2. Experimental setup

factors affecting the state of the environment are both natural (hydrometeors, volcanic ash, desert dust, smoke aerosols, etc.) and anthropogenic (industrial, fuel combustion, fires, heating, etc.) atmospheric aerosols or particulate matter (PM) [1]. The aerosols are fine solid or liquid particles, determining largely the climate, temperature, and dynamic structure of the atmosphere, the functioning of ecosystems, the microphysical properties of clouds and various chemical and photochemical processes in air [2]. The concentration of aerosols in the atmosphere determines the air quality, and in turn affects the human health [[3] and references therein]. The fine and ultrafine aerosol particles are particularly harmful for human health, as they more easily penetrate and accumulate in the human body and lead to an increase in cardiovascular and respiratory diseases and even to lung cancer [4, 5]. Because of the above facts and considerations, atmospheric aerosols have been the subject of more intensive research in recent decades [6, 7]. Particularly, near-surface atmospheric measurements over densely populated or industrial areas, purposed to help monitor air quality, have attained

Laser radars (lidars) are recognized to be a reliable and powerful instrument for investigating atmospheric objects and air parameters [7–9]. As compared to other measurement approaches, the lidar technique exhibits advantages such as possibilities for performing fast, highly sensitive and accurate monitoring of vast atmospheric domains with high spatial and temporal resolution. Lidar systems are mainly used to assess the vertical structure of the aerosol layers and determine the optical and microphysical properties of the vertical profiles of the aerosols. Scanning ground-based and airborne lidars are applied to produce threedimensional (3D) maps of earth's surface and man-made features [10], as well as for characterization of tropospheric wind profiles [11, 12] and temperature fields [13]. Along with the above applications, scanning elastic-scattering lidars are used to obtain maps of important atmospheric pollutants, particularly, of near-surface aerosol fields [14–17]. The aerosol lidar mapping represents a fast and effective approach to detect polluting aerosol loads over broad areas, as well as to characterize them in terms of local density, spatial distribution,

In this chapter, we present experimental results on lidar mapping of near-surface atmospheric aerosol fields over the city of Sofia, its suburbs and surrounding villages, obtained during an extensive 7-month experimental campaign in 2015 [19]. It was carried out in the framework of a common project with Sofia Municipality aimed to help the local authorities to improve the regional air-quality monitoring. Possibilities are also discussed to incorporate lidar mapping technologies synergistically into municipal air-quality monitoring systems. Aerosol lidar maps are considered to become basic components of such monitoring systems. Their advantages result from the high efficiency of the laser light interaction with the atmospheric particles, thus, providing better visualization of atmospheric motions in comparison with other remote sens-

The analysis of the lidar mapping experiments performed was focused on the following issues:

i. Two-dimensional (2D) aerosol density distribution above the city areas.

ii. Temporal and spatial dynamics of the near-surface aerosol fields.

increasing societal significance.

86 Aerosols - Science and Case Studies

and temporal dynamics [18–20].

ing techniques, such as the microwave or acoustic probing.

#### 2.1. Description of the area investigated

Sofia (the capital of Bulgaria) is located at about 550 m above sea level (a.s.l.), in a valley surrounded by hills and mountains, bordering Stara Planina to the northeast and Vitosha to the southwest. This topographic position, and the temperature inversion conditioned by it, is one of the factors determining the regional air quality. In the past years, the air pollution over the city of Sofia has become a serious ecological problem, provoked by the presence of different industrial facilities, a considerable decrease of green zones, as well as the accelerating growth of the population and the number of cars. The analysis of the PM amount and properties over Europe [21] showed that the aerosol concentrations have reached secondhighest levels in eastern and southern Europe.

The measurement site (42.65N, 23.38E; 590 m a.s.l.) is located at the Institute of Electronics, Bulgarian Academy of Sciences, in the southeast part of Sofia. Figure 1 shows a topographical map of Sofia area, overlaid by lidar sector scans in north-northwest (north-NW) (to the central parts of Sofia) and southwest (SW) (toward Vitosha Mountain) directions. In the former case, a horizontal lidar scanning (at a step of 1.7°) was performed in two close azimuth sectors of 8.5°(321–329.5° with respect to the north clockwise) and 17° (348–5°). In the latter case, the investigated sector was 50° (188–238°, at a step of 2°), whereas a low elevation angle (6–7°) was used to measure near-surface atmospheric aerosol fields because of the presence of high buildings in the angular sectors of measurements.

Under the conditions mentioned above, surface areas of about 100 km<sup>2</sup> were scanned and mapped over the central city zone, the north industrial zone and the south urban and suburb parts, including the north slopes of Vitosha Mountain. The results of the lidar aerosol mapping over these areas would allow one to detect and analyze the aerosols of different origin (natural, urban, industrial, etc.), as well as to contribute to the establishment of a modern city air-quality monitoring system.

Figure 1. Topographical map of Sofia region with outlined azimuth sectors of lidar scanning.

#### 2.2. Lidar systems

The measurements described in this chapter were conducted by scanning observation zones in horizontal and vertical directions over Sofia using lidar systems developed at the Laser Radars Laboratory of the Institute of Electronics, Bulgarian Academy of Sciences (LRL-IE). The LRL-IE working groups actively participate in the joint lidar research over the European continent in the framework of the European Aerosol Research Lidar Network, performing systematic lidar monitoring of atmospheric processes [18, 22], unusually high concentrations of aerosols in the troposphere [23], transport of mineral dust from Sahara desert [24], volcanic eruptions [25], and formation of smoke layers resulting from forest or industrial fires [26].

The lidar systems used in the experimental campaign considered are shown in Figure 2. The first one is based on a Cu-vapor laser emitting pulses with duration of 10 ns at a repetition rate of 5 kHz and wavelengths of 510.6 and 578.2 nm. Their mean powers are 1.2 and 0.8 W, respectively. The beam divergence is about 2 mrad. The laser beam is directed in parallel to

Figure 2. Photographs of the Cu-vapor (a) and Nd:YAG (b) laser-based lidars used in the aerosol lidar mapping experiments.

the axis of the receiving telescope, forming a lidar base of ~30 cm between the axes. A Cassegrain-type telescope with 20 cm aperture and 1 m focal length receives the backscattered laser emission from the atmosphere. Narrow-band interference filters are used to separate the lidar signals. Registration in a photon-counting mode is applied. The single electron pulses, produced by the photodetector (a photomultiplier), are accumulated by a photon-counting board in a computer. This board allows registration of the backscattered lidar signal with a spatial resolution of 15 or 30 m in 1024 samples and variable averaging time. In the present experiments, the lidar profiles of the laser emission backscattered in the atmosphere were registered with an accumulation time of 1 min. In addition, averaging was performed by summation of the data of 5–10 profiles; thus, the effective measurement time for each profile amounted to 5–10 min in a single azimuth direction.

The lidar used to perform lidar mapping experiments in the south-southwest direction is based on a solid-state frequency-doubled Nd:YAG laser (pulse energy of up to 600 mJ at 1064 nm, 80 mJ at 532 nm; fixed repetition rate of 2 Hz, FWHM pulse duration of 15 ns, beam divergence of 2 mrad), acting as a two-wavelength lidar transmitter. The optical part of the lidar receiver consists of a Cassegrain-type telescope (aperture 35 cm; focal distance 200 cm) and a threechannel spectrum analyzer based on narrow-band interference filters (1–3 nm FWHM). The receiver's electronic part comprises three compact photoelectronic modules, each including a photodetector, a 10 MHz 14-bit analog-to-digital converter (ADC), a high-voltage power supply, and controlling electronics. The signals backscattered in the atmosphere are digitized every 100 ns by the ADC, resulting in a 15 m range resolution. The system provides detection and storage of lidar returns from distances of up to 30 km. The lidar is mounted on a stable metal coaxial construction allowing reliable fixing and precise synchronized mutual motion of both the telescope and the output laser beam in horizontal and vertical direction with an angular resolution of about 1°.

2.2. Lidar systems

88 Aerosols - Science and Case Studies

The measurements described in this chapter were conducted by scanning observation zones in horizontal and vertical directions over Sofia using lidar systems developed at the Laser Radars Laboratory of the Institute of Electronics, Bulgarian Academy of Sciences (LRL-IE). The LRL-IE working groups actively participate in the joint lidar research over the European continent in the framework of the European Aerosol Research Lidar Network, performing systematic lidar monitoring of atmospheric processes [18, 22], unusually high concentrations of aerosols in the troposphere [23], transport of mineral dust from Sahara desert [24], volcanic eruptions [25],

The lidar systems used in the experimental campaign considered are shown in Figure 2. The first one is based on a Cu-vapor laser emitting pulses with duration of 10 ns at a repetition rate of 5 kHz and wavelengths of 510.6 and 578.2 nm. Their mean powers are 1.2 and 0.8 W, respectively. The beam divergence is about 2 mrad. The laser beam is directed in parallel to

and formation of smoke layers resulting from forest or industrial fires [26].

Figure 1. Topographical map of Sofia region with outlined azimuth sectors of lidar scanning.

#### 2.3. Lidar data processing: deriving the atmospheric aerosol backscatter profiles

The lidar remote sensing is based on the interaction (absorption and scattering) of the laser light with molecules and aerosols in the atmosphere. The detected backscattered lidar signals contain information concerning the state and composition of the probed atmospheric domain. The so-called lidar equation describes the power of the received backscattered signal as a range-resolved function of the lidar parameters and the atmospheric optical properties (aerosol backscattering and extinction coefficients). For a singlescattering elastic lidar (measuring backscattered light at the same wavelength as the sensing laser wavelength λ) the power P(r), detected at a time t after the instant of pulse emission, is written as [27]:

$$P(r) = P\_0 \frac{c\tau}{2} A \varepsilon \frac{\gamma'(r)}{r^2} [\beta\_a(r) + \beta\_m(r)] \exp\{-2j\_0'[a\_a(\rho) + a\_m(\rho)]d\rho\},\tag{1}$$

where P<sup>0</sup> is the average power of a single laser pulse, c is the speed of light, r = ct/2 is the distance along the laser beam path, τ is the pulse duration, A is the area of the receiver, ε is the overall system efficiency, γ(r) describes the overlap between the laser beam and the receiver field of view, and βaðrÞ, βmðrÞ and αaðrÞ, αmðrÞ are the backscattering and extinction coefficients for aerosols and molecules, respectively, at wavelength λ.

The determination of the aerosol extinction and backscattering coefficients (BSCs) on the basis of Eq. (1) requires the solution of a Bernoulli differential equation. A stable solution has been proposed by Klett [28] and Fernald [29], applying an inverse integration algorithm starting from the far end of the lidar sounding path. In the case of the backscattering coefficient (BSC), it has the following form:

$$\beta\_a(r) = -\beta\_m(r) + \frac{P(r)r^2 \exp\{-2[S\_a(r) - S\_m] \int\_r^{r\_{\text{ref}}} \beta\_m(\rho) d\rho\}}{\frac{P(r\_{\text{ref}})r\_{\text{ref}}^2}{\beta\_a(r\_{\text{ref}}) + \beta\_m(r\_{\text{ref}})} + 2\int\_r^{r\_{\text{ref}}} S\_a(\rho)P(\rho)\rho^2 \exp\left\{-2[S\_a(\rho) - S\_m] \int\_\rho^{r\_{\text{ref}}} \beta\_m(\rho') d\rho'\right\} d\rho} \tag{2}$$

where SaðrÞ ¼ αaðrÞ=βaðrÞ and Sm ¼ αmðrÞ=βmðrÞ ¼ 8π=3 are the aerosol and the molecular extinction-to-backscatter lidar ratios, respectively. The reference range rref is chosen so that the aerosol backscatter coefficient at that point is either negligible compared to the molecular backscatter coefficient or is known from other sources. The vertical profiles of the βmðrÞ could be determined from the Standard Atmosphere Model [30] and from meteorological data. This algorithm is now widely applied in practice, assuming also that the aerosol lidar ratio Sa is invariant along the laser beam path. The exact value of this ratio is determined depending on the laser wavelength and also on a priori assumptions about the atmospheric conditions and the type of the aerosols observed.

In the case of lidar measurements in vertical or quasi-vertical directions, aerosol-free atmospheric domains are usually reached at certain altitudes (as a rule higher than 5–6 km) in the free troposphere, where the total backscatter coefficient βðrÞ ¼ βaðrÞ þ βmðrÞ≈βmðrÞ is a priori known. In the case considered here of horizontal or quasi-horizontal lidar measurements, lidar paths pass through close-to-surface atmospheric parts rich in aerosols, for which molecular reference values of β(r) could not be used. In such cases, alternative and/or auxiliary approaches for retrieving the profiles of β(r) have to be applied, in order to characterize the aerosol content in the observation areas. Such an approach is the so-called "slope method" [27], applicable to characterizing atmospheric domains with relatively homogeneous aerosol composition and concentration.

A typical feature of the atmosphere is its vertical stratification, expressed in the formation of a vertical succession of horizontally extended layers of different thickness. Inside these layers, the atmospheric air content and parameters remain practically constant over considerable horizontal distances. The longer the reachable lidar range of horizontal sounding, the higher the probability such homogeneous air volumes to be present along the lidar line of sight, providing favorable conditions for the method to be used. Thus, the slope method appears to be very suitable for determining aerosol characteristics in horizontal lidar measurements or such performed at lowelevation angles. The accuracy of this method increases with increasing the aerosol concentration, favoring its application to the lidar measurements conducted in the near-surface atmospheric layers where the highest aerosol concentrations are usually observed.

Applying the slope method to solving the lidar equation, one can obtain the following expressions for the aerosol extinction and backscattering coefficients:

$$
\alpha\_a(r) = \alpha\_a = -0.5d \{ \ln[P(r)r^2] \}/dr \tag{3}
$$

and

2.3. Lidar data processing: deriving the atmospheric aerosol backscatter profiles

written as [27]:

90 Aerosols - Science and Case Studies

it has the following form:

the type of the aerosols observed.

PðrÞ ¼ P<sup>0</sup>

cτ 2 Aε γðrÞ

for aerosols and molecules, respectively, at wavelength λ.

<sup>P</sup>ðrrefÞr<sup>2</sup> ref <sup>β</sup>aðrrefÞþβmðrref<sup>Þ</sup> <sup>þ</sup> <sup>2</sup>∫

<sup>β</sup>aðrÞ ¼ <sup>−</sup>βmðrÞ þ <sup>P</sup>ðrÞr2expf−2½SaðrÞ−Sm�<sup>∫</sup>

rref

The lidar remote sensing is based on the interaction (absorption and scattering) of the laser light with molecules and aerosols in the atmosphere. The detected backscattered lidar signals contain information concerning the state and composition of the probed atmospheric domain. The so-called lidar equation describes the power of the received backscattered signal as a range-resolved function of the lidar parameters and the atmospheric optical properties (aerosol backscattering and extinction coefficients). For a singlescattering elastic lidar (measuring backscattered light at the same wavelength as the sensing laser wavelength λ) the power P(r), detected at a time t after the instant of pulse emission, is

<sup>r</sup><sup>2</sup> <sup>½</sup>βaðrÞ þ <sup>β</sup>mðrÞ�expf−2<sup>∫</sup>

where P<sup>0</sup> is the average power of a single laser pulse, c is the speed of light, r = ct/2 is the distance along the laser beam path, τ is the pulse duration, A is the area of the receiver, ε is the overall system efficiency, γ(r) describes the overlap between the laser beam and the receiver field of view, and βaðrÞ, βmðrÞ and αaðrÞ, αmðrÞ are the backscattering and extinction coefficients

The determination of the aerosol extinction and backscattering coefficients (BSCs) on the basis of Eq. (1) requires the solution of a Bernoulli differential equation. A stable solution has been proposed by Klett [28] and Fernald [29], applying an inverse integration algorithm starting from the far end of the lidar sounding path. In the case of the backscattering coefficient (BSC),

where SaðrÞ ¼ αaðrÞ=βaðrÞ and Sm ¼ αmðrÞ=βmðrÞ ¼ 8π=3 are the aerosol and the molecular extinction-to-backscatter lidar ratios, respectively. The reference range rref is chosen so that the aerosol backscatter coefficient at that point is either negligible compared to the molecular backscatter coefficient or is known from other sources. The vertical profiles of the βmðrÞ could be determined from the Standard Atmosphere Model [30] and from meteorological data. This algorithm is now widely applied in practice, assuming also that the aerosol lidar ratio Sa is invariant along the laser beam path. The exact value of this ratio is determined depending on the laser wavelength and also on a priori assumptions about the atmospheric conditions and

In the case of lidar measurements in vertical or quasi-vertical directions, aerosol-free atmospheric domains are usually reached at certain altitudes (as a rule higher than 5–6 km) in the free troposphere, where the total backscatter coefficient βðrÞ ¼ βaðrÞ þ βmðrÞ≈βmðrÞ is a priori known. In the case considered here of horizontal or quasi-horizontal lidar measurements, lidar

r

rref

<sup>r</sup> SaðρÞPðρÞρ<sup>2</sup> exp <sup>f</sup>−2½SaðρÞ−Sm�∫

<sup>r</sup> βmðρÞdρg

rref <sup>ρ</sup> βmðρ′

Þdρ′ gdρ (2)

<sup>0</sup>½αaðρÞ þ αmðρÞ�dρg, (1)

$$
\beta\_a(r) = \beta\_a = \alpha\_a / \mathcal{S}\_a \tag{4}
$$

Within the mentioned atmospheric parts of homogeneous aerosol parameters, Sa keeps empirically defined constant values depending on the aerosol types and densities. Both accuracy and reliability of the slope method increase proportionally to the lengths of the homogeneous parts present along the lidar beam path.

As an important advantage of the slope method, in comparison to other lidar approaches, determination of the aerosol extinction or backscattering is only or predominantly based on lidar measurement data, without the need of information or suppositions concerning relations between the analyzed quantities. In addition, the method makes use of simple mathematics, provides analytical solutions, and does not require numerical approaches and algorithms.

To implement the lidar mapping described here of the near-surface aerosol density distribution over Sofia region, a combination of the widely adopted and well-elaborated method of Klett-Fernald and the slope method was used. In this combination, the slope method is applied to determining the aerosol extinction and backscattering coefficients in appropriate parts of the lidar beam path by using the technology presented above (Eqs. (3) and (4)). Subsequently, the values of βaðrÞ obtained are used as reference (calibrating) ones in Eq. (2), in retrieving the whole range profiles of βaðrÞ by means of the Klett-Fernald approach. In this manner, the advantages of the two approaches are synergistically combined. As a result, the lidar range profiles of the aerosol extinction and backscatter coefficients are retrieved with relatively high

Figure 3. Range profiles of the aerosol backscattering coefficient at three different azimuth angles (a) and aerosol distribution lidar map based on a series of BCS profiles (b) as measured in the time interval 20:35–21:28 LT on 5 November 2015.

precision and reliability, which are accordingly transferred to the colormaps based on them of the near-surface aerosol density distribution.

#### 2.4. Lidar mapping of aerosol fields

Generally, the aerosol field could be described as a distribution of the aerosol mass concentration M (μg/m<sup>3</sup> ) defined as the mass of PM per unit volume. From the lidar measurements, the extinction and backscattering coefficients of the aerosol particles are determined that are directly proportional to the aerosol mass concentration:

$$M = k\alpha\_a = k\beta\_a S\_a.\tag{5}$$

The mass concentration could be retrieved from the lidar data combining different experimental and numerical approaches [31]. So, obtaining data about the distribution of the aerosol backscattering coefficient could be regarded as representative for the aerosol mass concentration distribution.

Figure 3 shows an example of the stages of formation of an aerosol lidar map using measurements performed on 5 November 2015, in the time interval 20:35–21:28 local time (LT). Aerosol backscattering profiles obtained at different azimuth angles along a fixed elevation angle are presented in Figure 3(a). On the basis of a series of such profiles, 2D color-coded sector maps of the near-surface aerosol density could be created. In Figure 3(b), an aerosol lidar map is displayed in Cartesian coordinates, based on the entire set of BSC profiles in the azimuth sector 190–220°, including the ones in Figure 3(a). Finally, the sector maps so-obtained are superposed on the satellite maps of the corresponding city region.
