types of warning, such as tornado or hail warnings may then be issued and generally after the

The key difference is that the watch is a forecast or very short range forecast service as strategic in nature whereas the warning is a nowcast (based on existing data, precise in time,

(Doswell et al, 1999; Moller, 1978).

This is concluded by a summary.

Meteorological Agency.

**2. Forecasting/Nowcasting/Severe weather warnings** 

more generic severe thunderstorm warning is issued.

location and weather element) and tactical in nature.

1998).

Severe weather is defined here as heavy rains, hail, strong winds including tornadoes and lightning. In the production of warnings, thresholds need to be defined. The thresholds are necessarily locally defined by climatology, local infrastructure and familiarity will dictate what is extreme. Table 1-4 show the warning criteria for Canada circa 1995. Canada is a very big country covering many different weather climatologies and therefore is illustrative of the variation of the severe weather thresholds (see also Galway, 1989). For example, Newfoundland on the east coast of Canada is a very windy location and hence strong winds are a common occurrence and the people have adapted to their environment and therefore it has the highest wind threshold in Canada. Each service needs to define these for them selves.

Fig. 1. The envisioned warning process from outlook to tornado watch. This is typical of the process that is used in most countries providing severe weather warning services. Getting the message out to and understood by the public is very important aspect of the utility of the warning service. Superimposing the warning on television, internet, mobile devices and directed messaging are critical to have the message heard.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 37

50 mm in 12 hrs; sodden ground/bare frozen ground: 25 mm in 24 hrs; spring: 25 mm in 24 hrs; slow moving thunderstorms: 50 mm/3

coastal regions 100 mm in 24 hr and interior of B.C. 25 mm in 24 hr

Quebec 50 mm in 24 hrs or 30 mm in 12 hrs during a spring thaw

hrs or 25 mm/3 hrs if ground is sodden.

Pacific 50 mm in 24 hr except in west Vancouver Island and northern

Quebec 50 km/h with gusts to 90 km/h or with only gust to 90 km/h

Alberta 60 km/h or gusts to 100 km/h except in Lethbridge Region: 70

Warnings for summer severe weather are for extreme or rare events - events that are at the high end of the spectrum of weather. In terms of statistics, rare events do not occur very often (by definition) and so statistical analyses are always suspect due to low numbers. It is difficult to easily demonstrate (using statistics) the efficacy of a warning program (Doswell et al, 1990; Ebert et al 2004). Qualitative analyses or case studies are required to understand the relationship between the provision of warnings and the saving of lives (Sills et al, 2004; Fox et al, 2004). The same applies to determining the efficacy of radar algorithms to the

This has a significant impact on statistics but also on the "cry wolf" syndrome (AMS, 2001; Barnes et al, 2007; Schumacher et al, 2010; Westefeld et al, 2006). An accurate but useless tornado forecast could be by stating that "next year there will be a tornado in the U.S." This statement is a climatological or statistical forecast. It has a very high probability of being true. However, the phenomenon is very small, perhaps 10-20 km in length and 500 m in width and so this particular prediction is not very useful. The information is highly accurate

Mandatory 90 km/h expected over adjacent marine areas; discretionary if gale force winds (63 to 89 km/h) expected over marine areas; discretionary for interior B.C. 65 km/h or gusts of 90

Ontario 60 km/h for 3 hours, or gusts of 90 km/h for 3 hrs

km/h or gusts to 120 km/h.

Prairie 60 km/h and/or gusts to 90 km/h for 1 hr

Table 4. Severe Weather Criteria in Canada: Strong Wind Warning

**Weather Centre Warning Criteria**  Newfoundland 50 mm in 24 hrs Maritimes 50 mm in 24 hrs

Alberta 50 mm in 24 hrs Arctic 50 mm in 24 hr Yukon 40 mm in 24 hr

**Weather Centre Warning Criteria** 

Prairie 80 mm in 24 hrs or 50 mm in 12 hrs

Newfoundland 75 km/h and/or gusts of 100 km/h Maritimes 65 km/h and/or gusts to 90 km/h

Arctic 60 km/h or gusts of 90 km/h

km/h

provision of weather warnings (Joe et al, 2004).

Yukon 60 km/h for 3 hr or gusts to 90 km/h

Table 3. Severe Weather Criteria in Canada: Heavy Rainfall Warning

Ontario

Pacific




Table 2. Severe Weather Criteria in Canada: Severe Thunderstorm Criteria

Doppler Radar Observations – 36 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Wind Strong winds that cause mobility problems and possible damage to vegetation and

Thunderstorm One or more of the following: strong winds causing mobility difficulty, damage to

Severe Weather Presence of tornado(es), damaging hail, heavy rain, strong winds, life and property

structures due to wind and hail, heavy rain that may cause local flooding and lightning

25 mm/h 20 mm No tornado criteria; no tornado warning; may

mention hurricanes in marine warning.

20 mm Tornado, water spout, funnel cloud, windfall;

no tornado warning

forecast or observation.

75 mm in 3 hr 20 mm Tornado or waterspout probable; tornado

km/h 25 mm/hr 12 mm Tornado, water spout, funnel cloud; tornado

Tornado or tornadic waterspout; no tornado warning; will issue hurricane prognostic message and information statements.

No tornado criteria in severe thunderstorm warning; tornado watch issued when confirmed tornadoes threaten to move into region or issued up to 6 hours in advance based on analysis or immediately for severe thunderstorms that indicate potential for becoming tornadic; tornado warning on

warning issued when expected or observed.

Tornado, waterspout or tornado warning when observed or expected or waterspout exists; cold air funnel cloud warning when cold air funnels expected but not tornadoes.

occurrence warning on a confirmed report.

Lightning intensity of 500 strikes in 1 hr over

latitude/longitude; no watches for severe thunderstorms or tornadoes; no tornado warnings; thunderstorm warning issued on a less than severe thunderstorm; will issue hurricane prognostic messages and

Potential of tornado; warning is for thunderstorms; no tornado warning.

an area of 1 degree x 1 degree

information statements.

Heavy Rainfall Heavy or prolonged rainfall accumulating on a scale sufficient to cause

15 mm

20 mm

**Type Description** 

gusts of 90 km/h

gusts of 90 km/h

gusts of 90 km/h

gusts of 90 km/h

km/h

gust of 90 km/h

gust to 90 km/h

gusts of 90 km/h

**Weather** 

Newfoundland

Maritimes

Quebec

Ontario

Alberta

Yukon

Pacific

Prairie <sup>90</sup>

Arctic <sup>90</sup>

structures.

local/widespread flooding.

exposed to real threat, lightning Tornado Public has real potential to be exposed to tornado(es). Table 1. Severe Weather Criteria in Canada: Warning Elements

**Centre Wind Rain Hail Remarks** 

25 mm in 1 hr or 50 mm in 3

25 mm in 1 hr or 50 mm in 12

50 mm/hr for 1 hr; 75 mm for 3

50 mm in 1 hr;

30 mm/h 20 mm

25 mm in 2 hr significant

25 mm in 1 hr 15 mm

Table 2. Severe Weather Criteria in Canada: Severe Thunderstorm Criteria

hail

hrs

hrs

hrs


Table 3. Severe Weather Criteria in Canada: Heavy Rainfall Warning


Table 4. Severe Weather Criteria in Canada: Strong Wind Warning

Warnings for summer severe weather are for extreme or rare events - events that are at the high end of the spectrum of weather. In terms of statistics, rare events do not occur very often (by definition) and so statistical analyses are always suspect due to low numbers. It is difficult to easily demonstrate (using statistics) the efficacy of a warning program (Doswell et al, 1990; Ebert et al 2004). Qualitative analyses or case studies are required to understand the relationship between the provision of warnings and the saving of lives (Sills et al, 2004; Fox et al, 2004). The same applies to determining the efficacy of radar algorithms to the provision of weather warnings (Joe et al, 2004).

This has a significant impact on statistics but also on the "cry wolf" syndrome (AMS, 2001; Barnes et al, 2007; Schumacher et al, 2010; Westefeld et al, 2006). An accurate but useless tornado forecast could be by stating that "next year there will be a tornado in the U.S." This statement is a climatological or statistical forecast. It has a very high probability of being true. However, the phenomenon is very small, perhaps 10-20 km in length and 500 m in width and so this particular prediction is not very useful. The information is highly accurate

Automated Processing of Doppler Radar Data for Severe Weather Warnings 39

al, 2004; Kessler and Wilson, 1971; Lakshmanan et al, 2003; Lakshmanan and Smith, 2009; Lakshmanan et al, 2009; Lenning et al, 1998; Mitchell et al, 1998; Stumpf et al, 1998; Winston, 1998; Witt et al, 1998a, Witt et al, 1998b). The efficacy of the detection depends on the radar scan strategy and quality of the radar (range, azimuth resolution, cycle time, sensitivity, elevation angles, number of elevation tilts, etc (Brown et al, 2000; Heinselman et al, 2008; Lakshmanan et al, 2006; Marshall and Ballantyne, 1975; McLaughlin et al, 2009; Vasiloff,

Watches are based on the concept that the juxtaposition of dynamics, thermodynamics and a mechanism to create upward motion and/or a mechanism to remove inhibition factors exists. This is often called the ingredients approach as one looks to see where the various ingredients come together and that is where severe weather will occur. Historically, this is based on the original Fawbush and Miller Technique (1953) but it has gone through significant evolution (Doswell, 1980, 1982, 1985, 2001; Johns and Doswell, 1992; Moller, 2001;

Fig. 2 shows the morphology of thunderstorms that theoretically develop under different wind shear and convective available potential energy (CAPE) situations (Brooks et al, 1993; Brooks et al, 1994; Markowski et al, 1998b; Weisman and Klemp, 1984; Weisman and Rotunno, 2000). Dynamics is represented by the 0-3 km magnitude of the wind shear. Other height limits may be used depending on the region and local operational usage. The atmospheric structure (low level moisture, mid level dry air, strength of inversions, etc) is important and the thermodynamics is represented by CAPE in this figure. While shear and CAPE are two basic indices that are often used, many other indices are investigated and

Moninger et al, 1991; Monteverdi et al, 2003; Rasmussen, 2003; Weiss et al 1980).

Fig. 2. Thunderstorm type as a function of CAPE and Shear.

2001).

used.

**2.2 Watches** 

but not very precise in terms of location or time. Most, if not all, people would ignore the warning and take the risk. Another form of the "cry wolf" syndrome is where warnings are issued indiscriminately for a very precise time and location and with considerable lead time. However, particularly for rare events (those at the extreme end of a distribution), this is accompanied by a high false alarm rate. If too many false alarms are issued, then these will also be ignored. So, for rare extreme hazardous events, high probability of detection is needed but the false alarms need to be mitigated (Bieringer and Ray, 1996; Black and Ashley, 2011; Glahn 2005; Hoekstra et al, 2011; Polger et al, 1994).

So the issuance of warnings requires a very fine balance of decision-making that takes into account lead time, climatology, societal risk behaviour, social-economic infrastructure, warning service capacity and many other regional, political and societal factors (Baumgart et al, 2008; Dunn, 1990; Hammer and Schmidlin, 2002; Mercer et al, 2009; Schmeits et al, 2008; Westefeld et al, 2006; Wilson et al, 2004). Nowcasts in general are user dependent (Baumgart et al, 2008). Warnings are an extreme kind of nowcasts in which the thresholds apply to a very broad range of users (the public). However, in the future, one can envision very specific warnings or nowcasts issued at lower thresholds that may affect specific users requiring tailored communication techniques and technologies (Keenan et al, 2004; Schumacher et al, 2010).

The wind hazard deserves an extended discussion (Doswell, 2001). There are various kinds of wind hazards that have distinctive life times and spatial features. Straight line winds can originate in synoptic systems or typhoons and are ubiquitous, broad in spatial scale (~100+ km) and extended in duration (~hours/days). Derechos1 are also straight line winds that originate out of mesoscale convective complexes (MCC; Davis et al, 2004; Evans and Doswell, 2001; Przybylinski, 1995; Weisman, 2001). The damaging portion exists at specific locations. They are smaller in size and temporal scale than the previous kind of winds. Gust fronts originate with the downdrafts of MCC's and depending on the nature of the MCC (isolated thunderstorm, multi- cellular, line echo wave pattern, bow echo, pulse storm); the gust front can take on many forms but generally emanate outwards from the MCC (Klingle et al, 1987). They can extend for a long time and there may be extreme winds in portions of the gust front.

The downdrafts can also generate quasi-circular outward flowing winds called downbursts (generic term). If the downbursts are over airports, small in diameter (<4km) and intense (>10 m/s velocity differential) then they are given a very specific term called the microburst (McCarthy et al, 1982; Wilson et al, 1988; Wilson and Wakimoto 2001). It is arbitrarily defined this way in order to be very clear to aviators that they are hazardous and should not be transected. They originate with a descending intense precipitation core and the wind intensity is enhanced by evaporative cooling (Byko et al, 2009). If evaporation is strong, by the time the downburst reaches the surface, there may not be any precipitation associated with it. In this case, the feature is called a dry downburst. If there is precipitation then it is called a wet downburst or microburst as the case may be.

There are algorithmic radar techniques for the identification of all of these severe weather features (Dance and Potts, 2002; Donaldson and Desrochers, 1990; Johnson et al, 1998; Joe et

<sup>1</sup>It is beyond the scope of this contribution to illustrate the various severe hazards in detail – see references for fourther information.

al, 2004; Kessler and Wilson, 1971; Lakshmanan et al, 2003; Lakshmanan and Smith, 2009; Lakshmanan et al, 2009; Lenning et al, 1998; Mitchell et al, 1998; Stumpf et al, 1998; Winston, 1998; Witt et al, 1998a, Witt et al, 1998b). The efficacy of the detection depends on the radar scan strategy and quality of the radar (range, azimuth resolution, cycle time, sensitivity, elevation angles, number of elevation tilts, etc (Brown et al, 2000; Heinselman et al, 2008; Lakshmanan et al, 2006; Marshall and Ballantyne, 1975; McLaughlin et al, 2009; Vasiloff, 2001).

#### **2.2 Watches**

Doppler Radar Observations – 38 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

but not very precise in terms of location or time. Most, if not all, people would ignore the warning and take the risk. Another form of the "cry wolf" syndrome is where warnings are issued indiscriminately for a very precise time and location and with considerable lead time. However, particularly for rare events (those at the extreme end of a distribution), this is accompanied by a high false alarm rate. If too many false alarms are issued, then these will also be ignored. So, for rare extreme hazardous events, high probability of detection is needed but the false alarms need to be mitigated (Bieringer and Ray, 1996; Black and

So the issuance of warnings requires a very fine balance of decision-making that takes into account lead time, climatology, societal risk behaviour, social-economic infrastructure, warning service capacity and many other regional, political and societal factors (Baumgart et al, 2008; Dunn, 1990; Hammer and Schmidlin, 2002; Mercer et al, 2009; Schmeits et al, 2008; Westefeld et al, 2006; Wilson et al, 2004). Nowcasts in general are user dependent (Baumgart et al, 2008). Warnings are an extreme kind of nowcasts in which the thresholds apply to a very broad range of users (the public). However, in the future, one can envision very specific warnings or nowcasts issued at lower thresholds that may affect specific users requiring tailored communication techniques and technologies (Keenan et al, 2004; Schumacher et al, 2010).

The wind hazard deserves an extended discussion (Doswell, 2001). There are various kinds of wind hazards that have distinctive life times and spatial features. Straight line winds can originate in synoptic systems or typhoons and are ubiquitous, broad in spatial scale (~100+ km) and extended in duration (~hours/days). Derechos1 are also straight line winds that originate out of mesoscale convective complexes (MCC; Davis et al, 2004; Evans and Doswell, 2001; Przybylinski, 1995; Weisman, 2001). The damaging portion exists at specific locations. They are smaller in size and temporal scale than the previous kind of winds. Gust fronts originate with the downdrafts of MCC's and depending on the nature of the MCC (isolated thunderstorm, multi- cellular, line echo wave pattern, bow echo, pulse storm); the gust front can take on many forms but generally emanate outwards from the MCC (Klingle et al, 1987). They can extend for a long time and there may be extreme winds in portions of

The downdrafts can also generate quasi-circular outward flowing winds called downbursts (generic term). If the downbursts are over airports, small in diameter (<4km) and intense (>10 m/s velocity differential) then they are given a very specific term called the microburst (McCarthy et al, 1982; Wilson et al, 1988; Wilson and Wakimoto 2001). It is arbitrarily defined this way in order to be very clear to aviators that they are hazardous and should not be transected. They originate with a descending intense precipitation core and the wind intensity is enhanced by evaporative cooling (Byko et al, 2009). If evaporation is strong, by the time the downburst reaches the surface, there may not be any precipitation associated with it. In this case, the feature is called a dry downburst. If there is precipitation then it is

There are algorithmic radar techniques for the identification of all of these severe weather features (Dance and Potts, 2002; Donaldson and Desrochers, 1990; Johnson et al, 1998; Joe et

1It is beyond the scope of this contribution to illustrate the various severe hazards in detail –

Ashley, 2011; Glahn 2005; Hoekstra et al, 2011; Polger et al, 1994).

called a wet downburst or microburst as the case may be.

see references for fourther information.

the gust front.

Watches are based on the concept that the juxtaposition of dynamics, thermodynamics and a mechanism to create upward motion and/or a mechanism to remove inhibition factors exists. This is often called the ingredients approach as one looks to see where the various ingredients come together and that is where severe weather will occur. Historically, this is based on the original Fawbush and Miller Technique (1953) but it has gone through significant evolution (Doswell, 1980, 1982, 1985, 2001; Johns and Doswell, 1992; Moller, 2001; Moninger et al, 1991; Monteverdi et al, 2003; Rasmussen, 2003; Weiss et al 1980).

Fig. 2 shows the morphology of thunderstorms that theoretically develop under different wind shear and convective available potential energy (CAPE) situations (Brooks et al, 1993; Brooks et al, 1994; Markowski et al, 1998b; Weisman and Klemp, 1984; Weisman and Rotunno, 2000). Dynamics is represented by the 0-3 km magnitude of the wind shear. Other height limits may be used depending on the region and local operational usage. The atmospheric structure (low level moisture, mid level dry air, strength of inversions, etc) is important and the thermodynamics is represented by CAPE in this figure. While shear and CAPE are two basic indices that are often used, many other indices are investigated and used.


Fig. 2. Thunderstorm type as a function of CAPE and Shear.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 41

This is a highly condensed version of the technique and there are many subtleties and morphological pathways as storms evolve. Severe storms begin as non-severe storms and algorithm developers and forecasters try very hard to extend lead times by trying to identify the severity of the future storm as early as possible. Note also that it is often in the collapsing stages of the storm (indicated by collapsing echo top or a descending core) when

Fig. 3. It is obligatory to show radar images of severe convective storms. Linear convective storms are show in (a) and (b) whereas isolated thunderstorms are shown in (c) and (d). Except for (d), reflectivity and radial velocity images are shown together. Fig. 3a shows double squall lines (1) with embedded cells and mesocyclones (2). (3) shows a shear line associated with a cold frontal passage, so the mesocyclones are pre-frontal and likely to have formed on a previously formed outflow boundary. Fig 3b shows embedded

thunderstorms on a bow echo. Note the boundaries (5) ahead of the bow echo. (8) shows a

Not discussed here is the identification of the initiation phase of convective weather (Wilson et al, 1998). Significant progress has been made in the warning of air mass thunderstorms. In the past, these were considered random and unforecastable. Wilson et al (1998) demonstrate that they are not random but form on boundaries (see the fine lines on Fig. 3c). Roberts et al

meso-scale intense straight line wind (nearing 48 m/s). Fig 3c show an isolated thunderstorm with a mesocyclone (4). Boundaries (5) can be seen and to be associated with the entire mesoscale convective complex and not just one individual cell. Fig. 3d shows the splitting of an isolated tornado producing storm. The yellow shading is the 40 dBZ contour. Often, cell identification thresholds are set lower (30 or 35 dBZ) in an attempt to get earlier cell detections but this demonstrates that this results in detecting

high low level reflectivity core displace towards the updraft

the severe weather reaches the surface (see Fig. 4).

deviant motion (right or left mover, depending on hemisphere)

concavity (hook echo)

different storm structures.

rotation

Watches are generally very broad in spatial nature due to the spatial density of the observations (soundings and surface observation), and models which are based on the observations, which is very sparse. The resulting analysis of severe weather potential is therefore necessarily broad. The situation is also very fluid and there can be many local factors such a topography or land-water boundaries or rural- urban differences, to name just a few (King et al, 2003; Wasula et al, 2002; Wilson et al, 2010). What are very difficult to identify are potential mechanisms to create upward motion (the trigger) or to overcome the convective inhibition (break the cap). On a synoptic scale, this could be lift generated by cold or warm fronts but on a smaller scale, they can be created by dry lines, thunderstorm outflows, lake-land breezes, urban hot spots, etc. Often they are very low level and therefore hard to observe. So forecasts of severe weather are indications that the potential ingredients exist. They are therefore very broad and strategic in nature.

#### **2.3 Warnings**

Weather warnings are issued when there is very high likelihood of severe weather. A broadly worded severe weather thunderstorm warning is most often first issued. If appropriate, it is followed by a more specific warning on a particular thunderstorm and specific severe weather element. This approach is not universal but is dependent on the climatology of severe weather and the level of the warning service that can or has been decided to provide. An important aspect of the detail of the warning is the ability to use the information by the end-user, which is often the public. The public may not know how to react. Given the "cry wolf" syndrome, there needs to be an education process (see Fig. 1). Often, a disaster is needed to get the attention of the public but the significance of the event can be lost in a few short years. Civil emergency services and hydro utilities can plan their post- event remediation actions/locations based on the warning areas and products. So, there can be many variations and underlying philosophies for the provision of warning services. This partially drives the design of the radar processing, visualization and warning preparations systems. It is one thing if severe weather is prevalent and there is a dedicated forecaster for a small area and the public is well attuned to the severity of the weather and have tornado shelters (Andra et al, 2002). It is another thing if the forecaster has to cover several radars and dealing with ill informed users (Leduc et al, 2002; Schumacher et al, 2010).

#### **3. Identifying severe thunderstorms**

#### **3.1 Lemon technique**

The specificity of the severe thunderstorm warning is primarily based on a radar feature identification technique attributed to Lemon (1977, 1980) and is based on a morphological approach (Moller et al, 1994). It is beyond the scope of this contribution to present or describe the various types of thunderstorms (Fig. 3 shows a small sample). As mentioned earlier, precipitation and precipitation cores form aloft and then descend.

The following features need to be identified:


Doppler Radar Observations – 40 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Watches are generally very broad in spatial nature due to the spatial density of the observations (soundings and surface observation), and models which are based on the observations, which is very sparse. The resulting analysis of severe weather potential is therefore necessarily broad. The situation is also very fluid and there can be many local factors such a topography or land-water boundaries or rural- urban differences, to name just a few (King et al, 2003; Wasula et al, 2002; Wilson et al, 2010). What are very difficult to identify are potential mechanisms to create upward motion (the trigger) or to overcome the convective inhibition (break the cap). On a synoptic scale, this could be lift generated by cold or warm fronts but on a smaller scale, they can be created by dry lines, thunderstorm outflows, lake-land breezes, urban hot spots, etc. Often they are very low level and therefore hard to observe. So forecasts of severe weather are indications that the potential ingredients

Weather warnings are issued when there is very high likelihood of severe weather. A broadly worded severe weather thunderstorm warning is most often first issued. If appropriate, it is followed by a more specific warning on a particular thunderstorm and specific severe weather element. This approach is not universal but is dependent on the climatology of severe weather and the level of the warning service that can or has been decided to provide. An important aspect of the detail of the warning is the ability to use the information by the end-user, which is often the public. The public may not know how to react. Given the "cry wolf" syndrome, there needs to be an education process (see Fig. 1). Often, a disaster is needed to get the attention of the public but the significance of the event can be lost in a few short years. Civil emergency services and hydro utilities can plan their post- event remediation actions/locations based on the warning areas and products. So, there can be many variations and underlying philosophies for the provision of warning services. This partially drives the design of the radar processing, visualization and warning preparations systems. It is one thing if severe weather is prevalent and there is a dedicated forecaster for a small area and the public is well attuned to the severity of the weather and have tornado shelters (Andra et al, 2002). It is another thing if the forecaster has to cover several radars and dealing with ill informed users (Leduc et al, 2002; Schumacher et al,

The specificity of the severe thunderstorm warning is primarily based on a radar feature identification technique attributed to Lemon (1977, 1980) and is based on a morphological approach (Moller et al, 1994). It is beyond the scope of this contribution to present or describe the various types of thunderstorms (Fig. 3 shows a small sample). As mentioned

earlier, precipitation and precipitation cores form aloft and then descend.

 tilted updraft, and/or weak or bound weak echo region displaced echo top relative to the low-mid level core

exist. They are therefore very broad and strategic in nature.

**2.3 Warnings** 

2010).

**3.1 Lemon technique** 

**3. Identifying severe thunderstorms** 

The following features need to be identified:

strong reflectivity gradients

This is a highly condensed version of the technique and there are many subtleties and morphological pathways as storms evolve. Severe storms begin as non-severe storms and algorithm developers and forecasters try very hard to extend lead times by trying to identify the severity of the future storm as early as possible. Note also that it is often in the collapsing stages of the storm (indicated by collapsing echo top or a descending core) when the severe weather reaches the surface (see Fig. 4).

Fig. 3. It is obligatory to show radar images of severe convective storms. Linear convective storms are show in (a) and (b) whereas isolated thunderstorms are shown in (c) and (d). Except for (d), reflectivity and radial velocity images are shown together. Fig. 3a shows double squall lines (1) with embedded cells and mesocyclones (2). (3) shows a shear line associated with a cold frontal passage, so the mesocyclones are pre-frontal and likely to have formed on a previously formed outflow boundary. Fig 3b shows embedded thunderstorms on a bow echo. Note the boundaries (5) ahead of the bow echo. (8) shows a meso-scale intense straight line wind (nearing 48 m/s). Fig 3c show an isolated thunderstorm with a mesocyclone (4). Boundaries (5) can be seen and to be associated with the entire mesoscale convective complex and not just one individual cell. Fig. 3d shows the splitting of an isolated tornado producing storm. The yellow shading is the 40 dBZ contour. Often, cell identification thresholds are set lower (30 or 35 dBZ) in an attempt to get earlier cell detections but this demonstrates that this results in detecting different storm structures.

Not discussed here is the identification of the initiation phase of convective weather (Wilson et al, 1998). Significant progress has been made in the warning of air mass thunderstorms. In the past, these were considered random and unforecastable. Wilson et al (1998) demonstrate that they are not random but form on boundaries (see the fine lines on Fig. 3c). Roberts et al

Automated Processing of Doppler Radar Data for Severe Weather Warnings 43

The Lemon technique implies that volume scanning radars are needed since many of the critical features originate aloft (see Fig. 4). Both, high data quality (Joe, 2009; Lakshmanan et al, 2007; Lakshmanan et al 2010; Lakshmanan et al, 2011) and rapid update cycles for the fast evolving thunderstorms (Crum and Alberty, 1993; Marshall and Ballantyne, 1975). In order to detect low level "clear air" boundaries important for the identification of convective initiation, high sensitivity is critical. Research literature often shows many examples of extensive clear air radar echoes that are not operationally observed. The operational question is whether it is a radar sensitivity issue or the lack of insect targets (the clear air targets have been identified as insects through dual-polarization signatures). Extensive clear air echoes are commonly reported observed on the WSR-88D and primarily in certain parts of the United States (Wilson et al, 1998). Table 5 shows the sensitivity of a small sample of radars including the WSR-88D, WSR-98D (S Band radars) and three C Band radars, one of which is a low powered (8 kW), travelling wave tube (TWT) solid sate pulse compression radar (Joe, 2009; Bech et al, 2004; O'Hora and Bech, 2007). In units of dBZ, the sensitivity is a function of range. Fifty kilometer range is arbitrarily chosen to compare the radar sensitivities. The table shows that all these state of the art radars can have comparable sensitivity. Therefore, the apparent lack of clear air echoes is due to the lack of local clear air radar targets and not due to radar sensitivity or wavelength (for example, see May et al, 2004). In addition, due to the dependendence of the backscatter on the inverse frequency

squared, C Band radars should observe insects better than S Band radars.

**Radar MDS at 50 km** 

Perhaps the most important consideration in the design of the operational radar processing, visualization and decision-making is the underlying philosophy of the weather service, existing systems and, of course, the capabilities and resources available (Joe et al 2002). In many cases, the warning service requirements are driven not only by the scientific capabilities or the needs but also by the political, societal and economic norms. Often a warning service is an ethical and moral reaction by NHMS's to a damaging event or events and hence it is also a political reaction by governments. This varies considerably from place to place. These requirements are tempered by existing observational infrastructure. Are

WSR-98D (TJ) -6.0dBZ WSR-98D (BJ) -5.5 dBZ WSR-88D (KTLX) -7.5 dBZ WSR-88D (KLCH) -8.5 dBZ WKR Conventional C Band (2 μs pulse) -11.0 dBZ WKR Conventional C Band (0.5 μs pulse) -5.0 dBZ CDV TWT (8kW) C Band (1 μs pulse) 6.0 dBZ CDV TWT (8kW) C Band (5 μs pulse) -7.0 dBZ CDV TWT (8kW) C Band (NLFM 30 μs pulse) -6.0 dBZ CDV TWT (8kW) C Band (NLFM 40 μs pulse) -9.0 dBZ INM Conventional C Band (2 μs pulse) -9.0 dBZ

Table 5. Minimum Detectable Signal of Various Radars

**4. Forecast process and system design** 

**3.3 Radar dependencies** 

(2006) discuss the tools to help bridge the convective initiation phase to the severe phase of thunderstorm nowcasting. The science or theory of thunderstorm is still evolving (Brooks et al, 1994; Brunner et al, 2007; Markowski, 2002; Rasmussen et al, 1994; Weisman and Rotunno, 2004).

Fig. 4. A time-height diagram through the core of a long lived thunderstorm with a mesocyclone. The "nose" on the left side of the shading indicates the precipitation and the mesocyclone originate at mid-levels of the atmosphere and develop vertically up and down. In the collapse phase of the storm or mesocyclone top, the severe weather reaches the ground (adapted from Burgess et al, 1993; Lemon and Doswell, 1979).

#### **3.2 Other data sets**

This contribution focuses on radar and its use in the preparation of warnings. In fact, all sources of observations and information are used to validate and enforce the conceptual models used to produce the warnings. Satellite imagery, such as provided by MSG and the future GOES-R, will be able to provide 5 minute updates over limited areas. Lightning networks are now prevalent and often used as surrogates for radar data where none is available. They also directly observe the lightning hazard (Branick et al, 1992; Gatlin et al, 2010; Goodman et al, 1988; Knupp et al, 2003; Lang et al, 2004; Schultz et al, 2011;). Even though a single lightning flash can cause serious harm or death, table 2 indicates that, in Canada, a propensity of lightning strikes is needed before a lightning warning will be issued. Surface wind reports can be also used. However, a tornado or a microburst is relatively small and most operational networks are too sparse to effectively sample the atmosphere for such a small feature. At some airports, a dense network of anemometers is established for this specific problem (Wilson et al, 1998). An important data set are eye witness reports (Doswell et al, 1999; Moller 1978; Smith, 1999). In the past, eye witness reports were required before a tornado warning would be issued. This made all tornado warnings "late" with negative lead times. This was done in order not to "cry wolf" and "alarm the public". An emerging source of information is the use of high resolution NWP (Hoekstra et al 2011; Li, 2010; Stensrud et al 2009). While phase errors exist (time and location of the thunderstorm), the models appear to be able to capture the morphology of the storm (see Fig. 2). While radar is the core observation system for severe weather warnings at the convective scale, these are not available everywhere. A warning service that does not include radar has yet to be effectively demonstrated.

#### **3.3 Radar dependencies**

Doppler Radar Observations – 42 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

(2006) discuss the tools to help bridge the convective initiation phase to the severe phase of thunderstorm nowcasting. The science or theory of thunderstorm is still evolving (Brooks et al, 1994; Brunner et al, 2007; Markowski, 2002; Rasmussen et al, 1994; Weisman and

Fig. 4. A time-height diagram through the core of a long lived thunderstorm with a mesocyclone. The "nose" on the left side of the shading indicates the precipitation and the mesocyclone originate at mid-levels of the atmosphere and develop vertically up and down. In the collapse phase of the storm or mesocyclone top, the severe weather reaches the

This contribution focuses on radar and its use in the preparation of warnings. In fact, all sources of observations and information are used to validate and enforce the conceptual models used to produce the warnings. Satellite imagery, such as provided by MSG and the future GOES-R, will be able to provide 5 minute updates over limited areas. Lightning networks are now prevalent and often used as surrogates for radar data where none is available. They also directly observe the lightning hazard (Branick et al, 1992; Gatlin et al, 2010; Goodman et al, 1988; Knupp et al, 2003; Lang et al, 2004; Schultz et al, 2011;). Even though a single lightning flash can cause serious harm or death, table 2 indicates that, in Canada, a propensity of lightning strikes is needed before a lightning warning will be issued. Surface wind reports can be also used. However, a tornado or a microburst is relatively small and most operational networks are too sparse to effectively sample the atmosphere for such a small feature. At some airports, a dense network of anemometers is established for this specific problem (Wilson et al, 1998). An important data set are eye witness reports (Doswell et al, 1999; Moller 1978; Smith, 1999). In the past, eye witness reports were required before a tornado warning would be issued. This made all tornado warnings "late" with negative lead times. This was done in order not to "cry wolf" and "alarm the public". An emerging source of information is the use of high resolution NWP (Hoekstra et al 2011; Li, 2010; Stensrud et al 2009). While phase errors exist (time and location of the thunderstorm), the models appear to be able to capture the morphology of the storm (see Fig. 2). While radar is the core observation system for severe weather warnings at the convective scale, these are not available everywhere. A warning service that

ground (adapted from Burgess et al, 1993; Lemon and Doswell, 1979).

does not include radar has yet to be effectively demonstrated.

Rotunno, 2004).

**3.2 Other data sets** 

The Lemon technique implies that volume scanning radars are needed since many of the critical features originate aloft (see Fig. 4). Both, high data quality (Joe, 2009; Lakshmanan et al, 2007; Lakshmanan et al 2010; Lakshmanan et al, 2011) and rapid update cycles for the fast evolving thunderstorms (Crum and Alberty, 1993; Marshall and Ballantyne, 1975). In order to detect low level "clear air" boundaries important for the identification of convective initiation, high sensitivity is critical. Research literature often shows many examples of extensive clear air radar echoes that are not operationally observed. The operational question is whether it is a radar sensitivity issue or the lack of insect targets (the clear air targets have been identified as insects through dual-polarization signatures). Extensive clear air echoes are commonly reported observed on the WSR-88D and primarily in certain parts of the United States (Wilson et al, 1998). Table 5 shows the sensitivity of a small sample of radars including the WSR-88D, WSR-98D (S Band radars) and three C Band radars, one of which is a low powered (8 kW), travelling wave tube (TWT) solid sate pulse compression radar (Joe, 2009; Bech et al, 2004; O'Hora and Bech, 2007). In units of dBZ, the sensitivity is a function of range. Fifty kilometer range is arbitrarily chosen to compare the radar sensitivities. The table shows that all these state of the art radars can have comparable sensitivity. Therefore, the apparent lack of clear air echoes is due to the lack of local clear air radar targets and not due to radar sensitivity or wavelength (for example, see May et al, 2004). In addition, due to the dependendence of the backscatter on the inverse frequency squared, C Band radars should observe insects better than S Band radars.


Table 5. Minimum Detectable Signal of Various Radars

#### **4. Forecast process and system design**

Perhaps the most important consideration in the design of the operational radar processing, visualization and decision-making is the underlying philosophy of the weather service, existing systems and, of course, the capabilities and resources available (Joe et al 2002). In many cases, the warning service requirements are driven not only by the scientific capabilities or the needs but also by the political, societal and economic norms. Often a warning service is an ethical and moral reaction by NHMS's to a damaging event or events and hence it is also a political reaction by governments. This varies considerably from place to place. These requirements are tempered by existing observational infrastructure. Are

Automated Processing of Doppler Radar Data for Severe Weather Warnings 45

Fig. 5. The flow of the radar data to warning product is much the same in all systems. But

Radar processing systems need quality controlled data. This can occur in a separate and independent process. In some cases, it is part of the adjustments and corrections that need to be made. Before the severe weather processing occurs (stage B in Fig. 5), it is assumed that the data is free of anomalous propagation, ground clutter and biases in power are adjusted. Second trip echoes and range folded may still be in the Doppler data (Joe 2009; Lakshmanan

In high shear environments the assumption that the radial velocities within a range volume are uniform may not be satisfied (Holleman and Beekhuis 2003; Joe and May 2003). Fig 6ab shows a simulated Doppler velocity spectrum (based on an example in Doviak and Zrnic, 1984) of a tornado contained within a single range volume. The spectrum is bi-modal and the peaks at located at the speed of the radial components of the tornado. Normally it is unimodal and Gaussian in shape. Fig. 6cd show the measured spectrum given two different Nyquist limits. The spectrum is aliased and overlaps with itself. The smaller the Nyquist limit, the greater the overlap. In highly sheared regions, the velocity data is noisy and can be non-sensical. The chapter on quantitative precipitation estimation addresses many of the

the contents of each stage can be different. Except for one system described in this contribution, all the others require human decision-making at stage C before the warning product is issued to the public. In the case of KONRAD (see section 6.10), the product goes

mainly to "sophisticated" users.

et al, 2010; Lakshmanan et al, 2011).

**5.1 Data quality** 

quality control issues.

there functioning radars or other data sources? Is there the capacity to design or even adopt a radar processing system? Is there the knowledge and capacity to interpret the data products to make effective warning decisions and issue warnings? And is there a way to reach the end-user in a timely fashion? It should not be forgotten that the end-user must be educated on the meaning of the warning and on how to react appropriately. Is there sufficient budget to develop a warning system? What is risk is acceptable? What level is the moral outrage?

An often overlooked design issue is the organization of the weather service. Warnings are provided for small areas (scale of the weather feature) in order to mitigate the "cry wolf" syndrome to be effective (Barnes et al, 2007; Hammer and Schmidlin, 2002). The critical issue is the capacity to provide the attention to the detail given the totality of the forecast responsibilities. The system design will be quite different if there are many forecast offices and few radars (one to one) compared to few offices and many radars (one office to ten radars as in Canada).

Of course, an overarching issue is the climatology of severe weather which ultimately is the core issue. For many countries, convective weather may occur year round and some only for the summer season. In the latter case, a design question is to determine the use case for the shoulder season where severe weather may occur unexpectedly and the warning service is seasonal.

Severe weather forecasting requires a unique forecasting skill set. In synoptic forecasting (for 12 hours and beyond), the forecaster compares current observations to numerical weather prediction models to evaluate the appropriateness of the model or to develop a conceptual model of the weather for the creation of the public forecast product (Doswell 2004). The product is usually produced on a fixed schedule. In severe weather forecasting, the observations need to be timely; there is urgency in the interpretation and the generation of the warning product. It is a "short fused" situation. These require different personality types and this also drives the design considerations. In order to mitigate the "cry wolf" situation while maintaining high probability of detection, a dedicated and separate warning forecaster function is required to be able to address the immediacy issues of the warning service. These are just some of the design considerations for a radar processing and visualization system and the forecast process for the provision of severe weather warnings. Forecast process refers to all components of the transformation of the data or observations into information used for decision- making and warning service production. It includes both the human and their tools and is often referred to as the man-machine mix. Given all the degrees of freedom in the chain, there are different models of the forecast process.

In the next section, a global survey (necessarily incomplete) is presented that will briefly examine the operational or near-operational systems that have been developed. Many have commonalities and only the underlying unique aspects will be highlighted.

#### **5. Components of a basic system**

In this section, the basic components or issues of severe weather radar processing/visualization are briefly discussed and a block diagram is provide in Fig. 5. The benefits of different radar types are discussed elsewhere (WMO, 2008).

Fig. 5. The flow of the radar data to warning product is much the same in all systems. But the contents of each stage can be different. Except for one system described in this contribution, all the others require human decision-making at stage C before the warning product is issued to the public. In the case of KONRAD (see section 6.10), the product goes mainly to "sophisticated" users.

#### **5.1 Data quality**

Doppler Radar Observations – 44 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

there functioning radars or other data sources? Is there the capacity to design or even adopt a radar processing system? Is there the knowledge and capacity to interpret the data products to make effective warning decisions and issue warnings? And is there a way to reach the end-user in a timely fashion? It should not be forgotten that the end-user must be educated on the meaning of the warning and on how to react appropriately. Is there sufficient budget to develop a warning system? What is risk is acceptable? What level is the

An often overlooked design issue is the organization of the weather service. Warnings are provided for small areas (scale of the weather feature) in order to mitigate the "cry wolf" syndrome to be effective (Barnes et al, 2007; Hammer and Schmidlin, 2002). The critical issue is the capacity to provide the attention to the detail given the totality of the forecast responsibilities. The system design will be quite different if there are many forecast offices and few radars (one to one) compared to few offices and many radars (one office to ten

Of course, an overarching issue is the climatology of severe weather which ultimately is the core issue. For many countries, convective weather may occur year round and some only for the summer season. In the latter case, a design question is to determine the use case for the shoulder season where severe weather may occur unexpectedly and the warning service is

Severe weather forecasting requires a unique forecasting skill set. In synoptic forecasting (for 12 hours and beyond), the forecaster compares current observations to numerical weather prediction models to evaluate the appropriateness of the model or to develop a conceptual model of the weather for the creation of the public forecast product (Doswell 2004). The product is usually produced on a fixed schedule. In severe weather forecasting, the observations need to be timely; there is urgency in the interpretation and the generation of the warning product. It is a "short fused" situation. These require different personality types and this also drives the design considerations. In order to mitigate the "cry wolf" situation while maintaining high probability of detection, a dedicated and separate warning forecaster function is required to be able to address the immediacy issues of the warning service. These are just some of the design considerations for a radar processing and visualization system and the forecast process for the provision of severe weather warnings. Forecast process refers to all components of the transformation of the data or observations into information used for decision- making and warning service production. It includes both the human and their tools and is often referred to as the man-machine mix. Given all the

degrees of freedom in the chain, there are different models of the forecast process.

commonalities and only the underlying unique aspects will be highlighted.

benefits of different radar types are discussed elsewhere (WMO, 2008).

**5. Components of a basic system** 

In the next section, a global survey (necessarily incomplete) is presented that will briefly examine the operational or near-operational systems that have been developed. Many have

In this section, the basic components or issues of severe weather radar processing/visualization are briefly discussed and a block diagram is provide in Fig. 5. The

moral outrage?

radars as in Canada).

seasonal.

Radar processing systems need quality controlled data. This can occur in a separate and independent process. In some cases, it is part of the adjustments and corrections that need to be made. Before the severe weather processing occurs (stage B in Fig. 5), it is assumed that the data is free of anomalous propagation, ground clutter and biases in power are adjusted. Second trip echoes and range folded may still be in the Doppler data (Joe 2009; Lakshmanan et al, 2010; Lakshmanan et al, 2011).

In high shear environments the assumption that the radial velocities within a range volume are uniform may not be satisfied (Holleman and Beekhuis 2003; Joe and May 2003). Fig 6ab shows a simulated Doppler velocity spectrum (based on an example in Doviak and Zrnic, 1984) of a tornado contained within a single range volume. The spectrum is bi-modal and the peaks at located at the speed of the radial components of the tornado. Normally it is unimodal and Gaussian in shape. Fig. 6cd show the measured spectrum given two different Nyquist limits. The spectrum is aliased and overlaps with itself. The smaller the Nyquist limit, the greater the overlap. In highly sheared regions, the velocity data is noisy and can be non-sensical. The chapter on quantitative precipitation estimation addresses many of the quality control issues.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 47

et al, 1980). Crane (1979) developed the cell identification techniques based on peak detection. These systems left a legacy for the development of the WSR88D algorithms. McGill developed SHARP (Bellon and Austin, 1978) for precipitation nowcasting and developed the crosscorrelation method for echo tracking which is still used today. It did not specifically address

Many of the innovations for the reflectivity-only algorithms of RADAP-II were adopted and significantly enhanced for the WSR88D (Crum and Alberty, 1993; Kitzmiller et al, 1995). Doppler algorithms were developed for mesocyclone and gust front detection (Hermes et al 1993; Uyeda and Zrnic, 1986; Zrnic et al, 1985). Considerable effort has been expended to improve upon these initial efforts. A search of the American Meteorological Society journal publications will illustrate that. Initially, the output from the WSR88D Radar Product Generator was displayed on a dedicated radar-only visualization system called the Principal User Product (PUP) display for the forecaster and later the forecaster workstation called AWIPS was used. This integrated all the data and products that the forecaster needed. WSR-88D algorithms were later deployed on the WSR-98D radars made by MetStar and used in

A fundamental question arose as to the role of automated guidance products versus manual interpretation (Andra et al, 2002). It is clear that automated generated products are for guidance and it should not be mistakenly interpreted that warnings were automatically generated and issued without an intervening well trained decision-maker. Initially, there was an extensive radar training program for forecasters, up to 6 weeks for specialists. Clearly, the expectation was that an expert level of training was needed to interpret Doppler radar data for severe weather warnings. This was re-enforced by the work of Pliske et al (1997) who analyzed how to achieve the expected benefits of a modernization program. This resulted in the development of an on-going training program for decision-making at the appropriately named, Warning Decision Training Branch of the National Severe Storms Laboratory. Professionally trained instructors on cognitive principles interactively have the skills to tailor the material to the appropriate knowledge level, abilities and learning styles

TITAN (Thunderstorm identification, tracking and nowcasting) was first developed in South Africa and then later at NCAR for support of weather modification programs. Dixon and Weiner (1993) described a simple but brilliant threshold technique for the identification of thunderstorm cell cores. This simplified the peak detection techniques of the Crane (1979) technique as the latter identified many weak cells and challenged the computing power of the day. It also described a methodology for tracking. It could be argued that this is the most widely used system in the world. It is freely available and requires some expertise to implement. It is used extensively in research environments (Lei et al, 2009). It is a stand alone system and integrating it into an operational environment has been done but there are capacity and support issues to consider. For example, it is used at the South African

severe weather algorithms, which is the focus of this contribution.

**6.2 WSR-88D, U.S.A., WSR-98D, China** 

China, Romania, India, Korea and other places.

of the student. It is a model for professional training.

**6.3 TITAN – NCAR** 

Weather Service.

Fig. 6. (a) Doppler velocity spectra at different ranges made with a radar with a very large Nyquist interval. The arrow points to the tornado. The spectrum is bi-model. (b) a simulation of the spectra. (c) and (d) are simulated measured spectra made with different Nyquist intervals. The spectrum overlaps and is aliased. In (c) the spectra is bi-modal still, would produce a radial velocity estimate near zero with a very broad variance. In (d), the mean is still zero, the spectra is uni-modal with a smaller variance.

#### **6. Global survey**

This section provides a necessarily brief global survey of various convective weather radar processing systems. In fact, there are only a few NHMS' that actually provide a severe weather warning service. The systems are presented in a sequence that approximately matches when they were developed and the reader can follow the progression of the system and philosophical developments.

#### **6.1 RADAP – II, U.S.A.**

The first radar processing system for severe weather was RADAP-II and it was built in the 1970's (Winston and Ruthi, 1986) and it followed from D/RADEX (Breidenbach et al, 1995; Saffle, 1976) within the National Weather Service. They used VIL (vertically integrated liquid water) and a significant innovation was the introduction of a SWP (Severe Weather Probability) product. They were using probabilistic and uncertainty concepts then! There were many innovations with RADAP-II but its deployment was curtailed due to the development of the Doppler upgrade called the WSR-88D (Crum and Alberty, 1993; Lemon et al, 1977; Wilson et al, 1980). Crane (1979) developed the cell identification techniques based on peak detection. These systems left a legacy for the development of the WSR88D algorithms. McGill developed SHARP (Bellon and Austin, 1978) for precipitation nowcasting and developed the crosscorrelation method for echo tracking which is still used today. It did not specifically address severe weather algorithms, which is the focus of this contribution.

#### **6.2 WSR-88D, U.S.A., WSR-98D, China**

Doppler Radar Observations – 46 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Fig. 6. (a) Doppler velocity spectra at different ranges made with a radar with a very large

This section provides a necessarily brief global survey of various convective weather radar processing systems. In fact, there are only a few NHMS' that actually provide a severe weather warning service. The systems are presented in a sequence that approximately matches when they were developed and the reader can follow the progression of the system

The first radar processing system for severe weather was RADAP-II and it was built in the 1970's (Winston and Ruthi, 1986) and it followed from D/RADEX (Breidenbach et al, 1995; Saffle, 1976) within the National Weather Service. They used VIL (vertically integrated liquid water) and a significant innovation was the introduction of a SWP (Severe Weather Probability) product. They were using probabilistic and uncertainty concepts then! There were many innovations with RADAP-II but its deployment was curtailed due to the development of the Doppler upgrade called the WSR-88D (Crum and Alberty, 1993; Lemon et al, 1977; Wilson

Nyquist interval. The arrow points to the tornado. The spectrum is bi-model. (b) a simulation of the spectra. (c) and (d) are simulated measured spectra made with different Nyquist intervals. The spectrum overlaps and is aliased. In (c) the spectra is bi-modal still, would produce a radial velocity estimate near zero with a very broad variance. In (d), the

mean is still zero, the spectra is uni-modal with a smaller variance.

**6. Global survey** 

and philosophical developments.

**6.1 RADAP – II, U.S.A.** 

Many of the innovations for the reflectivity-only algorithms of RADAP-II were adopted and significantly enhanced for the WSR88D (Crum and Alberty, 1993; Kitzmiller et al, 1995). Doppler algorithms were developed for mesocyclone and gust front detection (Hermes et al 1993; Uyeda and Zrnic, 1986; Zrnic et al, 1985). Considerable effort has been expended to improve upon these initial efforts. A search of the American Meteorological Society journal publications will illustrate that. Initially, the output from the WSR88D Radar Product Generator was displayed on a dedicated radar-only visualization system called the Principal User Product (PUP) display for the forecaster and later the forecaster workstation called AWIPS was used. This integrated all the data and products that the forecaster needed. WSR-88D algorithms were later deployed on the WSR-98D radars made by MetStar and used in China, Romania, India, Korea and other places.

A fundamental question arose as to the role of automated guidance products versus manual interpretation (Andra et al, 2002). It is clear that automated generated products are for guidance and it should not be mistakenly interpreted that warnings were automatically generated and issued without an intervening well trained decision-maker. Initially, there was an extensive radar training program for forecasters, up to 6 weeks for specialists. Clearly, the expectation was that an expert level of training was needed to interpret Doppler radar data for severe weather warnings. This was re-enforced by the work of Pliske et al (1997) who analyzed how to achieve the expected benefits of a modernization program. This resulted in the development of an on-going training program for decision-making at the appropriately named, Warning Decision Training Branch of the National Severe Storms Laboratory. Professionally trained instructors on cognitive principles interactively have the skills to tailor the material to the appropriate knowledge level, abilities and learning styles of the student. It is a model for professional training.

#### **6.3 TITAN – NCAR**

TITAN (Thunderstorm identification, tracking and nowcasting) was first developed in South Africa and then later at NCAR for support of weather modification programs. Dixon and Weiner (1993) described a simple but brilliant threshold technique for the identification of thunderstorm cell cores. This simplified the peak detection techniques of the Crane (1979) technique as the latter identified many weak cells and challenged the computing power of the day. It also described a methodology for tracking. It could be argued that this is the most widely used system in the world. It is freely available and requires some expertise to implement. It is used extensively in research environments (Lei et al, 2009). It is a stand alone system and integrating it into an operational environment has been done but there are capacity and support issues to consider. For example, it is used at the South African Weather Service.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 49

Fig. 8. In this system, there is a trend to go back to basic imagery products such as shear and

The CARDS (Canadian Radar Decision Support) system was developed as part of the radar upgrade (Joe et al, 2002; Lapczak et al, 1999) and built on the previous concepts. In Canada, a single severe weather forecaster is responsible for the provision of warnings for the area coverage of about ten radars. This is in contrast to other countries, where it is approximately one radar for one forecaster. While this may seem like a work overload situation, there are some interesting side benefits. It has been estimated that in a one radar for one forecaster situation, a forecaster will likely face only one significant event in his career. In the Canadian scenario, a severe weather forecaster will therefore experience ten big events. It can be argued that these experienced forecasters will be better at decision making and will therefore make better warnings (Doswell, 2004). Forecasting is a complex process and it remains to be seen whether this is a true. Given these constraints, the weather service of Canada is arguably the most reliant on automated guidance products. They are critical in aiding the forecaster to diagnose those cells which need detailed interrogation to upgrade

aggregated shear to aid in the interpretation and utility of the data.

from a severe weather warning to a more specific warning.

**6.5 CARDS – Canada** 

#### **6.4 WDSS-I and II – USA**

WDSS-I was a research analysis tool and made great strides in developing innovative algorithms and concepts. WDSS-I (Eilts et al, 1996) processed single radar data. A particular innovation was the Storm Cell Identification and Tracking algorithm (Johnson et al, 1998) which ranked the storms by severity. This extended the SWP product from RADAP-II. This system is commercially available from Weather Decision Technologies. WDSS-II was an enhanced version of WDSS-I (Lakshmanan et al, 2006). It has a multiradar capability and integrates other data. Fig. 7 shows a chart of the data processing flow and lists the algorithms. A technical innovation is in the handling of radar data in overlap regions. Radar cell identifications (and others such as mesocyclone detection) are first done along each PPI surface to identify 2D cell objects. Then these 2D objects are collated together into a 3D multi-radar object. A five minute window is used to aggregate the data and cells are time shifted to a common moment in time. A service innovation is that this extends the warning service capability to a regional level (more than the domain of single radar). WDSS-II saw the return to the display of more imagery to support experts in their decision-making (Fig. 8).

Fig. 8 shows shear fields and aggregated shear fields. While they were computed as part of the severe weather algorithms internal computations, they were not previously displayed. With the development of fast computers and display capabilities and the realization that expert forecasters can effectively use these products, they became in vogue.

Fig. 7. The data flow of the WDSS-II system. This system integrates "other" data (numerical weather prediction data) including model data into the radar processing. While this is common for QPE applications to help identify the bright band or melting level, this was an innovation in severe weather processing.

Fig. 8. In this system, there is a trend to go back to basic imagery products such as shear and aggregated shear to aid in the interpretation and utility of the data.

#### **6.5 CARDS – Canada**

Doppler Radar Observations – 48 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

WDSS-I was a research analysis tool and made great strides in developing innovative algorithms and concepts. WDSS-I (Eilts et al, 1996) processed single radar data. A particular innovation was the Storm Cell Identification and Tracking algorithm (Johnson et al, 1998) which ranked the storms by severity. This extended the SWP product from RADAP-II. This system is commercially available from Weather Decision Technologies. WDSS-II was an enhanced version of WDSS-I (Lakshmanan et al, 2006). It has a multiradar capability and integrates other data. Fig. 7 shows a chart of the data processing flow and lists the algorithms. A technical innovation is in the handling of radar data in overlap regions. Radar cell identifications (and others such as mesocyclone detection) are first done along each PPI surface to identify 2D cell objects. Then these 2D objects are collated together into a 3D multi-radar object. A five minute window is used to aggregate the data and cells are time shifted to a common moment in time. A service innovation is that this extends the warning service capability to a regional level (more than the domain of single radar). WDSS-II saw the return to the display of more imagery to support experts in their

Fig. 8 shows shear fields and aggregated shear fields. While they were computed as part of the severe weather algorithms internal computations, they were not previously displayed. With the development of fast computers and display capabilities and the realization that

Fig. 7. The data flow of the WDSS-II system. This system integrates "other" data (numerical weather prediction data) including model data into the radar processing. While this is common for QPE applications to help identify the bright band or melting level, this was an

expert forecasters can effectively use these products, they became in vogue.

**6.4 WDSS-I and II – USA** 

decision-making (Fig. 8).

innovation in severe weather processing.

The CARDS (Canadian Radar Decision Support) system was developed as part of the radar upgrade (Joe et al, 2002; Lapczak et al, 1999) and built on the previous concepts. In Canada, a single severe weather forecaster is responsible for the provision of warnings for the area coverage of about ten radars. This is in contrast to other countries, where it is approximately one radar for one forecaster. While this may seem like a work overload situation, there are some interesting side benefits. It has been estimated that in a one radar for one forecaster situation, a forecaster will likely face only one significant event in his career. In the Canadian scenario, a severe weather forecaster will therefore experience ten big events. It can be argued that these experienced forecasters will be better at decision making and will therefore make better warnings (Doswell, 2004). Forecasting is a complex process and it remains to be seen whether this is a true. Given these constraints, the weather service of Canada is arguably the most reliant on automated guidance products. They are critical in aiding the forecaster to diagnose those cells which need detailed interrogation to upgrade from a severe weather warning to a more specific warning.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 51

machine or the mono or color graphics terminal. Each innovation increased the capacity to deliver better products. In today's technology, every button press or mouse click that is eliminated delivers "a big bang for the buck". This key innovation allowed the data to be effectively used in the Sydney Olympic Command Centre (Joe et al, 2004; Keenan et al,

Similar to the SCIT of WDSS-I, CARDS implemented a fuzzy logic technique to rank storms. The technique is configurable (see Table 6). It shows the parameters that the users decided to use and the thresholds that they considered as weak, moderate, strong and severe (see

In the overlap region, cells are selected from one radar or the other, unlike WDSS-II. Due to attenuation concerns, lack of experience with the fuzzy logic storm severity technique and that reflectivity (and reflectivity based products) was still the prime parameter for determining storm severity; users selected the cell detection with the maximum reflectivity as the cell for visualization. However, this would likely not be the case anymore as nearest radar or maximum information or maximum severity ranking would

count m/s/km cm m/s kg m^2 /

Rank: 5-6 means a value of 5 or more but less than 7. WER: The number of directions where reflectivity

 determines a BWER (with low reflectivity below) Meso: Average Pattern Vector Shear (see Zrnic et al, 1985)

VIL if VIL for classification then 10 20 30 40 are the thresholds

Table 6. Fuzzy Logic Membership Functions for Parameters Used to Rank Storms.

increases

 Gust potential in m/s

Max Z: Max reflectivity in the cell 45 dBZ Echotop Ht: Reliable echo top parameter

Vil Density: Similar to WDRAFT in pattern

Hail: Average Hail Size

Minimum 1 0-2 5-11 4 0.5 10 2.2 30 5.5 Weak 2 3-4 12-17 6 1.3 15 3 45 8.5 Moderate 3 5-6 18-21 8 2.3 20 3.5 50 10.5 Severe 4 7+ 22-26 10 5 25 4 60 12.5

(0-8) ETop

0 0 0 0 0 0 0 0

density

km

Max Z 45 dBZ

dBZ km

Thresholds Rank BWER Meso Hail Wdraft Vil

2004).

also Doswell et al, 2006).

be chosen today.

Notes:

 WDRAFT :

Fig. 9. An example of a CARDS composite, SCIT and cell view. The size of the forecast domain is about ~2000 km x 1600 km. The image shows a zoomed image of the cells, tracks and lightning strikes. Eight Canadian radars and 12 US radars contribute to the image. The composite and the SCIT table products are invoked and displayed at the same time. The forecaster can either drill down to a CELL VIEW via the composite or via the SCIT table. They can also rapidly survey the cells from the SCIT table without invoking the CELL VIEW products. The colour coding indicates the categorical ranking. On the right is an example of a cell view. This shows a variety of images that allows the forecaster to quickly make a decision as to the severity of the storm. The product shows an ensemble product of the algorithms (upper left hand corner, not described), automatically determined cross-sections, four CAPPIs (1.5, 3.0, 7.0, 9.0 km), reflectivity gradient, MAXR, echo top, VIL density, Hail, BWER and 45 dBZ echo top and time graphs.

In an envisioned future exercise for the design of CARDS, it was identified that there was actually no hard requirement for single radar products. One of the main reasons for missed warnings was that the forecaster was so intent on one thunderstorm that they forgot about the others. There was a loss of situational awareness. This happens even with experienced forecasters or analysts and is common in many fields where critical decisions are made. A regional composite that could display and overlay the most popular products (CAPPI, EchoTop, etc) is the main product to maintain situational awareness. Thunderstorms cell locations are identified, ranked, color coded and displayed on the composite and in a table similar to the SCIT table. Selecting the cell of interest in the composite or in the table, the user is able to quickly and rapidly drill down to reveal a cell view product (Fig. 9) that contain all the products that the user would use to interrogate a cell and make decisions. The cell view has a legacy from Chisholm and Renick (1972). The design exercise also identified the critical reliance on automated guidance products.

Another important innovation is that the visualization tool for the image and data products is based on hypertext transfer protocol (http) which means that any computer regardless of operating system can access the full functionality of the radar data. Analyzing breakthroughs in the use of radar, access to the data and the products has been "the" key innovation. Recall the days of radar operators who hand drew radar maps or the facsimile Doppler Radar Observations – 50 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Fig. 9. An example of a CARDS composite, SCIT and cell view. The size of the forecast domain is about ~2000 km x 1600 km. The image shows a zoomed image of the cells, tracks and lightning strikes. Eight Canadian radars and 12 US radars contribute to the image. The composite and the SCIT table products are invoked and displayed at the same time. The forecaster can either drill down to a CELL VIEW via the composite or via the SCIT table. They can also rapidly survey the cells from the SCIT table without invoking the CELL VIEW products. The colour coding indicates the categorical ranking. On the right is an example of a cell view. This shows a variety of images that allows the forecaster to quickly make a decision as to the severity of the storm. The product shows an ensemble product of the algorithms (upper left hand corner, not described), automatically determined cross-sections, four CAPPIs (1.5, 3.0, 7.0, 9.0 km), reflectivity gradient, MAXR, echo top, VIL density, Hail,

In an envisioned future exercise for the design of CARDS, it was identified that there was actually no hard requirement for single radar products. One of the main reasons for missed warnings was that the forecaster was so intent on one thunderstorm that they forgot about the others. There was a loss of situational awareness. This happens even with experienced forecasters or analysts and is common in many fields where critical decisions are made. A regional composite that could display and overlay the most popular products (CAPPI, EchoTop, etc) is the main product to maintain situational awareness. Thunderstorms cell locations are identified, ranked, color coded and displayed on the composite and in a table similar to the SCIT table. Selecting the cell of interest in the composite or in the table, the user is able to quickly and rapidly drill down to reveal a cell view product (Fig. 9) that contain all the products that the user would use to interrogate a cell and make decisions. The cell view has a legacy from Chisholm and Renick (1972). The design exercise also

Another important innovation is that the visualization tool for the image and data products is based on hypertext transfer protocol (http) which means that any computer regardless of operating system can access the full functionality of the radar data. Analyzing breakthroughs in the use of radar, access to the data and the products has been "the" key innovation. Recall the days of radar operators who hand drew radar maps or the facsimile

BWER and 45 dBZ echo top and time graphs.

identified the critical reliance on automated guidance products.

machine or the mono or color graphics terminal. Each innovation increased the capacity to deliver better products. In today's technology, every button press or mouse click that is eliminated delivers "a big bang for the buck". This key innovation allowed the data to be effectively used in the Sydney Olympic Command Centre (Joe et al, 2004; Keenan et al, 2004).

Similar to the SCIT of WDSS-I, CARDS implemented a fuzzy logic technique to rank storms. The technique is configurable (see Table 6). It shows the parameters that the users decided to use and the thresholds that they considered as weak, moderate, strong and severe (see also Doswell et al, 2006).

In the overlap region, cells are selected from one radar or the other, unlike WDSS-II. Due to attenuation concerns, lack of experience with the fuzzy logic storm severity technique and that reflectivity (and reflectivity based products) was still the prime parameter for determining storm severity; users selected the cell detection with the maximum reflectivity as the cell for visualization. However, this would likely not be the case anymore as nearest radar or maximum information or maximum severity ranking would be chosen today.





Table 6. Fuzzy Logic Membership Functions for Parameters Used to Rank Storms.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 53

SWIRLS updates and outputs nowcast products at 6-minute intervals. For severe thunderstorms, the major results are visualized as an image product called the Severe Weather Map on its client workstation in the forecasting office, as well as a web page named SPIDASS (**S**WIRLS **P**anel for **I**ntegrated **D**isplay of **A**lerts on **S**evere **S**torms) dedicated for

Fig. 11. (a) shows an example of the Severe Weather Map. Textual alerts with quantitative details were printed at the bottom. (b) , the main panel of SPIDASS web page provides a compact view of all alerts arranged in rows and colour-coded for different severity levels.

The Hong Kong Observatory has also developed separate multi-sensor thunderstorm nowcasting systems for the aviation community and the public utilities services (Li, 2009). A lightning nowcasting system, named the Airport Thunderstorm and Lightning Alerting System (ATLAS), covers the Hong Kong International Airport (HKIA). It combines rapidly updated CG lightning strike information, radar reflectivity and TREC wind information to nowcast lightning strikes using a modified Semi-Lagrangian advection scheme. Depending on the predicted distance from HKIA, ATLAS will automatically generate RED (1km) or

ATLAS is equipped with two ensemble algorithms, to take into account the possible rapid development nature of lightning (transient and sporadic). The Weighted Ensemble (WE) algorithm sums all available 12-minute CG forecasts with decreasing weight with time. If the sum exceeds an optimized threshold, alerts are created. WE has proved to be effective for alerting persistent and wide-spread thunderstorms. The Time Lagged Ensemble (TLE) algorithm sums the 1-minute forecasts valid at the same time from the twelve 1-minute forecasts provided in the past 12 minutes with decreasing weight over time. TLE is proved to be more skilful in predicting rapidly developing, small or wide-spread thunderstorms

than WE. Figure 12 shows a snapshot of the ATLAS product.

severe weather alerts (Fig. 11).

AMBER (5 km) alerts.

#### **6.6 SWIRLS and its variants – Hong Kong, China**

In Hong Kong, lightning strikes and damaging squalls are major threats accompanying thunderstorms. In support of the Thunderstorm Warning operations, SWIRLS (**S**hort-range **W**arning of **I**ntense **R**ainstorms in **L**ocalized **S**ystems) was developed to track and predict severe weather including rainstorms, cloud-to-ground (CG) lightning, damaging thunderstorm squalls and hail for the general public. The warning decision and message preparation are made by the Observatory's duty forecaster. Once issued, the warning message are disseminated automatically through various channels including radio and television broadcast automatic telephone enquiry system, Internet web page, as well as mobile apps for smart phones and social networking platforms such as Twitter.

An innovation is the DELITE (**D**etection of cloud **E**lectrification and **L**ightning based on **I**sothermal **T**hunderstorm **E**choes) algorithm for lightning warning. It selects radar and other parameters most relevant to the microphysical processes leading up to the electrification of a cumulus cloud (Fig. 10). This includes radar reflectivity at constant temperature levels (0C, -10C, and -20C), the thermal profile of the troposphere (from either numerical weather model analysis or the latest available radiosonde data), the echo top height and the vertically integrated liquid (VIL). CG lightning initiation is expected if prescribed thresholds are exceeded.

The above severe weather analyses are performed on a cell basis and the threat areas are identified as elliptical cells in the corresponding interest fields with values greater than or equal to prescribed thresholds. For example, the detailed cell identification technique follows the GTrack algorithm of SWIRLS. For lightning and downburst, the interest fields are 3-km CAPPI and 0-5 km VIL respectively. The thresholds are 25 dBZ and 5 mm respectively.

MOVA **(M**ulti-scale **O**ptical flow by **V**ariational **A**nalysis) is a gridded echo-motion field that is derived from consecutive radar reflectivity fields by solving an optical-flow equation with a smoothness constraint. To capture multi-scale echo motions, the optical-flow equation is solved iteratively for a cascade of grids from coarse to fine resolutions (about 512 to 3 km).

Fig. 10. (a) Conceptual model of CG lightning. The main source of electric charges is assumed to be located in the mixed-phase layer between 0 and -20C. Prior to electrification, the updraft is expected to separate the charge carriers vertically. Negative charge carriers (i.e. graupel) are expected to reside mainly in the mixed-phase layer. The updraft pumps super-cooled rain water into this layer and wet the carriers. (b) Flow chart of the logic of the algorithm.

Doppler Radar Observations – 52 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

In Hong Kong, lightning strikes and damaging squalls are major threats accompanying thunderstorms. In support of the Thunderstorm Warning operations, SWIRLS (**S**hort-range **W**arning of **I**ntense **R**ainstorms in **L**ocalized **S**ystems) was developed to track and predict severe weather including rainstorms, cloud-to-ground (CG) lightning, damaging thunderstorm squalls and hail for the general public. The warning decision and message preparation are made by the Observatory's duty forecaster. Once issued, the warning message are disseminated automatically through various channels including radio and television broadcast automatic telephone enquiry system, Internet web page, as well as

An innovation is the DELITE (**D**etection of cloud **E**lectrification and **L**ightning based on **I**sothermal **T**hunderstorm **E**choes) algorithm for lightning warning. It selects radar and other parameters most relevant to the microphysical processes leading up to the electrification of a cumulus cloud (Fig. 10). This includes radar reflectivity at constant temperature levels (0C, -10C, and -20C), the thermal profile of the troposphere (from either numerical weather model analysis or the latest available radiosonde data), the echo top height and the vertically integrated liquid (VIL). CG lightning initiation is expected if

The above severe weather analyses are performed on a cell basis and the threat areas are identified as elliptical cells in the corresponding interest fields with values greater than or equal to prescribed thresholds. For example, the detailed cell identification technique follows the GTrack algorithm of SWIRLS. For lightning and downburst, the interest fields are 3-km

MOVA **(M**ulti-scale **O**ptical flow by **V**ariational **A**nalysis) is a gridded echo-motion field that is derived from consecutive radar reflectivity fields by solving an optical-flow equation with a smoothness constraint. To capture multi-scale echo motions, the optical-flow equation is solved iteratively for a cascade of grids from coarse to fine resolutions (about 512 to 3 km).

Fig. 10. (a) Conceptual model of CG lightning. The main source of electric charges is assumed to be located in the mixed-phase layer between 0 and -20C. Prior to electrification, the updraft is expected to separate the charge carriers vertically. Negative charge carriers (i.e. graupel) are expected to reside mainly in the mixed-phase layer. The updraft pumps super-cooled rain water into this layer and wet the carriers. (b) Flow chart of the logic of the algorithm.

CAPPI and 0-5 km VIL respectively. The thresholds are 25 dBZ and 5 mm respectively.

mobile apps for smart phones and social networking platforms such as Twitter.

**6.6 SWIRLS and its variants – Hong Kong, China** 

prescribed thresholds are exceeded.

SWIRLS updates and outputs nowcast products at 6-minute intervals. For severe thunderstorms, the major results are visualized as an image product called the Severe Weather Map on its client workstation in the forecasting office, as well as a web page named SPIDASS (**S**WIRLS **P**anel for **I**ntegrated **D**isplay of **A**lerts on **S**evere **S**torms) dedicated for severe weather alerts (Fig. 11).

Fig. 11. (a) shows an example of the Severe Weather Map. Textual alerts with quantitative details were printed at the bottom. (b) , the main panel of SPIDASS web page provides a compact view of all alerts arranged in rows and colour-coded for different severity levels.

The Hong Kong Observatory has also developed separate multi-sensor thunderstorm nowcasting systems for the aviation community and the public utilities services (Li, 2009). A lightning nowcasting system, named the Airport Thunderstorm and Lightning Alerting System (ATLAS), covers the Hong Kong International Airport (HKIA). It combines rapidly updated CG lightning strike information, radar reflectivity and TREC wind information to nowcast lightning strikes using a modified Semi-Lagrangian advection scheme. Depending on the predicted distance from HKIA, ATLAS will automatically generate RED (1km) or AMBER (5 km) alerts.

ATLAS is equipped with two ensemble algorithms, to take into account the possible rapid development nature of lightning (transient and sporadic). The Weighted Ensemble (WE) algorithm sums all available 12-minute CG forecasts with decreasing weight with time. If the sum exceeds an optimized threshold, alerts are created. WE has proved to be effective for alerting persistent and wide-spread thunderstorms. The Time Lagged Ensemble (TLE) algorithm sums the 1-minute forecasts valid at the same time from the twelve 1-minute forecasts provided in the past 12 minutes with decreasing weight over time. TLE is proved to be more skilful in predicting rapidly developing, small or wide-spread thunderstorms than WE. Figure 12 shows a snapshot of the ATLAS product.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 55

to the nowcasts in the early stages and towards the model at the longer lead times. Figure 13 shows the comparison between the effects of ATNS using SWIRLS simple extrapolation (1-6 hours forecasts) and ATNS using SWIRLS-NHM blended forecasts for the case of 4 Jun 2009 (1-6 hour forecasts). The simple TREC extrapolation (left column) overpredicts the

Fig. 13. An example showing the comparison between the effects of SWIRLS simple

column) forecasts and the radar-based QPE (different scale).

extrapolation and blending of SWIRLS and NHM rainfall. Figures from top to bottom are 1 hr, 2-hr and 6hr simple extrapolation (left column), AANS blended precipitation (middle

rainfall intensities in this case at long lead times (6 hours).

Fig. 12. A snapshot of ATLAS webpage. The image shows the actual position of the CGs (ellipses with solid line), the predicted CGs (ellipses with dashed line), the 12-minute forecast in blue and the 30-minute forecast in grey.

The Aviation Thunderstorm Nowcasting System (ATNS) has been developed to predict the movement of thunderstorms to help local Air Traffic Management to better manage the flight traffic over the Hong Kong Flight Information Region for the next few hours (Li and Wong, 2010). A blending approach is adopted to extend the forecast range and to capture the development and dissipation of thunderstorms. The NWP model used is a high resolution non-hydrostatic model with horizontal resolution of 5 km (Li et al. 2005; Wong et al. 2009). Volume radar reflectivity data are ingested into the model via the LAPS data assimilation system (Albers et al. 1996) and radar Doppler radial wind and 3D radar winds are assimilated via the JNoVA-3DVAR data assimilation system (Honda et al. 2005) to improve the initial moisture field and wind fields, respectively.

The blending algorithm is as follows: (i) SWIRLS radar forecast reflectivity is converted into surface precipitation using a dynamic reflectivity-rainfall (Z-R) relation; (ii) precipitation forecasts are extracted from the NHM; and (iii) then they are blended. The latter blending process involves: (i) Phase correction where a variational technique minimizes the root mean square error of the forecast rainfall field from a previous model run (usually initialized at 1-2 hours before) and the actual radar-raingauge derived precipitation distribution (Wong et al. 2009). (ii) Calibration of the QPF rainfall intensities is based on the observed radar-based quantitative precipitation estimate (QPE), and (iii) blending of calibrated model QPF with the radar nowcast out to 6 hours where the weighting is biased Doppler Radar Observations – 54 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Fig. 12. A snapshot of ATLAS webpage. The image shows the actual position of the CGs (ellipses with solid line), the predicted CGs (ellipses with dashed line), the 12-minute

The Aviation Thunderstorm Nowcasting System (ATNS) has been developed to predict the movement of thunderstorms to help local Air Traffic Management to better manage the flight traffic over the Hong Kong Flight Information Region for the next few hours (Li and Wong, 2010). A blending approach is adopted to extend the forecast range and to capture the development and dissipation of thunderstorms. The NWP model used is a high resolution non-hydrostatic model with horizontal resolution of 5 km (Li et al. 2005; Wong et al. 2009). Volume radar reflectivity data are ingested into the model via the LAPS data assimilation system (Albers et al. 1996) and radar Doppler radial wind and 3D radar winds are assimilated via the JNoVA-3DVAR data assimilation system (Honda et al. 2005) to

The blending algorithm is as follows: (i) SWIRLS radar forecast reflectivity is converted into surface precipitation using a dynamic reflectivity-rainfall (Z-R) relation; (ii) precipitation forecasts are extracted from the NHM; and (iii) then they are blended. The latter blending process involves: (i) Phase correction where a variational technique minimizes the root mean square error of the forecast rainfall field from a previous model run (usually initialized at 1-2 hours before) and the actual radar-raingauge derived precipitation distribution (Wong et al. 2009). (ii) Calibration of the QPF rainfall intensities is based on the observed radar-based quantitative precipitation estimate (QPE), and (iii) blending of calibrated model QPF with the radar nowcast out to 6 hours where the weighting is biased

forecast in blue and the 30-minute forecast in grey.

improve the initial moisture field and wind fields, respectively.

to the nowcasts in the early stages and towards the model at the longer lead times. Figure 13 shows the comparison between the effects of ATNS using SWIRLS simple extrapolation (1-6 hours forecasts) and ATNS using SWIRLS-NHM blended forecasts for the case of 4 Jun 2009 (1-6 hour forecasts). The simple TREC extrapolation (left column) overpredicts the rainfall intensities in this case at long lead times (6 hours).

Fig. 13. An example showing the comparison between the effects of SWIRLS simple extrapolation and blending of SWIRLS and NHM rainfall. Figures from top to bottom are 1 hr, 2-hr and 6hr simple extrapolation (left column), AANS blended precipitation (middle column) forecasts and the radar-based QPE (different scale).

Automated Processing of Doppler Radar Data for Severe Weather Warnings 57

Fig. 14. (a) The detected thunderstorm is the ellipse oriented south-west to north-east. A motion to the south east is shown. The contours and shading show the probability that the thunderstorm will advect or propagate into those locations. The probabilities were verified for a season of storms around Sydney and Beijing, with excellent reliability, with a Brier skill score of between 0.36 and 0.44 with respect to an advected threat area forecast. (b) An example of a prototype TIFS strike probability product. Three cells are identified as A, B and C and represented as ellipses. The tracks of B and C are indicted by the partial ellipses and the colours indicate the strike probability, marked as E and F and appear consistent. The track for cell A is marked as D1 and appears anomalous. D2 is the track that the analyst has

Fig. 15. This figure shows a nowcast of the reflectivity and lightning from NoCAWS. The

modified to produce the final strike probability map (c).

plus signs are nowcasts of lightning strikes.

#### **6.7 SIGOONS – France**

Significant Weather Object Oriented Nowcast System (SIGOONS) is a component of the Synergie workstation (Brovelli et al, 2005). Thunderstorm cells are identified using the RDT (Rapidly Developing Thunderstorms) technique by Hering et al (2005) and are represented as objects. This database is updated every five minutes and is automatically quality controlled against other observational data. The objects may have deterministic and probabilistic attributes and have a time dimension – they can grow and decay. Products are automatically generated and tailored according to pre-defined customer requirements. Discrepancies are brought to the attention of the forecaster who can select persistence over linear extrapolation nowcasts. The forecaster can take additional initiative. The attributes of the weather objects can be manipulated and altered by forecasters.

#### **6.8 THESPA and TIFS, Australia**

Within the Bureau of Meteorology, forecasters use RAPIC to interactively interrogate the data. The innovation is the radar data is loaded on the graphics memory of the client computer and extremely rapid response of the display is achieved. To avoid dual-PRF dealiasing errors, only single PRF data is used resulting in a Nyquist interval of 16 m/s. This implies considerable forecaster training is required to interpret highly aliased Doppler data.

Thunderstorm Strike Probability (THESPA, Dance et al, 2010) generates probabilistic nowcasts. Using the historical statistics of the nowcast position errors as a function of lead time and detected storm properties, storm motion is modeled as a bivariate Gaussian distribution on storm speed and direction. For a given geographical point, the strike probability from all possible thunderstorms is computed for the forecast period (Fig. 14).

The algorithm is embedded in the Thunderstorm Interactive Forecast System (TIFS, Bally 2004). The Beijing Olympics provided an opportunity to explore and prototype new nowcasting techniques (Wang et al, 2010). TIFS was modified to ingest the storm locations and tracks from the CARDS, SWIRLS, WDSS and TITAN to create a poor man's ensemble. From each of the storms and tracks, THESPA was used to compute a consensus or ensemble strike probability (Fig. 14b). A warning product would be automatically generated. The analyst (B08 Forecast Demonstration Project team member) would evaluate the product and determine if intervention was needed. The analyst could then use the graphical interface and add, delete or modify cells or tracks. The analyst could view and modify any of the ensemble members and the strike probability display would update. Accepting the change would regenerate the automated warning product, be disseminated and overwriting the fully automated product.

#### **6.9 NoCAWS – SMB**

The Shanghai Meteorological Bureau's NoCAWS system was one of the nowcast systems used for the World Expo on Nowcasting Services (WENS) component of the Multi-hazard Early Warning Service project (MHEWS). It integrates observations, mesoscale models and nowcasts to host data displays; analysis tools, severe weather alerting tools to generate automatic forecasts and warning for forecasters. It covers the scales from outlooks to warnings. An innovative feature is lightning forecasts. COTREC winds are used to nowcast cell motions. Advection and statistical relationships between lightning and reflectivity are used to nowcast lightning (Fig. 15)

Doppler Radar Observations – 56 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Significant Weather Object Oriented Nowcast System (SIGOONS) is a component of the Synergie workstation (Brovelli et al, 2005). Thunderstorm cells are identified using the RDT (Rapidly Developing Thunderstorms) technique by Hering et al (2005) and are represented as objects. This database is updated every five minutes and is automatically quality controlled against other observational data. The objects may have deterministic and probabilistic attributes and have a time dimension – they can grow and decay. Products are automatically generated and tailored according to pre-defined customer requirements. Discrepancies are brought to the attention of the forecaster who can select persistence over linear extrapolation nowcasts. The forecaster can take additional initiative. The attributes of

Within the Bureau of Meteorology, forecasters use RAPIC to interactively interrogate the data. The innovation is the radar data is loaded on the graphics memory of the client computer and extremely rapid response of the display is achieved. To avoid dual-PRF dealiasing errors, only single PRF data is used resulting in a Nyquist interval of 16 m/s. This implies considerable forecaster training is required to interpret highly aliased Doppler data. Thunderstorm Strike Probability (THESPA, Dance et al, 2010) generates probabilistic nowcasts. Using the historical statistics of the nowcast position errors as a function of lead time and detected storm properties, storm motion is modeled as a bivariate Gaussian distribution on storm speed and direction. For a given geographical point, the strike probability from all possible thunderstorms is computed for the forecast period (Fig. 14).

The algorithm is embedded in the Thunderstorm Interactive Forecast System (TIFS, Bally 2004). The Beijing Olympics provided an opportunity to explore and prototype new nowcasting techniques (Wang et al, 2010). TIFS was modified to ingest the storm locations and tracks from the CARDS, SWIRLS, WDSS and TITAN to create a poor man's ensemble. From each of the storms and tracks, THESPA was used to compute a consensus or ensemble strike probability (Fig. 14b). A warning product would be automatically generated. The analyst (B08 Forecast Demonstration Project team member) would evaluate the product and determine if intervention was needed. The analyst could then use the graphical interface and add, delete or modify cells or tracks. The analyst could view and modify any of the ensemble members and the strike probability display would update. Accepting the change would regenerate the automated warning product, be disseminated and overwriting the fully automated product.

The Shanghai Meteorological Bureau's NoCAWS system was one of the nowcast systems used for the World Expo on Nowcasting Services (WENS) component of the Multi-hazard Early Warning Service project (MHEWS). It integrates observations, mesoscale models and nowcasts to host data displays; analysis tools, severe weather alerting tools to generate automatic forecasts and warning for forecasters. It covers the scales from outlooks to warnings. An innovative feature is lightning forecasts. COTREC winds are used to nowcast cell motions. Advection and statistical relationships between lightning and reflectivity are

the weather objects can be manipulated and altered by forecasters.

**6.7 SIGOONS – France** 

**6.8 THESPA and TIFS, Australia** 

**6.9 NoCAWS – SMB** 

used to nowcast lightning (Fig. 15)

Fig. 14. (a) The detected thunderstorm is the ellipse oriented south-west to north-east. A motion to the south east is shown. The contours and shading show the probability that the thunderstorm will advect or propagate into those locations. The probabilities were verified for a season of storms around Sydney and Beijing, with excellent reliability, with a Brier skill score of between 0.36 and 0.44 with respect to an advected threat area forecast. (b) An example of a prototype TIFS strike probability product. Three cells are identified as A, B and C and represented as ellipses. The tracks of B and C are indicted by the partial ellipses and the colours indicate the strike probability, marked as E and F and appear consistent. The track for cell A is marked as D1 and appears anomalous. D2 is the track that the analyst has modified to produce the final strike probability map (c).

Fig. 15. This figure shows a nowcast of the reflectivity and lightning from NoCAWS. The plus signs are nowcasts of lightning strikes.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 59

Fig. 17. An example of NowcastMix. It combines and merges the output from several

consolidated sets of most-probable short-term forecasts (Fig. 17).

Since there are several nowcasting systems avaliable, NowCastMIX processes these available nowcast products together in an integrated grid-based analysis, providing a generic, optimal warning solution with a 5-minute update cycle. The products are combined using a fuzzy logic approach (James et al 2011). The method includes estimates for the storm cell motion by combining raw cell tracking inputs from the KONRAD and CellMOS systems with vector fields derived from comparing consecutive radar images. Finally, the resulting gridded warning fields are spatially filtered to provide regionallyoptimized warning levels for differing thunderstorm severities for forecasters. NowCastMIX delivers a synthesis of the various nowcasting and forecast model system inputs to provide

Japan Meteorological Agency initiated their hazardous wind warning program in 2007. A hazardous-wind-possibility-index is calculated based on the NWP prediction of wind and radar reflectivity exceeding a threshold. An innovation is the use of a template matching technique for the detection of mesocyclones. Rankine vortex and divergence flow field templates of different intensity and spatial scale are generated and matched to the radial velocity field. This is done every five minutes. Detections on two consecutive time steps are required as a quality controlled metric. Then the two estimates are combined every ten minutes to estimate a hazardous wind potential. Nowcasting is based on a motion analysis. Different thresholds are statistically established and the success ratio (1-FAR) and the probability of detection (POD) are used to categorize the hazard level (Table 7). If level 2 is exceeded (see Fig. 18), then it alerts a forecaster to issue Hazardous Wind Watch. A forecaster may ignore the level 2 information, when: (i) the storm is near the boundary of a warning area and it will be out before the time of warning or (ii) the quality of radar data seems poor (e.g. AP or sea clutter). A forecaster can issue a warning at level 1 when (i) reliable report of a tornado/tornadoes and/or and (ii) strong gust (say, greater than 30 m/s)

nowcasting systems into a hazard map.

**6.11 Japan – JMA** 

caused by a convective cloud.

#### **6.10 KONRAD/NinJo/NowCastMIX – DWD**

There are several tools in the German Weather Service and include KONRAD (Lang et al, 2001), Mesocyclone detection (Hengstebeck et al, 2011), AutoWARN, EPM (editing, prediction, monitoring), Cellviews (Joe et al, 2003). All of these are integrated into the Ninjo system (Koppert et al, 2004). KONRAD was developed as a research prototype and uses a variable elevation angle PPI reflectivity product for the identification and warning potential of cells. The 10 minute volume scan product is used for further classification. The cells are displayed as abstractions and only a >28dBZ contour is displayed in the end user product (Fig. 16). Of all the systems discussed, it is the only truly automated system where the products go directly out to the end-user without human oversight. However, it targets sophisticated end-users such as emergency authorities, county administrators, fire departments and the military and not the public. One could argue that these are guidance products for external versus internal decision-makers for planning but not warning service. So the "cry wolf" syndrome is not a significant issue. This does demonstrate the potential use of fully authomated products.

Fig. 16. An example of the abstraction from reflectivity to symbolic representation of thunderstorms from the KONRAD system. It is the only system described in the contribution that is totally automated. It is directed to "sophisticated users" for planning purposes.

The AutoWARN system in NinJo integrates various meteorological data and products in a warning decision support process, generating real-time warning proposals for assessment and possible modification by the duty forecasters. These warnings finally issued by the forecaster are then exported to a system generating textual and graphical warning products for dissemination to customers. On very short, nowcasting timescales, several systems are continuously monitored. These include the radar-based storm-cell identification and tracking methods, KONRAD and CellMOS; 3D radar volume scans yielding vertically integrated liquid water (VIL) composites; precise lightning strike locations; the precipitation prediction system, RadVOR-OP as well as synoptic reports and the latest high resolution numerical analysis and forecast data.

Fig. 17. An example of NowcastMix. It combines and merges the output from several nowcasting systems into a hazard map.

Since there are several nowcasting systems avaliable, NowCastMIX processes these available nowcast products together in an integrated grid-based analysis, providing a generic, optimal warning solution with a 5-minute update cycle. The products are combined using a fuzzy logic approach (James et al 2011). The method includes estimates for the storm cell motion by combining raw cell tracking inputs from the KONRAD and CellMOS systems with vector fields derived from comparing consecutive radar images. Finally, the resulting gridded warning fields are spatially filtered to provide regionallyoptimized warning levels for differing thunderstorm severities for forecasters. NowCastMIX delivers a synthesis of the various nowcasting and forecast model system inputs to provide consolidated sets of most-probable short-term forecasts (Fig. 17).

#### **6.11 Japan – JMA**

Doppler Radar Observations – 58 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

There are several tools in the German Weather Service and include KONRAD (Lang et al, 2001), Mesocyclone detection (Hengstebeck et al, 2011), AutoWARN, EPM (editing, prediction, monitoring), Cellviews (Joe et al, 2003). All of these are integrated into the Ninjo system (Koppert et al, 2004). KONRAD was developed as a research prototype and uses a variable elevation angle PPI reflectivity product for the identification and warning potential of cells. The 10 minute volume scan product is used for further classification. The cells are displayed as abstractions and only a >28dBZ contour is displayed in the end user product (Fig. 16). Of all the systems discussed, it is the only truly automated system where the products go directly out to the end-user without human oversight. However, it targets sophisticated end-users such as emergency authorities, county administrators, fire departments and the military and not the public. One could argue that these are guidance products for external versus internal decision-makers for planning but not warning service. So the "cry wolf" syndrome is not a significant issue. This does demonstrate the potential

Fig. 16. An example of the abstraction from reflectivity to symbolic representation of

thunderstorms from the KONRAD system. It is the only system described in the contribution that is totally automated. It is directed to "sophisticated users" for planning purposes.

The AutoWARN system in NinJo integrates various meteorological data and products in a warning decision support process, generating real-time warning proposals for assessment and possible modification by the duty forecasters. These warnings finally issued by the forecaster are then exported to a system generating textual and graphical warning products for dissemination to customers. On very short, nowcasting timescales, several systems are continuously monitored. These include the radar-based storm-cell identification and tracking methods, KONRAD and CellMOS; 3D radar volume scans yielding vertically integrated liquid water (VIL) composites; precise lightning strike locations; the precipitation prediction system, RadVOR-OP as well as synoptic reports and the latest high resolution

**6.10 KONRAD/NinJo/NowCastMIX – DWD** 

use of fully authomated products.

numerical analysis and forecast data.

Japan Meteorological Agency initiated their hazardous wind warning program in 2007. A hazardous-wind-possibility-index is calculated based on the NWP prediction of wind and radar reflectivity exceeding a threshold. An innovation is the use of a template matching technique for the detection of mesocyclones. Rankine vortex and divergence flow field templates of different intensity and spatial scale are generated and matched to the radial velocity field. This is done every five minutes. Detections on two consecutive time steps are required as a quality controlled metric. Then the two estimates are combined every ten minutes to estimate a hazardous wind potential. Nowcasting is based on a motion analysis. Different thresholds are statistically established and the success ratio (1-FAR) and the probability of detection (POD) are used to categorize the hazard level (Table 7). If level 2 is exceeded (see Fig. 18), then it alerts a forecaster to issue Hazardous Wind Watch. A forecaster may ignore the level 2 information, when: (i) the storm is near the boundary of a warning area and it will be out before the time of warning or (ii) the quality of radar data seems poor (e.g. AP or sea clutter). A forecaster can issue a warning at level 1 when (i) reliable report of a tornado/tornadoes and/or and (ii) strong gust (say, greater than 30 m/s) caused by a convective cloud.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 61

The client refreshes real-time observations cycled in 5 min from radars and AWS (automatic weather stations) and provides real-time alerts (sounding, flashing) for indication of severe weather events (meeting certain thresholds such as wind speeds or rainfall amounts). It also provides an interactive tool for preparing, editing and issuing Nowcast and warning for

Based on quality control, a regional 3D reflectivity mosaic is produced by trying to fill the gaps that are generated by terrain blockage or AP. Products such as vertically integrated liquid (VIL) , echo top (ET) and COTREC winds are then derived. QPE algorithm involves extraction of convective echoes from stratiform echoes by texture and horizontal gradient properties. Different Z-R relations are used for convective rain and stratiform rain. COTREC (continuous tracking radar echo by correlation) vectors are echo motion vectors that are derived from moving radar reflectivity patterns through grid-to-grid cross-correlation and then adjusted by a horizontal non-divergence constraint for hourly nowcasts of rainfall (Li et al, 1995). This is blended with mesoscale numerical prediction model output for 2-3 hour

SWAN provides real time verifications for storm tracking and reflectivity nowcasts. Storm track errors are shown as distance differences between observed storm tracks and predicted storm tracks (1h). Observed radar reflectivity are also verified against extrapolated

Severe weather warnings can be prepared and issued through SWAN by graphical interface by circling an area on the screen, clicking an icon and doing some minor wording (Fig. 19). A web-based version of SWAN has been developed and deployed in Guangdong

Fig. 19. A SWAN display showing cells/tracks (main screen), SCIT (bottom) and time histories of critical parameters (right). There are similarities with WDSS, NinJo and CARDS

severe weathers

nowcasts.

displays.

forecasted reflectivity.

Meteorological Bureau.


Table 7. Hazardous Criteria Level

Fig. 18. The processing steps for hazardous wind potential at JMA. It is typical of current systems where mesoscale NWP predictions are assumed to be good enough to match with the observations.

#### **6.12 SWAN – CHINA/CMA**

In 2008, the China Meteorological Administration (CMA) launched a campaign on the development of its first version of integrated nowcasting system SWAN (Severe Weather Analysis and Nowcast system). This system aims at providing an integrated, state-of-the-art and timely severe weather nowcast platform for operational forecasters at all levels over China. SWAN ingests data from China's new generation Doppler radars (both S-band and C-band), automatic weather station, satellite, and mesoscale numerical weather prediction model. It offers a tool for severe weather monitoring, analysis, nowcasting and warnings such as flashing a real-time alert, driving next algorithm processes and sending a warning via SMS, etc.

The server application includes several modules, such as providing log files for monitoring system behavior, configuring network environment, setting data acquisition parameters, performing quality control for radar data and AWS data, generating 3D radar reflectivity mosaic, running algorithm for nowcast products, analyzing observation data and providing message for alerting the forecasters.

Doppler Radar Observations – 60 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

1 1-FAR= 1-5% and POD=60-70%

Fig. 18. The processing steps for hazardous wind potential at JMA. It is typical of current systems where mesoscale NWP predictions are assumed to be good enough to match with

In 2008, the China Meteorological Administration (CMA) launched a campaign on the development of its first version of integrated nowcasting system SWAN (Severe Weather Analysis and Nowcast system). This system aims at providing an integrated, state-of-the-art and timely severe weather nowcast platform for operational forecasters at all levels over China. SWAN ingests data from China's new generation Doppler radars (both S-band and C-band), automatic weather station, satellite, and mesoscale numerical weather prediction model. It offers a tool for severe weather monitoring, analysis, nowcasting and warnings such as flashing a real-time alert, driving next algorithm processes and sending a warning

The server application includes several modules, such as providing log files for monitoring system behavior, configuring network environment, setting data acquisition parameters, performing quality control for radar data and AWS data, generating 3D radar reflectivity mosaic, running algorithm for nowcast products, analyzing observation data and providing

2 Success Ratio = 1- FAR = 5- 10% with POD=20-30%

**Warning Level Criteria** 

Table 7. Hazardous Criteria Level

the observations.

via SMS, etc.

**6.12 SWAN – CHINA/CMA** 

message for alerting the forecasters.

The client refreshes real-time observations cycled in 5 min from radars and AWS (automatic weather stations) and provides real-time alerts (sounding, flashing) for indication of severe weather events (meeting certain thresholds such as wind speeds or rainfall amounts). It also provides an interactive tool for preparing, editing and issuing Nowcast and warning for severe weathers

Based on quality control, a regional 3D reflectivity mosaic is produced by trying to fill the gaps that are generated by terrain blockage or AP. Products such as vertically integrated liquid (VIL) , echo top (ET) and COTREC winds are then derived. QPE algorithm involves extraction of convective echoes from stratiform echoes by texture and horizontal gradient properties. Different Z-R relations are used for convective rain and stratiform rain. COTREC (continuous tracking radar echo by correlation) vectors are echo motion vectors that are derived from moving radar reflectivity patterns through grid-to-grid cross-correlation and then adjusted by a horizontal non-divergence constraint for hourly nowcasts of rainfall (Li et al, 1995). This is blended with mesoscale numerical prediction model output for 2-3 hour nowcasts.

SWAN provides real time verifications for storm tracking and reflectivity nowcasts. Storm track errors are shown as distance differences between observed storm tracks and predicted storm tracks (1h). Observed radar reflectivity are also verified against extrapolated forecasted reflectivity.

Severe weather warnings can be prepared and issued through SWAN by graphical interface by circling an area on the screen, clicking an icon and doing some minor wording (Fig. 19). A web-based version of SWAN has been developed and deployed in Guangdong Meteorological Bureau.

Fig. 19. A SWAN display showing cells/tracks (main screen), SCIT (bottom) and time histories of critical parameters (right). There are similarities with WDSS, NinJo and CARDS displays.

Automated Processing of Doppler Radar Data for Severe Weather Warnings 63

offer the possibility for the NHMS to add their own specialized products into their systems. Many of the countries mentioned above in fact use a combination of products from their own systems and those of the manufacturers. Information is readily available in trade shows or on their web sites. The ideal requirement is a seamless, user-friendly integrated visualization, decision-making and production system to cover all scales (the seamless prediction concept) and this is the trend in many NHMS' for all data, products and so radar only processing or visualization systems are an interim step towards this and requires investment, resources, time and effort to achieve. NinJo and AWIPS (not described here) provides an example of how radar is expected to be integrated into a comprehensive

The purpose of this contribution was to illustrate the issues faced by NHMS's. There is a push to use meteorological technology as much as possible and to automate as much as possible. Computing technology is still a limiting factor – computers, telecommunications and data/product storage are all continuing issues that can always be faster and bigger. If there is the time, the resources and the expertise, manual interpretation of basic radar products is still the best way to provide severe weather warning services and to optimally utilize the considerable capabilities of the forecasters. However, tools are needed to streamline and accelerate the process but this is highly dependent on organizational factors. Automated products introduce another level of complexity and knowledge requirement. They can be black boxes that bewilder the user. However, creating black boxes without diagnostic capabilities, providing poor tools and denying access to basic products and information, is self-defeating. It is a sure way of making smart people (appear) "dumb". The algorithms aren't perfect given the need for high POD. They never will be and they can be better and substantial work on data quality, feature detection and prediction are needed. The systems described exhibit the great efforts and resources are expended to do this. Saving a single button click or a mouse movement can make the difference between a bad and a good system. This is difficult to describe as a requirement and prototyping and

While reliable weather radars and expertise play a central role in the warning process, this is still a challenge for many countries. Satellite and lightning systems are now available that have minimal support requirements. Stand alone applications for severe weather can and are being developed for these system. In the absence of radars, there is no question that they will provide benefits but their efficacy, the forecast process and the service level for severe weather warnings need to be demonstrated. No doubt that they should also enhance existing systems that rely on weather radar networks. This is occurring but beyond the scope of this contribution. No convective scale warning service has been soley developed without radar and so this is a new area to investigate. Understanding the technology, interpretation of the data and the products will require more development, enhanced

For the convective weather problem, dual-polarization radar will have benefits in data quality, hail detection and rainfall estimation but this is again beyond the scope of this contribution (Frame et al, 2009). Earth curvature and beam propagation preclude low level detection and so many of the hazardous phenomena are not actually measured beyond a few tens of kilometer from the radar site and must therefore be inferred from measurements aloft. The CASA (Cooperative Adapting and Sensing of the Atmosphere) is a network of X

forecast analysis, diagnosis, prognosis and production tool.

demonstration projects are the only way to appreciate this.

expertise, demonstration and decision-making skills.

#### **7. Conclusion**

The objective of this contribution was to provide a broad overview of the use of radar and radar networks for the provision of severe weather warnings and to very briefly describe historical legacies and current practice. The target audience are those NHMS' who might be contemplating developing or enhancing such a service. Weather radar clearly plays a central role in this application. Not discussed are important applications such as nowcasting precipitation, quantitiative precipitation estimation, wind retrieval, data assimilation for numerical weather prediction, etc. It also does not address the convective initiation aspects (Roberts et al, 2006; Sun et al, 1991; Sun and Crook, 1994). For a reliable warning service, design, infrastructure (reliable power and telecommunications), support and maintenance are critical and were not discussed in this contribution. These are major considerations but out of scope for this contribution.

The level and nature of the service will be determined by both meteorological and nonmeteorological factors. The prevalence of severe weather, climatology and a defining event determine the impact, the exposure and the opportunity to develop a warning service. Socio-economic factors, risk persona, as well as the organizational structure, are particularly important in the design and expectations for the radar processing, visualization and dissemination systems. This contribution provided a short global survey of radar based systems to illustrate the commonality but also the differences in implementation. One solution does not fit all. Underlying these systems is the forecast process and it is emphasized that they all rely on human expertise in the decision-making process and so the human-machince mix is a critical item. This will drive the expertise and therefore the training requirements for the severe weather analyst.

This contribution highlighted the use of automation in the production of guidance products. Some systems rely on very little automation and totally rely on manual interpretation. All systems, except one, default to this mode. One of most highly automated systems is CARDS (Canada). Automation is necessary because of the need for look at details for warning preparation purposes while maintaining situational awareness in the situation where one forecaster is responsible for about ten radars. It processes radar data for identifying and ranking thunderstorm cells and features. It also creates highly processed image products to streamline and to guide the decision-making process. It still relies on human decisionmaking for the final preparation of the warning. KONRAD is the only system that produces totally automated products. However, it could be argued that these products are directed to "sophisticated users" for their specific planning and decision-making purposes and not warning purposes.

Given the limited space and time, all radar processing systems were inadequately described. There is room for improvement in describing all aspects of the processing chain from better algorithms (e.g. hail, hook echoes; Lemon, 1998; Wang et al, 2011) to advanced concepts where thermodynamic diagnostic fields, useful for understanding, are retrieved (Sun et al, 1991; Sun and Crook, 1994). Through the description of specific innovative aspects of individual systems, and since there are commonalities amongst them, the intent was to provide the reader with an overview of the capabilities of all the systems. There is fine work being done elsewhere that is not represented; to name a few, Italy, Switzerland and Finland. Another glaring oversight is the lack of description of systems by manufacturers. Some even Doppler Radar Observations – 62 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

The objective of this contribution was to provide a broad overview of the use of radar and radar networks for the provision of severe weather warnings and to very briefly describe historical legacies and current practice. The target audience are those NHMS' who might be contemplating developing or enhancing such a service. Weather radar clearly plays a central role in this application. Not discussed are important applications such as nowcasting precipitation, quantitiative precipitation estimation, wind retrieval, data assimilation for numerical weather prediction, etc. It also does not address the convective initiation aspects (Roberts et al, 2006; Sun et al, 1991; Sun and Crook, 1994). For a reliable warning service, design, infrastructure (reliable power and telecommunications), support and maintenance are critical and were not discussed in this contribution. These are major considerations but

The level and nature of the service will be determined by both meteorological and nonmeteorological factors. The prevalence of severe weather, climatology and a defining event determine the impact, the exposure and the opportunity to develop a warning service. Socio-economic factors, risk persona, as well as the organizational structure, are particularly important in the design and expectations for the radar processing, visualization and dissemination systems. This contribution provided a short global survey of radar based systems to illustrate the commonality but also the differences in implementation. One solution does not fit all. Underlying these systems is the forecast process and it is emphasized that they all rely on human expertise in the decision-making process and so the human-machince mix is a critical item. This will drive the expertise and therefore the

This contribution highlighted the use of automation in the production of guidance products. Some systems rely on very little automation and totally rely on manual interpretation. All systems, except one, default to this mode. One of most highly automated systems is CARDS (Canada). Automation is necessary because of the need for look at details for warning preparation purposes while maintaining situational awareness in the situation where one forecaster is responsible for about ten radars. It processes radar data for identifying and ranking thunderstorm cells and features. It also creates highly processed image products to streamline and to guide the decision-making process. It still relies on human decisionmaking for the final preparation of the warning. KONRAD is the only system that produces totally automated products. However, it could be argued that these products are directed to "sophisticated users" for their specific planning and decision-making purposes and not

Given the limited space and time, all radar processing systems were inadequately described. There is room for improvement in describing all aspects of the processing chain from better algorithms (e.g. hail, hook echoes; Lemon, 1998; Wang et al, 2011) to advanced concepts where thermodynamic diagnostic fields, useful for understanding, are retrieved (Sun et al, 1991; Sun and Crook, 1994). Through the description of specific innovative aspects of individual systems, and since there are commonalities amongst them, the intent was to provide the reader with an overview of the capabilities of all the systems. There is fine work being done elsewhere that is not represented; to name a few, Italy, Switzerland and Finland. Another glaring oversight is the lack of description of systems by manufacturers. Some even

**7. Conclusion** 

out of scope for this contribution.

warning purposes.

training requirements for the severe weather analyst.

offer the possibility for the NHMS to add their own specialized products into their systems. Many of the countries mentioned above in fact use a combination of products from their own systems and those of the manufacturers. Information is readily available in trade shows or on their web sites. The ideal requirement is a seamless, user-friendly integrated visualization, decision-making and production system to cover all scales (the seamless prediction concept) and this is the trend in many NHMS' for all data, products and so radar only processing or visualization systems are an interim step towards this and requires investment, resources, time and effort to achieve. NinJo and AWIPS (not described here) provides an example of how radar is expected to be integrated into a comprehensive forecast analysis, diagnosis, prognosis and production tool.

The purpose of this contribution was to illustrate the issues faced by NHMS's. There is a push to use meteorological technology as much as possible and to automate as much as possible. Computing technology is still a limiting factor – computers, telecommunications and data/product storage are all continuing issues that can always be faster and bigger. If there is the time, the resources and the expertise, manual interpretation of basic radar products is still the best way to provide severe weather warning services and to optimally utilize the considerable capabilities of the forecasters. However, tools are needed to streamline and accelerate the process but this is highly dependent on organizational factors. Automated products introduce another level of complexity and knowledge requirement. They can be black boxes that bewilder the user. However, creating black boxes without diagnostic capabilities, providing poor tools and denying access to basic products and information, is self-defeating. It is a sure way of making smart people (appear) "dumb". The algorithms aren't perfect given the need for high POD. They never will be and they can be better and substantial work on data quality, feature detection and prediction are needed. The systems described exhibit the great efforts and resources are expended to do this. Saving a single button click or a mouse movement can make the difference between a bad and a good system. This is difficult to describe as a requirement and prototyping and demonstration projects are the only way to appreciate this.

While reliable weather radars and expertise play a central role in the warning process, this is still a challenge for many countries. Satellite and lightning systems are now available that have minimal support requirements. Stand alone applications for severe weather can and are being developed for these system. In the absence of radars, there is no question that they will provide benefits but their efficacy, the forecast process and the service level for severe weather warnings need to be demonstrated. No doubt that they should also enhance existing systems that rely on weather radar networks. This is occurring but beyond the scope of this contribution. No convective scale warning service has been soley developed without radar and so this is a new area to investigate. Understanding the technology, interpretation of the data and the products will require more development, enhanced expertise, demonstration and decision-making skills.

For the convective weather problem, dual-polarization radar will have benefits in data quality, hail detection and rainfall estimation but this is again beyond the scope of this contribution (Frame et al, 2009). Earth curvature and beam propagation preclude low level detection and so many of the hazardous phenomena are not actually measured beyond a few tens of kilometer from the radar site and must therefore be inferred from measurements aloft. The CASA (Cooperative Adapting and Sensing of the Atmosphere) is a network of X

Automated Processing of Doppler Radar Data for Severe Weather Warnings 65

Brooks, H. E., C. A. Doswell III, J. Cooper, 1994: On the Environments of Tornadic and

Brooks, H. E., C. A. Doswell III, L. J. Wicker, 1993: STORMTIPE: A Forecasting Experiment

Brooks, H. E., C. A. Doswell III, R. B. Wilhelmson, 1994: The Role of Midtropospheric Winds

Brovelli, P., S. Sénési, E. Arbogast, P. Cau, S. Cazabat, M. Bouzom, J. Reynaud, 2005:

Brown, R. A., J. M. Janish, V. T. Wood, 2000: Impact of WSR-88D Scanning Strategies on

Brunner, J. C., S. A. Ackerman, A. S. Bachmeier, R. M. Rabin, 2007: A Quantitative Analysis

Burgess, D. W., R. J. Donaldson JR., and P. R. Desrochers, 1993: Tornado detection and

Byko, Z., P. Markowski, Y. Richardson, J. Wurman, E. Adlerman, 2009: Descending

Chisholm, A. J. and J. Renick, 1972: The kinematics of multicell and supercell Alberta

Crane, R. K., 1979: Automatic cell detection and tracking. IEEE., Trans. Geosci. Electron., GE-

Crum, T. D. and R. L. Alberty, 1993: The WSR-88D and the WSR-88D Operational Support

Dance, S., E. Ebert, D. Scurrah, 2010: Thunderstorm Strike Probability Nowcasting, Journal

Dance, S., R. Potts, 2002: Microburst Detection Using Agent Networks, Journal of

Davis, C., N. Atkins, D. Bartels, L. Bosart, M. Coniglio, G. Bryan, W. Cotton, D. Dowell, B.

Dixon, M. and G. Weiner, 1993: TITAN, Thunderstorm Identification, Tracking, Analysis

and Nowcasting - A Radar-based Methodology, JAOT, 10, 785-797.

of Atmospheric and Oceanic Technology, Volume 27, Issue 1) 79-9.

Atmospheric and Oceanic Technology, Volume 19, Issue 5) 646-65.

1994) 606-61.

2000) 90-10.

Issue 3 (September 1993) 352-36.

Review, Volume 122, Issue 1 (January 1994) 126-13.

Volume 22, Issue 4 (August 2007) 853-87.

1 (February 2009) 155-18.

Rep. No. 72-2, 24-31.

(August 2004) 1075-1093

17, 250–262.

1687.

nowcasting system, Météo-France, Toulouse, France, WSN05

Geophys. Monogr., No. 79, Amer. Geophys. Union, 203–221.

Nontornadic Mesocyclones, Weather and Forecasting, Volume 9, Issue 4 (December

Using a Three-Dimensional Cloud Model, Weather and Forecasting, Volume 8,

in the Evolution and Maintenance of Low-Level Mesocyclones, Monthly Weather

Nowcasting thunderstoms with SIGOONS. A significant weather object oriented

Severe Storm Algorithms, Weather and Forecasting, Volume 15, Issue 1 (February

of the Enhanced-V Feature in Relation to Severe Weather, Weather and Forecasting,

warning by radaR. The Tornado: Its Structure, Dynamics, Prediction, and Hazards,

Reflectivity Cores in Supercell Thunderstorms Observed by Mobile Radars and in a High-Resolution Numerical Simulation, Weather and Forecasting, Volume 24, Issue

hailstorms, Alberta Hail Studies, 1972, Reseach Council of Alberta Hail Studies

Facility, Bulletin of the American Meteorological Society, Volume 74, Issue 9, 1669-

Jewett, R. Johns, D. Jorgensen, J. Knievel, K. Knupp, W. C. Lee, G. Mcfarquhar, J. Moore, R. Przybylinski, R. Rauber, B. Smull, R. Trapp, S. Trier, R. Wakimoto, M. Weisman, C. Ziegler, 2004: The Bow Echo and MCV Experiment: Observations and Opportunities, Bulletin of the American Meteorological Society, Volume 85, Issue 8

Band radars that address this issue but it is in early-transitional development (McLaughlin et al, 2009; Ruzanski et al, 2011). It also addresses the issue of rapid or adapting scan strategies (Heinselman et al, 2008) which is being investigated now but beyond the scope of this contribution. In any case, with increasing computing power, telecommunications, additional observations and new technology, these are exciting times.

#### **8. References**


Doppler Radar Observations – 64 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Band radars that address this issue but it is in early-transitional development (McLaughlin et al, 2009; Ruzanski et al, 2011). It also addresses the issue of rapid or adapting scan strategies (Heinselman et al, 2008) which is being investigated now but beyond the scope of this contribution. In any case, with increasing computing power, telecommunications,

AMS, 2001: Expectations Concerning Media Performance during Severe Weather

Albers S., J. McGinley, D. Birkenheuer, and J. Smart 1996: The Local Analysis and Prediction

Andra Jr., D. L., E. M. Quoetone, W. F. Bunting, 2002: Warning Decision Making: The

Bally, J., 2004: The Thunderstorm Interactive Forecast System: Turning Automated

Barnes, L. R., E. C. Gruntfest, M. H. Hayden, D. M. Schultz, C. Benight, 2007: False Alarms

Baumgart, L. A., E. J. Bass, B. Philips, K. Kloesel, 2008: Emergency Management Decision

Bech, J., Vilaclara E., Pineda, N., Rigo, T., Lopez, J., O'Hora, F., Lorente, J., Sempere, D.,

Bieringer, P., P. S. Ray, 1996: Comparison of Tornado Warning Lead Times with and

Black, A. W., W. S. Ashley, 2011: The Relationship between Tornadic and Nontornadic

Branick, M. L., C. A. Doswell III, 1992: Polarity, An Observation of the Relationship between

Breidenbach, J. P., D. H. Kitzmiller, W. E. McGovern, R. E. Saffle, 1995: The Use of

service: description and applications, ERAD, Visby, Sweden, 416-420. Bellon, A. and G. L. Austin, 1978: The Evaluation of Two Years of Real-Time Operation of a

American Meteorological Society, 2001: Volume 82, Issue 4 705-70.

Emergencies (Adopted by the AMS Council 14 January 2001), Bulletin of the

System (LAPS): Analyses of clouds, precipitation, and temperature, Wea.

Relative Roles of Conceptual Models, Technology, Strategy, and Forecaster Expertise on 3 May 1999, Weather and Forecasting, Volume 17, Issue 3 (June 2002)

Thunderstorm Tracks into Severe Weather Warnings, Weather and Forecasting,

and Close Calls: A Conceptual Model of Warning Accuracy, Weather and

Making during Severe Weather, Weather and Forecasting, Volume 23, Issue 6

Fabregas F.X., 2004: The weather radar network of the Catalan meteorological

Short-Term Precipitation Forecasting Procedure (SHARP), Journal of Applied

without NEXRAD Doppler Radar, Weather and Forecasting, Volume 11, Issue 1

Convective Wind Fatalities and Warnings, Weather, Climate, and Society, Volume

Supercell Structure and Lightning Ground-Strike, Weather and Forecasting,

Volumetric Radar Reflectivity Predictors in the Development of a Second-Generation Severe Weather Potential Algorithm, Weather and Forecasting, Volume

additional observations and new technology, these are exciting times.

**8. References** 

Forecasting, 11, 273-287.

(December 2008) 1268-127.

(March 1996) 47-5.

3, Issue 1 (January 2011) 31-4.

Volume 7, Issue 1) 143-14.

10, Issue 2 (June 1995) 369-37.

Volume 19, Issue 1 (February 2004) 64-7.

Meteorology, Volume 17, Issue 12, 1778-1787.

Forecasting, Volume 22, Issue 5 (October 2007) 1140-114.

559-566.


Automated Processing of Doppler Radar Data for Severe Weather Warnings 67

Galway, J. G., 1989: The Evolution of Severe Thunderstorm Criteria within the Weather Service, Weather and Forecasting, Volume 4, Issue 4 (December 1989) 585-59. Gatlin, P. N., S. J. Goodman, 2010: A Total Lightning Trending Algorithm to Identify Severe

Glahn, B., 2005: Tornado-Warning Performance in the Past and Future—Another

Goodman, S. J., D. E. Buechler, P. J. Meyer, 1988: Convective Tendency Images Derived from

Hammer, B., T. W. Schmidlin, 2002: Response to Warnings during the 3 May 1999 Oklahoma

Heinselman, P. L., D. L. Priegnitz, K. L. Manross, T. M. Smith, R. W. Adams, 2008: Rapid

Radar, Weather and Forecasting, Volume 23, Issue 5 (October 2008) 808-824. Hengstebeck, T., D. Heizenreder, P. Joe, P. Lang, 2011: The Mesocyclone Detection

Hermes, L. G., A. Witt, S. D. Smith, D. Klingle-Wilson, D. Morris, G. J. Stumpf, M. D. Eilts,

Hering, A. M., S. Senesi, P. Ambrosetti and I. Bernard-Bouissieres, 2005: Nowcasting

Holleman, I. and H. Beekhuis, 2003: Analysis and correction of dual PRF velocity data,

Honda, Y., M. Nishijima, K. Koizumi, Y. Ohta, K. Tamiya, T. Kawabata and T. Tsuyuki,

James, P. M., S. Treple, D. Heizenreder and B. K. Reichert, 2011: NowCastMIX – A fuzzy

Joe, P., 2009: A First Look at Radar Data Quality for the Beijing 2008 Forecast Demonstration

Climate, and Society, Volume 3, Issue 2 (April 2011) 128-14.

1 (January 2010) 3-22.

(August 2005) 1135-114.

2011, Palma de Mallorca

JAOT, 20(4), 443-453.

Sept 2011., EMS2011-234,

Xaioding Yu (invited lead paper).

Issue 3 (September 1988) 173-188.

Volume 17, Issue 3 (June 2002) 577-58.

Volume 10, Issue 5 (October 1993) 693-70.

Quart. J. Roy. Meteor. Soc., 131, 3465-3475.

Thunderstorms, Journal of Atmospheric and Oceanic Technology, Volume 27, Issue

Perspective, Bulletin of the American Meteorological Society, Volume 86, Issue 8

a Combination of Lightning and Satellite Data, Weather and Forecasting, Volume 3,

City Tornado: Reasons and Relative Injury Rates, Weather and Forecasting,

Sampling of Severe Storms by the National Weather Radar Testbed Phased Array

Algorithm of DWD, 6th European Conference on Severe Storms, ECSS, 3-7 October

1993: The Gust-Front Detection and Wind-Shift Algorithms for the Terminal Doppler Weather Radar System, Journal of Atmospheric and Oceanic Technology,

thunderstorms in complex cases using radar data, WMO Symposium on Nowcasting and Very Short Range Forecasting, Toulouse France, paper 2. 14. Hoekstra, S., K. Klockow, R. Riley, J. Brotzge, H. Brooks, S. Erickson, 2011: A Preliminary

Look at the Social Perspective of Warn-on-Forecast: Preferred Tornado Warning Lead Time and the General Public's Perceptions of Weather Risks, Weather,

2005: A pre-operational variational data assimilation system for a non-hydrostatic model at the Japan Meteorological Agency: Formulation and preliminary results.

logic based tool for providing automatic integrated nowcasting systems, 11th EMS Annual Meeting, 10th European Conference on Applications of Meteorology, 12-16

Project, in Collection of Papers on the New Generation of China Radars, (Xin Yi Dai, Tian Qi Leida Yewu Yingyoung Lunwenji, ISBN 978-7-5029-4468-1), edited by


Doppler Radar Observations – 66 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Donaldson, Jr., Ralph J., P. R. Desrochers, 1990: Improvement of Tornado Warnings by

Doswell III, C. A., 1980: Synoptic-Scale Environments Associated with High Plains Severe

Doswell, C. A. III, 1982: The operational meteorology of convective weather, Volume I: Operational Mesoanalysis, NOAA Tech Memo ERL NSSFC-5, 168p. Doswell, C. A. III, 1985: The operational meteorology of convective weather, Volume II:

Doswell III, C. A., 2001: Severe Convective Storms, Meteorological Monographs, Volume 28,

Doswell III, C. A. 2004: Weather Forecasting by Humans—Heuristics and Decision Making.

Doswell III, C. A., R. Davies-Jones, D. L. Keller, 1990: On Summary Measures of Skill in Rare

Doswell III, C. A., R. Edwards, R. L. Thompson, J. A. Hart, K. C. Crosbie, 2006: A Simple and

Doswell III, C. A., A. R. Moller, H. E. Brooks, 1999: Storm Spotting and Public Awareness

Doviak, R. J., and D. S. Zrnic, 1984: Doppler Radar and weather observations, Academic

Dunn, L. B., 1990: Two Examples of Operational Tornado Warnings Using Doppler Radar

Ebert, E. E., L. J. Wilson, B. G. Brown, P. Nurmi, H. E. Brooks, J. Bally, M. Jaeneke, 2004:

Project, Weather and Forecasting, Volume 19, Issue 1 (February 2004) 73-96 Eilts, M. D., and Coauthors, 1996: Severe weather warning decision support system.

Evans, J. S., C. A. Doswell III, 2001: Examination of Derecho Environments Using Proximity Soundings, Weather and Forecasting, Volume 16, Issue 3 (June 2001) 329-34. Fawbush, E. J. and R. C. Miller, 1953: Forecasting Tornadoes, USAF Air University Quarterly

Fox, Neil I., R. Webb, J. Bally, M. W. Sleigh, C. E. Pierce, D. M. L. Sills, P. I. Joe, J. Wilson, C.

2000, Weather and Forecasting, Volume 19, Issue 1 (February 2004) 97-114 Frame, J., P. Markowski, Y. Richardson, J. Straka, J. Wurman, 2009: Polarimetric and Dual-

Event Forecasting Based on Contingency Tables, Weather and Forecasting, Volume

Flexible Method for Ranking Severe Weather Events, Weather and Forecasting,

since the First Tornado Forecasts of 1948, Weather and Forecasting, Volume 14,

Data, Bulletin of the American Meteorological Society, Volume 71, Issue 2

Verification of Nowcasts from the WWRP Sydney 2000 Forecast Demonstration

Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, AmeR. MeteoR.

G. Collier, 2004: The Impact of Advanced Nowcasting Systems on Severe Weather Warning during the Sydney 2000 Forecast Demonstration Project: 3 November

Doppler Radar Observations of the Lipscomb County, Texas, Supercell Thunderstorm on 23 May 2002, Monthly Weather Review, Volume 137, Issue 2

and Forecasting, Volume 5, Issue 2 (June 1990) 247-25.

Storm Scale Analysis, NOAA Tech Memo ERL ESG-15, 240p.

11 (November 1980) 1388-1400.

Weather and Forecasting 19:6, 1115-112.

Volume 21, Issue 6 (December 2006) 939-95.

5, Issue 4 (December 1990) 576-58.

Issue 4 (August 1999) 544-557.

(February 1990) 145-15.

561pp.

Press, 458p.

Soc., 536–540.

Review, 1, 108-11.

(February 2009) 544-56.

Doppler Radar Measurement of Mesocyclone Rotational Kinetic Energy, Weather

Thunderstorms, Bulletin of the American Meteorological Society, Volume 61, Issue


Automated Processing of Doppler Radar Data for Severe Weather Warnings 69

Lakshmanan, V. and T. Smith, 2010: An Objective Method of Evaluating and Devising Storm-Tracking Algorithms, Weather and Forecasting, Volume 25, 701-709. Lakshmanan, V., T. Smith, K. Hondl, G. J. Stumpf, A. Witt, 2006: A Real-Time, Three-

Lakshmanan, V., T. Smith, G. Stumpf, K. Hondl, 2007: The Warning Decision Support

Lakshmanan, V., J. Zhang, K. Hondl, and C. Langston, 2011: A statistical approach to

Lang, P., 2001: Cell tracking and warning indicators derived from operational radar products, 30th International Radar Conference, AMS, Munich, Germany, 245-247. Lang, T. J., L. J. Miller, M. Weisman, S. A. Rutledge, L. J. Barker III, V. N. Bringi, V.

Applied Earth Observations and Remote Sensing, vol. s, p. accepted. Lakshmanan, V., J. Zhang, and K. Howard, 2010, A technique to censor biological echoes in

radar reflectivity data, J. Applied Meteorology, vol. 49, 435-462.

Conf. Radar Met., Montreal, AMS, 327-330.

SELS, San Antonio, 12-16 August 2002, 170-173.

criteria, NWS NSSFC-1, PB 231409 60p.

Issue 9 (September 1979) 1184-119.

13, Issue 4 (December 1998) 1029-104.

Issue 5, 802-82.

61.

Dimensional, Rapidly Updating, Heterogeneous Radar Merger Technique for Reflectivity, Velocity, and Derived Products, Weather and Forecasting, Volume 21,

System–Integrated Information, Weather and Forecasting, Volume 22, Issue 3, 596-

mitigating persistent clutter in radar reflectivity data, IEEE J. Selected Topics in

Chandrasekar, A. Detwiler, N. Doesken, J. Helsdon, C. Knight, P. Krehbiel, Walter A. Lyons, D. Macgorman, E. Rasmussen, W. Rison, W. D. Rust, Ronald J. T., 2004: The Severe Thunderstorm Electrification and Precipitation Study, Bulletin of the American Meteorological Society, Volume 85, Issue 8 (August 2004) 1107-1125 Lapczak, S., E. Aldcroft, M. Stanley-Jones, J. Scott, P. Joe, P. Van Rijn, M. Falla, A. Gagne, P.

Ford, K. Reynolds and D. Hudak, 1999: The Canadian National Radar Project, 29th

2002: The July 4 2001 Severe Weather Outbreak in Southern Ontario as Diagnosed by the New Radar Data Processing System of the National Radar Project of Canada,

Identification, Tracking, and Forecasting—An Enhanced TITAN Algorithm, Journal

Application to Severe Thunderstorm Study and Potential Real-Time Warning, Bulletin of the American Meteorological Society, Volume 58, Issue 11, 1187-119. Lemon, L. R., C. A. Doswell III, 1979: Severe Thunderstorm Evolution and Mesocyclone

Structure as Related to Tornadogenesis, Monthly Weather Review, Volume 107,

Algorithms along the Northeastern Gulf Coast, Weather and Forecasting, Volume

Leduc, M., P. Joe, M. Falla, P. Van Rijn, S. Lapczak, I. Ruddick, A. Ashton and R. Alsen,

Lei H., S. Fu, L. Zhao, Y. Zheng, H. Wang and Y. Lin, 2009: 3D Convective Storm

of Atmospheric and Oceanic Technology, 2009: Volume 26, Issue 4) 719-73. Lemon, L. R., 1977: new severe thunderstorm radar identification techniques are warning

Lemon, L. R., 1980: new severe thunderstorm radar identification techniques are warning

Lemon, L. R., 1998: The Radar "Three-Body Scatter Spike": An Operational Large-Hail Signature, Weather and Forecasting, Volume 13, Issue 2 (June 1998) 327-34. Lemon, L. R., Ralph J. Donaldson, Jr., D. W. Burgess, R. A. Brown, 1977: Doppler Radar

Lenning, E., H. E. Fuelberg, A. I. Watson, 1998: An Evaluation of WSR-88D Severe Hail

criteria: a preliminary report, NWS NSSFC-1, PB 273049 60p.


Doppler Radar Observations – 68 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Joe, P., D. Burgess, R. Potts, T. Keenan, G. Stumpf, A. Treloar, 2004: The S2K Severe Weather

Joe, P., M. Falla, P. Van Rijn, L. Stamadianos, T. Falla, D. Magosse, L. Ing and J. Dobson,

Joe, P., P. T. May, 2003: Correction of Dual PRF Velocity Errors for Operational Doppler

Johns, R. H., C. A. Doswell III, 1992: Severe Local Storms Forecasting, Weather and

Johnson, J. T., P. L. MacKeen, A. Witt, E. De W. Mitchell, G. J. Stumpf, M. D. Eilts, K. W.

Keenan, T., P. Joe, J. Wilson, C. Collier, B. Golding, D. Burgess, P. May, C. Pierce, J. Bally, A.

Kessler, E., J. W. Wilson, 1971: Radar in an Automated National Weather System, Bulletin of the American Meteorological Society, Volume 52, Issue 11, 1062-106. King, P. W. S., M. J. Leduc, D. M. L. Sills, N. R. Donaldson, D. R. Hudak, P. Joe, B. P.

Canada, SELS, San Antonio, 12-16 August 2002, 221-224.

Forecasting, Volume 7, Issue 4 (December 1992) 588-61.

19, Issue 1, 43-63.

27.

Issue 4 (April 2003) 429-44.

Volume 84, Issue 8, 1041-1054.

2003) 172-188

288-305.

523-37.

forecast, J. Atm. Res., vol. 67, 367-380.

and Atmos. Tech., vol. 26, no. 11, 2353-2365.

Detection Algorithms and Their Performance, Weather and Forecasting, Volume

2002: Radar Data Processing for Severe Weather in the National Radar Project of

Weather Radars, Journal of Atmospheric and Oceanic Technology, Volume 20,

Thomas., 1998: The Storm Cell Identification and Tracking Algorithm: An Enhanced WSR-88D Algorithm, Weather and Forecasting, Volume 13, Issue 2, 263-

Crook, A. Seed, D. Sills, L. Berry, R. Potts, I. Bell, N. Fox, E. Ebert, M. Eilts, K. O'Loughlin, R. Webb, R. Carbone, K. Browning, R. Roberts, C. Mueller, 2004: The Sydney 2000 World Weather Research Programme Forecast Demonstration Project: Overview and Current Status, Bulletin of the American Meteorological Society,

Murphy, 2003: Lake Breezes in Southern Ontario and Their Relation to Tornado Climatology, Weather and Forecasting, Volume 18, Issue 5 (October 2003) 795-807. Kitzmiller, D. H., W. E. McGovern, R. F. Saffle, 1995: The WSR-88D Severe Weather Potential Algorithm, Weather and Forecasting, Volume 10, Issue 1 (March 1995) 141-15. Klingle, D. L., D. R. Smith, M. M. Wolfson, 1987: Gust Front Characteristics as Detected by Doppler Radar, Monthly Weather Review, Volume 115, Issue 5 (May 1987) 905-91. Knupp, K. R., S. Paech, S. Goodman, 2003: Variations in Cloud-to-Ground Lightning

Characteristics among Three Adjacent Tornadic Supercell Storms over the Tennessee Valley Region, Monthly Weather Review, Volume 131, Issue 1 (January

technique to quality control radar reflectivity data, J. Applied Meteorology, vol. 46,

identifying storm cells in geospatial images, J. Ocean. Atmos. Tech., vol. 26, no. 3,

Koppert, H. -J., Pedersen, T. S., Zuercher, B., Joe, P., 2004: How to make an international

Lakshmanan, J. T. ., A. Fritz, T. Smith, K. Hondl, and G. J. Stumpf, 2007: An automated

Lakshmanan, V., K. Hondl, and R. Rabin, 2009: An efficient, general-purpose technique for

Lakshmanan, V., R. Rabin, and V. DeBrunner, 2003: Multiscale storm identification and

Lakshmanan, V. and T. Smith, 2009: Data mining storm attributes from spatial grids, J. Ocea.

Meteorological Workstation project successful, BAMS, 1087-109.


Automated Processing of Doppler Radar Data for Severe Weather Warnings 71

Mercer, A. E., Chad M. Shafer, C. A. Doswell III, Lance M. L., M. B. Richman, 2009: Objective

Mitchell, E. De W., S. V. Vasiloff, G. J. Stumpf, A. Witt, M. D. Eilts, J. T. Johnson, K. W. T.,

Moller, A. R., 1978: The Improved NWS Storm Spotters' Training Program at Ft. Worth,

Moller, A. R., 2001: Severe Local Storms Forecasting, Meteorological Monographs, Volume

Moller, A. R., C. A. Doswell III, M. P. Foster, G. R. Woodall, 1994: The Operational

Moninger, W. R., C. Lusk, W. F. R. s, J. Bullas, B. de Lorenzis, J. C. McLeod, E. Ellison, J.

Monteverdi, J. P., C. A. Doswell III, G. S. Lipari, 2003: Shear Parameter Thresholds for

O'Hora, Fritz and Joan Bech, 2007: Improving weather radar observations using pulse-

Pliske, R., D. Klinger, R. Hutton, B. Crandall, B. Knight, and G. Klein, 1997: Understanding

Polger, P. D., B. S. Goldsmith, R. C. Przywarty, J. R. Bocchieri, 1994: National Weather

Przybylinski, R. W., 1995: The Bow Echo: Observations, Numerical Simulations, and Severe Weather Detection Methods, Weather and Forecasting, Volume 10 (2) 203-21. Rasmussen. E. N., 2003: Refined Supercell and Tornado Forecast Parameters. Weather and

Rasmussen, E. N., J. M. Straka, R. Davies-Jones, C. A. Doswell III, F. H. Carr, M. D. Eilts, D.

Ruzanski, E., V. Chandrasekar, Y. Wang, 2011: The CASA Nowcasting System. Journal of

Roberts, Rita D., D. Burgess, M. Meister, 2006: Developing Tools for Nowcasting Storm

Saffle, R. E., 1976: D/RADEX products and field operation. 17th Conf. on Radar

compression techniques, Meteorological Applications, 14, 389-401.

Meteorological Society, Volume 75, Issue 2 (February 1994) 203-21.

Weather Review, Volume 137, Issue 12 (December 2009) 4355-436.

Weather and Forecasting, Volume 13, Issue 2, 352-36.

Weather and Forecasting, Volume 9, Issue 3, 327-347

Society, Volume 72, Issue 9 (September 1991) 1339-1354.

and Forecasting, Volume 18, Issue 2 (April 2003) 357-37.

(December 1978) 1574-1582

28, Issue 50 (November 2001) 433-480

Laboratory, U. S. Air Force, 122

Forecasting 18:3, 530-53.

Issue 6 (June 1994) 995-100.

Atmospheric and Oceanic Technology, 28, 640-65.

Meteorology, Seattle, AMS, Boston, MA., 555-559.

Severity, Weather and Forecasting, Volume 21, 540-55.

Classification of Tornadic and Nontornadic Severe Weather Outbreaks, Monthly

1998: The National Severe Storms Laboratory Tornado Detection Algorithm,

Tex., Bulletin of the American Meteorological Society, Volume 59, Issue 12

Recognition of Supercell Thunderstorm Environments and Storm Structures,

Flueck, P. D. Lampru, K. C. Young, J. Weaver, R. S. Philips, R. Shaw, T. R. Stewart, S. M. Zubrick, 1991: Shootout-89, A Comparative Evaluation of Knowledge-based Systems That Forecast Severe Weather, Bulletin of the American Meteorological

Forecasting Tornadic Thunderstorms in Northern and Central California, Weather

skilled weather forecasting: Implications for training and the design of forecasting tools. Contractor Rep. AL/HR-CR-1997-003, Material Command, Armstrong

Service Warning Performance Based on the WSR-88D, Bulletin of the American

R. MacGorman, 1994: Verification of the Origins of Rotation in Tornadoes Experiment: VORTEX, Bulletin of the American Meteorological Society, Volume 75,


Doppler Radar Observations – 70 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Li, P.W. 2009: Development of a thunderstorm nowcasting system for Hong Kong

Li, P. W. and W.K. Wong, 2010: Development of an Advanced Aviation Nowcasting System

Li, L., W. Schmid and J. Joss, 1995: Nowcasting of motion and growth of precipitation with

Li, P. W. and S. T. Lai, 2004: Applications of radar-based nowcasting techniques for

Li, P. W. and S. T. Lai, 2004a: Short-range quantitative precipitation forecasting in Hong

Li, P.W., W.K. Wong and E.S.T. Lai, 2005: A New Thunderstorm Nowcasting System in

Markowski, P. M., 2002: Hook Echoes and Rear-Flank Downdrafts: A Review, Monthly

Markowski, P. M., E. N. Rasmussen, J. M. Straka, 1998a: The Occurrence of Tornadoes in

Markowski, P. M., J. M. Straka, E. N. Rasmussen, D. O. Blanchard, 1998b: Variability of

Marshall, J. S. and E. H. Ballantyne, 1975: Weather Surveillance Radar, J. A. M., 14, 1317-

May, P. T., T. D. Keenan, R. Potts, J. W. Wilson, R. Webb, A. Treloar, E. Spark, S. Lawrence,

Processing Issues, Weather and Forecasting, Volume 19, Issue 1, 115-130. McCarthy, J., J. W. Wilson, T. T. Fujita, 1982: The Joint Airport Weather Studies Project, Bulletin of the American Meteorological Society, Volume 63, Issue 1) 15-1. McLaughlin, D., D. Pepyne, B. Philips, J. Kurose, M. Zink, D. Westbrook, E. Lyons, E.

Melnikov, V. M., D. W. Burgess, D. L. Andra JR., M. P. Foster, J. M. Krause, 2005: K. A.

Atlanta, Georgia, USA, 17-21 January 2010.

Kong, Journal of Hydrology, 288, 189-209.

Issue 11 (November 1998) 2959-297.

radar over a complex orography, JAM, 34(6), 1286-1300

short-range Forecasting, Toulouse, France, 5-9 Sep. 2005.

Weather Review, Volume 130, Issue 4 (April 2002) 852-87.

Forecasting, Volume 13, Issue 3 (September 1998) 852-85.

Meteorological Society, Volume 90, Issue 12, 1797-1817.

Decision Making, Weather and Forecasting, Volume 20, 775-78.

15 Jan 2009.

264.

1338.

International Airport, AMS Aviation, Range, Aerospace Meteorology Special Symposium on Weather-Air Traffic Management Integration, Phoenix, Arizona, 11-

by Including Rapidly Updated NWP Model in Support of Air Traffic Management, Proceedings 14th Conference on Aviation, Range and Aerospace Meteorology,

mesoscale weather forecasting in Hong Kong, Meteorological Applications, 11, 253-

Hong Kong, WMO/WWRP International Symposium on Nowcasting and Very-

Supercells Interacting with Boundaries during VORTEX-95, Weather and

Storm-Relative Helicity during VORTEX, Monthly Weather Review, Volume 126,

E. Ebert, J. Bally, P. Joe, 2004: The Sydney 2000 Olympic Games Forecast Demonstration Project: Forecasting, Observing Network Infrastructure, and Data

Knapp, A. Hopf, A. Defonzo, R. Contreras, T. Djaferis, E. Insanic, S. Frasier, V. Chandrasekar, F. Junyent, N. Bharadwaj, Y. Wang, Y. Liu, B. Dolan, K. Droegemeier, J. Brotzge, M. Xue, K. Kloesel, K. Brewster, F. Carr, S. Cruz-Pol, K. Hondl, P. Kollias, 2009: Short-Wavelength Technology and the Potential For Distributed Networks of Small Radar Systems, Bulletin of the American

Scharfenberg, D. J. Miller, Terry J. Schuur, P. T. Schlatter, Scott E. Giangrande, 2005: The Joint Polarization Experiment: Polarimetric Radar in Forecasting and Warning


Automated Processing of Doppler Radar Data for Severe Weather Warnings 73

Wasula, A. C., L. F. Bosart, K. D. LaPenta, 2002: The Influence of Terrain on the Severe

Weather and Forecasting, Volume 17, Issue 6 (December 2002) 1277-1289. Weiss, S. J., C. A. Doswell III, F. P. Ostby, 1980: Comments on Automated 12–36 Hour

Weisman, M. L., 2001: Bow Echoes: A Tribute to T. T. Fujita, Bulletin of the American

Weisman, M. L., J. B. Klemp, 1984: The Structure and Classification of Numerically

Weisman, M. L., R. Rotunno, 2000: The Use of Vertical Wind Shear versus Helicity in

Weisman, M. L., R. Rotunno, 2004: "A Theory for Strong Long-Lived Squall Lines"

Westefeld, J. S., A. R. Less, T. Ansley, H.S. Yi, 2006: Severe-Weather Phobia, Bulletin of the

Wilson, J. W., E. A. Brandes, 1979: Radar Measurement of Rainfall—A Summary, Bulletin of

Wilson, J., R. Carbone, H. Baynton, R. Serafin, 1980: Operational Application of

Wilson, J. W., N. A. Crook, C. K. Mueller, J. Sun, M. Dixon, 1998: Nowcasting

Wilson, J. W., E. E. Ebert, T. R. Saxen, R. D. Robertss, C. K. Mueller, M. Sleigh, C. E. Pierce,

Wilson, J. W., Y. Feng, Min Chen, Rita D. R. s, 2010: Nowcasting Challenges during the

Wilson, J. W., J. A. Moore, G. B. Foote, B. Martner, T. Uttal, J. M. Wilczak, A. R. Rodi, 1988:

Wilson, J. W., R. M. Wakimoto, 2001: The Discovery of the Downburst: T. T. Fujita's Contribution, Bulletin of the American Meteorological Society, Volume 82, 49-6. Winston, H. A., 1998: A Comparison of Three Radar-Based Severe-Storm-Detection

Winston, H. A., L. J. Ruthi, 1986: Evaluation of RADAP II Severe-Storm-Detection Algorithms, Bulletin of the American Meteorological Society, Volume 67, 145-15. Witt, A., M. D. Eilts, G. J. Stumpf, J. T. Johnson, E. De W. Mitchell, K. W. T., 1998a: An

Weather Review, Volume 112, Issue 12 (December 1984) 2479-2498

American Meteorological Society, 2006: Volume 87, 747-74.

the American Meteorological Society, Volume 60, 1048-105.

Nowcasting, Weather and Forecasting, Volume 19, 131-15.

Systems, Weather and Forecasting, Volume 25, 1691-171.

American Meteorological Society, Volume 69, 1328-134.

Applied MeteorologyVolume 19, Issue 11, 1328-1333.

Meteorological Society, Volume 82, 97-116

Issue 9 (May 2000) 1452-1472

Volume 61, 1154-116.

Volume 79, 2079-209.

Volume 3, 131-140.

Volume 13, 286-30.

361-382

Weather Distribution across Interior Eastern New York and Western New England,

Probability Forecasts of Thunderstorms and Severe Local Storms, Journal of

Simulated Convective Stormsin Directionally Varying Wind Shears, Monthly

Interpreting Supercell Dynamics, Journal of the Atmospheric Sciences, Volume 57,

Revisited, Journal of the Atmospheric Sciences, Volume 61, Issue 4 (February 2004)

Meteorological Doppler Radar, Bulletin of the American Meteorological Society,

Thunderstorms: A Status Report, Bulletin of the American Meteorological Society,

A. Seed, 2004: Sydney 2000 Forecast Demonstration Project: Convective Storm

Beijing Olympics: Successes, Failures, and Implications for Future Nowcasting

Convection Initiation and Downburst Experiment (CINDE), Bulletin of the

Algorithms on Colorado High Plains Thunderstorms, Weather and Forecasting,

Enhanced Hail Detection Algorithm for the WSR-88D, Weather and Forecasting,


Doppler Radar Observations – 72 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Schaefer, J. T., 1990: The Critical Success Index as an Indicator of Warning Skill, Weather

Schmeits, M. J., Kees J. Kok, D. H. P. Vogelezang, R. M. van Westrhenen, 2008: Probabilistic

Schultz, C. J., W. A. Petersen, L. D. Carey, 2011: Lightning and Severe Weather: A

Schumacher, R. S., D. T. Lindsey, A. B. Schumacher, J. Braun, S. D. Miller, J. L. Demuth,

Serafin, R. J., J. W. Wilson, 2000: Operational Weather Radar in the United States: Progress

Sills, D. M. L., J. W. Wilson, P. I. Joe, D. W. Burgess, R. M. Webb, N. I. Fox, 2004: The 3

Smith, P. L., 1999: Effects of Imperfect Storm Reporting on the Verification of Weather Warnings, Bulletin of the American Meteorological Society, Volume 80, 1099-110. Stensrud, D. J., L. J. Wicker, K. E. Kelleher, M. Xue, M. P. Foster, J. T. Schaefer, R. S.

Stumpf, G. J., A. Witt, E. DeW. Mitchell, P. Spencer, J. T. Johnson, M. D. Eilts, K. W. Thomas,

Sun, J., and N. A. Crook, 1994: Wind and thermodynamic retrieval from single-Doppler

Sun, J., D. Flicker, and D. K. Lilly, 1991: Recovery of three-dimensional wind and

Uyeda, H. and D. S. Zrnic, 1986: Automatic Detection of Gust Fronts, Journal of

Vasiloff, S. V., 2001: Improving Tornado Warnings with the Federal Aviation

Wang, H. K., R. Mercer, J. Baron and P. Joe, 2011: Skeleton-based hook echo detection in

Wang, J. J., T. Keenan, P. Joe, J. Wilson, E. S. T. Lai, F. Liang, Y. Wang, B. Ebert, Q. Ye, J.

Atmospheric and Oceanic Technology, Volume 3, 36-5.

Meteorological Society, Volume 82, 861-87.

radar reflectivity data, J. Tech, submitted.

and Opportunity, Bulletin of the American Meteorological Societ.

Forecasting, Volume 26, Issue 5 (October 2011) 744-75.

Forecasts of (Severe) Thunderstorms for the Purpose of Issuing a Weather Alarm in the Netherlands, Weather and Forecasting, Volume 23, Issue 6 (December 2008)

Comparison between Total and Cloud-to-Ground Lightning Trends, Weather and

2010: Multidisciplinary Analysis of an Unusual Tornado: Meteorology, Climatology, and the Communication and Interpretation of Warnings, Weather

November Tornadic Event during Sydney 2000: Storm Evolution and the Role of Low-Level Boundaries, Weather and Forecasting, Volume 19, Issue 1, February

Schneider, S. G. Benjamin, S. Weygandt, J. T. Ferree, J. P. Tuell, 2009: Convective-Scale Warn-on-Forecast System, Bulletin of the American Meteorological Society,

D. W. Burgess, 1998: The National Severe Storms Laboratory Mesocyclone Detection Algorithm for the WSR-88D, Weather and Forecasting, Volume 13, 304-

measurements of a gust front observed during Phoenix-II. Mon. Wea. Rev., 122,

temper¬ature from simulated single-Doppler radar data. J. Atmos. Sci., 48, 876-890.

Administration's Terminal Doppler Weather Radar, Bulletin of the American

Bally, A. Seed, M. X. Chen, J. Xue, B. Conway, 2010: Overview of the Beiijing 2008 Olympics Forecast Demonstration Project, China Meteorological Press, 145pp.

and Forecasting, Volume 5, 570-57.

and Forecasting, Volume 25, 1412-142.

1253-126.

2004) 22-42

32.

1075-1091.

Volume 90, 1487-149.


**3** 

*1China 2USA* 

**Aviation Applications of Doppler Radars in the** 

Doppler radars are indispensable nowadays in the assurance of aviation safety. In particular, many airports in the world are equipped with Terminal Doppler Weather Radar (TDWR) in the alerting of low-level windshear and turbulence. The microburst alerts from certain TDWR are taken as "sky truth" and the aircraft may not fly when microburst alerts are in force.

This chapter summarizes some recent developments on the aviation applications of TDWR in Hong Kong (Figure 1). It first starts with a case study of a typical event of microburst alert associated with severe thunderstorms. The applications of TDWR in the alerting of windshear and turbulence are then described, namely, in the calculation of windshear hazard factor using the radial velocity data from the radar, and the calculation of eddy dissipation rate based on the spectrum width data of the radar. It is hoped that this chapter could serve as an introduction to the aviation applications of TDWR, for the reference of the

**2. A typical microburst event leading to missed approaches of aircraft** 

change in the prevailing winds from southwesterlies to easterlies.

The missed approaches at the Hong Kong International Airport (HKIA) took place during the overnight period of 8 to 9 September 2010 when intense thunderstorm activity brought heavy rain and frequent lightning to the whole Hong Kong. During the period, an intense rain band with north-south orientation swept from east to west across Hong Kong. More than 50 millimeters of rainfall in an hour were generally recorded over the territory and a record-breaking number of 13,102 cloud-to-ground lightning strokes were registered during the hour just after midnight. When the thunderstorms edged close to the HKIA which is situated at the western part of the territory, gusty strong easterlies from the downdraft of the thunderstorm first affected the flight paths east of the airport resulting in an abrupt

Two flights, which tried to land as the thunderstorms approached HKIA, aborted landing and diverted to Macao eventually. Both flights approached the HKIA from the east under the prevailing southwesterly winds (Figure 2). The first aircraft went around twice. The first

**1. Introduction** 

weather services of other airports.

**Alerting of Windshear and Turbulence** 

P.W. Chan1 and Pengfei Zhang2 *1Hong Kong Observatory, Hong Kong, 2University of Oklahoma, Norman, OK,* 


Guide/CIMO%20Guide%207th%20Edition,%202008/Part%20II/Chapter%209. pdf


### **Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence**

P.W. Chan1 and Pengfei Zhang2 *1Hong Kong Observatory, Hong Kong, 2University of Oklahoma, Norman, OK, 1China 2USA* 

#### **1. Introduction**

Doppler Radar Observations – 74 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Witt, A., M. D. Eilts, G. J. Stumpf, E. De W. Mitchell, J. T. Johnson, K. W. T., 1998b:

WMO, 2008: Commission on Instruments, Methods and Observations Guide, Chapter 9, 7th

Wong, W. K., L. H. Y. Yeung, Y. C. Wang & M. Chen, 2009: Towards the Blending of NWP

Wong K. Y., C. L. Yip and P. W. Li, 2007: A Novel Algorithm for Automatic Tropical

Yeung, L. H. Y., E. S. T. Lai & P. K. Y. Chan, 2008: Thunderstorm Downburst and Radar-

Yeung, L. H. Y., E. S. T. Lai & S. K. S. Chiu, 2007: Lightning Initiation and Intensity

Zipser, E. J., C. Liu, D. J. Cecil, S. W. Nesbitt, D. P. Yorty, 2006: Where are the most intense

Zrnic, D. S., D. W. Burgess, L. D. Hennington, 1985: Automatic Detection of Mesocyclonic

Meteorology and Hydrology, Helsinki, Finland 30 June - 4 July 200. Yeung, L. H. Y., W. K. Wong, P. K. Y. Chan & E. S. T. Lai, 2009: Applications of the Hong

http://www.wmo.int/pages/prog/www/IMOP/publications/CIMO-

Nowcasting, 30 Aug-4 Sep 2009, Whistler, B. C., Canada.

thunderstorms on Earth?, BAMS, 87 (8), 1057-1071.

Weather and Forecasting, Volume 13, 513-51.

edition, available at

49-5.

B. C., Canada.

August 2007.

Volume 2, 425-438.

Evaluating the Performance of WSR-88D Severe Storm Detection Algorithms,

Guide/CIMO%20Guide%207th%20Edition,%202008/Part%20II/Chapter%209. pdf

with Nowcast — Operation Experience in B08FDP, WMO Symposium on

Cyclone Eye Fix Using Doppler Weather Radar, Meteorological Applications, 14,

based Nowcasting of Squalls, the Fifth European Conference on Radar in

Kong Observatory nowcasting system SWIRLS-2 in support of the 2008 Beijing Olympic Games. WMO Symposium on Nowcasting, 30 Aug-4 Sep 2009, Whistler,

Nowcasting Based on Isothermal Radar Reflectivity – A Conceptual Model, the 33rd International Conference on Radar Meteorology, Cairns, Australia, 6-10

Shear with Doppler Radar, Journal of Atmospheric and Oceanic Technology,

Doppler radars are indispensable nowadays in the assurance of aviation safety. In particular, many airports in the world are equipped with Terminal Doppler Weather Radar (TDWR) in the alerting of low-level windshear and turbulence. The microburst alerts from certain TDWR are taken as "sky truth" and the aircraft may not fly when microburst alerts are in force.

This chapter summarizes some recent developments on the aviation applications of TDWR in Hong Kong (Figure 1). It first starts with a case study of a typical event of microburst alert associated with severe thunderstorms. The applications of TDWR in the alerting of windshear and turbulence are then described, namely, in the calculation of windshear hazard factor using the radial velocity data from the radar, and the calculation of eddy dissipation rate based on the spectrum width data of the radar. It is hoped that this chapter could serve as an introduction to the aviation applications of TDWR, for the reference of the weather services of other airports.

#### **2. A typical microburst event leading to missed approaches of aircraft**

The missed approaches at the Hong Kong International Airport (HKIA) took place during the overnight period of 8 to 9 September 2010 when intense thunderstorm activity brought heavy rain and frequent lightning to the whole Hong Kong. During the period, an intense rain band with north-south orientation swept from east to west across Hong Kong. More than 50 millimeters of rainfall in an hour were generally recorded over the territory and a record-breaking number of 13,102 cloud-to-ground lightning strokes were registered during the hour just after midnight. When the thunderstorms edged close to the HKIA which is situated at the western part of the territory, gusty strong easterlies from the downdraft of the thunderstorm first affected the flight paths east of the airport resulting in an abrupt change in the prevailing winds from southwesterlies to easterlies.

Two flights, which tried to land as the thunderstorms approached HKIA, aborted landing and diverted to Macao eventually. Both flights approached the HKIA from the east under the prevailing southwesterly winds (Figure 2). The first aircraft went around twice. The first

Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence 77

aborted landing was due to technical consideration. In the second approach at around 00:08 HKT (=UTC + 8 hours), it encountered strong tailwind. Landing was subsequently aborted and the aircraft diverted to Macao thereafter. Four minutes later, the second aircraft followed the same glide path of the first aircraft but also failed to land at the HKIA because of the same

Flight data retrieved from the flight data recorders of the two aircraft was analyzed to reveal the meteorological conditions encountered by aircrafts. It appeared that the missed approach was attributable to the strong tailwind which exceeded the airline pre-defined

According to the flight data, the first aircraft experienced more than 15 knots tailwind after it descended to below 1600 feet (Figure 3(a)) in its second approach. The tailwind increased from 25 knots when the aircraft descended to 780 feet (labeled 'A' in Figure 3(a)) and strengthened to 37 knots at 708 feet at 00:08 HKT (labeled 'B' in Figure 3(a)), which far exceed the limit for tailwind landing. As a result, diversion to other airport was

The second aircraft also experienced the tailwind of around 15 knots when it descended to around 1600 feet. The tailwind increased and reached 19 knots when the aircraft descended to 1423 feet (labeled 'C' in Figure 3(b)) but then decreased and fluctuated between 7 to 12 knots when the aircraft further descended to 1028 feet (labeled 'D' in Figure 3(b)). At around 00:12 HKT, the tailwind started to strengthen again and exceeded 15 knots. The maximum tailwind experienced by the aircraft was 22 knots, which also exceeded the limit for tailwind landing, at 859 feet above the runway (labeled 'E' in Figure 3(b)). Similar to the first aircraft, the second aircraft executed a missed approach due to the strong tailwind and

The TDWR also captured the wind conditions when the two aircraft conducted missed approaches. Figures 4(a) and 4(b) showed the radial velocity measured by TDWR at 0008 HKT and 0012 HKT 9 September respectively. Gusts reaching 27 m/s (i.e. around 50 knots) were captured by the TDWR over the eastern part of the HKIA. The zero isotach, which marked the leading edge of the shear line, agreed well with that identified based on

The HKO Windshear and Turbulence Alerting System (WTWS) integrates windshear and turbulence alerts generated by different algorithms such as Anemometer-based Windshear Alerting Rules-Enhanced (AWARE) (Lee, 2004), LIDAR Windshear Alerting System (LIWAS) (Shun and Chan, 2008), TDWR alerts and other algorithms. Alerts are then generated for 8 runway corridors (north runway and south runway have two arrival and

At 0008 HKT, the zero isotach over the HKIA detected by the TDWR was analyzed as a gust front and was shown on the WTWS display (Figure 5(a)). In addition, microburst alerts, which represent windshear loss of 30 knots or more with precipitation, were provided by TDWR to the east of the HKIA; windshear alerts were generated from AWARE over the runways; turbulence alerts were in force due to the thunderstorm to the north of the HKIA. Over the 8 corridors of the HKIA, all had windshear alerts with magnitude ranging from +25 to +30 knots. At 0012 HKT, although the gust front was not detected by the TDWR

two departure corridors each) and shown on a graphical display, the WTWS display.

reason, i.e. the strong tailwind. The aircraft was also diverted to Macao at 00:12 HKT.

threshold, namely 15 knots for tailwind landing.

conducted.

was diverted to Macao.

anemometer data.

Fig. 1. The locations of the Hong Kong TDWR (red dot) radar and Hong Kong International Airport (HKIA). The blue beams illustrate the radar beams over the runways corridor 07LA of the airport with 1o azimuth interval. Three yellow lines indicate the approach paths and their names are marked.

Fig. 2. Flight paths of the two aircraft which had to conduct missed approach. Red line indicated the flight path for the first aircraft and yellow for the second aircraft. Orange wind barbs showed the locations of aircraft when tailwind was encountered. The 1st and 2nd aircraft recorded tailwind of 37 and 22 knots respectively.

Doppler Radar Observations – 76 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

Fig. 1. The locations of the Hong Kong TDWR (red dot) radar and Hong Kong International Airport (HKIA). The blue beams illustrate the radar beams over the runways corridor 07LA of the airport with 1o azimuth interval. Three yellow lines indicate the approach paths and

Fig. 2. Flight paths of the two aircraft which had to conduct missed approach. Red line indicated the flight path for the first aircraft and yellow for the second aircraft. Orange wind barbs showed the locations of aircraft when tailwind was encountered. The 1st and 2nd

aircraft recorded tailwind of 37 and 22 knots respectively.

their names are marked.

aborted landing was due to technical consideration. In the second approach at around 00:08 HKT (=UTC + 8 hours), it encountered strong tailwind. Landing was subsequently aborted and the aircraft diverted to Macao thereafter. Four minutes later, the second aircraft followed the same glide path of the first aircraft but also failed to land at the HKIA because of the same reason, i.e. the strong tailwind. The aircraft was also diverted to Macao at 00:12 HKT.

Flight data retrieved from the flight data recorders of the two aircraft was analyzed to reveal the meteorological conditions encountered by aircrafts. It appeared that the missed approach was attributable to the strong tailwind which exceeded the airline pre-defined threshold, namely 15 knots for tailwind landing.

According to the flight data, the first aircraft experienced more than 15 knots tailwind after it descended to below 1600 feet (Figure 3(a)) in its second approach. The tailwind increased from 25 knots when the aircraft descended to 780 feet (labeled 'A' in Figure 3(a)) and strengthened to 37 knots at 708 feet at 00:08 HKT (labeled 'B' in Figure 3(a)), which far exceed the limit for tailwind landing. As a result, diversion to other airport was conducted.

The second aircraft also experienced the tailwind of around 15 knots when it descended to around 1600 feet. The tailwind increased and reached 19 knots when the aircraft descended to 1423 feet (labeled 'C' in Figure 3(b)) but then decreased and fluctuated between 7 to 12 knots when the aircraft further descended to 1028 feet (labeled 'D' in Figure 3(b)). At around 00:12 HKT, the tailwind started to strengthen again and exceeded 15 knots. The maximum tailwind experienced by the aircraft was 22 knots, which also exceeded the limit for tailwind landing, at 859 feet above the runway (labeled 'E' in Figure 3(b)). Similar to the first aircraft, the second aircraft executed a missed approach due to the strong tailwind and was diverted to Macao.

The TDWR also captured the wind conditions when the two aircraft conducted missed approaches. Figures 4(a) and 4(b) showed the radial velocity measured by TDWR at 0008 HKT and 0012 HKT 9 September respectively. Gusts reaching 27 m/s (i.e. around 50 knots) were captured by the TDWR over the eastern part of the HKIA. The zero isotach, which marked the leading edge of the shear line, agreed well with that identified based on anemometer data.

The HKO Windshear and Turbulence Alerting System (WTWS) integrates windshear and turbulence alerts generated by different algorithms such as Anemometer-based Windshear Alerting Rules-Enhanced (AWARE) (Lee, 2004), LIDAR Windshear Alerting System (LIWAS) (Shun and Chan, 2008), TDWR alerts and other algorithms. Alerts are then generated for 8 runway corridors (north runway and south runway have two arrival and two departure corridors each) and shown on a graphical display, the WTWS display.

At 0008 HKT, the zero isotach over the HKIA detected by the TDWR was analyzed as a gust front and was shown on the WTWS display (Figure 5(a)). In addition, microburst alerts, which represent windshear loss of 30 knots or more with precipitation, were provided by TDWR to the east of the HKIA; windshear alerts were generated from AWARE over the runways; turbulence alerts were in force due to the thunderstorm to the north of the HKIA. Over the 8 corridors of the HKIA, all had windshear alerts with magnitude ranging from +25 to +30 knots. At 0012 HKT, although the gust front was not detected by the TDWR

Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence 79

(a)

(b) Fig. 4. Velocity measured by TDWR on 9 September 2010. The cool/warm colors represent winds towards/away from the TDWR. Area with gusts reaching 27 m/s was circled in black. The zero isotach (gust front) was in purple. (a) TDWR image at 0008 HKT; (b) TDWR images at 0012 HKT. The zero isotach (gust front) moved westwards to the western end of

HKIA.

(Figure 5(b)) any more, using the surface anemometers and TDWR base data, windshear alerts with magnitude ranging from +15 to +25 knots were issued for the four western corridors. Meanwhile, areas with the microburst alerts shifted westwards and affected the eastern corridors. WTWS issued microburst alerts of -35 knots to the four eastern corridors. During the event, the WTWS functioned properly and was able to provide adequate warning to the aircraft of the windshear to be expected due to the thundery weather.

Fig. 3. Time series in HKT of tailwind in knots (red square) and aircraft altitude in feet (blue diamond) retrieved from the flight data recorders. (a) Flight data for the first aircraft. Tailwind reached 37 knots at 00:08 HKT. (b) Flight data for the second aircraft.

Doppler Radar Observations – 78 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

(Figure 5(b)) any more, using the surface anemometers and TDWR base data, windshear alerts with magnitude ranging from +15 to +25 knots were issued for the four western corridors. Meanwhile, areas with the microburst alerts shifted westwards and affected the eastern corridors. WTWS issued microburst alerts of -35 knots to the four eastern corridors. During the event, the WTWS functioned properly and was able to provide adequate

(a)

(b) Fig. 3. Time series in HKT of tailwind in knots (red square) and aircraft altitude in feet (blue diamond) retrieved from the flight data recorders. (a) Flight data for the first aircraft. Tailwind reached 37 knots at 00:08 HKT. (b) Flight data for the second aircraft.

warning to the aircraft of the windshear to be expected due to the thundery weather.

(a)

Fig. 4. Velocity measured by TDWR on 9 September 2010. The cool/warm colors represent winds towards/away from the TDWR. Area with gusts reaching 27 m/s was circled in black. The zero isotach (gust front) was in purple. (a) TDWR image at 0008 HKT; (b) TDWR images at 0012 HKT. The zero isotach (gust front) moved westwards to the western end of HKIA.

Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence 81

In aviation meteorology, windshear refers to a sustained change of wind speed and/or wind direction that causes the aircraft to deviate from the intended flight path. Low-level windshear (below 1600 feet) could be hazardous to the arriving/departing aircraft. Hong Kong is situated in a subtropical coastal area and it is common to have intense convective weather in the spring and summer. To alert low-level windshear associated with microburst and gust front, a TDWR is operated by the Hong Kong Observatory (HKO) in the vicinity of HKIA (Figure 1). It is a C-band radar with 0.5-degree half-power beam width scanning over the airport and determines convergence/divergence features along the runway orientation from the Doppler velocities. Windshear alerts are generated when the velocity change is 15

Another index that quantifies the windshear threat is the F-factor (Proctor et al., 2000). It is based on the fundamentals of flight mechanics and the understanding of windshear phenomena. The F-factor could also be calculated from the Quick Access Recorder (QAR) data recorded on the commercial jets (Haverdings, 2000). In this study, an attempt is made to calculate F-factor for some typical microburst events at HKIA based on the TDWR measurements and the results are compared with the F-factor determined from the QAR

F-factor is calculated from TDWR's radial velocity data in two steps. First of all, convergence/divergence features are identified from the TDWR data. Then F-factor is determined from each convergence/divergence feature by assuming a wind field model of

To compute convergence/divergence features, the method described in Merritt (1987) is adopted. The TDWR microburst detection algorithm identifies microburst by searching for significant velocity difference along a radial in a search window of 4 range gates (4 x 150 metres per gate = 600 metres in length, and one degree in azimuth). If the windshear along a search window is divergent (i.e. radial wind generally increases with increasing distance from the radar), the search window is taken to be a divergence shear segment. Likewise,

Two divergence/convergence segments are associated as a divergence/convergence shear features if their minimum overlap in range is 0.5 km or if their maximum angular spacing is 2 degrees azimuth. A divergence/convergence region contains at least 4 shear segments with a minimum length of 0.95 km and a minimum area of 1 km2. Moreover, the maximum velocity difference among the shear segments inside a divergence region should be at least 5 m/s. As such, the shear within a divergence region is at least 5 m/s per 600 m, i.e. 0.008

F-factor is related to the total aircraft energy and its rate of change, and is defined to be:

*<sup>W</sup> <sup>w</sup> <sup>F</sup>*

where *Wx* is the component of atmospheric wind directed horizontally along the flight path

*a*

its rate of change, *g* the acceleration due to gravity, *w* the updraft of

*g V* (1)

**3. Windshsear hazard factor based on TDWR** 

microburst. The two steps are briefly described below.

convergence shear segment is also identified.

*<sup>x</sup>*

knots or more.

data.

m/s/m.

(direction *x*) and *Wx*

(b) Fig. 5. WTWS display on 9 September 2010. Gust front analyzed by TDWR (purple line) over the HKIA; microburst alerts generated by TDWR (red solid band-aids); windshear alerts generated by AWARE (red hollow rectangles), by TDWR (red hollow irregular polygons); by LIDAR (red arrows, over the runways only); turbulence alert generated by TDWR (brown polygon with dots). Black numbers were the windshear magnitude in knots. (a) 0008 HKT on 9 September 2010. A gust front was over the HKIA. Windshear alerts were issued by the WTWS for all runway corridors. LIDAR data was highly attenuated by precipitation and could only detect windshear over the runway. (b) 0012 HKT on 9 September 2010. Microburst alerts of -35 knots were issued to the four eastern corridors. Windshear alerts with magnitude ranging from +15 to +25 knots were issued for the four western corridors.

#### **3. Windshsear hazard factor based on TDWR**

Doppler Radar Observations – 80 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

(a)

(b) Fig. 5. WTWS display on 9 September 2010. Gust front analyzed by TDWR (purple line) over the HKIA; microburst alerts generated by TDWR (red solid band-aids); windshear alerts generated by AWARE (red hollow rectangles), by TDWR (red hollow irregular polygons); by LIDAR (red arrows, over the runways only); turbulence alert generated by TDWR (brown polygon with dots). Black numbers were the windshear magnitude in knots. (a) 0008 HKT on 9 September 2010. A gust front was over the HKIA. Windshear alerts were issued by the WTWS for all runway corridors. LIDAR data was highly attenuated by precipitation and could only detect windshear over the runway. (b) 0012 HKT on 9 September 2010. Microburst alerts of -35 knots were issued to the four eastern corridors. Windshear alerts with magnitude ranging from +15 to +25 knots were issued for the four

western corridors.

In aviation meteorology, windshear refers to a sustained change of wind speed and/or wind direction that causes the aircraft to deviate from the intended flight path. Low-level windshear (below 1600 feet) could be hazardous to the arriving/departing aircraft. Hong Kong is situated in a subtropical coastal area and it is common to have intense convective weather in the spring and summer. To alert low-level windshear associated with microburst and gust front, a TDWR is operated by the Hong Kong Observatory (HKO) in the vicinity of HKIA (Figure 1). It is a C-band radar with 0.5-degree half-power beam width scanning over the airport and determines convergence/divergence features along the runway orientation from the Doppler velocities. Windshear alerts are generated when the velocity change is 15 knots or more.

Another index that quantifies the windshear threat is the F-factor (Proctor et al., 2000). It is based on the fundamentals of flight mechanics and the understanding of windshear phenomena. The F-factor could also be calculated from the Quick Access Recorder (QAR) data recorded on the commercial jets (Haverdings, 2000). In this study, an attempt is made to calculate F-factor for some typical microburst events at HKIA based on the TDWR measurements and the results are compared with the F-factor determined from the QAR data.

F-factor is calculated from TDWR's radial velocity data in two steps. First of all, convergence/divergence features are identified from the TDWR data. Then F-factor is determined from each convergence/divergence feature by assuming a wind field model of microburst. The two steps are briefly described below.

To compute convergence/divergence features, the method described in Merritt (1987) is adopted. The TDWR microburst detection algorithm identifies microburst by searching for significant velocity difference along a radial in a search window of 4 range gates (4 x 150 metres per gate = 600 metres in length, and one degree in azimuth). If the windshear along a search window is divergent (i.e. radial wind generally increases with increasing distance from the radar), the search window is taken to be a divergence shear segment. Likewise, convergence shear segment is also identified.

Two divergence/convergence segments are associated as a divergence/convergence shear features if their minimum overlap in range is 0.5 km or if their maximum angular spacing is 2 degrees azimuth. A divergence/convergence region contains at least 4 shear segments with a minimum length of 0.95 km and a minimum area of 1 km2. Moreover, the maximum velocity difference among the shear segments inside a divergence region should be at least 5 m/s. As such, the shear within a divergence region is at least 5 m/s per 600 m, i.e. 0.008 m/s/m.

F-factor is related to the total aircraft energy and its rate of change, and is defined to be:

$$F = \frac{\dot{W}\_x}{g} - \frac{w}{V\_a} \tag{1}$$

where *Wx* is the component of atmospheric wind directed horizontally along the flight path (direction *x*) and *Wx* its rate of change, *g* the acceleration due to gravity, *w* the updraft of the atmosphere, and *Va* the airspeed of the aircraft. By estimating the updraft from mass continuity constraint, it is shown to be equivalent to:

$$F = \frac{\partial \mathcal{W}\_x}{\partial x} \left[ \frac{V\_g}{g} + \frac{\mathbf{2h}}{V\_a} \right] \tag{2}$$

Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence 83

southwest orientation moved southeastwards from inland areas across the coast. At HKIA, the TDWR issued microburst alerts of 30 knots headwind loss for the aircraft between 09:20

Figure 6(a) shows the moment when a microburst associated with the thunderstorms affected the runway corridors to the east of HKIA. Divergent flow feature was found at 0.6 degree conical scan of TDWR. For an aircraft arriving at the north runway of HKIA (location in Figure 1) from the east, the windshear associated with the microburst is performance decreasing (due to decreasing headwind). Using the formulae above, the Ffactor for the microburst is determined to be about 0.14, which exceeds the must alert threshold and the windshear associated with the microburst is considered to be hazardous to the aircraft. Flight data are obtained for an aircraft arriving at the north runway from the east at that time. They are processed by the algorithm in Haverdings (2000) and the variation of F-factor along the glide path is shown in Figure 6(b). At about the location of the microburst (near the eastern threshold of the north runway), the F-factor is found to be about 0.13, which is generally consistent with the value determined from TDWR data. Thus for microburst associated with the thunderstorm, the F-factor determined from TDWR measurements and that from QAR data of the aircraft are comparable with each other. The other peaks/troughs of F-factor from the QAR data (Figure 6(b)) are not revealed in the TDWR measurements. They may not be properly handled by the microburst model for F-

Fig. 6. (a) Divergence features (highlighted in lighter colours) associated with microburst on 18 May 2007, overlaid on the radial velocity from the TDWR (colour scale on the right). Ffactor of each feature is given as a number next to the box indicating the location of the feature. (b) F-factor as recorded on an aircraft flying at about the same time as in (a) along the glide path shown as a red arrow in (a). The red arrow in (b) is the approximate location

To study the change in the F-factor following the evolution of the microburst, the intense convective event on 8 June 2007 is considered. Severe gusts associated with thunderstorms and microburst with a recorded maximum of 35.9 m/s affected HKIA in the morning of that day. A helicopter parked on the apron toppled in strong winds during the passage of the intense storm cells. We just focus on the windshear hazard associated with the microburst. The divergence features determined from the radial velocity of the TDWR at 0.6-degree conical scans are shown in Figure 7. Stronger winds associated with the microburst got

of the windshear feature encountered by the aircraft.

and 09:27 UTC.

factor calculation.

where *Vg* is the ground speed of the aircraft, and h the altitude above ground.

For each convergence/divergence feature captured by the TDWR, the velocity change Δ*U* and the distance over which this change occurs Δ*R* are calculated. It is shown in Hinton (1993) with reference to a microburst model that F-factor could be calculated from:

$$F = K \frac{\Delta II}{\Delta R} \left[ \left( \frac{\Delta R}{L} \right)^2 - \left( \frac{\Delta R}{L} \right)^3 \frac{\sqrt{\pi}}{2a} \text{erf} \left( \frac{aL}{\Delta R} \right) \right] \bullet \left[ \frac{V\_g}{g} + \frac{2h\_r}{V\_a} \right] \tag{3}$$

where *K* = 4.1925, *α* = 1.1212, *hr* the above-ground-level (AGL) altitude of the TDWR radar beam, *L* the characteristic shear length of 1000 m, and *erf*(*y*) the error function.

The microburst model in Hinton (1993) includes a shaping function which describes the change in microburst outflow with altitude. This function is given by:

$$p(h) = \frac{e^{-0.22h/H} - e^{-2.75h/H}}{0.7386} \tag{4}$$

where *h* is the altitude above ground and *H* the altitude of maximum outflow speed (assumed to be 90 m). The F-factor *F1* from the TDWR at the radar beam altitude *h1* is then related to the F-factor *F2* of the aircraft at the altitude *h2* by the following equation:

$$F\_2 = F\_1 \frac{p(h\_2)\left(\frac{V\_g}{g} + \frac{2h\_2}{V\_a}\right)}{p(h\_1)\left(\frac{V\_g}{g} + \frac{2h\_1}{V\_a}\right)}\tag{5}$$

Combining (3) – (5) and with Δ*U* and Δ*R* determined, the F-factor associated with a divergence/convergence feature at the altitude of the aircraft along the glide path could be calculated.

For the formulation in (1), F-factor is positive if the windshear is performance decreasing (headwind decreasing or downdraft) and negative if the windshear is performance increasing (headwind increasing or updraft). As discussed in Proctor et al. (2000), for onboard windshear systems, the windshear is considered to be hazardous if *F* is greater than 0.1, and a *must alert* threshold is set to be 0.13. The must alert threshold means a wind shear alert must be issued when that threshold is reached/exceeded.

A microburst event that affected HKIA on 18 May 2007 is considered here as an illustration of the method. In the evening of that day, a surface trough of low pressure lingered around the south China coast, bringing unsettled weather to the region. Between 09 and 10 UTC (5 and 6 p.m. of 18 May 2007), a band of strong radar echoes with east-northeast to westDoppler Radar Observations – 82 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

the atmosphere, and *Va* the airspeed of the aircraft. By estimating the updraft from mass

*<sup>W</sup> <sup>V</sup> <sup>h</sup> <sup>F</sup> x gV* 

For each convergence/divergence feature captured by the TDWR, the velocity change Δ*U* and the distance over which this change occurs Δ*R* are calculated. It is shown in Hinton

2

where *K* = 4.1925, *α* = 1.1212, *hr* the above-ground-level (AGL) altitude of the TDWR radar

The microburst model in Hinton (1993) includes a shaping function which describes the

where *h* is the altitude above ground and *H* the altitude of maximum outflow speed (assumed to be 90 m). The F-factor *F1* from the TDWR at the radar beam altitude *h1* is then

2

*<sup>V</sup> <sup>h</sup> p h g V F F*

<sup>2</sup> ( )

*g*

 

 

<sup>2</sup> ( )

*<sup>V</sup> <sup>h</sup> p h g V*

*g*

1

Combining (3) – (5) and with Δ*U* and Δ*R* determined, the F-factor associated with a divergence/convergence feature at the altitude of the aircraft along the glide path could be

For the formulation in (1), F-factor is positive if the windshear is performance decreasing (headwind decreasing or downdraft) and negative if the windshear is performance increasing (headwind increasing or updraft). As discussed in Proctor et al. (2000), for onboard windshear systems, the windshear is considered to be hazardous if *F* is greater than 0.1, and a *must alert* threshold is set to be 0.13. The must alert threshold means a wind shear

A microburst event that affected HKIA on 18 May 2007 is considered here as an illustration of the method. In the evening of that day, a surface trough of low pressure lingered around the south China coast, bringing unsettled weather to the region. Between 09 and 10 UTC (5 and 6 p.m. of 18 May 2007), a band of strong radar echoes with east-northeast to west-

( ) 0.7386 *hH hH e e p h*

related to the F-factor *F2* of the aircraft at the altitude *h2* by the following equation:

2 1

alert must be issued when that threshold is reached/exceeded.

calculated.

0.22 / 2.75 /

 

*UR R L <sup>V</sup> <sup>h</sup> F K erf RL L R gV* 

where *Vg* is the ground speed of the aircraft, and h the altitude above ground.

beam, *L* the characteristic shear length of 1000 m, and *erf*(*y*) the error function.

change in microburst outflow with altitude. This function is given by:

(1993) with reference to a microburst model that F-factor could be calculated from:

*a*

2 3 2

 

2

*a*

1

*a*

(2)

(3)

*g r a*

(4)

(5)

continuity constraint, it is shown to be equivalent to:

*<sup>g</sup>* <sup>2</sup> *<sup>x</sup>*

southwest orientation moved southeastwards from inland areas across the coast. At HKIA, the TDWR issued microburst alerts of 30 knots headwind loss for the aircraft between 09:20 and 09:27 UTC.

Figure 6(a) shows the moment when a microburst associated with the thunderstorms affected the runway corridors to the east of HKIA. Divergent flow feature was found at 0.6 degree conical scan of TDWR. For an aircraft arriving at the north runway of HKIA (location in Figure 1) from the east, the windshear associated with the microburst is performance decreasing (due to decreasing headwind). Using the formulae above, the Ffactor for the microburst is determined to be about 0.14, which exceeds the must alert threshold and the windshear associated with the microburst is considered to be hazardous to the aircraft. Flight data are obtained for an aircraft arriving at the north runway from the east at that time. They are processed by the algorithm in Haverdings (2000) and the variation of F-factor along the glide path is shown in Figure 6(b). At about the location of the microburst (near the eastern threshold of the north runway), the F-factor is found to be about 0.13, which is generally consistent with the value determined from TDWR data. Thus for microburst associated with the thunderstorm, the F-factor determined from TDWR measurements and that from QAR data of the aircraft are comparable with each other. The other peaks/troughs of F-factor from the QAR data (Figure 6(b)) are not revealed in the TDWR measurements. They may not be properly handled by the microburst model for Ffactor calculation.

Fig. 6. (a) Divergence features (highlighted in lighter colours) associated with microburst on 18 May 2007, overlaid on the radial velocity from the TDWR (colour scale on the right). Ffactor of each feature is given as a number next to the box indicating the location of the feature. (b) F-factor as recorded on an aircraft flying at about the same time as in (a) along the glide path shown as a red arrow in (a). The red arrow in (b) is the approximate location of the windshear feature encountered by the aircraft.

To study the change in the F-factor following the evolution of the microburst, the intense convective event on 8 June 2007 is considered. Severe gusts associated with thunderstorms and microburst with a recorded maximum of 35.9 m/s affected HKIA in the morning of that day. A helicopter parked on the apron toppled in strong winds during the passage of the intense storm cells. We just focus on the windshear hazard associated with the microburst. The divergence features determined from the radial velocity of the TDWR at 0.6-degree conical scans are shown in Figure 7. Stronger winds associated with the microburst got

Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence 85

data along the glide path is given in Figure 8(b). At the location of the microburst, the Ffactor from the aircraft is comparable with that calculated from the TDWR data, even for this case of terrain-disrupted airflow. As discussed in the first case study, the other peaks/troughs of F-factor from the QAR data (Figure 8(b)) are not revealed in the TDWR measurements. They may not be properly handled by the microburst model for F-factor

Fig. 8. (a) Divergence features (highlighted in lighter colours) associated with windshear in terrain-disrupted airflow on 3 August 2006, overlaid on the radial velocity from the TDWR (colour scale given in Figure 6). F-factor of each feature is given as a number next to the box indicating the location of the feature. (b) F-factor as recorded on an aircraft flying at about the same time as in (a) along the glide path shown as a blue arrow in (a). The red arrow in (b) is the approximate location of the windshear feature encountered by the aircraft.

The measurement of spectrum width is determined not only by the Doppler velocity distribution and density distribution of the scatterers within the resolution volume, but also radar observation parameters like beamwidth, pulse width, antenna rotation rate, etc. According to Doviak and Zrnic (2006), there are five major spectral broadening mechanisms that contribute to the spectrum width measurements, which can be written as

222222

(6)

2 . (7)

 *vst do* 

where s represents mean wind shear contribution, t represents turbulence, represents antenna motion, d represents different terminal velocities of hydrometeors of different sizes, and o represents variations of orientations and vibrations of hydrometeors. Except <sup>s</sup> andt, the rest of the terms on the right hand side of the Eq.(6) are considered to be negligible for the measurements of v in this paper (Brewster and Zrnic, 1986). Thus the

<sup>2</sup>v2 -<sup>s</sup>

In the Eq.(7), mean wind shear width term s can be decomposed into three terms due to mean radial velocity shear at three orthogonal directions in radar coordinate(Doviak and

t

calculation.

follow

Zrnic, 2006):

**4. Calculation of turbulence intensity** 

turbulence term s can be obtained,

closer to the ground level (about 260 m above mean sea level at the location of the microburst) in a short time interval within 3 minutes, with the maximum value of towardsthe-radar velocity increasing from 18 m/s (dark blue in Figure 7) to 23 m/s (magenta in Figure 7). As a result, the F-factor increases in magnitude from 0.14 to 0.23, which exceeds the must alert threshold. The TDWR-based F-factor provides a good indication about the level of hazard associated with an evolving microburst.

Fig. 7. Time series of the divergence feature associated with a microburst on 8 June 2007. The feature is highlighted in lighter colour and enclosed in a box. The number next to the box is the F-factor calculated for the feature. The background is the radial velocity from the TDWR, with the colour scale given in Figure 6.

Besides intense convective weather, the windshear hazard in terrain-disrupted airflow is also studied. The Typhoon Prapiroon case on 3 August 2006 is considered. On that day, Prapiroon was located at about 200 km to the southwest of Hong Kong over the South China Sea and tracked northwest towards the western coast of southern China. This typhoon brought about gale-force east to southeasterly airflow to Hong Kong. Due to complex terrain to the south of the airport, airflow disturbances occurred inside and around HKIA. Divergent flow features were observed near the airport from time to time. Figure 8(a) shows such a feature at 0.6-degree conical scan of the TDWR at about 4:47 a.m., 3 August. The F-factor associated with this feature is about 0.22, which exceeds the must alert threshold for windshear. An aircraft landed at the north runway of HKIA from the west at about that time (within one minute). The variation of the F-factor determined from QAR data along the glide path is given in Figure 8(b). At the location of the microburst, the Ffactor from the aircraft is comparable with that calculated from the TDWR data, even for this case of terrain-disrupted airflow. As discussed in the first case study, the other peaks/troughs of F-factor from the QAR data (Figure 8(b)) are not revealed in the TDWR measurements. They may not be properly handled by the microburst model for F-factor calculation.

Fig. 8. (a) Divergence features (highlighted in lighter colours) associated with windshear in terrain-disrupted airflow on 3 August 2006, overlaid on the radial velocity from the TDWR (colour scale given in Figure 6). F-factor of each feature is given as a number next to the box indicating the location of the feature. (b) F-factor as recorded on an aircraft flying at about the same time as in (a) along the glide path shown as a blue arrow in (a). The red arrow in (b) is the approximate location of the windshear feature encountered by the aircraft.

#### **4. Calculation of turbulence intensity**

Doppler Radar Observations – 84 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

closer to the ground level (about 260 m above mean sea level at the location of the microburst) in a short time interval within 3 minutes, with the maximum value of towardsthe-radar velocity increasing from 18 m/s (dark blue in Figure 7) to 23 m/s (magenta in Figure 7). As a result, the F-factor increases in magnitude from 0.14 to 0.23, which exceeds the must alert threshold. The TDWR-based F-factor provides a good indication about the

Fig. 7. Time series of the divergence feature associated with a microburst on 8 June 2007. The feature is highlighted in lighter colour and enclosed in a box. The number next to the box is the F-factor calculated for the feature. The background is the radial velocity from the

Besides intense convective weather, the windshear hazard in terrain-disrupted airflow is also studied. The Typhoon Prapiroon case on 3 August 2006 is considered. On that day, Prapiroon was located at about 200 km to the southwest of Hong Kong over the South China Sea and tracked northwest towards the western coast of southern China. This typhoon brought about gale-force east to southeasterly airflow to Hong Kong. Due to complex terrain to the south of the airport, airflow disturbances occurred inside and around HKIA. Divergent flow features were observed near the airport from time to time. Figure 8(a) shows such a feature at 0.6-degree conical scan of the TDWR at about 4:47 a.m., 3 August. The F-factor associated with this feature is about 0.22, which exceeds the must alert threshold for windshear. An aircraft landed at the north runway of HKIA from the west at about that time (within one minute). The variation of the F-factor determined from QAR

level of hazard associated with an evolving microburst.

TDWR, with the colour scale given in Figure 6.

The measurement of spectrum width is determined not only by the Doppler velocity distribution and density distribution of the scatterers within the resolution volume, but also radar observation parameters like beamwidth, pulse width, antenna rotation rate, etc. According to Doviak and Zrnic (2006), there are five major spectral broadening mechanisms that contribute to the spectrum width measurements, which can be written as follow

$$
\sigma\_v^2 = \sigma\_s^2 + \sigma\_t^2 + \sigma\_a^2 + \sigma\_d^2 + \sigma\_o^2 \tag{6}
$$

where s represents mean wind shear contribution, t represents turbulence, represents antenna motion, d represents different terminal velocities of hydrometeors of different sizes, and o represents variations of orientations and vibrations of hydrometeors. Except <sup>s</sup> andt, the rest of the terms on the right hand side of the Eq.(6) are considered to be negligible for the measurements of v in this paper (Brewster and Zrnic, 1986). Thus the turbulence term s can be obtained,

$$
\Box \mathfrak{o}\_{\mathsf{t}}^{2} = \Box \mathfrak{o}\_{\mathsf{v}}^{2} \mathsf{-} \mathsf{o}\_{\mathsf{s}}^{2} \,. \tag{7}
$$

In the Eq.(7), mean wind shear width term s can be decomposed into three terms due to mean radial velocity shear at three orthogonal directions in radar coordinate(Doviak and Zrnic, 2006):

$$
\sigma\_s^2 = \sigma\_{s\theta}^2 + \sigma\_{s\phi}^2 + \sigma\_{sr}^2 = \left(r\_0 \sigma\_\theta k\_\theta\right)^2 + \left(r\_0 \sigma\_\phi k\_\phi\right)^2 + \left(\sigma\_r k\_r\right)^2,\tag{8}
$$

where <sup>r</sup> 2 = (0.35c/2)2, 2 = 12/16ln2, and 2 = <sup>1</sup> 2/16ln2. Here c/2 is range resolution, and 1is the one-way angular resolution (i.e., beamwidth). *k*, *k*, and *k*r are the components of shear along the three orthogonal directions.

In order to use t to estimate eddy dissipation rate (EDR) , it must be assumed that within radar resolution volume turbulence is isotropic and its outer scale is larger than the maximum dimension of the radar's resolution volume (which is indicated as V6). Under these assumptions, in the case of *<sup>r</sup> r* the relation between turbulence spectrum width t and EDR can be approximately written as (Labitt, 1981)

$$
\varepsilon \approx \frac{0.72 \sigma\_t^3}{r \sigma\_\theta A^{3/2}} \,\,\,\,\tag{9}
$$

Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence 87

Following international practice, EDR values are classified into four categories in terms of the intensity of turbulence. For convenience and in line with alerting purpose of low-level turbulence, EDRs in the following figures and context will be labeled or indicated as insignificant (LL), light (L), moderate (M), and severe (S) instead of its value. It is also worth mentioning that EDR values presented in this paper are derived from the spectrum width data after smoothing by using a 9 point median filter along the radar beams in order to suppress the fluctuations in the determination of spectrum width values. This kind of fluctuation is expected, for instance, to arise from the limited and finite number of data

The spectrum width errors are large in region of low SNR. Here we selected a case to demonstrate the importance of the SNR threshold in the quality control of EDR data. Around 21 UTC on 6 June 2008, the TDWR radar observed thunderstorms over HKIA. Without SNR threshold, estimated EDR suggested severe turbulence region (red color; Figure 9(a)) in the region about azimuth of 270o and centered at about 25 km. High spectrum widths (~4.5 m/s) are indeed measured in this region (see Figure 9(c)). But reflectivity (Figure 9(e)) and SNR (Figure 9(d)) are around -8 dBZ and 10 dB respectively. The relatively large spectrum widths in this region can be caused by incorrect noise power estimates (Fang et al., 2004). To avoid

such biases, we use a SNR threshold of 20 dB as recommended by Fang et al., (2004).

Fig. 9. (a) EDR, (b) EDR with SNR> 20 dB, (c) spectrum width, (d) SNR, (e) reflectivity, and (f) Doppler velocity at elevation angle of 0.6o at 21:28 UTC on 6 June 2008. Range ring is 50

km and azimuths are every 30o.

points in the digitization of the spectrum of the return signal.

where A is constant (i.e., about 1.6). When *<sup>r</sup> r* , the relation can be approximated by

$$\varepsilon \approx \left[ \frac{<\sigma\_t^2 >^{3/2}}{\sigma\_r (1.35A)^{3/2}} \right] \frac{11}{15} + \frac{4}{15} \frac{r^2 \sigma\_\theta^2}{\sigma\_r^2} \text{ ${}^{-3/2}$ }\tag{10}$$

Eqs. (9) and (10) are used to estimate EDR using Hong Kong TDWR observed spectrum width.

In hazardous weather mode, the Hong Kong TDWR conducts sector scans from azimuth 182o to 282o (i.e., confined to the approach and departure paths). Each sector scan takes about 4 minutes. Thus, the low altitude wind shear can be detected within a minute. The range and angular resolutions of the radar are 150 m and 0.5o respectively. The maximum range reaches 90 km. The radar data includes reflectivity, Doppler velocity, spectrum width, and signal-to-noise ratio (SNR) recorded with the azimuth interval of 1o.

Based on the Eqs. (9) and (10), EDR can be estimated when spectrum width observation is available. In this feasibility study, EDR estimation is only performed at the lowest elevation angle of 0.6o. The vertical wind shear contribution to the EDR is calculated by using spatially averaged mean Doppler velocity at two lowest elevation angles. Because the closest two elevation angles at lowest level are 0.6o and 1.0o at scans 11 and 12, vertical wind shear is calculated by using the Doppler velocity fields at these two scans. For simplicity, EDR is estimated at scan 17 with elevation angle of 0.6o. Azimuthal and radial wind shear is also calculated at this scan. So in the current algorithm, one EDR field at elevation angle of 0.6o will be generated for each volume scan.

The control of the TDWR spectrum width data quality is very important for EDR estimation. It has been found that there is a variety of sources of errors in spectrum width measurements in previous studies (Fang et al. 2004). Especially if signal to noise ratio (SNR) is low, spectrum width measurements have large variance. In this study, SNR > 20 dB is assigned as a simple and straightforward threshold for the EDR estimates. In other words, EDR at the gate with SNR < 20 dB is marked as missing data (MD) in our algorithm. In the future, more comprehensive quality control processor will be designed and implemented in our algorithm to deal with other error sources.

Doppler Radar Observations – 86 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

and 1is the one-way angular resolution (i.e., beamwidth). *k*, *k*, and *k*r are the components

In order to use t to estimate eddy dissipation rate (EDR) , it must be assumed that within radar resolution volume turbulence is isotropic and its outer scale is larger than the maximum dimension of the radar's resolution volume (which is indicated as V6). Under

> 3 3/2 0.72 *<sup>t</sup> r A*

2 3/2 2 2

3/2 2 11 4 [ ]( ) (1.35 ) 15 15

*r*

 

 

*r r*

Eqs. (9) and (10) are used to estimate EDR using Hong Kong TDWR observed spectrum

In hazardous weather mode, the Hong Kong TDWR conducts sector scans from azimuth 182o to 282o (i.e., confined to the approach and departure paths). Each sector scan takes about 4 minutes. Thus, the low altitude wind shear can be detected within a minute. The range and angular resolutions of the radar are 150 m and 0.5o respectively. The maximum range reaches 90 km. The radar data includes reflectivity, Doppler velocity, spectrum width,

Based on the Eqs. (9) and (10), EDR can be estimated when spectrum width observation is available. In this feasibility study, EDR estimation is only performed at the lowest elevation angle of 0.6o. The vertical wind shear contribution to the EDR is calculated by using spatially averaged mean Doppler velocity at two lowest elevation angles. Because the closest two elevation angles at lowest level are 0.6o and 1.0o at scans 11 and 12, vertical wind shear is calculated by using the Doppler velocity fields at these two scans. For simplicity, EDR is estimated at scan 17 with elevation angle of 0.6o. Azimuthal and radial wind shear is also calculated at this scan. So in the current algorithm, one EDR field at elevation angle of 0.6o

The control of the TDWR spectrum width data quality is very important for EDR estimation. It has been found that there is a variety of sources of errors in spectrum width measurements in previous studies (Fang et al. 2004). Especially if signal to noise ratio (SNR) is low, spectrum width measurements have large variance. In this study, SNR > 20 dB is assigned as a simple and straightforward threshold for the EDR estimates. In other words, EDR at the gate with SNR < 20 dB is marked as missing data (MD) in our algorithm. In the future, more comprehensive quality control processor will be designed and implemented in

 

2 22

*r r rk rk k* , (8)

the relation between turbulence spectrum width

, (9)

3/2

(10)

, the relation can be approximated by

2 = 12/16ln2. Here c/2 is range resolution,

0 0 ( )( )( )

 

22 2 2

 

2 = 12/16ln2, and

 

*A*

*t*

and signal-to-noise ratio (SNR) recorded with the azimuth interval of 1o.

t and EDR can be approximately written as (Labitt, 1981)

 *s s s sr* 

of shear along the three orthogonal directions.

where A is constant (i.e., about 1.6). When *<sup>r</sup> r*

these assumptions, in the case of *<sup>r</sup> r*

will be generated for each volume scan.

our algorithm to deal with other error sources.

2 = (0.35c/2)2,

where <sup>r</sup>

width.

Following international practice, EDR values are classified into four categories in terms of the intensity of turbulence. For convenience and in line with alerting purpose of low-level turbulence, EDRs in the following figures and context will be labeled or indicated as insignificant (LL), light (L), moderate (M), and severe (S) instead of its value. It is also worth mentioning that EDR values presented in this paper are derived from the spectrum width data after smoothing by using a 9 point median filter along the radar beams in order to suppress the fluctuations in the determination of spectrum width values. This kind of fluctuation is expected, for instance, to arise from the limited and finite number of data points in the digitization of the spectrum of the return signal.

The spectrum width errors are large in region of low SNR. Here we selected a case to demonstrate the importance of the SNR threshold in the quality control of EDR data. Around 21 UTC on 6 June 2008, the TDWR radar observed thunderstorms over HKIA. Without SNR threshold, estimated EDR suggested severe turbulence region (red color; Figure 9(a)) in the region about azimuth of 270o and centered at about 25 km. High spectrum widths (~4.5 m/s) are indeed measured in this region (see Figure 9(c)). But reflectivity (Figure 9(e)) and SNR (Figure 9(d)) are around -8 dBZ and 10 dB respectively. The relatively large spectrum widths in this region can be caused by incorrect noise power estimates (Fang et al., 2004). To avoid such biases, we use a SNR threshold of 20 dB as recommended by Fang et al., (2004).

Fig. 9. (a) EDR, (b) EDR with SNR> 20 dB, (c) spectrum width, (d) SNR, (e) reflectivity, and (f) Doppler velocity at elevation angle of 0.6o at 21:28 UTC on 6 June 2008. Range ring is 50 km and azimuths are every 30o.

Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence 89

Fig. 10. EDR maps (a) at 21:32 UTC on 27 April 2006 and (b) 13:17 UTC on 13 June 2008. The

Another issue of the comparison is the contribution of mean wind shear to the measured spectrum width. For the estimation of EDR, the contribution of wind shear has to be extracted from the radar measured spectrum width. But for the comparison with aircraft measured EDR or even turbulence alert for aviation safety, wind shear might not need to be removed. For example if the aircraft experiences a sharp change in altitude, this may not be caused by isotropic turbulence but it is a measure of aircraft response to vertical shear of mean wind. As such, the aircraft estimated EDR based on vertical velocity may be slightly higher. Pilots and passengers in aircraft may also experience severe "turbulence", which is a

Scatterplots of median and maximum EDR along the 5 nm of flight paths estimated by aircraft and radar are shown in Figure 11. Two plots for each are shown; one in which mean wind shear contributions to the observed spectrum widths are removed and a second plot in

mountainous Lantau Island is located to the south of the radar scans.

combination of the effects of both turbulence and wind shear.

On the other hand, there are two small regions near the radar at the range of 6 km where EDR is also high. But in this region there is relatively strong horizontal shear of the radial wind component (Figure 9(f); green color identifies the wind has a component toward the radar and red color indicates wind is away from the radar). Furthermore, the reflectivity is about 10 dBZ and SNR is around 35 dB. Because this region is on the downwind side of Lantau Island, the ambient flow (green in Figure 9(f)) is blocked by the Island and back flow (red in Figure 9(f)) is induced. The wind shear contributions, computed using Eq. (8), have been removed from the calculation EDR presented in Figure 9(a). Thus the EDR should not be biased by strong shear of mean radial wind. Thunderstorm outflow may be another reason for the severe turbulence in this region. Because there is no strong horizontal shear of the Doppler velocity field in the region 270o and 25 km, we conclude that the large EDRs presented in in that region of Figure 9(a) are unrealistic. After a threshold SNR> 20 dB is applied, it can be seen that these large EDR values are removed (Figure 9(b)).

Using the Hong Kong TDWR observations in 2006 and 2008, many EDR maps were produced and examined. Here wind shear contribution has been removed from spectrum width measurements. Here the mean wind shears in horizontal and vertical directions are calculated by using mean radial velocity field smoothed by a 9 points median filter along the radar beam in the Eq.(8). Figure 10 shows two typical EDR maps during light rain at 21:32 UTC on 27 April 2006 (Figure 10(a)) and during a thunderstorm at 13:17 UTC on 13 June 2008 (Figure 10(b)). For most of the scanned area, EDR is low and turbulence is classified as insignificant or light (green and light blue). Small pockets of moderate and severe turbulence (yellow and red) are scattered in the scanned area. Near the Lantau Island, moderate and severe levels of turbulence are frequently observed in the cases we studied. The blockage of the Island on the ambient flow may be a reason for the occurrence of the turbulent airflow. Based on the numerical simulations, Clark et al. (1997) and Chan (2009) found that mechanical effect of a mountainous island is a source of the generation of the turbulence.

Clear air cases have been investigated as well, but we found that SNR of the Hong Kong TDWR is too low to provide reliable and meaningful EDR maps.

After the EDR maps were generated, EDR profiles along the flight paths can be compared with aircraft measured EDR. A total of 14 cases are selected to make the comparison. The aircraft EDRs are estimated based on the vertical wind measured by aircraft (Cornman et al., 2004).

Radar derived EDR profile is constructed by selecting the EDR in a resolution volume V6 closest to the flight path and at an elevation angle of 0.6o. There are still differences in the measurement heights between the aircraft and the radar beam for these two EDR datasets. Only a part of the flight path is covered by the radar beam. For example, aircraft approaching runway 25RA is in the radar beam only at the distance between 0.5 and 1.5 nm from the end of runway. From this point of view, EDRs estimated by aircraft and the radar would be compared within this distance interval. It should also be mentioned that radar estimated EDR is based on the spectrum width of the Doppler velocity, i.e. velocity in the radial direction along a radar beam. On the other hand, the aircraft estimated EDR is based on the vertical wind. As such, the two EDR datasets are derived from different components of the wind. Put aside errors in measurement, in order to have agreement turbulence must be isotropic.

Doppler Radar Observations – 88 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

On the other hand, there are two small regions near the radar at the range of 6 km where EDR is also high. But in this region there is relatively strong horizontal shear of the radial wind component (Figure 9(f); green color identifies the wind has a component toward the radar and red color indicates wind is away from the radar). Furthermore, the reflectivity is about 10 dBZ and SNR is around 35 dB. Because this region is on the downwind side of Lantau Island, the ambient flow (green in Figure 9(f)) is blocked by the Island and back flow (red in Figure 9(f)) is induced. The wind shear contributions, computed using Eq. (8), have been removed from the calculation EDR presented in Figure 9(a). Thus the EDR should not be biased by strong shear of mean radial wind. Thunderstorm outflow may be another reason for the severe turbulence in this region. Because there is no strong horizontal shear of the Doppler velocity field in the region 270o and 25 km, we conclude that the large EDRs presented in in that region of Figure 9(a) are unrealistic. After a threshold SNR> 20 dB is

Using the Hong Kong TDWR observations in 2006 and 2008, many EDR maps were produced and examined. Here wind shear contribution has been removed from spectrum width measurements. Here the mean wind shears in horizontal and vertical directions are calculated by using mean radial velocity field smoothed by a 9 points median filter along the radar beam in the Eq.(8). Figure 10 shows two typical EDR maps during light rain at 21:32 UTC on 27 April 2006 (Figure 10(a)) and during a thunderstorm at 13:17 UTC on 13 June 2008 (Figure 10(b)). For most of the scanned area, EDR is low and turbulence is classified as insignificant or light (green and light blue). Small pockets of moderate and severe turbulence (yellow and red) are scattered in the scanned area. Near the Lantau Island, moderate and severe levels of turbulence are frequently observed in the cases we studied. The blockage of the Island on the ambient flow may be a reason for the occurrence of the turbulent airflow. Based on the numerical simulations, Clark et al. (1997) and Chan (2009) found that mechanical effect of a mountainous island is a source of the generation of the

Clear air cases have been investigated as well, but we found that SNR of the Hong Kong

After the EDR maps were generated, EDR profiles along the flight paths can be compared with aircraft measured EDR. A total of 14 cases are selected to make the comparison. The aircraft EDRs are estimated based on the vertical wind measured by aircraft (Cornman et al., 2004).

Radar derived EDR profile is constructed by selecting the EDR in a resolution volume V6 closest to the flight path and at an elevation angle of 0.6o. There are still differences in the measurement heights between the aircraft and the radar beam for these two EDR datasets. Only a part of the flight path is covered by the radar beam. For example, aircraft approaching runway 25RA is in the radar beam only at the distance between 0.5 and 1.5 nm from the end of runway. From this point of view, EDRs estimated by aircraft and the radar would be compared within this distance interval. It should also be mentioned that radar estimated EDR is based on the spectrum width of the Doppler velocity, i.e. velocity in the radial direction along a radar beam. On the other hand, the aircraft estimated EDR is based on the vertical wind. As such, the two EDR datasets are derived from different components of the wind. Put aside errors in measurement, in order to have agreement turbulence must

TDWR is too low to provide reliable and meaningful EDR maps.

applied, it can be seen that these large EDR values are removed (Figure 9(b)).

turbulence.

be isotropic.

Fig. 10. EDR maps (a) at 21:32 UTC on 27 April 2006 and (b) 13:17 UTC on 13 June 2008. The mountainous Lantau Island is located to the south of the radar scans.

Another issue of the comparison is the contribution of mean wind shear to the measured spectrum width. For the estimation of EDR, the contribution of wind shear has to be extracted from the radar measured spectrum width. But for the comparison with aircraft measured EDR or even turbulence alert for aviation safety, wind shear might not need to be removed. For example if the aircraft experiences a sharp change in altitude, this may not be caused by isotropic turbulence but it is a measure of aircraft response to vertical shear of mean wind. As such, the aircraft estimated EDR based on vertical velocity may be slightly higher. Pilots and passengers in aircraft may also experience severe "turbulence", which is a combination of the effects of both turbulence and wind shear.

Scatterplots of median and maximum EDR along the 5 nm of flight paths estimated by aircraft and radar are shown in Figure 11. Two plots for each are shown; one in which mean wind shear contributions to the observed spectrum widths are removed and a second plot in

Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence 91

turbulence only occurs at one radar gate at the distance of 0 nm, closest to the end of the runway. It is noted that at this location, the radar beam is higher than the flight path by

We have also compared aircraft and radar estimated EDR profiles including wind shear contribution along the aircraft flight path. For this case, aircraft B777 flew through a storm

Figure 12 shows the EDR estimated by aircraft and the radar along the flight path 25RA around 13:05 UTC on 19 April 2008. It is one of the two cases in which severe turbulence was encountered by the aircraft. Blue dots in Figure 13 represent the EDR estimated by the aircraft as it was landing at HKIA. Three peaks over 0.5 2/3 *m s* / , classified as severe turbulence, are recorded at distance of 0.77, 3.65, and 4.90 nm away from the runway end. EDR profiles estimated by using radar data at an elevation angle of 0.6o with the wind shear contribution included in the volume scans around 13:05 UTC are overlaid onto the aircraft estimated EDR in Figure 13. The radar estimated EDR profiles at 13:01, 13:05, and 13:09 UTC (brown dots, red squares, and green dots in Figure 13) matches well with aircraft EDR between distance of 0.5 and 1.5 nm, shaded in green color in Figure 13, where the aircraft was in a region common to the 0.6o radar beam. It means that radar and aircraft were

with maximum reflectivity of 42 dBZ and landed in clouds and light rain at HKIA.

measuring turbulence in approximately the same region at nearly the same time.

Fig. 12. (a) EDR, (b) spectrum width, (c) reflectivity factor, and (d) Doppler velocity at

The peaks of these 3 EDR profiles at 13:01, 13:05, and 13:09 UTC are in the green shaded interval and the maximum value is 0.48 2/3 *m s* / , just slightly smaller than 0.5 2/3 *m s* / . In order to find if there are higher EDR near the flight time (13:05 UTC), we examined the EDR for the two scans one minute before and after the passage of the aircraft at 13:05 UTC in the same volume scan at 13:05 UTC. The profiles are shown with light and dark purple dots in Figure 13. High EDRs with values of 0.69 and 0.76 2/3 *m s* / are found within the shaded

elevation angle of 0.6o at 13:05 UTC on 19 April 2008. Range ring is at 10 km.

about 160m.

which mean wind shear contribution has been retained. All median EDRs are smaller than 0.4 2/3 *m s* / (i.e., moderate or light turbulence). 13 of 14 median EDRs indicate turbulences are light. Based on maximum EDRs, two severe turbulent patches (EDR > 0.5 2/3 *m s* / ) are detected by both aircraft and radar with wind shear, but they are not on the same flight paths. With wind shear contribution, median and maximum radar EDRs evidently increase.

Fig. 11. Scatterplots of median and maximum EDR estimated by aircraft and radar along the 5 nm of flight paths for the selected 14 cases.

Comparing maximum intensity between aircraft and radar without wind shear, 8 of the 14 cases are in the same category. Seven of them are moderate turbulence. For 4 aircraft estimated light turbulence cases, the radar tends to overestimate them as moderate (3 cases) and severe (1 case) with wind shear contribution. After closer examination of the overestimation case at 07:17 UTC on 25 June 2008, it is found that the maximum severe Doppler Radar Observations – 90 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

which mean wind shear contribution has been retained. All median EDRs are smaller than 0.4 2/3 *m s* / (i.e., moderate or light turbulence). 13 of 14 median EDRs indicate turbulences are light. Based on maximum EDRs, two severe turbulent patches (EDR > 0.5 2/3 *m s* / ) are detected by both aircraft and radar with wind shear, but they are not on the same flight paths. With wind shear contribution, median and maximum radar EDRs evidently increase.

Fig. 11. Scatterplots of median and maximum EDR estimated by aircraft and radar along the

Comparing maximum intensity between aircraft and radar without wind shear, 8 of the 14 cases are in the same category. Seven of them are moderate turbulence. For 4 aircraft estimated light turbulence cases, the radar tends to overestimate them as moderate (3 cases) and severe (1 case) with wind shear contribution. After closer examination of the overestimation case at 07:17 UTC on 25 June 2008, it is found that the maximum severe

5 nm of flight paths for the selected 14 cases.

turbulence only occurs at one radar gate at the distance of 0 nm, closest to the end of the runway. It is noted that at this location, the radar beam is higher than the flight path by about 160m.

We have also compared aircraft and radar estimated EDR profiles including wind shear contribution along the aircraft flight path. For this case, aircraft B777 flew through a storm with maximum reflectivity of 42 dBZ and landed in clouds and light rain at HKIA.

Figure 12 shows the EDR estimated by aircraft and the radar along the flight path 25RA around 13:05 UTC on 19 April 2008. It is one of the two cases in which severe turbulence was encountered by the aircraft. Blue dots in Figure 13 represent the EDR estimated by the aircraft as it was landing at HKIA. Three peaks over 0.5 2/3 *m s* / , classified as severe turbulence, are recorded at distance of 0.77, 3.65, and 4.90 nm away from the runway end. EDR profiles estimated by using radar data at an elevation angle of 0.6o with the wind shear contribution included in the volume scans around 13:05 UTC are overlaid onto the aircraft estimated EDR in Figure 13. The radar estimated EDR profiles at 13:01, 13:05, and 13:09 UTC (brown dots, red squares, and green dots in Figure 13) matches well with aircraft EDR between distance of 0.5 and 1.5 nm, shaded in green color in Figure 13, where the aircraft was in a region common to the 0.6o radar beam. It means that radar and aircraft were measuring turbulence in approximately the same region at nearly the same time.

Fig. 12. (a) EDR, (b) spectrum width, (c) reflectivity factor, and (d) Doppler velocity at elevation angle of 0.6o at 13:05 UTC on 19 April 2008. Range ring is at 10 km.

The peaks of these 3 EDR profiles at 13:01, 13:05, and 13:09 UTC are in the green shaded interval and the maximum value is 0.48 2/3 *m s* / , just slightly smaller than 0.5 2/3 *m s* / . In order to find if there are higher EDR near the flight time (13:05 UTC), we examined the EDR for the two scans one minute before and after the passage of the aircraft at 13:05 UTC in the same volume scan at 13:05 UTC. The profiles are shown with light and dark purple dots in Figure 13. High EDRs with values of 0.69 and 0.76 2/3 *m s* / are found within the shaded

Aviation Applications of Doppler Radars in the Alerting of Windshear and Turbulence 93

This chapter discusses the aviation applications of TDWR. This radar issues microburst alerts which are crucial in the assurance of aviation safety. A typical case of microburst detection by TDWR in association with intense thunderstorms is described first in this chapter. Then the applications of TDWR in the alerting of windshear and turbulence are described. Windshear is alerted through the calculation of windshear hazard factor, which is a rather well established technology. On the other hand, the use of spectrum width data from the radar in the alerting of turbulence has a relatively shorter development history,

Study is underway in Hong Kong to use X-band radar in the alerting of windshear and turbulence on experimental basis at the Hong Kong International Airport. The use of longrange S band radar in the alerting of turbulence for enroute aircraft is also under study.

Brewster, K.A. and D.S. Zrnic, 1986: Comparison of eddy dissipation rate from spatial

Chan, P.W., 2009: Atmospheric turbulence in complex terrain: verifying numerical model

Clark, T.L., T. Keller, J. Coen, P. Neilley, H. Hsu, and W.D. Hall, 1997: Terrain-induced

Cornman, L. B., G. Meymaris, and M. Limber, 2004: An update on the FAA Aviation

Doviak, R. J., and D. S. Zrnic, 2006: Doppler radar and weather observations. Dover

Fang, M., R.J. Doviak, and Melnikov, 2004: Spectrum width measured by the WSR-88D

Gilbert, D., L.B. Cornman, A.R. Rodi, R.G. Frechlich, and R.K. Goodrich, 2004: Calculating

Haverdings, H., 2000: Updated specification of the WINDGRAD algorithm, NLR TR-2000-

spectra of Doppler velocities and Doppler spectrum widths. J. Atmos. Oceanic

results with observations by remote-sensing instruments. Meteorology and

turbulence over Lantau Island: 7 June 1994 Tropical Storm Russ case study. J.

Weather Research Program's in situ turbulence measurement and reporting system. Preprints, Eleventh Conf. on Aviation, Range, and Aerospace Meteorology,

Publications Inc., Mineola, New York, 562 pp. (except for the preface with links to online errata and supplements, this is an exact copy of the 1st and 2nd printing of

radar: Error sources and statistics of various weather phenomena. J. Atmos.

EDR from aircraft wind data during flight in and out of Juneau AK: Techniques and challenges associated with non-straight and level flight patterns. Preprints, 11th Conf. on Aviation, Range and Aerospace Meteorology. Hyannis, MA, Amer.

and the technology is under exploration in Hong Kong.

Atmospheric Physics, 103, 145–157.

Hyannis, MA, Amer. Meteor. Soc., P4.3.

the 1993 Academic Press edition).

Oceanic Technol., 21, 888-904.

Meteor. Soc., CD-ROM, 4.4.

63, National Aerospace Laboratory, 2000.

Atmos. Sci., 54, 1795-1814.

Such progress of these studies would be reported in the future.

**5. Conclusion** 

**6. References** 

Technol., 3, 440-452.

interval. This convinces us that the EDR peak is not caused by random error of radar measurements.

Fig. 13. EDR along the flight path estimated by the aircraft B777 (blue dots) at 13:05 UTC and by the TDWR radar at the time indicated in the legend on 19 April 2008. X axis is the distance between aircraft and the end of runway. The distance interval shaded by the green color indicates where the aircraft passes through the altitude interval observed with the 0.6o elevated beam.

It raises another question: the aircraft may contaminate the radar measurements of the atmospheric status, since the aircraft disturbs the atmosphere and changes the original atmospheric condition in the measurement region as it flies by. In addition, aircraft itself as a target embedded in other scatterers, such as raindrops, may contaminate the spectrum width measurements as well. Both of the two factors could affect spectrum width and EDR value.

It could also be seen that the radar EDR profiles do not match the two aircraft estimated EDR peaks at the distance of 3.65 and 4.90 nm. It might be caused by the spatial difference between the aircraft and the radar beams. The flight heights at the distance of 3.65 and 4.90 nm are higher than the radar beams by about 260 m and 400 m respectively.

Wind shear contribution to spectrum width measurement for this case has been examined. After removing wind shear contribution, the EDR peak at the distance of 0.69 nm is reduced from 0.48 to 0.46 2/3 *m s* / (not shown) at 13:05 UTC. It means that wind shear contribution is small in this region. Because wind shear of the large scale mean wind should be persistent over the 4 minute for entire volume scan, the EDR peaks without wind shear contribution at 13:04 and 13:06 UTC at the distance of 0.69 nm are reduced to 0.67 and 0.74 2/3 *m s* / respectively. It indicates severe turbulence that is matched with aircraft estimate at 13:05 UTC.

Note that the aircraft estimated EDR is considered as ground truth in the above analysis, but it also contains errors and requires significant QC effort, especially as airplane is climbing or descending (Gilbert et al., 2004).

#### **5. Conclusion**

Doppler Radar Observations – 92 Weather Radar, Wind Profiler, Ionospheric Radar, and Other Advanced Applications

interval. This convinces us that the EDR peak is not caused by random error of radar

Fig. 13. EDR along the flight path estimated by the aircraft B777 (blue dots) at 13:05 UTC and

It raises another question: the aircraft may contaminate the radar measurements of the atmospheric status, since the aircraft disturbs the atmosphere and changes the original atmospheric condition in the measurement region as it flies by. In addition, aircraft itself as a target embedded in other scatterers, such as raindrops, may contaminate the spectrum width measurements as well. Both of the two factors could affect spectrum width and EDR

It could also be seen that the radar EDR profiles do not match the two aircraft estimated EDR peaks at the distance of 3.65 and 4.90 nm. It might be caused by the spatial difference between the aircraft and the radar beams. The flight heights at the distance of 3.65 and 4.90

Wind shear contribution to spectrum width measurement for this case has been examined. After removing wind shear contribution, the EDR peak at the distance of 0.69 nm is reduced from 0.48 to 0.46 2/3 *m s* / (not shown) at 13:05 UTC. It means that wind shear contribution is small in this region. Because wind shear of the large scale mean wind should be persistent over the 4 minute for entire volume scan, the EDR peaks without wind shear contribution at 13:04 and 13:06 UTC at the distance of 0.69 nm are reduced to 0.67 and 0.74 2/3 *m s* / respectively. It indicates severe turbulence that is matched with aircraft estimate at 13:05

Note that the aircraft estimated EDR is considered as ground truth in the above analysis, but it also contains errors and requires significant QC effort, especially as airplane is climbing or

nm are higher than the radar beams by about 260 m and 400 m respectively.

by the TDWR radar at the time indicated in the legend on 19 April 2008. X axis is the distance between aircraft and the end of runway. The distance interval shaded by the green color indicates where the aircraft passes through the altitude interval observed with the 0.6o

measurements.

elevated beam.

value.

UTC.

descending (Gilbert et al., 2004).

This chapter discusses the aviation applications of TDWR. This radar issues microburst alerts which are crucial in the assurance of aviation safety. A typical case of microburst detection by TDWR in association with intense thunderstorms is described first in this chapter. Then the applications of TDWR in the alerting of windshear and turbulence are described. Windshear is alerted through the calculation of windshear hazard factor, which is a rather well established technology. On the other hand, the use of spectrum width data from the radar in the alerting of turbulence has a relatively shorter development history, and the technology is under exploration in Hong Kong.

Study is underway in Hong Kong to use X-band radar in the alerting of windshear and turbulence on experimental basis at the Hong Kong International Airport. The use of longrange S band radar in the alerting of turbulence for enroute aircraft is also under study. Such progress of these studies would be reported in the future.

#### **6. References**


**Part 2** 

**Precipitation Estimation and Nowcasting** 

