Optical and Physical Characteristics of Aerosol Layers in Australia Based on CALIPSO

: Atmospheric aerosols have important impacts on global radiative forcing, air pollution, and human health. This study investigated the optical and physical properties of aerosol layers over Australia from 2007 to 2019 using the Cloud-Aerosol Lidar and Infrared Pathﬁnder Satellite Observations (CALIPSO) Level 2 aerosol products. Australia was divided into three sub-regions (western highlands, central plains, and eastern ranges). Interannual and seasonal optical property variations in aerosol layers in the three sub-regions were analyzed and compared. Results showed that annual mean values of AOD L (lowest aerosol layer AOD) and AOD T (total AOD of all aerosol layers) were always higher in the eastern ranges region than the other two regions from 2007 to 2019. The reason could be that Australian population was predominantly located in the eastern ranges region, where more human activities could bring signiﬁcant aerosol loadings. B L (base height of the lowest aerosol layer), H L (top height of the lowest aerosol layer), and H H (top height of the highest aerosol layer) all showed trends of “western highlands > eastern mountains > central plains”, indicating that the higher the elevation, the higher the B L , H L , and H H . T L (thickness of the lowest aerosol layer) was higher during the day than at night, which might account for increased diurnal atmospheric convection and nocturnal aerosol deposition. DR L (depolarization ratio of the lowest aerosol layer) was higher in the western highlands and central plains than the eastern mountains, probably because these two regions have large deserts with more irregularly shaped dust aerosols. CR L (color ratio of the lowest aerosol layer) had slightly higher values in the eastern ranges than the other two regions, probably due to the wet climate of the eastern ranges, where aerosols were more hygroscopic and had larger particle sizes. This study can provide technical support for the control and management of regional air pollutants.


Introduction
Atmospheric aerosols are suspended microscopic particles in the air, typically with aerodynamic diameters of 10 −3 -10 2 µm [1][2][3]. Atmospheric aerosols have a strong impact on global radiative forcing, air pollution, and human health [4][5][6][7]. Atmospheric aerosols can absorb or scatter solar short-wave and long-wave radiation, altering the radiation balance between the Earth and the atmosphere. Aerosols can also act as cloud condensation nuclei, altering the life cycle of clouds and affecting local or regional weather conditions [8][9][10][11][12]. Therefore, studies of atmospheric aerosols are of great importance, which can help scientists and government departments to understand global radiation changes and formulate air pollution control policies.
Many aerosol observation stations have been established around the world, which are helpful for the study of aerosols. For example, the AERONET (Aerosol Robotic Net-the number of aerosol layers (N), the AOD proportion for the lowest aerosol layer (PAOD L ), the receding ratio of the lowest aerosol layer (DR L ), and the color ratio of the lowest aerosol layer (CR L ). AOD L , H L , B L , H H , DR L , CR L , and N can be obtained directly from L2, while AOD T , TL, and PAOD L are obtained indirectly, using the following equations: where N = 1 represents the lowest aerosol layer, i.e.: AOD L = AOD 1 (4) Atmosphere 2023, 14, x FOR PEER REVIEW 3 of 15

Materials and Methods
CALIPSO (Cloud-Aerosol Lidar and Infrared Path and Satellite Observation) carries three instruments (CALIOP, IIR, and WFC). The CALIOP is a visible and near-infrared polarimetric lidar used for observing the phase state of Earth s aerosols and clouds [38]. The CALIOP has one 1064 nm channel and two channels of 532 nm horizontal polarization

Materials and Methods
CALIPSO (Cloud-Aerosol Lidar and Infrared Path and Satellite Observation) carries three instruments (CALIOP, IIR, and WFC). The CALIOP is a visible and near-infrared polarimetric lidar used for observing the phase state of Earth s aerosols and clouds [38]. The CALIOP has one 1064 nm channel and two channels of 532 nm horizontal polarization   To extrapolate temporal and spatial variabilities of aerosol characteristics conveniently in Australia, we divided Australia into three regions based on topography ( Figure 1): western highlands (A), central plains (B), and eastern ranges (C). Seasons were defined as spring (September to November), summer (December to February), autumn (March to May), and winter (June to August). Annual and seasonal statistical information of aerosol variables were calculated over three subregions from 2007 to 2019. Null and negative values were removed first in this study; AOD L and AOD T values greater than 10, B L , H L , H H , and T L greater than 100 and DR L , and CR L greater than 10 were removed second, because these datasets were false noise interferences; then, the remaining datasets were utilized in the next step.

Interannual Variations in Aerosol Layer Optical Properties over Australia
Annual mean AOD T values for the A, B, and C subregions were analyzed from 2007 to 2019. Figure 3b (daytime) and Figure 4b (nighttime) revealed that a slight increase in the annual mean AOD T over Australia. The C region had a higher AOD T than the A and B regions. The annual mean AOD T (daytime: 0.20; nighttime: 0.16) was higher in the eastern mountains region, followed by the central plains (daytime: 0.16; nighttime: 0.13), and finally the western highlands (daytime: 0.14; nighttime: 0.13). This might be because Australia was controlled by the subtropical high-pressure belt and the effect of narrow eastern ranges, which could result in wet conditions in the east and arid conditions in the mid-west parts of Australia all year round [35,36]. Previous studies had shown that humidity had a positive correlation with AOD; therefore, AOD T and AOD L were higher in the eastern ranges than in the other two regions [33,40]. It might also be related to the distribution of population (Figure 2), economic development, and industrial activity in Australia. Figure 2 revealed that the population density in the east was high, and more populous cities existed in the eastern ranges region, such as Sydney, Melbourne, and Brisbane. High populations could lead to high anthropogenic particulate emissions, and therefore, high aerosol AOD T and AOD L values existed in the east [33]. of organic matter and aerosols into the atmosphere. During this time, Australia experienced several dust storms resulting in large amounts of dust also being released into the atmosphere resulting in increased AOD values [33,35,36]. It might also be due to the rapid economic development of Australia during this period and the increase in human activities such as industrial production and transport, which may have led to increased concentrations of aerosols in the atmosphere and hence higher AOD values [33]. From 2012 to 2014, international cooperation might have contributed to reduced emissions of pollutants in the atmosphere, resulting in lower AOD values. Between 2014 and 2016, Australia experienced continued economic growth. Increased human activities, such as energy consumption, industrial production, and transport, led to increased aerosol concentrations in the atmosphere and consequently higher AOD values. It should be noted that decreases in AODL in 2012-2014 and in 2016-2018, as well as the increase in 2013-2016 in the C region (Eastern Australia), could be due to interannual variability in air humidity in this region. Also, it may be one of the reasons for the lower interannual variation in nighttime AODL compared to daytime in the C region. Analysis was carried out on the AODL, which was strongly influenced by natural and anthropogenic activities [41,42]. Figure 3a,b (daytime) and Figure 4a,b (nighttime) showed that the annual mean AODL over Australia has a similar annual variability pattern to AODT, i.e., the annual mean AODL was higher in the eastern ranges (daytime: 0.16; nighttime: 0.11), followed by the central plains (daytime: 0.12; nighttime: 0.09), and finally the western highlands (daytime: 0.11; nighttime: 0.08). This might be because Australia is surrounded by sea, with cold currents in the west and warm currents in the east, creating a spatial distribution of a wetter climate in the east and a drier climate in the west, resulting in a higher AODL in the eastern ranges [33].     Figure 4c (nighttime) showed the interannual variation in BL, which is mainly influenced by elevations [41]. Results showed that BL was highest in the western highlands (daytime: 1.07 km; nighttime: 1.00 km), followed by the eastern mountains region (daytime: 0.78 km; nighttime: 0.87 km) and the central plains (daytime: 0.72 km; nighttime: 0.74 km). The main reason was the influence of elevation, with higher BL values in higher elevation areas. Annual mean HL and HH values (Figures 3d,e and 4d,e) were also higher in the western highlands (daytime: 2.41 km; nighttime: 2.24 km), followed by the eastern mountains (daytime: 2.23 km; nighttime: 1.93 km) and the central plains (daytime: 2.01 km; nighttime: 1.91 km). This result showed that HL and HH were also more related to topography, with higher elevation regions having higher HL and HH values. This result was similar to that of Zhang s study [41][42][43]. Figure 3f (daytime) and 4f (nighttime) showed the interannual variation in TL in the western highlands (daytime: 1.34 km; nighttime: 1.24 km), the central plains (daytime: 1.29 km; nighttime: 1.17 km), and the eastern mountains (daytime: 1.46 km; nighttime 1.06 km). As shown, values at night were relatively small compared to the daytime. This should be due to the mixed layer that was thicker in the daytime due to convective processes driven by solar radiation [44]. At the same time, low temperatures and weak vertical convection in the atmosphere at night could lead to thinner aerosol layers [41,42,45]. Figure 3g (daytime) and Figure 4g (nighttime) showed that there was little difference between annual averaged N values for the eastern ranges (daytime: 1.26; nighttime: 1.70), central plains (daytime: 1.25; nighttime: 1.58), and western highlands (daytime: 1.23; nighttime: 1.60) regions of Australia. However, annual averaged N values were slightly greater in the eastern ranges region than in the central plains and western highlands regions. One plausible reason for this was that the Australian population is predominantly distributed in the eastern ranges region, where anthropogenic factors contribute to the large atmospheric aerosol loadings, resulting in significant vertical stratifications of atmospheric aerosols [33].
According to Figures 3h and 4h, it can be seen that there was little difference in PAODL values between the eastern mountains region (daytime: 89%; nighttime: 75%), the central plain region (daytime: 89%; nighttime: 77%), and the western plateau region (day- The increase in AOD T values between 2010 and 2012 might be due to forest fires (Margaret River Bushfires and Carnarvon Bushfires) resulting in the release of large amounts of organic matter and aerosols into the atmosphere. During this time, Australia experienced several dust storms resulting in large amounts of dust also being released into the atmosphere resulting in increased AOD values [33,35,36]. It might also be due to the rapid economic development of Australia during this period and the increase in human activities such as industrial production and transport, which may have led to increased concentrations of aerosols in the atmosphere and hence higher AOD values [33]. From 2012 to 2014, international cooperation might have contributed to reduced emissions of pollutants in the atmosphere, resulting in lower AOD values. Between 2014 and 2016, Australia experienced continued economic growth. Increased human activities, such as energy consumption, industrial production, and transport, led to increased aerosol concentrations in the atmosphere and consequently higher AOD values. It should be noted that decreases in AOD L in 2012-2014 and in 2016-2018, as well as the increase in 2013-2016 in the C region (Eastern Australia), could be due to interannual variability in air humidity in this region. Also, it may be one of the reasons for the lower interannual variation in nighttime AOD L compared to daytime in the C region.
Analysis was carried out on the AOD L , which was strongly influenced by natural and anthropogenic activities [41,42]. Figure 3a,b (daytime) and Figure 4a,b (nighttime) showed that the annual mean AOD L over Australia has a similar annual variability pattern to AOD T , i.e., the annual mean AOD L was higher in the eastern ranges (daytime: 0.16; nighttime: 0.11), followed by the central plains (daytime: 0.12; nighttime: 0.09), and finally the western highlands (daytime: 0.11; nighttime: 0.08). This might be because Australia is surrounded by sea, with cold currents in the west and warm currents in the east, creating a spatial distribution of a wetter climate in the east and a drier climate in the west, resulting in a higher AOD L in the eastern ranges [33].  Figure 4c (nighttime) showed the interannual variation in B L , which is mainly influenced by elevations [41]. Results showed that B L was highest in the western highlands (daytime: 1.07 km; nighttime: 1.00 km), followed by the eastern mountains region (daytime: 0.78 km; nighttime: 0.87 km) and the central plains (daytime: 0.72 km; nighttime: 0.74 km). The main reason was the influence of elevation, with higher B L values in higher elevation areas. Annual mean H L and H H values (Figures 3d,e and 4d,e) were also higher in the western highlands (daytime: 2.41 km; nighttime: 2.24 km), followed by the eastern mountains (daytime: 2.23 km; nighttime: 1.93 km) and the central plains (daytime: 2.01 km; nighttime: 1.91 km). This result showed that H L and H H were also more related to topography, with higher elevation regions having higher H L and H H values. This result was similar to that of Zhang's study [41][42][43]. Figure 3f (daytime) and 4f (nighttime) showed the interannual variation in T L in the western highlands (daytime: 1.34 km; nighttime: 1.24 km), the central plains (daytime: 1.29 km; nighttime: 1.17 km), and the eastern mountains (daytime: 1.46 km; nighttime 1.06 km). As shown, values at night were relatively small compared to the daytime. This should be due to the mixed layer that was thicker in the daytime due to convective processes driven by solar radiation [44]. At the same time, low temperatures and weak vertical convection in the atmosphere at night could lead to thinner aerosol layers [41,42,45]. Figure 3g (daytime) and Figure 4g (nighttime) showed that there was little difference between annual averaged N values for the eastern ranges (daytime: 1.26; nighttime: 1.70), central plains (daytime: 1.25; nighttime: 1.58), and western highlands (daytime: 1.23; nighttime: 1.60) regions of Australia. However, annual averaged N values were slightly greater in the eastern ranges region than in the central plains and western highlands regions. One plausible reason for this was that the Australian population is predominantly distributed in the eastern ranges region, where anthropogenic factors contribute to the large atmospheric aerosol loadings, resulting in significant vertical stratifications of atmospheric aerosols [33].
According to Figures 3h and 4h, it can be seen that there was little difference in PAOD L values between the eastern mountains region (daytime: 89%; nighttime: 75%), the central plain region (daytime: 89%; nighttime: 77%), and the western plateau region (daytime: 90%; nighttime: 76%). PAOD L was the ratio of AOD L to AOD T and was mainly influenced by N. Results showed that PAOD L had a negative correlation with N. The larger the value of N, the smaller the PAOD L . This was similar to the results of the Yellow River basin in China and Pakistan [41,43]. The DR L reflected the aerosol particle degree of non-sphericity in the lowest aerosol layer. The larger DR L values mean more non-spherical particles; smaller DR L values mean more spherical particles [41]. Results showed that annual mean DR L values were higher in the western highlands (daytime: 0.06; nighttime: 0.04) and the central plains (daytime: 0.06; nighttime: 0.04) than in the eastern mountains (daytime: 0.05; nighttime: 0.03), which indicated that more non-spherical particles were present in the A region and the B region, as these two regions had large areas of deserts with irregularly shaped dust aerosols. These results also indicated that more non-spherical particles were present during the daytime compared to the nighttime, which may be due to nighttime aerosol depositions. In addition, it may also be caused by nighttime water condensation on non-spherical particles. The water coating will make a particle more spherical, increasing the total sphericity of the particles [34,41]. CR L values represented particle sizes; the higher the CR L values, the larger the particle sizes. Figures 3j and 4j reflected the annual mean variations in CR L in the lowest aerosol layer during the daytime and at nighttime. Results showed that larger CR L values existed in the eastern mountains region (daytime: 0.85; nighttime: 0.58) and western highlands region (daytime: 0.76; nighttime: 0.54) compared to the central plain region (daytime: 0.61; nighttime: 0.49). The larger size of aerosol particles in C region may be caused by sea salt [37]. The medium-sized aerosol particles in the A region may be due to sandy and dusty aerosols (the Great Sandy Desert, Great Victoria Desert, Gibson Desert, and Simpson Desert are all located in the western highlands region of Australia) [34].

Seasonal Variations in Aerosol Layer Optical Properties over Australia
Seasonal variations in aerosol layer optical properties over Australia were conducted during the daytime and nighttime (Figures 5-8). Results showed that AOD T was more variable in the eastern mountains region, reaching its highest value in summer (daytime: 0.23; nighttime: 0.20). And AOD T values were highest in the western highlands (daytime: 0.17; nighttime: 0.16) and the central plains (daytime: 0.18; nighttime: 0.14) in spring because of dust. This may be due to the influence of southeasterly trade winds in summer and the tall eastern mountains blocking the warm and humid airflow, resulting in more rain in the east and less rain in the mid-west. Hence, the highest AOD T values in the eastern mountains was in summer, which was similar to the study results of Mitchell et al., where AOD and scattering coefficients peaked in the central plains of Australia in spring and summer [35]. Overall, AOD T values were lowest in winter (daytime: 0.15; nighttime: 0.11) over most of the Australian continent, because high pressures can lead to a dry atmosphere and low humidity over most of the continent [35]. From spatial distribution maps (Figures 5 and 6), AOD values were relatively high in the eastern ranges region of Australia compared to the other two regions, which may be because the Australian population was predominantly located in the eastern ranges region, leading to high anthropogenic pollution emissions, which was consistent with the findings of Mitchell et al. [36].
As shown in Figures 5 and 6, H L and H H values were higher in the western highlands compared to the other two regions. Notably, H L and H H values were higher in spring in the western plateau region compared to the other three seasons, which may be the reason that the western plateau region had large areas of desert and was significantly affected by sand and dust aerosols in spring [33]. The seasonal plots showed gradual decreases in H L and H H values from spring to winter. These can be attributed to the intensification of vertical convection caused by seasonal temperature gradients, which can result in the enhanced vertical movement of the aerosol layers [37].
As shown in Figures 7g and 8g, the highest N values were observed in spring (daytime: 1.31; nighttime: 1.76), and the lowest N values were observed in winter (daytime: 1.19; nighttime: 1.45). This may be due to a stronger vertical convection in spring and a weaker vertical convection in winter [37]. The nighttime N values were higher than those in the daytime (~0.37); the reason for this could be that atmosphere convection increased and fully mixed during the daytime, leading to a decrease in N and an increase in T L ; during the night, atmospheric convection weakened and residual layer appeared, resulting in the stratification of the atmosphere and an increase in N [45]. High values of N were mainly found in the eastern ranges region, which is characterized by a dense population, high economic activity, and more industrial units and agricultural activities, resulting in more aerosol stratification [36]. The seasonal mean of T L was highest in spring (daytime: 1.42 km; nighttime: 1.22 km) and lowest in winter (daytime: 1.27 km; nighttime: 1.05 km). As T L was calculated from B L and H L , its variation mainly depended on the distribution of B L and H L (Figures 7d,e and 8d,e). In addition, the T L values were slightly lower at nighttime than during the daytime (~0.2 km).
As shown in Figures 7h and 8h, PAOD L mean values for all seasons were almost the same for the whole of Australia (daytime: 89%; nighttime: 76%). The higher PAOD L values in winter compared to the other seasons may be related to the properties of AOD T (AOD L ) and N as discussed above. In addition, PAOD L values were higher during the daytime than at nighttime, which may be due to higher aerosol loadings in the lowest aerosol layer, where coarse particles dominated [34,41]. Seasonal mean values of DR L were slightly higher in spring (daytime: 0.06; nighttime: 0.04) and summer (daytime: 0.07; nighttime: 0.04) than in autumn (daytime: 0.05; nighttime: 0.03) and winter (daytime: 0.05; nighttime: 0.03). This might be due to relatively heavy dust transports during these two seasons, resulting in a higher number of non-spherical aerosol particles [37]. DR L values were higher in the western highlands and central plains regions than in the eastern ranges region, which may be due to the high number of deserts and dust aerosols in the western highlands and central plains regions. Seasonal mean values of CR L did not vary significantly between the seasons (daytime: 0.74; nighttime: 0.54), and its spatial distribution showed higher CR L values in the eastern ranges and western highlands regions, indicating that larger aerosol particle sizes existed in these two regions. Mitchell et al. showed that as the atmospheric humidity increased, the aerosol particle sizes increased [33]. The eastern mountainous region was wet and rainy leading to larger CR L values. The larger CR L values in the western highlands may be due to the influence of sand and dust aerosols.    As shown in Figures 7g and 8g, the highest N values were observed in spring (daytime: 1.31; nighttime: 1.76), and the lowest N values were observed in winter (daytime: 1.19; nighttime: 1.45). This may be due to a stronger vertical convection in spring and a weaker vertical convection in winter [37]. The nighttime N values were higher than those in the daytime (~0.37); the reason for this could be that atmosphere convection increased  As shown in Figures 7g and 8g, the highest N values were observed in spring (daytime: 1.31; nighttime: 1.76), and the lowest N values were observed in winter (daytime: 1.19; nighttime: 1.45). This may be due to a stronger vertical convection in spring and a weaker vertical convection in winter [37]. The nighttime N values were higher than those in the daytime (~0.37); the reason for this could be that atmosphere convection increased

Correlation of Aerosol Properties in Australia
To better understand the spatial and temporal distribution of aerosol properties over Australia, seasonal correlations between some aerosol layer parameters during the daytime and at nighttime were carried out (Figures 9-11). We used the Pearson correlation coefficients in this part [46]. The results showed a more significant positive correlation between AOD L and T L . And these relationships varied during the daytime and at nighttime and also changed at different seasons. Daytime correlations were mostly higher than nighttime correlations, which may be related to reduced anthropogenic factors at night. In terms of seasonal distribution, the highest correlations for the three Australian regions were found in spring, followed by summer and autumn, and the lowest correlations were found in winter, because aerosol contents in spring and summer were relatively high compared to those in autumn and winter. This was consistent with the results of the Yellow River basin in China and Pakistan [41,43]. Figure 10 depicts the correlation between N and H H in Australia. N was mainly distributed between 1 and 7 in regions A, B, and C, all of which had strong positive correlations with H H . This phenomenon revealed that the higher the H H , the greater the N values. This result was consistent with the existing knowledge reported for the Yellow River Basin and the Qinghai-Tibet Plateau [42,43]. Figure 11 depicts the correlation between N and PAOD L in Australia. The results showed that N exhibited a significant negative correlation with PAOD L . Based on the current knowledge, the larger the N value, the smaller the PAOD L value, which was consistent with the previous research [41,42].

Conclusions
In this article, the optical and physical properties of aerosol layers over Australia were investigated using CALIPSO level 2 products from 2007 to 2019. The interannual, seasonal variabilities of AOD L , AOD T , B L , H L , H H , T L , N, PAOD L , DR L , and CR L were analyzed and discussed for the daytime and the nighttime. The main conclusions are as follows: (1) The annual mean values of AOD T were highest in the eastern ranges. This may be related to Australian climatic characteristics (wet and rainy in the east, dry and low rainfall in the mid-west) and population distributions.
(2) AOD T values reached maximum levels in the summer in the eastern range region and in the spring in the other two regions. This may be due to high atmospheric humidity in the summer over the eastern range region, while the other two regions had more frequent dusty weather in spring.
( (4) T L values were smaller at nighttime than during the daytime. This may be related to nocturnal aerosol depositions.
(5) DR L values were higher in the western highlands and central plains than in the eastern mountain ranges. This was because the central and western regions were dotted with large areas of deserts and because there were large amounts of non-spherical dust aerosols. More non-spherical particles were present during the daytime than at nighttime.
(6) CR L values were higher in the eastern mountainous regions and the western highland regions than in the central plain regions. This indicated that aerosol particles in the first two regions were of a larger particle size.
In future work, we will investigate dust emissions in the A and B regions using CALIOP data. Meanwhile, we will use more in situs datasets to observe more optical and physical parameters over Australia.