**3. Resources**

statistical model for the estimating PM mass concentrations using AOT measured from remote sensing techniques and meteorological parameters. Furthermore, the ground truth observa‐ tions collected within AIRSPACE project were used to assess qualitative and quantitative

daily average concentrations there was a 0.9% (95%CI: 0.6%, 1.2%) increase in all-cause and 1.2% (95%CI: -0.0%, 2.4%) increase in cardiovascular admissions. A recent study regarding dust storm events in Nicosia, Cyprus, found a 2.43% (95% CI: 0.53, 4.37) increase in daily

Air pollution in large cities is one of the major issues to be addressed by local and global communities due to its widespread presence, and deleterious impact on human life (Hadji‐ mitsis, 2009). As air pollution is a major environmental health risk, by reducing the levels of air pollutants, countries will reduce the incidence of disease from respiratory infections, heart disease and lung cancer (WHO, 2011). Actions by policy makers and public authorities at the national, regional and international levels are required in order to control the exposure to air pollutants (EEA, Air Quality in Europe, 2012 - Report). Transboundary domestic air pollution is of high concern among the EU member states. In 2010, about 21% of the EU urban population was exposed to concentrations of PM10 above the limit value established by the European Environmental Agency (EEA, Air Quality in Europe, 2012 - Report). The WHO, USEPA (U.S. Environmental Protection Agency) and EEA have established an extensive body of legislation which establishes standards and objectives for a number of air pollutants such as PM10 (coarse particles), PM2.5 (fine particles) and O3 (WHO, Fact sheet No 313, 2011; USEPA, NAAQS, 2012;

Current research focused on the study of regional and intercontinental transport of air pollutants, such as particulate matter (PM10, 2.5), points to a need for additional data sources to monitor air pollution in multiple dimensions, both spatially and temporally. To address this issue, earth observations from satellite sensors can be a valuable tool for monitoring air

Although air quality monitoring stations have been established in major cities, there is an increased need to establish mobile stations for additional coverage, as such stations provide a means for alerting the public regarding air quality. However, measuring stations are localised and do not provide sufficient geographic coverage, since air quality is highly variable spatially. The use of earth observations to monitor air pollution in different geographical areas, espe‐ cially cities, has received considerable attention from researchers (see Wald et al., 1999; Grosso and Paronis, 2012; Hadjimitsis, 2009; Hadjimitsis et al., 2010; Jones and Christopher, 2007; Michaelides et al., 2011; Nisantzi et al., 2012; Retalis and Sifakis, 2010; Retalis et al., 2003; Retalis et al., 1999; Vidot et al., 2007). Several researchers (Chudnovsky et al., 2013; Gupta et al., 2006; Koelemeijer et al., 2006; van Donkelaar et al., 2010) have focused on the use of satellite sensors on air pollution studies, especially their ability for systematic monitoring and synoptic coverage. The use of sunphotometers and LIDAR systems are found to be suitable tools for

pollution due to their ability to provide complete and synoptic views of large areas.

increase in PM10

increase in PM10 concentrations on

performance of a chemical model forecast of PM concentrations throughout Cyprus.

Midletton et al. (2008) reported that in Nicosia, Cyprus for every 10-μg/m3

dust days in comparison with non-dust days (Neophytou et al., 2013).

cardiovascular mortality associated with each 10-μg/m3

182 Remote Sensing of Environment: Integrated Approaches

**2. Background**

EEA, AAS, 2012).

### **3.1. CIMEL Sunphotometer**

At the main study site in Limassol, the sunphotometer observations were performed by a CIMEL sun-sky radiometer, which is part of the AERONET Global Network (http://aero‐ net.gsfc.nasa.gov). The CIMEL sunphotometer allows for measurements of direct solar irradiance and sky radiance at 8 wavelengths; 340, 380, 440, 500, 670, 870, 1020 and 1640 nm. The technical specifications of the instrument are given in detail by Holben et al. (1998).

**Figure 1.** CUT-TEPAK AERONET station

The instrument is located on the roof of the building of the Department of Civil Engineering and Geomatics of Cyprus University of Technology (CUT) (34.675ºN, 33.043ºE elevation: 10 m). The CUT\_TEPAK AERONET station is located in the center of Limassol, 500m away from the sea. The sunphotometric station has been in operation since April 2010. Figure 1 features the CUT-TEPAK AERONET Cimel sun-photometer.

#### **3.2. MICROTOPS Sunphotometer**

At the study sites in Nicosia, Larnaka and Paphos where CIMEL's data were not available a handheld MICROTOPS II sunphotometer was used in order to retrieve AOT measurements. The sun-photometer is equipped with five accurately aligned optical collimators and internal baffles to eliminate internal reflections. Microtops II provides AOT and water vapor retrievals at five channels, which are determined using the Bouguer-Lambert-Beer law. In order to achieve measurements with great accuracy, the sunphotometer was mounted on a tripod at the same location each time. To avoid cloud contamination, measurements were taken during cloud-free daylight hours. Figure 2 shows the MICROTOPS II handheld sunphotometer used.

**Figure 2.** MICROTOPS II handheld sunphotometer

#### **3.3. CUT LIDAR System**

For the vertical distribution of aerosols, the LIDAR system located at CUT, in Limassol, Cyprus (34.675ºN, 33.043ºE, 10 m above sea level) was used. The LIDAR records daily measurements between 08:00 UTC and 09:00 UTC (consistent with MODIS overpass) and provides continuous measurements for the retrieval of the aerosol optical properties over Limassol, Cyprus inside the Planetary Boundary Layer (PBL) and the lower free troposphere, thus providing informa‐ tion for the load, the size and the sphericity of the aerosols.

whole range starting at the full overlap of the LIDAR (~300 m) up to tropopause level. Three channels are detected, one for the wavelength 1064 nm and two for 532 nm. The two polari‐ zation components at 532nm are separated in the receiver by means of polarizing beamsplitter cubes (PBC). A special optomechanical design allows the manual ±45°-rotation of the whole depolarization detector module with respect to the laser polarization for evaluating the depolarization calibration constant of the system. The CUT depolarization LIDAR operates at 532nm and it is possible to rotate the detection box including the polarization beam-splitter cube in order to calibrate the instrument (Freudenthaler et al., 2009). Firstly, the backscattered LIDAR signals (P and S) were recorded using the normal orientation of the LIDAR detection box. For the next two steps, the LIDAR detection box is rotated by ±45º, and the P and S signals are recorded. The operation principal of this method is based on the fact that same amount of energy is sent to P and S channels, at "opposite" directions (Freudenthaler et al., 2009). Photomultiplier tubes (PMTs) are used as detectors at all wavelengths except for the signals at 1064 nm (avalanche photodiode, APD). A transient recorder that combines a powerful A/D converter (12 bit at 20 MHz) with a 250 MHz fast photon counting system (Licel, Berlin) is used for the detection of 532 nm radiation, while only analog detection is used at 1064nm. The raw

Air Pollution from Space http://dx.doi.org/10.5772/39310 185

**Figure 3.** CUT's Depolarization Lidar System

signal spatial resolution is 7.5 meters. The CUT LIDAR system is featured in Figure 3.

The LIDAR transmits laser pulses at 532 and 1064 nm simultaneously and collinear with a repetition rate of 20 Hz. This system is based on a small, rugged, flashlamp-pumped Nd-YAG laser with pulse energies around 25 and 56 mJ at 1064 and 532 nm, respectively. An achromatic beam expander reduces the divergence to less than 0.15 mrad. Elastically backscatter signals at two wavelengths (532nm, 1064nm) are collected with a Newtonian telescope with primary mirror diameter of 200 mm and an overall focal length of 1000 mm. The field of view (FOV) of the telescope is 2 mrad. The mirror and cover plate coatings are optimized for the wavelength range from 532 nm to 1064 nm. A plain cover plate protects the mirrors. Behind the field stop two plano-convex with a focal length of 80 mm output parallel rays. The LIDAR covers the

**Figure 3.** CUT's Depolarization Lidar System

The sun-photometer is equipped with five accurately aligned optical collimators and internal baffles to eliminate internal reflections. Microtops II provides AOT and water vapor retrievals at five channels, which are determined using the Bouguer-Lambert-Beer law. In order to achieve measurements with great accuracy, the sunphotometer was mounted on a tripod at the same location each time. To avoid cloud contamination, measurements were taken during cloud-free daylight hours. Figure 2 shows the MICROTOPS II handheld sunphotometer used.

For the vertical distribution of aerosols, the LIDAR system located at CUT, in Limassol, Cyprus (34.675ºN, 33.043ºE, 10 m above sea level) was used. The LIDAR records daily measurements between 08:00 UTC and 09:00 UTC (consistent with MODIS overpass) and provides continuous measurements for the retrieval of the aerosol optical properties over Limassol, Cyprus inside the Planetary Boundary Layer (PBL) and the lower free troposphere, thus providing informa‐

The LIDAR transmits laser pulses at 532 and 1064 nm simultaneously and collinear with a repetition rate of 20 Hz. This system is based on a small, rugged, flashlamp-pumped Nd-YAG laser with pulse energies around 25 and 56 mJ at 1064 and 532 nm, respectively. An achromatic beam expander reduces the divergence to less than 0.15 mrad. Elastically backscatter signals at two wavelengths (532nm, 1064nm) are collected with a Newtonian telescope with primary mirror diameter of 200 mm and an overall focal length of 1000 mm. The field of view (FOV) of the telescope is 2 mrad. The mirror and cover plate coatings are optimized for the wavelength range from 532 nm to 1064 nm. A plain cover plate protects the mirrors. Behind the field stop two plano-convex with a focal length of 80 mm output parallel rays. The LIDAR covers the

**Figure 2.** MICROTOPS II handheld sunphotometer

184 Remote Sensing of Environment: Integrated Approaches

tion for the load, the size and the sphericity of the aerosols.

**3.3. CUT LIDAR System**

whole range starting at the full overlap of the LIDAR (~300 m) up to tropopause level. Three channels are detected, one for the wavelength 1064 nm and two for 532 nm. The two polari‐ zation components at 532nm are separated in the receiver by means of polarizing beamsplitter cubes (PBC). A special optomechanical design allows the manual ±45°-rotation of the whole depolarization detector module with respect to the laser polarization for evaluating the depolarization calibration constant of the system. The CUT depolarization LIDAR operates at 532nm and it is possible to rotate the detection box including the polarization beam-splitter cube in order to calibrate the instrument (Freudenthaler et al., 2009). Firstly, the backscattered LIDAR signals (P and S) were recorded using the normal orientation of the LIDAR detection box. For the next two steps, the LIDAR detection box is rotated by ±45º, and the P and S signals are recorded. The operation principal of this method is based on the fact that same amount of energy is sent to P and S channels, at "opposite" directions (Freudenthaler et al., 2009). Photomultiplier tubes (PMTs) are used as detectors at all wavelengths except for the signals at 1064 nm (avalanche photodiode, APD). A transient recorder that combines a powerful A/D converter (12 bit at 20 MHz) with a 250 MHz fast photon counting system (Licel, Berlin) is used for the detection of 532 nm radiation, while only analog detection is used at 1064nm. The raw signal spatial resolution is 7.5 meters. The CUT LIDAR system is featured in Figure 3.

#### **3.4. Surface monitoring**

#### *3.4.1. PM10 concentration monitoring*

Two approached were used to monitor near-surface levels of particulate matter. DustTrack monitors were used at all sites to provide continuous monitoring of PM10. Harvard Impac‐ tors were usedto collect 24 hour samples ofPM10 andPM2.5 which couldbe analyzedfor mass, elemental composition, and other physical-chemical properties of the aerosol. The surface monitoringofparticulatematter(PM)concentrations,TheDustTrack(TSI,Model8533)monitors were located in each of the four sampling sites and were selected to provide weekly monitor‐ ing of PM10 concentrations during morning hours from 08:00 to 13:00 UTC. It records the PM temporal variability with satisfactory time resolution.DustTrak's nominalflow rate of 1.7 l/min is obtainedby an internalpumpintegralto the sampler.The monitoris factory calibratedforthe respirable fraction of standard ISO12103-1, A1 test dust (Arizona Test Dust), which is represen‐ tative of a wide variety of aerosols. It measures concentrations in the range of 0.001– 100 mg/m3 , with a resolution of 0.1% of the reading or 0.001 mg/m3 . Before each measurement, the instru‐ mentis zeroedandits flowrate is checked.PM10 concentrationshavebeenrecordedcontinuous‐ ly since March 2011. The instrument is located, on the roof of the Cyprus International Institute (CII) in Limassol, at 10 m above ground level in order to avoid the measurements being affected by localized pollution such as passing cars. PM10 concentrations were also recorded by Dust‐ Trak (TSI, Model 8520) at Nicosia, Larnaca and Paphos. One TSI DustTrack has been operated by Frederick University since July 2011 and is located at the top of the Frederick University librarybuildinginNicosia,at10mabovegroundlevel.The secondDustTrackhasbeenoperated by CUT's scientific team during 15-day campaigns at Larnaca and Paphos. All sampling points were selected to ensure exposure to wind and to be free of other obstacles. Figure 4 features the TSI Dust Trak. Harvard Impactors were operated each third day at the primary sampling site in Limassol and every sixth day at the other sampling sites.

providing comprehensive and reliable data on the air pollution throughout Cyprus based on

Air Pollution from Space http://dx.doi.org/10.5772/39310 187

Air pollution near ground level measurement sites were established in the four cities of Cyprus: Nicosia, Larnaca, Limassol and Paphos. These sites were located at positions thought to be representative of air pollution in each city. In Nicosia, the site is located on the roof of the Frederick University library building, on the same site where the DustTrak and sunphotometer were operated. The Larnaca site is located in the center of the city, on the roof of the tax agency building. The Limassol site is located on the roof of the CII building in the center of the city and Paphos site is on the roof of the economics department of Paphos Municipality. In Figure

The Moderate Resolution Imaging Spectro-Radiometer (MODIS) observations from the TERRA and AQUA satellites both measuring spectral radiance in 36 channels (412–14200 nm), in with resolutions between 250 m and 1 km (at nadir) were used to provide a climatology for Cyprus. In polar orbit, approximately 700 km above the Earth, MODIS views a swath of approximately 2300 km resulting in near daily global coverage of Earth's land/ocean/atmos‐ phere system. The swath is broken into 5-min ''granules'', each approximately 2,030 km long.

ground level measurements.

**Figure 5.** Harvard Samplers

*3.4.3. Satellite observations*

5 the setup of the Harvard samplers is presented.

**Figure 4.** TSI DUST-Track

#### *3.4.2. PM10 sampling and elemental composition determinations*

Under the AIRSPACE project, the Harvard School of Public Health (HSPH) and Cyprus International Institute for Environmental and Public Health (CII) were responsible for providing comprehensive and reliable data on the air pollution throughout Cyprus based on ground level measurements.

Air pollution near ground level measurement sites were established in the four cities of Cyprus: Nicosia, Larnaca, Limassol and Paphos. These sites were located at positions thought to be representative of air pollution in each city. In Nicosia, the site is located on the roof of the Frederick University library building, on the same site where the DustTrak and sunphotometer were operated. The Larnaca site is located in the center of the city, on the roof of the tax agency building. The Limassol site is located on the roof of the CII building in the center of the city and Paphos site is on the roof of the economics department of Paphos Municipality. In Figure 5 the setup of the Harvard samplers is presented.

**Figure 5.** Harvard Samplers

**3.4. Surface monitoring**

**Figure 4.** TSI DUST-Track

*3.4.1. PM10 concentration monitoring*

186 Remote Sensing of Environment: Integrated Approaches

with a resolution of 0.1% of the reading or 0.001 mg/m3

in Limassol and every sixth day at the other sampling sites.

*3.4.2. PM10 sampling and elemental composition determinations*

Two approached were used to monitor near-surface levels of particulate matter. DustTrack monitors were used at all sites to provide continuous monitoring of PM10. Harvard Impac‐ tors were usedto collect 24 hour samples ofPM10 andPM2.5 which couldbe analyzedfor mass, elemental composition, and other physical-chemical properties of the aerosol. The surface monitoringofparticulatematter(PM)concentrations,TheDustTrack(TSI,Model8533)monitors were located in each of the four sampling sites and were selected to provide weekly monitor‐ ing of PM10 concentrations during morning hours from 08:00 to 13:00 UTC. It records the PM temporal variability with satisfactory time resolution.DustTrak's nominalflow rate of 1.7 l/min is obtainedby an internalpumpintegralto the sampler.The monitoris factory calibratedforthe respirable fraction of standard ISO12103-1, A1 test dust (Arizona Test Dust), which is represen‐ tative of a wide variety of aerosols. It measures concentrations in the range of 0.001– 100 mg/m3

mentis zeroedandits flowrate is checked.PM10 concentrationshavebeenrecordedcontinuous‐ ly since March 2011. The instrument is located, on the roof of the Cyprus International Institute (CII) in Limassol, at 10 m above ground level in order to avoid the measurements being affected by localized pollution such as passing cars. PM10 concentrations were also recorded by Dust‐ Trak (TSI, Model 8520) at Nicosia, Larnaca and Paphos. One TSI DustTrack has been operated by Frederick University since July 2011 and is located at the top of the Frederick University librarybuildinginNicosia,at10mabovegroundlevel.The secondDustTrackhasbeenoperated by CUT's scientific team during 15-day campaigns at Larnaca and Paphos. All sampling points were selected to ensure exposure to wind and to be free of other obstacles. Figure 4 features the TSI Dust Trak. Harvard Impactors were operated each third day at the primary sampling site

Under the AIRSPACE project, the Harvard School of Public Health (HSPH) and Cyprus International Institute for Environmental and Public Health (CII) were responsible for

,

. Before each measurement, the instru‐

#### *3.4.3. Satellite observations*

The Moderate Resolution Imaging Spectro-Radiometer (MODIS) observations from the TERRA and AQUA satellites both measuring spectral radiance in 36 channels (412–14200 nm), in with resolutions between 250 m and 1 km (at nadir) were used to provide a climatology for Cyprus. In polar orbit, approximately 700 km above the Earth, MODIS views a swath of approximately 2300 km resulting in near daily global coverage of Earth's land/ocean/atmos‐ phere system. The swath is broken into 5-min ''granules'', each approximately 2,030 km long. Aerosol products are reported at 10 km resolution (at nadir). Details of file specification of MODIS L2 aerosol products can be found at the website http://modis.gsfc.nasa.gov/.

**2. Vertical profile of the aerosol backscatter:** A light detection and ranging (LIDAR) system was established in Limassol in April 2010, consisting of a laser capable of measuring aerosol backscatter and aerosol depolarization ratio in the atmosphere as a function of height. This allows the AOT, and hence the scaling to aerosol concentration, to be quantified below the boundary layer since this fraction best represents the PM measure‐

1) **Satellite data products from the MODIS sensor**: Aerosol optical thickness (AOT) and

**Figure 7 :** Overall Methodology of the Airspace project

**LIDAR data [aerosol vertical distribution]**

size distribution DAILY

Backscatter coefficient and particle depolarization ratio profiles

**TSI Dust track PM10 concentration**

> PM10 concentration

**Meteo data Surface T, P, RH and wind** 

> **Chemical Model PM estimation**

**AIRSPACE Outputs Instrumentation Retrievals** 

**Statistical Model PM estimation** 

Air Pollution from Space http://dx.doi.org/10.5772/39310 189

**CIMEL/Microtops II Columnar aerosol parameters**

AOT, AE

**Climatology Natural-Anthropogenic** 

**MODIS VALIDATION** 

**Figure 7.** Overall Methodology of the Airspace project

2) **Vertical profile of the aerosol backscatter**: A light detection and ranging (LIDAR) system was established in Limassol in April 2010, consisting of a laser capable of measuring aerosol backscatter and aerosol depolarization ratio in the atmosphere as a function of height. This allows the AOT, and hence the scaling to aerosol concentration, to be quantified below the boundary layer since this fraction best represents the PM measurements in a well-mixed

**3. Integrated aerosol optical thickness for the entire atmospheric column:** A sunphotom‐ eter station was installed in the centre of Limassol (at the CUT premises), where pollution from both industrial and urban sources exist. This further assists in the calibration and verification of satellite derived AOT data. Moreover, two hand-held sunphotometers were

3) **Integrated aerosol optical thickness for the entire atmospheric column**: A sunphotometer station was installed in the centre of Limassol (at the CUT premises), where pollution from both industrial and urban sources exist. This will further assist calibration and verification of satellite derived AOT data. Moreover, two hand-held sunphotometers

4) **Measurements of particulate matter (PM) concentration levels**: PM10 and PM2.5 hourly concentration values were collected using portable PM samplers available from the CUT Remote Sensing Laboratory and the Harvard University. Additionally, the HARVARD samplers used and located in Limassol, Nicosia, Larnaca and Paphos regions were used to

5) **Meteorological data from the entire area of Cyprus:** Relative humidity measurements combined with the AOT fraction below the boundary layer derived by the LIDAR were incorporated into the statistical PM-AOT models, for improving the PM concentration estimation. Classification of the synoptic situations in Cyprus was also taken into account.

**4. Measurements of particulate matter (PM) concentration levels:** Ground level PM was monitored using two methods. Continuous measurements of PM10 were taken using portable monitors (DustTrack model 8533). These continuous measurements were supplemented with measurements of PM10 and PM2.5 taken using Harvard Impactors. The material collected by the Harvard Impactor was analyzed for chemical composition.

**5. Meteorological data from the entire area of Cyprus:** Relative humidity measurements combined with the AOT fraction below the boundary layers, derived by the LIDAR, were incorporated into the statistical PM-AOT models, for improving the PM concentration estimation. Classification of the synoptic situations in Cyprus was also taken into account.

**6. Simulation results from dispersion/air pollution model:** A modeling system that incorporates a fully interactive coupling between the chemistry-aerosol and meteorology (radiation and cloud-physics) portions of the model was created, allowing real-time

ments in a well-mixed boundary layer.

boundary layer.

**4.1 Method** 

**Satellite aerosol Products [AOT, AE]** 

AOT & AE

**10 years Aerosol** 

used to measure urban, industrial and dust pollution.

were used to measure urban, industrial and dust pollution.

provide chemical characterization of the collected aerosols.

The overall methodology is described in the following (see Fig. 7):

aerosol size/type data were collected for the years 2002-2010 over Cyprus.

**Aerosol separation** 

**Figure 6.** MODIS image for Eastern Mediterranean region
