Linear and wavelet analysis of aerosol optical depth (AOD) and prevailing meteorological factors during summer (2003–2016) over Turkey using Remote Sensing

Abstract

The summer season has a high influence on the aerosol load of Turkey. Atmospheric parameters produce a radiative climatic response when exposed to aerosols from multiple sources. The main aim of this research is to observe the spatio-temporal characteristics of aerosol optical depth (AOD) with atmospheric and surface covariates during the summer of 2003–2016 in Turkey. The AOD products for this study were retrieved at level 3 from Moderate Resolution Imaging Spectroradiometer (MODIS) and Multiangle Imaging Spectroradiometer (MISR) on a daily timescale. Prevailing meteorological parameters such as precipitation, air temperature, relative humidity, surface soil moisture, and enhanced vegetation index (EVI) were used on a daily and monthly timescale. The spatio-temporal distribution of aerosols and meteorological parameters were plotted and analyzed from which each variable exhibited a unique spatial distribution. The Mann–Kendall (MK) trend test was used to analyze the increasing and decreasing trends of AOD throughout the years. MODIS AOD had an increasing trend in the eastern, northeastern, southeastern, southwestern, and northern regions of Turkey whereas a significant decreasing trend was observed in the northwestern regions. MISR AOD had a strong significantly increasing trend in the eastern, northeastern, southern, and central regions whereas the western, northern, and central regions of Turkey expanded under a decreasing trend. Furthermore, correlation analysis was performed to analyze the relationship between AOD and the prevailing covariates from which the correlation between EVI and AOD was extremely variable in both datasets. Wavelet analysis (both continuous and coherence) was performed to analyze the periodic nature of AOD and meteorological parameters. Both MODIS and MISR exhibited significant coherencies in the 16–256 periodic for all covariates except for EVI which was plotted on a monthly time scale. In this study, a notable variation was observed in the space–time dynamics between MODIS and MISR datasets which can be used for future studies that attempt to analyze the relationship between AOD and covariates during the summer season.