Remote Sensing (RS)
Hosein Nesari; Reza Shah-Hosseini; Amirreza Goodarzi; Soheil Sobhan Ardakani; Saeed Farzaneh
Abstract
Extended Abstract
Introduction
Atmospheric aerosols are a colloid of solid particles or liquid droplets suspended in the atmosphere. Their diameter is between 10-2 to 10-3 micrometers. They directly and indirectly affect the global climate by absorbing and scattering solar radiation, and they also ...
Read More
Extended Abstract
Introduction
Atmospheric aerosols are a colloid of solid particles or liquid droplets suspended in the atmosphere. Their diameter is between 10-2 to 10-3 micrometers. They directly and indirectly affect the global climate by absorbing and scattering solar radiation, and they also have a serious impact on human health by emitting harmful substances. In addition, high concentrations of aerosols on a local scale due to natural or human activities have adverse effects on human health, including cancers, pulmonary inflammation, and cardiopulmonary mortality. Monitoring the temporal and spatial variability of high concentrations of aerosols requires regular measurement of their optical properties such as aerosol optical depth (AOD).
Materials & Methods
Algeria is a large country with little knowledge of the spatial and temporal diversity of AOD, and the low spatial resolution of existing products makes it very difficult to predict aerosols (airborne particles) at the local scale, especially in arid southern regions. As a result, AOD recovery with data with higher spatial resolution is crucial for determining air pollution and air quality information. Several AERONET stations have been installed in Algeria. The Tamanrasset_INM station has been selected based on its location and the availability of historical AOD data for the period (2015-2016).
In this study, Landsat-8 / OLI image from tile 192/44 was used for satellite images. To this end, 23 TOA-corrected L1G-level Landsat-8 / OLI cloudless scenes were downloaded from January 2015 to December 2016 in the study area. DN values are converted to TOA reflections using the scaling factor coefficients in the OLI Landsat-8 metadata file. In this study, the minimum monthly reflectance technique was used to recover AOD in this area. As a result, LSR images were used in the recovery process in different months of 2015 and 2016. The process of selecting reference LSRs was initially based on the selection of clear, foggy / cloudless sky images. The selected images were then used to construct artificial images in which each pixel corresponds to the second lowest surface reflection of all selected monthly images to be the LSR pixel for the respective month. The AOD retrieval method developed in this study is based on a LUT, using the 6S radiative transfer model. The advantage of using the 6S model is its ability to estimate direct components and scattering using a limited number of inputs for each spectral band in the entire solar domain. The effect of the viewing angle is limited because Landsat data are usually obtained with a fixed viewing angle. Surface reflectance can be estimated from a pre-calculated LSR database. The accuracy of AOD recovery depends on the use of the appropriate aerosol model. A continental model was selected from the available aerosol models. Other atmospheric parameters such as ozone, carbon dioxide, carbon monoxide and water vapor are considered by default. The AOD values used to make LUT are set as follows: 0.0, 0.05, 0.1, 1.5, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.2 and 1.5. The zenith angles of the sun and the sensor range from 0 to 70 degrees with a step of 5 degrees and the range of azimuth angles from 0 to 180 degrees with a step of 12 degrees. Using these parameters, the radiative transfer equation was run in forward to obtain the TOA reflection. Different combinations of input and TOA output parameters are stored in LUT. AOD retrieval is based on a comparison between the TOAs estimated with the model and the observed items using the best fit approach. Using such an approach, the estimated AODs are simulated in accordance with those used in the production of TOAs, using a competency function that minimizes the distance.
Results & Discussion
In this study, the AODs recovered at 550 nm in a 5-by-5-pixel window around the AERONET site were averaged. The considered AERONET values are the average of all measurements taken within ± 30 minutes of image acquisition time. Observation regression results (AOD from Landsat 8 images and AERONET stations) showed that the correlation coefficient is about 84%. This study shows a good fit of the model on the research data and shows the high capability of the model. This study showed a strong recovery of AOD against AERONET data of more than 70% at . The differences can be attributed to a limited number of points or hypotheses related to the aerosol model used in this study. The assumption of using a pre-calculated LSR does not limit the accuracy of this method because we have shown that in arid regions where the change in land cover in different months of the year is small, a pre-calculated LSR image can be representation used the share of surface reflection in the radiative transfer model throughout the month.
Conclusion
In this study, an AOD derived from a high-resolution satellite at an urban scale was produced in the city of Tamanrasset, Algeria. The developed method assumes that the change in land cover is minimal and the temporal change in LSR is not significant. A pre-calculated LSR image is created to show the surface reflection in the retrieval process. Based on the 6S radiative transfer model, an LUT was constructed to simulate the TOA reflection of the built-in LSRs and a set of geometric and atmospheric parameters. The retrieved AODs were compared with the AERONET ground data. The results show that this approach can achieve reasonable accuracy in AOD recovery, which reaches about 70.9% at . In addition, this approach is suitable for estimating AOD in urban areas compared to existing AOD products with low spatial resolution. The results of this study show a 4% improvement compared to the results of Omari et al. (2019). The results of this study showed that ignoring the monthly changes in LSR values leads to good results in AOD recovery.
Hossein Bagheri; Mohammad Hassan Zali
Abstract
Extended Abstract
Introduction
The concentration of particulate matters has recently increased in the metropolitan area of Tehran resulting in many severe hazards for both the environment and citizens. Particulate matters (PM) with a diameter less than 2.5 microns (PM2.5) are considered to be one ...
Read More
Extended Abstract
Introduction
The concentration of particulate matters has recently increased in the metropolitan area of Tehran resulting in many severe hazards for both the environment and citizens. Particulate matters (PM) with a diameter less than 2.5 microns (PM2.5) are considered to be one of the most dangerous types of pollution. Estimating the concentration of these particles in Tehran is challenging due to the existence of various sources of pollution and the lack of sufficient ground stations. Aerosol optical depth (AOD) data retrieved from satellite imagery can be an alternative. However, AOD are not easily convertible into surface pollution and requires the development of appropriate models such as those based on data-driven approaches and machine learning techniques. Thus, the present study seeks to create a model to estimate the concentration of PM2.5 in Tehran employing deep generative models and in-situ measurements, meteorological data, and AOD data extracted from MODIS satellite imagery. Reviewed literature has proved the ability of deep learning techniques to solve regression and classification problems. Deep learning techniques are divided into various categories, one of which is based on the generative models seeking to reconstruct the input features. In this way, high-level and efficient features can be employed to explore the relationship between PM2.5 and AOD. Thus, the present study has investigated the potential of deep generative models for estimating PM2.5 concentration from high resolution AOD data retrieved from satellite imagery.
Materials and Study Area
As a metropolitan area suffering from air pollution particularly in winters, the capital city of Iran, Tehran was selected as the study area. PM2.5, the main source of pollution in Tehran, is mainly emitted from vehicles and especially old urban public transport fleet.
Aerosol data collected by Aqua and Terra sensors of MODIS and retrieved by Multi-angle Implementation of Atmospheric Correction (MAIAC) algorithm were used in the present study. Meteorological data were obtained from the global ECMWF climate model, and the concentration of PM2.5 was measured at air quality monitoring stations. Data were collected for a time interval of January 2013 to January 2020.
Methods
The present study has investigated the potential of deep generative models used to provide an estimate of PM2.5 concentration based on satellite AOD data. To reach such an aim, three types of deep generative neural networks, deep autoencoder (DAE), deep belief network (DBN) and conditional generative adversarial network (CGAN) were developed. Moreover, the performance of deep generative modes was compared with linear regression techniques as typical models used to explore the relation between PM2.5 and AOD data. Finally, the most accurate model for the generation of high resolution (1km) PM2.5 maps from AOD data was selected based on the performance of models.
Results and Discussion
The accuracy of each developed model was evaluated using the test data and the obtained results were compared with results obtained from other basic linear regression models. Accuracy evaluation indicated that the developed deep autoencoder (DAE) combined with support vector regression led to the highest correlation (R2 = 0.69) and lowest RMSE (10.34) and MAE (7.95) and thus, can be potentially used for high resolution estimation of PM2.5 concentration. Next was the developed deep belief network which with a performance close to DAE demonstrated its potential capability to estimate PM2.5 concentration from satellite AOD data. The CGAN network acted less accurately in the estimation of PM2.5 concentration as compared to other deep generative models, but outperformed the linear regression algorithms on the test data. To sum up, findings indicated that deep generative models have outperformed classical linear regression techniques used for high resolution estimation of PM2.5 from satellite AOD data. Among the linear methods, the highest accuracy was achieved by the Lasso algorithm with an RSME of 12.14 and MAE of 9.46 on the test data which showed the significance of regularization for the improvement of performance in linear regression algorithms. Nevertheless, the accuracy of linear regression techniques was much lower than deep generative models.
Conclusion
Finally, DAE was selected as the best model for the estimation of PM2.5 concentration across the study area and high resolution maps of PM2.5 concentration were generated using the developed model. Investigating the daily PM2.5 maps generated for two days with different air quality conditions (clean and polluted) demonstrated the efficiency of the developed DAE for PM2.5 modeling.