Meteorological Drought Prediction in Karoon Watershed using Meteorological Variables

Abstract:The drought prediction, its severity and occurrence time are among important issues of meteorology and hydrology researchers. In this respect, the importance of weather and meteorological variables for estimating the drought severity is important for researchers. In this research by using variables such as air temperature and geopotential height in various layers of atmosphere, the relation between these variables in 31 years was investigated by standardized precipitation index (SPI) as regional drought index in Karoon watershed. After estimating average precipitation by statistical index of mutual information of point- meteorological variables that have a stronger relation with SPI profile in investigated area, meteorological variables were selected as suitable variables. Then along with prediction of regional drought class SPI, the support vector machine (SVM) was used as one of statistical learning methods. The results showed that this method is suitable for prediction of drought occurrence. Also by location searching for relation between meteorological variables and target values, the best combination of variables was selected and in most of cases this prediction with a suitable precision is applicable for real usages. The prediction precision in autumn and winter was higher than other seasons.