Investigating the Relationship between Drought and Water Quality Reduction Using Remote Sensing and Neural Network Methods

Due to the fact that droughts can affect both water quality and quantity, the purpose of this study is to determine the effect of droughts on water quality and quantity in Northern Fars province, Iran, based on drought indicators. The drought indices PCI, TVDI, and NDVI are used to study drought from 2000 to 2020. Also, the kriging method is used to generate zoning maps of elements in water (Ca, Cl, EC, K, Na, Mg). Then, using the neural network (MLP) method, the amount of elements in the water is predicted based on drought indices. Based on the values of the drought indicators, the trend of drought changes in the region is increasing from 2000 to 2020, with the southern areas of the region experiencing a more acute drought than the rest of the region. In addition, the zoning map of the elements in water indicated that salt concentrations are higher in the southern parts than in the northern parts. Correlation between drought indices and the amounts of elements in water showed that Ca has a high correlation (R2= 0.820) with TVDI index, and also Cl, EC, K, Na, and Mg have significant correlations (R > 0.8) with the index. Using drought indicators, MLP results for predicting water quality status show that southern regions have more solutes and lower water quality. Furthermore, the R2 values of the model for predicting the elements Cl, EC, K, Na, Mg, TDS, TH using PCI index equal to 0.85 and for Ca using TVDI index equal to 0.71, which indicates high accuracy.