MERRA-2 and OMI aided assessment of the effect of major air pollutants on surface temperatures

doi:10.21203/rs.3.rs-3598757/v1

Download PDF

Research Article

MERRA-2 and OMI aided assessment of the effect of major air pollutants on surface temperatures

https://doi.org/10.21203/rs.3.rs-3598757/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

We attempted to show the effect of lockdown on the aerosol, black carbon, dust, and organic carbon loads and their influence on moderating surface temperatures. This study produced and interpreted Time-averaged maps for three phases, i.e., pre-lockdown, lockdown, and post-lockdown phases. We observed a 4% increase in black carbon loads during the lockdown period compared to pre-lockdown and a 27% increase in the post-lockdown period. It was observed that there was a 6% increase in dust loads during the lockdown period and a 5% increase in the post-lockdown period when compared with pre-lockdown levels. Approximately a 48% increase in organic carbon loads during the lockdown period and a 24% increase in the post-lockdown period was observed. We observed a 12% increase in surface temperatures during the lockdown period and almost no considerable increase in the post-lockdown period.

Climate Analysis and Modeling

Geographic Information Systems

Atmospheric Sciences

Climatology

Aerosols

Dust

COVID-19 lockdown

Organic and black carbon

Surface

The emission of black carbon from the partial combustion of biofuels in the southern part of Asia has affected radiative forcing, resulting in ambiguity in climate change and health impacts (Velis & Cook, 2021). Dark carbonaceous aerosols can be classified as black carbon, and they significantly absorb Ultraviolet, Visible, and NIR (near-infrared) radiations due to graphitic microcrystal structures (Cheng et al., 2019). Fossil fuels contribute about 66%, and biofuels contribute approximately 34% of black carbon emissions (Dumka et al., 2018). The dust loaded with minerals can seriously impact global and regional climatic systems through the absorption and scattering of solar radiations over terrain (Adebiyi et al., 2023). Dust loads can affect human health, socioeconomic factors, and air quality (Ali et al., 2021). Aerosols are important vehicles that can be screened to know the environmental quality of a region and can be attributed to pollution (Robichaud, 2020). COVID-19 has trained us in several ways to face future pandemics and also allowed us to study the anthropogenic effect on air quality (Ravindra et al., 2022). Studying aerosols at a regional scale has become routine in monitoring air quality (Sokhi et al., 2022). Aerosol optical depth can be used to understand the regional troposphere better, and satellite products provide reliable information over a more extended period (Wei et al., 2020). The aerosol load can be attributed to population growth, and it was observed through earlier studies conducted in Asia (Ramachandran, Rupakheti & Lawrence, 2020). Ozone monitoring instrument (NASA) has provided information about environmental quality since 2004 through various bands and resolutions, and several studies highlighted its importance (Griffin et al., 2019). Organic carbon is crucial in regulating the secondary production of CO₂, particulate matter, and Ozone (Ho et al., 2022). In industrial and urban areas, elemental and organic carbon are the dominating species in the particulate matter (Kuzu et al., 2020). This work was planned based on the assumption that the COVID pandemic has altered the air quality parameters and might have reduced the aerosols, black carbon, dust, and organic carbon loads. These reduction patterns would have changed the dynamics of surface temperature. During the lockdown, there were severe restrictions on industrial routines and transit mechanisms over terrain and air, leading us to observe the changes in the atmosphere better. This has been attributed to reduced anthropogenic emissions during the lockdown. It was felt that the change detection study to know the actual variation in the air quality with surface temperatures would help weigh the anthropogenic influence on the environment. In this study, three phases, i.e., pre-lockdown phase (2004-10-01 to 2020-03-24), 2. lockdown phase (2020-04-25 to 2020-05-31), and 3. post-lockdown phase (2020-06-01 to 2023-01-29) were selected.

The climatic condition in India constitutes a diverse range of weather conditions, and this can be attributed to the uneven topography and other factors. It is obvious that the climatic conditions of this area are regulated by the Thar desert and the Himalayas. The Hindu Kush mountains mostly disallow katabatic winds (Central Asian), leaving the subcontinent relatively warm. The year 2022 was recorded as one of the hottest years since 1901, i.e., 0.15^oC > 1981 to 2010 average (India Meteorological Department). It is of great concern that eleven out of fifteen years that were hot falls in the time range from 2008 to 2022 (IMD). The high temperatures affected many crop species, and the wheat crop was affected the most. India spreads from 8°4’ to 37°6’ N latitudes and 68°7' to 97°25' E longitudes and is close to the equator.

The datasets needed for this study are from

OMI (Ozone monitoring instrument)

MERRA-2 or Modern-Era Retrospective Analysis for Research and Applications, Version 2

To conduct a change detection study with the COVID-19 period as a reference point

To demonstrate the elasticity of the occurrence of high surface temperatures soon after a brief impasse.

To provide some statistical insights and inter-variable interactions during the three periods that are considered in this study.

The softwares used in this study are

ArcGIS

MS Excel

R and Python

The datasets needed for this study are collected from the Giovanni website of NASA (https://giovanni.gsfc.nasa.gov/giovanni/), which provides data free of cost. The data from the Ozone monitoring instrument (OMI) with 1-degree spatial resolution is used in this study. The OMI instrument can easily demarcate aerosol types like sulfates, dust, and smoke (Filonchyk et al., 2020). It employs push-broom mode, enabling us to observe radiation backscattering via UV and visible channels (Lamsal et al., 2021). It has hyperspectral properties that help in classifying the obtained data as precise and accurate (Lee, Ahn & Kang, 2020). The accuracy of the instrument is calibrated using long-term data. The main reason for selecting OMI is because it gives accuracy better than Total Ozone Mapping Spectrometer (TOMS) and Global Ozone Monitoring Experiment (GOME) instruments (Ziemke et al., 2019). Two wavelengths, i.e., 388 nm and 500 nm, are considered in this study to estimate the aerosols in three phases. The phases are: 1. Pre-lockdown phase (2004-10-01 to 2020 − 0324), 2. Lockdown phase (2020-04-25 to 2020-05-31), and 3. post-lockdown phase (2020-06-01 to 2023-01-29). In this study, the aerosol optical depth data at 388 nm and 500 nm is selected. The datasets needed to estimate black carbon (kg/m³), dust loads (kg/m³), organic carbon load (kg/m³), and surface temperature (^oC) are obtained from MERRA-2 or Modern-Era Retrospective analysis for Research and Applications, Version 2. This data is controlled by the Global Modeling and Assimilation Office (GMAO) of NASA. MERRA-2 has a resolution of 0.625^o × 0.5^o (Gelaro et al. 2017). The time-averaged maps (TAMs) for each variable across the time frames were prepared for each phase as stated above. From the TAMs, one thousand random pixels (points) were extracted with their values intact. Preprocessing of the data was done using normalization, and missing values were checked for consistency. Using the processed datasets, correlation analysis was performed to determine the variables that are positively correlated. The principal component analysis, or PCA, can be vital in analyzing data with more or higher dimensions (Kherif & Latypova, 2020). It can be used to get the key/important information from the data. This can be achieved by transforming the native/original data into a space with low dimensions.

The highly correlated variables will be assembled in one place. The important steps involved with PCA are: 1. Data normalization, 2. computation of covariance matrix, 3. eigenvalues/eigenvectors, 4. Choosing principal components, and 5. Transformation of data in new space (Gewers et al., 2021). The data used in this study are of different scales, and it has become essential to normalize them to prevent biased results. This process can be done by subtracting observation values from the mean and dividing them by the standard deviation value. The normalized data can be used to build a covariable matrix, and this can be classified as a symmetric matrix (Mao, Chen & Wong, 2019). Eigenvalue explains the quantum of variance in a dataset for a direction under consideration, i.e., eigenvector (Marukatat, 2023). Every eigenvector will have an eigenvalue. The number of pairs (eigenvectors and values) is equal to the number of variables that are in the dataset. All these pairs may not be relevant. In the first principal component, the eigenvector with maximum eigenvalue will be present (Aït-Sahalia & Xiu, 2019). Similarly, the second principal component will have an eigenvector second maximum eigenvalue, which continues. In the new dimensional space, the original data will be re-oriented and placed on a new subspace determined by principal components. The reorientation process is generally done by taking the product of original data and eigenvectors that are previously computed. The transformation process does not necessarily alter the original data but allows us to view/represent data with a more interpretable visualization (Wu et al., 2023). The interactions between the variables are also presented to know the inter-variable interactions. The study area and methodology adopted are shown in Fig. 1 and Fig. 2.

The satellite data is often affected by diverse environmental conditions, and air pollution is one such factor that is important. The datasets procured from the sensors onboard the satellite need to be cross-checked with the available ground data, but in this case, limited data is available. The installation of air monitoring stations all over the country that may provide high spatial resolution is an arduous task.

Aerosol optical depth (388 nm)

Some of the regions in the northern part of India, i.e., Punjab, Haryana, Delhi, Uttar Pradesh, half portion of Bihar, and a tip of Jammu and Kashmir, showed high aerosol density during the pre-lockdown phase. During the lockdown period, these portions exhibited a steep decrease in aerosol loads. After the lockdown period, the intensity of aerosol loads returned to the normal condition (even higher). This can be attributed to enhanced anthropogenic activities and the ingress of aerosol loads from other territories nearby. This shows that the brief lockdown period has effectively cleared the lower air space, but similar aerosol loads were observed post-lockdown periods, strengthening the long-standing view of anthropogenic lowering of air quality over the Indian subcontinent (Fig. 3a. to 3c). The highest AOD_388nm value observed is 0.073575, and the lowest is 0.02288 during the pre-lockdown period. During the lockdown period, the highest AOD_388nm value observed is 0.16035 (higher than the pre-lockdown period), and the lowest value observed is 0.0099 (lower than the pre-lockdown). During the post-lockdown period, the AOD_388nm maximum value observed is 0.0864333 (close to the pre-lockdown period), and the AOD_388nm minimum value observed is 0.0209 (close to the pre-lockdown period).

Some portions of NE J &K exhibited high aerosol load before lockdown, but during the lockdown, there was no data available; however, there is a reduced load post-lockdown. Andaman and Nicobar Islands showed a limited aerosol load differential, and a similar pattern is observed over Lakshadweep islands. Madhya Pradesh, Chandigarh, and Orissa showed an increase in aerosol loads after the lockdown. The northeastern portions of Andhra Pradesh, north Tamil Nadu, and Pondicherry showed steep increments in aerosol loads after the lockdown.

Aerosol optical depth (500 nm)

When observed at 500 nm, the states of Punjab, Haryana, Delhi, Uttar Pradesh, Bihar, and some portion of Rajasthan exhibited high aerosol intensity along with the coastal belt across West Bengal, Orissa, Pondicherry, Andhra Pradesh, Tamil Nadu, and Lakshadweep Islands during the pre-lockdown period. During the lockdown period, some portions of Manipur. Mizoram, Lakshadweep Islands, and Andaman Nicobar Islands showed increased aerosol intensity. Post lockdown, the aerosol loads reached the earlier state (Fig. 4a. to 4c). The highest AOD_500nm value observed is 0.0350225, and the lowest is 0.01066 during the pre-lockdown period. During the lockdown period, the highest AOD_500nm value observed is 0.0803 (higher than the pre-lockdown period), and the lowest value observed is 0.0043 (lower than the pre-lockdown). During the post-lockdown period, the AOD_500nm maximum value observed is 0.0406294 ( higher than the pre-lockdown period), and the AOD_500nm minimum value observed is 0.00915 (lower than the pre-lockdown period).

Aerosol Index (UV)

The highest Aerosol Index (UV) value observed is 1.27291, and the lowest is 0.753027 during the pre-lockdown period (Data gaps were observed). During the lockdown period, the highest Aerosol Index (UV) value observed was 1.89674 (higher than the pre-lockdown period), and the lowest value observed was 0.5 (lower than the pre-lockdown). During the post-lockdown period, the Aerosol Index (UV) maximum value observed is 1.72333 (higher than the pre-lockdown period), and the Aerosol Index (UV) minimum value observed is 0.752182 (close to the pre-lockdown period) (Fig. 5a. to 5c).

Black carbon (BC)

The highest black carbon value observed is 3.10e^− 009 kg m^− 3, and the lowest is 5.90e^− 011 kg m^− 3 during the pre-lockdown period. During the lockdown period, the black carbon value observed is 3.22e^− 009 kg m^− 3 (higher than the pre-lockdown period), and the lowest value observed is 4.90e^− 011 kg m^− 3 (lower than the pre-lockdown). During the post-lockdown period, the black carbon maximum value observed is 4.27e^− 009 kg m^− 3 (higher than the pre-lockdown period), and the black carbon minimum value observed is 7.24e^− 011 kg m^− 3 (higher than the pre-lockdown period) (Fig. 6a. to 6c).

Dust

The highest dust load value observed is 6.43e^− 008 kg m^− 3, and the lowest is 9.42e^− 010 kg m^− 3 during the pre-lockdown period. During the lockdown period, the dust load value observed is 6.81e^− 008 kg m^− 3 (slightly higher than the pre-lockdown period), and the lowest value observed is 1.33e^− 009 kg m^− 3. During the post-lockdown period, the dust load maximum value observed is 6.78e^− 008 kg m^− 3 (higher than the pre-lockdown period), and the dust load minimum value observed is 1.30e^− 009 kg m^− 3 (Fig. 7a. to 7c).

Organic carbon

The highest organic load value observed is 1.58e^− 008 kg m^− 3, and the lowest is 2.63e^− 010 kg m^− 3 during the pre-lockdown period. During the lockdown period, the organic carbon load value observed is 3.04e^− 008 kg m^− 3 (higher than the pre-lockdown period), and the lowest value observed is 2.24e^− 010 kg m^− 3. During the post-lockdown period, the organic carbon load maximum value observed is 2.07e^− 008 kg m^− 3 (higher than the pre-lockdown period), and the organic carbon load minimum value observed is 2.99e^− 010 kg m^− 3 (Fig. 8a. to 8c).

Surface temperature

The highest surface temperature observed is 28.57 ^oC, and the lowest is -12.72 ^oC during the pre-lockdown period. During the lockdown period, the highest surface temperature value observed was 32.46 ^oC (higher than the pre-lockdown period), and the lowest value observed was − 14.10 ^oC. During the post-lockdown period, the surface temperature maximum value observed is 28.69 ^oC (close to the pre-lockdown period), and the surface temperature minimum value observed is -11.60 ^oC (Fig. 9a. to 9c).

Table 1

Change in percentage
Variables	Prelockdown (max) (A)	Lockdown (max) (B)	Postlockdown (max) (C)	Change (percentage) (A & B)	Change (percentage) (A & C)
AOD388nm	0.073575	0.16035	0.086433	54% ↑	15% ↑
AOD500nm	0.035023	0.0803	0.040629	56% ↑	14% ↑
Aerosol index (UV)	1.27291	1.89674	1.72333	33% ↑	26% ↑
Black carbon	3.10E-09	3.22E-09	4.27E-09	4% ↑	27% ↑
Dust	6.43E-08	6.81E-08	6.78E-08	6% ↑	5% ↑
Organic carbon	1.58E-08	3.04E-08	2.07E-08	48% ↑	24% ↑
Surface temperature	28.57	32.46	28.69	12% ↑	0% (app)

There is a 54% increase in AOD_388nm loads during the lockdown period and a 15% increase in the post-lockdown period when compared with pre-lockdown AOD_388nm concentrations. There is a 56% increase in AOD_500nm loads during the lockdown period and a 14% increase in the post-lockdown period when compared with pre-lockdown AOD_500nm concentrations. There is a 33% increase in Aerosol index (UV) loads during the lockdown period and a 26% increase in the post-lockdown period when compared with pre-lockdown Aerosol index (UV) concentrations. There is a 4% increase in Black carbon loads during the lockdown period and a 27% increase in the post-lockdown period when compared with pre-lockdown Black carbon concentrations. There is a 6% increase in Dust loads during the lockdown period and a 5% increase in the post-lockdown period when compared with pre-lockdown Dust concentrations. There is a 48% increase in Organic carbon loads during the lockdown period and a 24% increase in the post-lockdown period when compared with pre-lockdown Organic carbon concentrations. There is a 12% increase in surface temperatures during the lockdown period and almost no considerable increase in the post-lockdown period compared to pre-lockdown surface temperatures. The change in the percentage of the variables in all three phases is given in table 1.

Correlation analysis and inter-variable interactions

Pre-lockdown

The surface temperature is highly correlated with dust (0.497) (DUSMASS) but negatively correlated with Aerosol index (UV) (UVAI). The organic carbon loads are highly correlated with black carbon (BCSMASS) and negatively correlated with the Aerosol index (UV). AOD388nm and AOD500nm loads are correlated (0.99). The correlation plot with values is given in Fig. 10, and inter-variable interactions in the pre-lockdown variables are shown in Figs. 11a to 11f.

Lockdown

The surface temperature is highly correlated with the Aerosol index (UV) (UVAI) (0.422). The organic carbon loads are highly correlated with black carbon (BCSMASS). Dust loads are highly correlated with Aerosol index (UV) (UVAI). AOD388nm and AOD500nm loads are correlated (0.99). The correlation plot with values is given in Fig. 12, and intervariable interactions in the lockdown variables are shown in Figs. 13a to 13f.

Post-lockdown

The surface temperature is highly correlated with dust (0.564). The organic carbon loads are highly correlated with black carbon (BCSMASS). Dust loads are highly correlated with Aerosol index (UV) (UVAI). AOD388nm and AOD500nm loads are correlated (0.98). The correlation plot with values is given in Fig. 14, and inter-variable interactions in the lockdown variables are shown in Figs. 15a to 15f.

Principal component analysis

Pre-lockdown period

The surface temperature in the pre-lockdown period has exhibited positive correlations with organic carbon, black carbon, and dust loads (Fig. 16a). The 1st principal component (PC) explained 67% of the variance, and the seco^nd principal component exhibited 18% of the variance (Fig. 16 b and e). The 1st and 2nd PCs explained about 85% of the total variance. The loading matrix showed that the fir^st principal component (PC) has positive values for all variables except the Aerosol Index (UV). Organic carbon (0.4365) and black carbon (0.4338) yielded higher values than others (Fig. 16f). This can be attributed to the data scarcity of the Aerosol Index (UV) variable across the time frame. In the 2nd PC, high negative values were observed for dust loads (-0.6051) and AOD₃₈₈ (-0.3288) (Fig. 16f). The scree plot shows the eigenvalues within 1st and 2nd PCs contributing 67% and 18.4%, respectively of the information in the data (Fig. 16b). The Cos2 or square cosine plot shows the quantum of each variable in each component. In this case, organic carbon and Aerosol Index (UV) variables are represented better on the component than the surface temperature variable (Fig. 16c). High Cos2 attributes are organic carbon, black carbon, and aerosol index (UV) and are shown in green color. Mid-Cos2 attributes are AOD_388nm, AOD_500nm, and dust (shown in orange). The low Cos2 attribute is the surface temperature (shown in black color) (Fig. 16d).

Lockdown period

The surface temperature in the lockdown period has exhibited positive correlations with black carbon, AOD_388nm, AOD_500nm, and Aerosol index (UV) loads (Fig. 17a). The 1st principal component (PC) explained 64.9% of the variance, and the second principal component exhibited 26.1% of the variance (Fig. 17 b and e). The 1st and 2nd PCs explained about 91% of the total variance. The loading matrix obtained showed that the first principal component (PC) has positive values for dust (0.5753) (Fig. 17f). In the 2nd PC, high negative values were observed for black carbon (-0.2467) (Fig. 17f). The scree plot shows the eigenvalues within 1st and 2nd PCs contributing 64.9% and 26.1% respectively of the information in the data (Fig. 17b). Organic carbon and dust variables represented well on the component than surface temperature variable (Fig. 17c). High Cos2 attributes are organic carbon and dust (shown in green color). Mid-Cos2 attributes are AOD_388nm, AOD_500nm, black carbon, and aerosol index (UV) (shown in orange color). The low Cos2 attribute is the surface temperature (shown in black color) (Fig. 17d).

Post-Lockdown period

The surface temperature in the post-lockdown period has exhibited positive correlations with organic carbon, black carbon, dust, and Aerosol index (UV) loads (Fig. 18a). The 1st principal component (PC) explained 59.3% of the variance, and the second principal component exhibited 33% of the variance (Fig. 18 b and e). The 1st and 2nd PCs explained about 92.3% of the total variance. The loading matrix obtained showed that the first principal component (PC) has positive values for surface temperature, dust, and aerosol index (UV) (Fig. 18f). In the 2nd PC, high negative values were observed for dust (-0.3397) (Fig. 18f). The scree plot shows the eigenvalues within 1st and 2nd PCs contributing 59.3% and 33% respectively of the information in the data (Fig. 18b). AOD_388nm and AOD_500nm variables represented well on the component than surface temperature variable (Fig. 18c). High Cos2 attributes are organic carbon, AOD_388nm, AOD_500nm, and dust (shown in green color). Mid-Cos2 attributes are black carbon and surface temperature variables (shown in orange color). The low Cos2 attribute is the aerosol index (UV) (shown in black color) (Fig. 17d).

The aerosol, black carbon, dust, and organic carbon loads regulate the opaqueness/transparency of an area, making it a necessary parameter to study radiative transfer mechanisms, thereby affecting surface temperatures. This study observes that aerosol, black carbon, dust, and organic carbon loads over the Indian subcontinent have slightly reduced due to the lockdown imposed to combat the COVID-19 pandemic. This research supports the observations made by several researchers that lockdown has reduced pollution load to some extent. Through this study, it became evident that due to the revival of anthropogenic activities, the aerosol loads have reached the previous level (Pre-lockdown). It is also observed that there is a noticeable change in surface temperatures (increase) in the lockdown phase.

Conflict of interest

The authors declare no conflict of interest

Acknowledgements

The author sincerely thanks NASA and Government of India for their support in this work (data providers).

Adebiyi, A., Kok, J.F., Murray, B.J., Ryder, C.L., Stuut, J.-B.W., et al. (2023) A review of coarse mineral dust in the Earth system. Aeolian Research. 60, 100849. doi:10.1016/j.aeolia.2022.100849.
Aït-Sahalia, Y. & Xiu, D. (2019) Principal Component Analysis of High-Frequency Data. Journal of the American Statistical Association. 114 (525), 287–303. doi:10.1080/01621459.2017.1401542.
Ali, N., Eqani, S.A.M.A.S., Nazar, E., Alhakamy, N.A., Rashid, M.I., Shahzad, K., Zeb, J., Shen, H., Ismail, I.M.I. & Albar, H.M.S.A. (2021) Arsenic and lead in the indoor residential settings of different socio-economic status; assessment of human health risk via dust exposure. Environmental Science and Pollution Research. 28 (11), 13288–13299. doi:10.1007/s11356-020-11546-w.
Cheng, Z., Atwi, K., Onyima, T. & Saleh, R. (2019) Investigating the dependence of light-absorption properties of combustion carbonaceous aerosols on combustion conditions. Aerosol Science and Technology. 53 (4), 419–434. doi:10.1080/02786826.2019.1566593.
Dumka, U.C., Kaskaoutis, D.G., Tiwari, S., Safai, P.D., Attri, S.D., Soni, V.K., Singh, N. & Mihalopoulos, N. (2018) Assessment of biomass burning and fossil fuel contribution to black carbon concentrations in Delhi during winter. Atmospheric Environment. 194, 93–109. doi:10.1016/j.atmosenv.2018.09.033.
Filonchyk, M., Hurynovich, V., Yan, H. & Yang, S. (2020) Atmospheric pollution assessment near potential source of natural aerosols in the South Gobi Desert region, China. GIScience & Remote Sensing. 57 (2), 227–244. doi:10.1080/15481603.2020.1715591.
Gewers, F.L., Ferreira, G.R., Arruda, H.F.D., Silva, F.N., Comin, C.H., Amancio, D.R. & Costa, L.D.F. (2021) Principal Component Analysis: A Natural Approach to Data Exploration. ACM Computing Surveys. 54 (4), 70:1-70:34. doi:10.1145/3447755.
Griffin, D., Zhao, X., McLinden, C.A., Boersma, F., Bourassa, A., et al. (2019) High-Resolution Mapping of Nitrogen Dioxide With TROPOMI: First Results and Validation Over the Canadian Oil Sands. Geophysical Research Letters. 46 (2), 1049–1060. doi:10.1029/2018GL081095.
Ho, K.-F., Lee, Y.-C., Niu, X., Xu, H., Zhang, R., Cao, J.-J., Tsai, C.-Y., Hsiao, T.-C. & Chuang, H.-C. (2022) Organic carbon and acidic ions in PM2.5contributed to particle bioreactivity in Chinese megacities during haze episodes. Environmental Science and Pollution Research. 29 (8), 11865–11873. doi:10.1007/s11356-021-16552-0.
Kherif, F. & Latypova, A. (2020) Chapter 12 - Principal component analysis. In: A. Mechelli & S. Vieira (eds.). Machine Learning. Academic Press. pp. 209–225. doi:10.1016/B978-0-12-815739-8.00012-2.
Kuzu, S.L., Yavuz, E., Akyüz, E., Saral, A., Akkoyunlu, B.O., Özdemir, H., Demir, G. & Ünal, A. (2020) Black carbon and size-segregated elemental carbon, organic carbon compositions in a megacity: a case study for Istanbul. Air Quality, Atmosphere & Health. 13 (7), 827–837. doi:10.1007/s11869-020-00839-1.
Lamsal, L.N., Krotkov, N.A., Vasilkov, A., Marchenko, S., Qin, W., Yang, E.-S., Fasnacht, Z., Joiner, J., Choi, S., Haffner, D., Swartz, W.H., Fisher, B. & Bucsela, E. (2021) Ozone Monitoring Instrument (OMI) Aura nitrogen dioxide standard product version 4.0 with improved surface and cloud treatments. Atmospheric Measurement Techniques. 14 (1), 455–479. doi:10.5194/amt-14-455-2021.
Lee, Y., Ahn, M.-H. & Kang, M. (2020) The New Potential of Deep Convective Clouds as a Calibration Target for a Geostationary UV/VIS Hyperspectral Spectrometer. Remote Sensing. 12 (3), 446. doi:10.3390/rs12030446.
Mao, X., Chen, S.X. & Wong, R.K.W. (2019) Matrix Completion With Covariate Information. Journal of the American Statistical Association. 114 (525), 198–210. doi:10.1080/01621459.2017.1389740.
Marukatat, S. (2023) Tutorial on PCA and approximate PCA and approximate kernel PCA. Artificial Intelligence Review. 56 (6), 5445–5477. doi:10.1007/s10462-022-10297-z.
Ramachandran, S., Rupakheti, M. & Lawrence, M.G. (2020) Aerosol-induced atmospheric heating rate decreases over South and East Asia as a result of changing content and composition. Scientific Reports. 10 (1), 20091. doi:10.1038/s41598-020-76936-z.
Ravindra, K., Singh, T., Vardhan, S., Shrivastava, A., Singh, S., Kumar, P. & Mor, S. (2022) COVID-19 pandemic: What can we learn for better air quality and human health? Journal of Infection and Public Health. 15 (2), 187–198. doi:10.1016/j.jiph.2021.12.001.
Robichaud, A. (2020) An overview of selected emerging outdoor airborne pollutants and air quality issues: The need to reduce uncertainty about environmental and human impacts. Journal of the Air & Waste Management Association. 70 (4), 341–378. doi:10.1080/10962247.2020.1723738.
Sokhi, R.S., Moussiopoulos, N., Baklanov, A., Bartzis, J., Coll, I., et al. (2022) Advances in air quality research – current and emerging challenges. Atmospheric Chemistry and Physics. 22 (7), 4615–4703. doi:10.5194/acp-22-4615-2022.
Velis, C.A. & Cook, E. (2021) Mismanagement of Plastic Waste through Open Burning with Emphasis on the Global South: A Systematic Review of Risks to Occupational and Public Health. Environmental Science & Technology. 55 (11), 7186–7207. doi:10.1021/acs.est.0c08536.
Wei, X., Chang, N.-B., Bai, K. & Gao, W. (2020) Satellite remote sensing of aerosol optical depth: advances, challenges, and perspectives. Critical Reviews in Environmental Science and Technology. 50 (16), 1640–1725. doi:10.1080/10643389.2019.1665944.
Wu, Z., Chen, W., Ma, Y., Xu, T., Yan, F., Lv, L., Qian, Z. & Xia, J. (2023) Explainable data transformation recommendation for automatic visualization. Frontiers of Information Technology & Electronic Engineering. 24 (7), 1007–1027. doi:10.1631/FITEE.2200409.
Ziemke, J.R., Oman, L.D., Strode, S.A., Douglass, A.R., Olsen, M.A., et al. (2019) Trends in global tropospheric ozone inferred from a composite record of TOMS/OMI/MLS/OMPS satellite measurements and the MERRA-2 GMI simulation. Atmospheric Chemistry and Physics. 19 (5), 3257–3269. doi:10.5194/acp-19-3257-2019.

Download PDF

Version 1

posted

You are reading this latest preprint version

MERRA-2 and OMI aided assessment of the effect of major air pollutants on surface temperatures

Status:

Version 1

Abstract

Figures

1. Introduction

2. Study area

3. Data requirements

4. Objectives

5. Software used

6. Methodology

7. Constraints

8. Results and discussion

Aerosol optical depth (388 nm)

Aerosol optical depth (500 nm)

Aerosol Index (UV)

Black carbon (BC)

Dust

Organic carbon

Surface temperature

Correlation analysis and inter-variable interactions

Pre-lockdown

Lockdown

Post-lockdown

Principal component analysis

Lockdown period

Post-Lockdown period

9. Conclusion

Declarations

References

Status:

Version 1