Evaluation of remote sensing indicators in drought monitoring using machine learning algorithms (Case study: Marivan city)

Remote sensing indices are used to analyze the Spatio-temporal distribution of drought conditions and to identify the severity of drought. This study, using various drought indices generated from Madis and TRMM satellite data extracted from Google Earth Engine (GEE) platform. Drought conditions in Marivan city from February to November for the years 2001 to 2017 were analyzed based on spatial and temporal distribution. Various factors in drought monitoring based on satellite data including maps of normalized vegetation difference index, vegetation condition index, temperature condition index, improved vegetation index, evapotranspiration index, precipitation index, rainfall measurement mission Tropical (TRMM) is considered as an independent variable. Also, the standard rainfall index obtained from meteorological data is calculated as a dependent variable to assess drought conditions. Stochastic forest and support vector regression methods were used to compare remote sensing data and terrestrial data and to examine the correlation between them and to combine several drought factors from remote sensing indicators and build a comprehensive drought monitoring model. The results of the random forest model had explanatory coefficient values ​​of 0.909 and also the squared values ​​of the mean squared error between the actual values ​​and the prediction were 0.259. The stochastic forest model was more accurate than the results of the support vector regression algorithm with the values ​​of the coefficient of determination of 0.88 and the mean squared error of 0.313.