**3. TC structure analysis**

Soon after the first satellite weather observations became available, routine TC intensity studies from this imagery began [23]. Determining the TC's center position was the only reliably produced operational use of TC satellite analysis until the influential work of Dvorak [22] in 1972. Dvorak developed an empirical model to diagnose TC intensity based on cloud organi‐ zation in the visible channel. Factors that affected strength of the storm were accounted by structural features such as magnitude of the brightness, temperatures, and curvature of the banding or distortion of the cloud pattern. A flowchart of rules in the Dvorak method was created to consistently generate subjective estimates of TC intensity from the structural cues in satellite imagery. Dvorak refined his method with time to incorporate additional rules and constrictions as well as use of infrared channels, aiming to improve the accuracy and reduce subjectivity of the technique.

The Dvorak technique continues as a critical part of operational analysis today. It is still the most common method to diagnose TC intensity, with further attempts at intensity analysis heavily borrowing for this work's legacy. Some efforts continue to hone the technique's accuracy, such as updating the wind‐pressure relationship used to define the intensity [24]. The application and customization of Dvorak analysis worldwide described in [23] help demonstrate its versatility and robustness as a tool for TC forecasting. Efforts to make the process more objective led to the creation of a fully automated TC intensity analysis [25], which itself continues to have updates as the Advanced Dvorak Technique (ADT) [26].

Beyond the Dvorak technique, there are several other methods to diagnose TC structure and intensity from infrared and visible channels. The correlations between infrared TBs and operational storm size metrics (i.e., radius of maximum winds (RMW), and radii of the 34‐, 50‐, and 64‐kt winds) were presented in 2007 [27]. A similar use of infrared imagery to resolve TC sizes, as defined by the radius of the 5‐kt 850 hPa wind, that relate to the TC lifecycle were studied [28]. A statistical analysis of the distribution of temperatures with respect to TC axisymmetry (the deviation angle variance technique) continues to show great promise in not only diagnosing structure, but also helping with TC centering, genesis, and intensity [29]. Another potential relationship between TC intensity and structure changes can be seen by differencing the GEO water vapor and infrared channels [30]. This methodology leverages information in each channel (due to the weighting function representing different altitudes and chemical profiles) to emphasize specific structural features such as overshooting convec‐ tive tops. Finally, there are efforts to relate the rotational speed of IR and visible cloud tops about the TC center to the intensity of major typhoons in the northwestern Pacific basin [31].

center. The bottom‐left panel is an image of the AMSR‐E 89H GHz channel, which clearly shows positions of the TC eyewall and center as well as convection cells. The bottom‐right panel is a composite image of the PMW polarization corrected temperature (PCT) in red, 85 GHz or 89 GHz vertical polarization in green, and horizontal polarization in blue [23]. It can provide additional information of the TC cloud patterns important to TC's overall structure and organization, especially for a potential low‐level circulation center. Therefore, this multi‐ sensor‐panel imagery can be applied for better analysis of these important TC characteristics such as eyewall development, formation of concentric eyewall, convective zones and cells, rain

Soon after the first satellite weather observations became available, routine TC intensity studies from this imagery began [23]. Determining the TC's center position was the only reliably produced operational use of TC satellite analysis until the influential work of Dvorak [22] in 1972. Dvorak developed an empirical model to diagnose TC intensity based on cloud organi‐ zation in the visible channel. Factors that affected strength of the storm were accounted by structural features such as magnitude of the brightness, temperatures, and curvature of the banding or distortion of the cloud pattern. A flowchart of rules in the Dvorak method was created to consistently generate subjective estimates of TC intensity from the structural cues in satellite imagery. Dvorak refined his method with time to incorporate additional rules and constrictions as well as use of infrared channels, aiming to improve the accuracy and reduce

The Dvorak technique continues as a critical part of operational analysis today. It is still the most common method to diagnose TC intensity, with further attempts at intensity analysis heavily borrowing for this work's legacy. Some efforts continue to hone the technique's accuracy, such as updating the wind‐pressure relationship used to define the intensity [24]. The application and customization of Dvorak analysis worldwide described in [23] help demonstrate its versatility and robustness as a tool for TC forecasting. Efforts to make the process more objective led to the creation of a fully automated TC intensity analysis [25], which

Beyond the Dvorak technique, there are several other methods to diagnose TC structure and intensity from infrared and visible channels. The correlations between infrared TBs and operational storm size metrics (i.e., radius of maximum winds (RMW), and radii of the 34‐, 50‐, and 64‐kt winds) were presented in 2007 [27]. A similar use of infrared imagery to resolve TC sizes, as defined by the radius of the 5‐kt 850 hPa wind, that relate to the TC lifecycle were studied [28]. A statistical analysis of the distribution of temperatures with respect to TC axisymmetry (the deviation angle variance technique) continues to show great promise in not only diagnosing structure, but also helping with TC centering, genesis, and intensity [29]. Another potential relationship between TC intensity and structure changes can be seen by differencing the GEO water vapor and infrared channels [30]. This methodology leverages

itself continues to have updates as the Advanced Dvorak Technique (ADT) [26].

band formation, central dense overcast, shear, and low‐level center.

142 Recent Developments in Tropical Cyclone Dynamics, Prediction, and Detection

**3. TC structure analysis**

subjectivity of the technique.

Besides the traditional infrared and visible channels, other remote sensing frequencies have proven useful for observing TCs. The use of satellite active microwave radars (scatterometers) to estimate storm size and intensity via surface wind analysis has been very beneficial to operational efforts at the National Hurricane Center (NHC) [32]. Scatterometers transmit pulses that bounce off the ocean surface; the backscatter's variability due to wind roughening enables retrievals of the surface wind vector estimates. QuikSCAT scatterometer data were used to derive climatology of storm sizes at the 23 kt radius as well as an outer radius using a wind structure model [33]. Similar studies were done to create a QuikSCAT climatology of storm sizes for the 34‐kt radius and outer‐core strength (OCS) intensity [34–36]. TC structure and intensity are also investigated using other satellite sensor datasets such as the TRMM precipitation radar (PR) and TMI in tandem with lightning flash density to compare differences in frequency thresholds between regions of the TC [37]. The synthetic aperture radar (SAR) was used to visualize extremely small mesoscale details of TC and subjectively catalogue characteristics of the eye, including spatial area, shape, and wavenumber [38]. SAR TC retrievals are currently poorly tied to physical processes and the radar retrievals from space occur infrequently, while passive microwave can directly characterize TC structure by penetrating non‐raining cloud tops, unlike in the visible and infrared channels [39].

While the ultimate purpose and mechanism of the TC eye formation remain uncertain, there are many characteristic modes associated with TC eye. In general, due to dynamical consid‐ erations of the eyewall, the TC eye tends to get smaller. Eventually, a new outer eyewall may form and produce an outer eye that encompasses the previous smaller eye. In rare cases of an extremely intense and compact TC, a "pinhole" eye will form [27]. In some other cases, a large and stable eye with multiple embedded meso‐vorticies can form [40, 41]. The TC eye shape is not always circular, as polygonal eyewalls seem to be due to vorticity instability [42, 43]. Recent observations show eye‐like features developing in the lower troposphere before being observed in the upper troposphere [44].

The strongest vertical motion in TCs was found just inward of the RMW [45, 46]. In addition, the eyewall tended to slope outward with height due to the TC warm core. The TC intensity and eyewall slope relationship shows a great deal of case‐to‐case variability [47]. Due to this slope, the updraft itself was also tilted so that most of the falling hydrometeors in the eyewall (and thus the radar reflectivity maximum) lie outside the RMW [46]. The high‐resolution airborne Doppler radar was recently used to update and extend these results [48, 49]. Regions of the TC core, defined by normalization with respect to the RMW, are shown to exhibit modes of radar reflectivity, convergence/divergence, and vorticity that correspond to the previously cited work; particularly, there is an outer peak in upper‐level divergence and low‐level convergence that occurs in the vicinity of secondary eyewall formation.

The nature of TC convection occurring in spiral bands was not known until their first obser‐ vations on radar [50]. However, the TC spiral band was quantitatively and qualitatively characterized shortly thereafter [51, 52]. A logarithmic spiral based on radar observations was introduced to start at an inner radial circle (i.e., the inner‐core or eyewall) rather than the center itself [51]. An analysis of the logarithmic spiral length has been quite useful in diagnosing intensity [22]. The idea that low‐latitude TCs have lower spiral crossing angles than higher latitude storms possibly due to storm motion was also introduced [51]. The TC spiral bands propagate outward [52], while some rain bands actually propagate inward when taking into account storm motion [53]. In general, there may be three types of spiral bands based on movement: stationary (non‐propagating), apparent propagation (stationary with respect to the TC center), and intrinsic propagation [50].

**Figure 5.** Two schematics of TC structures from [109]. (a) Horizontal cross‐section of structural features as presented by radar. (b) Vertical cross‐section of the same structural features and their relation to the secondary circulation described in Eliassen [110] and Shapiro and Willoughby [111] (adapted from Willoughby (1995)).

These structural characteristics are summarized in **Figure 5** for a well‐organized double eyewall TC. The inner eyewall, outer eyewall, principal convective band, and secondary convective band are clearly presented with radar reflectivity (**Figure 5a**). The stratiform precipitation occurs largely in the moat areas. These unique TC features are clearly captured by the PMW sensor's measurements at high‐frequency channels as shown in **Figure 4**. The TC's tilted updraft at the eyewall, forced descent air at the eye, lower level inflow and upper level outflow, brightband associated with stratiform precipitation as well as mesoscale updraft (downdraft) above (below) the brightband are demonstrated in **Figure 5b**. Detailed TC vertical temperature profiles can also be observed by PMW sensors. **Figure 6** shows cross section of the AMSU‐retrieved temperature anomalies through hurricane Bonnie at 1200 UTC August 25, 1998. The TC warm core near 250 hPa and the vertical temperature profiles match well with observations. Thus, this unique warm core feature can be applied for TC intensity estimates discussed in the next section.

**Figure 6.** Cross section of temperature anomalies through Hurricane Bonnie at 1200 UTC 25 Aug 1998 retrieved from AMSU data (adapted from Kidder et al. (2010). ©American Meteorological Society. Used with permission).

#### **4. TC intensity estimation**

characterized shortly thereafter [51, 52]. A logarithmic spiral based on radar observations was introduced to start at an inner radial circle (i.e., the inner‐core or eyewall) rather than the center itself [51]. An analysis of the logarithmic spiral length has been quite useful in diagnosing intensity [22]. The idea that low‐latitude TCs have lower spiral crossing angles than higher latitude storms possibly due to storm motion was also introduced [51]. The TC spiral bands propagate outward [52], while some rain bands actually propagate inward when taking into account storm motion [53]. In general, there may be three types of spiral bands based on movement: stationary (non‐propagating), apparent propagation (stationary with respect to the

**Figure 5.** Two schematics of TC structures from [109]. (a) Horizontal cross‐section of structural features as presented by radar. (b) Vertical cross‐section of the same structural features and their relation to the secondary circulation described

These structural characteristics are summarized in **Figure 5** for a well‐organized double eyewall TC. The inner eyewall, outer eyewall, principal convective band, and secondary convective band are clearly presented with radar reflectivity (**Figure 5a**). The stratiform precipitation occurs largely in the moat areas. These unique TC features are clearly captured

in Eliassen [110] and Shapiro and Willoughby [111] (adapted from Willoughby (1995)).

TC center), and intrinsic propagation [50].

144 Recent Developments in Tropical Cyclone Dynamics, Prediction, and Detection

Generally, there are two types of available polar‐orbiting microwave sensors: imagers and sounders. The microwave imagers typically consist of sensors with frequencies that measure surface properties as well as organization of various water phases in the atmosphere. The microwave sounders aim to provide profiles of atmospheric thermal structure and moisture estimates. Depending on the mission goals of a particular sensor, there can be overlap between available channels of a microwave imager and sounder.

**Figure 7.** A multi‐panel comparison of the AMSU temperature structure at the upper levels of Atlantic Hurricane Rita (2005). "X" is the TC center position. The strength of the temperature anomalies represents deepening of the warm core structure and corresponds well with increasing intensity (adapted from UW Madison/CIMSS: http://trop‐ ic.ssec.wisc.edu/real‐time/amsu/).

Attempts to diagnose TC structure and intensity from microwave sounders occurred shortly after the first sounder was launched [54]. Although the coarse resolution of sounders has traditionally created an analysis barrier due to smoothing over the storm features, the more advanced sensors such as AMSU are starting to resolve the magnitude of thermal anomalies as well as core/eye size more faithfully [19, 55]. **Figure 7** presents the AMSU‐retrieved temperature anomaly distributions at 4000 ft at different stages of Hurricane Rita during 18– 21 September, 2015. It clearly shows increased amplitude of the unique TC upper level warm core feature corresponding well with intensification of the hurricane intensity. Multiple linear regressions of the AMSU channels also can estimate features such as maximum sustained wind (MSW), minimum sea level pressure (MSLP), and wind radii at the 34‐, 50‐, and 64‐kt thresh‐ olds [56]. Use of AMSU data as part of an ensemble shows great promise for more accurate TC structure retrieval; a combination of the ADT and two different AMSU intensity estimates makes up the satellite consensus (SATCON) method of TC intensity estimation [57], which has the highest skill of all satellite‐based intensity estimation methods [58]. A comparison of the TC sustained 1 minute wind estimates from different techniques is displayed for an example storm in **Figure 8**. It demonstrates that the PMW sensor‐based measurements can be used to accurately estimate the TC intensity. Although every method with individual PMW sensor in general agrees with each other, differences are still obvious. Results indicate that SATCON performs well against the TC's best track dataset. These techniques have been continuously evolved to create a better sounder TC intensity algorithm with new sensors such as SSMIS and ATMS, which have improved spatial resolution to depict the TC's warm core [18].

estimates. Depending on the mission goals of a particular sensor, there can be overlap between

**Figure 7.** A multi‐panel comparison of the AMSU temperature structure at the upper levels of Atlantic Hurricane Rita (2005). "X" is the TC center position. The strength of the temperature anomalies represents deepening of the warm core structure and corresponds well with increasing intensity (adapted from UW Madison/CIMSS: http://trop‐

Attempts to diagnose TC structure and intensity from microwave sounders occurred shortly after the first sounder was launched [54]. Although the coarse resolution of sounders has traditionally created an analysis barrier due to smoothing over the storm features, the more advanced sensors such as AMSU are starting to resolve the magnitude of thermal anomalies as well as core/eye size more faithfully [19, 55]. **Figure 7** presents the AMSU‐retrieved temperature anomaly distributions at 4000 ft at different stages of Hurricane Rita during 18– 21 September, 2015. It clearly shows increased amplitude of the unique TC upper level warm core feature corresponding well with intensification of the hurricane intensity. Multiple linear regressions of the AMSU channels also can estimate features such as maximum sustained wind (MSW), minimum sea level pressure (MSLP), and wind radii at the 34‐, 50‐, and 64‐kt thresh‐ olds [56]. Use of AMSU data as part of an ensemble shows great promise for more accurate TC structure retrieval; a combination of the ADT and two different AMSU intensity estimates makes up the satellite consensus (SATCON) method of TC intensity estimation [57], which has

available channels of a microwave imager and sounder.

146 Recent Developments in Tropical Cyclone Dynamics, Prediction, and Detection

ic.ssec.wisc.edu/real‐time/amsu/).

**Figure 8.** Satellite consensus (SATCON) intensity analysis of Typhoon Champi (2015) in the West Pacific. The solid black line shows the best track intensity from JTWC, the black dots show the subjective Dvorak satellite estimates, and all the other plots show objective satellite‐based estimates (adapted from Derrick Herndon and CIMSS, http://trop‐ ic.ssec.wisc.edu/real‐time/satcon/).

To contrast, microwave imagers have a more recent appearance in TC analysis, with perhaps more potential for added value. The near real‐time access to digital microwave imagery was the largest impediment near the turn of the millennium [39], in which authors describe efforts at NRL‐MRY to provide near real‐time access to high‐resolution TC images, and to support a temporal return frequency favorable for operational TC forecasting. A more detailed discus‐ sion about the utility of microwave imagers indicated that despite the ability to create the PMW‐derived physical quantities (e.g., sea surface wind magnitude, precipitable water, and cloud liquid water) using multiple frequencies which utilize different spatial resolutions, they smooth over important structural features of the TC [59]. These features include a TB depression at the high frequency channel (85, 89 or 91 GHz) due to ice scattering and lower frequencies such as 37 GHz principally showing liquid hydrometeor emissions near and below the freezing level. Both of these previously mentioned channels are measured at horizontal (H) and vertical (V) polarizations. Near the interface of the outer TC and the environment, interpretations at either polarization become muddled due to multiple competing influences (e.g., water vapor, cloud water, and sea surface). The polarization correction temperature (PCT) can improve the representation of atmospheric features, allowing them to stand out from surface background [60].

Despite its relatively new arrival, some progresses are apparent in using microwave imagers to examine tropical cyclones. For example, the NHC extensively uses microwave imagery to better locate a TC center and subjectively diagnose changes in structure [32]. The Morphed Integrated Microwave Imagery at CIMSS (MIMIC), a technique to create "morphed" anima‐ tions of passive microwave imagery using an advection function between satellite passes, was introduced to allow a visually appealing depiction of TC structure changes [61]. Other studies also revealed relationship between microwave imager data to TC intensity [62, 63]. The microwave data have been used to improve TC intensity estimates through early detection of a forming eyewall [64], while a color composite of the H‐pol, V‐pol, and PCT data at 37 GHz developed at NRL‐MRY has shown particular promise in diagnosing TC inner core formation [39]. A symmetrical and closed TB threshold ("cyan ring") was applied to predict the TC onset rapid intensification [44].

Some efforts focused on cataloging TCs through an extended climatology of microwave imagers. The microwave data interpolated onto an 8‐km grid in the hurricane satellite (HURSAT) archive was created in 2008 [65]. The HURSAT‐microwave consists of data from the SSM/I platforms between 1987 and 2009, using global best track data from the International Best Track Archive for Climate Stewardship (IBTrACS) to search for TCs [24]. Based on this dataset, the TCs composited by their intensification rate and environmental wind shear were analyzed to compare microwave signatures during different intensity regimes [66]. Recently, a new study on eyewall size estimates using the HURSAT‐microwave data compared to the aircraft reconnaissance measurements demonstrates the similarity of in‐situ and satellite‐ derived structural profiles [67]. A more advanced TC PMW TB database at 1 km spatial resolution has been developed at NRL‐MRY from all available PMW sensors in 1987–2012 with more consistent inter‐calibrated TBs at 89 GHz, better eye fixing, and high quality interpolation scheme [68, 69].
