Evaluation of drought events in various climatic conditions using data-driven models and a reliability-based probabilistic model

Due to a wide range of socio-economic losses caused by drought over the past decades, having reliable insight properties plays key role in monitoring and forecasting situations, finally generating robust methodologies for adapting various vulnerability situations. The most important factor causing is rainfall, but increasing or decreasing temperature consequently, evapotranspiration can intensify moderate severity events. Standardized Precipitation Evaporation Index (SPEI), as one well-known indices definition situation, applied based on potential precipitation, evapotranspiration, water balance. In this study, values SPEI are formulated climates three Artificial Intelligence (AI) models: Gene Expression Programming (GEP), Model Tree (MT), Multivariate Adaptive Regression Spline (MARS). Meteorological variables including maximum (Tmax), minimum (Tmin), average (Tmean), relative humidity (RH), 24-h rainfall (P24) wind speed (U2) were used perform AI models. Dataset reported from four synoptic stations through Iran, dating back 58-year period beginning 1957. Each technique was run all climatic situations: Temperate-Warm (T-W), Wet-Warm (W-W), Arid-Cold (A-C), Arid-Warm (A-W). Results models development indicated that M5 version MT provided accurate prediction situations comparison with GEP MARS techniques. conditions evaluated reliability-based probabilistic framework take into account influence any uncertainty randomness associated meteorological variables. way, Monte-Carlo scenario sampling approach has been assess limit state function models-based-SPEI. Based reliability analysis stations, probability exceedance declined below 75%, varied “Normal” “Very Extreme Humidity”.