Estimation of Aerosol Optical Depth Using MODIS sensor Images over Persian Gulf surfaces

Introduction
Atmospheric aerosol plays a significant role in the Earth's radiation budget through radiative forcing and chemical perturbations. Aerosols are intricately linked to the climate system and the hydrologic cycle. The net effect of aerosols is to cool the climate system by reflecting sunlight. Quantifying the net effect requires accurate information on the global distribution of aerosol properties that have to be estimated from satellite observations. Estimating aerosol properties is also one of the first steps in generating high-level land surface products from satellite observations. Effective aerosol retrieval information is also essential to satellite imagery atmospheric correction. Satellite remote sensing has been employed to supplement the prediction of ground-level dust concentration. Satellites are able to cover vast spaces at a relatively low cost. For aerosol studies, the launch of the Moderate Resolution Imaging Spectro-radiometer (MODIS) has enabled the retrieval of aerosol optical depth (AOD) data globally from the satellite's spectral observation. MODIS AOD is a measure of light transmittal by aerosols in an atmospheric column during the satellite overpass. With the evolution of the retrieval algorithm, MODIS AOD has become increasingly important in the role of producing more accurate estimation for the aerosol. Estimation of aerosol loadings is of great importance to the studies on global climate changes. Meteorological numerical models and ground stations are not able to tracking and detection of dust storms and in many cases have significant errors. This issue show necessitates use of reconstruction ways dust according to remote sensing techniques. The purpose of this study, use of remote sensing technology and MODIS images to estimate dust concentration on the Persian Gulf surface and estimating the linear correlation relationship between the dust measurements in ground and atmospheric. In this study, we develop a new algorithm for estimating the aerosol optical depths using MODIS data over Persian Gulf surfaces. This algorithm is validated using AERONET measurements.
Materials & Methods
In the current study we analyze atmospheric aerosol optical properties over Persian Gulf. Annually, Dust storms are imported into Persian Gulf from the West and North West and South West. Data corresponding to the station was extracted for channels of 0.644, 0.855, 0.466, 0.553, 1.243 and 1.643 µm of satellite image. The ocean algorithm was designed to retrieve only over Dark Ocean, (i.e. away from glint). There is a special case when we retrieve over glint, and that is described below. The algorithm calculates the glint angle, which denotes the angle of reflection, compared with the specular reflection angle. The glint angle is defined as:
Θ_glint=〖cos〗^(-1) ((cosθ_s cosθ_v )+(sinθ_s sinθ_v cosϕ)) (1)
Where θ_s, θ_v and ϕ are the solar zenith, the satellite zenith and the relative azimuth angles (between the sun and satellite), respectively. The retrieval requires a single fine mode and a single coarse mode. The trick, however, is to determine which of the (4 x 5) twenty combinations of fine and coarse modes and their relative optical contributions that best mimics the MODIS-observed spectral reflectance. The reflectance from each mode is combined using η as the weighting parameter:
ρ_λ^LUT (τ_0.55^tot )=ηρ_λ^f (τ_0.55^tot )+(1-η)ρ_λ^c (τ_0.55^tot) (2)
Where ρ_λ^LUT(τ_0.55^tot) is a weighted average reflectance of an atmosphere with a pure fine mode 'f' and optical thickness τ_0.55^tot and the reflectance of an atmosphere with a pure coarse mode 'c' also with the same τ_0.55^tot. Before the final results are output, additional consistency checks are employed. Our primary means of true ‘validation’ is comparison with ocean-based sunphotometer measurements, specifically, those of AERONET. The AERONET measured AOD is easily interpolated to the exact MODIS wavelengths (for example 0.55 µm) by quadratic interpolation in log reflectance/log AOD space. The AERONET ‘sun-measured’ definition of FW differs from either of the MODIS (land or ocean) definitions, but should be correlated with either. AOD images were compared with values obtained in AERONET stations, in finally AOD values were evaluated using statistical indexes of Average, standard deviation, correlation, Mean Squared Error (RMSE) and Mean Difference Square Error (RMSD). In the following validation, we use AERONET Level 1.5 data of the Dalma, Bahrain, Abu Al Bukhoosh, Sir Bu Nuair, Umm Al Quwain, MAARCO and Mussafa station when available.
Discussion of Results & Conclusions
The evaluation results showed that good correlation exists between the AOD simulation and AERONET data, with the correlation coefficient exceeding 0.90. The best and most suitable mode demonstrated for 1.243 and 1.643 bonds with the correlation coefficient equal to 0.94 and 0.99 and RMSE and RMSD index equal to 0. 2 and 0.02 for band of 1.243 and 0.1 and 0.01 for band of 1.643, respectively. We conclude that significant limitations exist for aerosol retrieval using marine AERONET stations. The number of matching points between the two datasets may become sufficient toattempt the reduction of the current uncertainties. Given the identified uncertainties, the results of this study do not contradict these previous validation efforts. Future research may reduce these uncertainties and require modifications to the retrieval algorithm. We used AOT data from a comprehensive set of Persian Gulf AERONET stations to evaluate the dust retrieval algorithm. Data evaluation was performed by using the Pearson correlation coefficient, root mean square error index (RMSE) and root mean square deviation index(RMSD). The evaluation results showed that good correlation exists between the AOD simulation and AERONET data, with the correlation coefficient exceeding 0.90. The regression analysis of AOD data revealed similar limitations. We found that the AOD simulation are on average 5%–25% more than the corresponding AERONET values, depending on the regression weighting assumptions for the comparison dataset. To further evaluate the performance of the algorithm in comparison with the AERONET measurements, we used the RGB image from MODIS. Overall, the comparison with the AERONET data has revealed similar performance of the two satellite datasets with a tendency of the simulation AOTs to underestimate and the MODIS over-ocean AOTs to overestimate the AERONET values. The range of these discrepancies is comparable to the uncertainties associated with the limited number of ocean stations and natural aerosol variability. The comparison of AOTs in the Persian Gulf AERONET stations showed that AOD values in AERONET stations are, on average, 5%–25% lower than the corresponding simulation ones. While the biases in ocean retrieval algorithms and cloud screening procedures may not be excluded, these results indicate that aerosol loading in marine locations may differ significantly from that in adjacent land areas, thereby limiting the achievable validation accuracy. We conclude that significant limitations exist for aerosol retrieval using marine AERONET stations. These limitations can be linked to the extreme sparsity of the marine AERONET data, uncertainties associatedwith local conditions at the marine stations, and regression accuracy limits imposed by natural aerosol variability. The number of matching points between the two datasets may become sufficient toattempt the reduction of the current uncertainties. Given the identified uncertainties, the results of this study do not contradict these previous validation efforts. Future research may reduce these uncertainties and require modifications to the retrieval algorithm.