Presenting a combined numerical-machine learning model Galerkin-MARS for modeling meteorological drought (Case study: Khuzestan Province)

Document Type : Research Paper

Authors

1 Department of Arid and Mountainous Areas Restoration Engineering, Faculty of Natural Resources, University of Tehran, Karaj, Iran

2 Assistant professor, Department of reclamation of arid and mountainous regions Engineering, Faculty of Natural Resources, University of Tehran, Karaj, Iran.

10.22059/jwim.2025.392401.1213

Abstract

Drought is one of the natural hazards, especially in arid and semi-arid regions. Analyzing the conditions, characteristics, and status of drought as a type of natural hazard in different regions with a focus on gathering solutions to cope drought and manage its risks is of great importance. In the present study, the analysis and prediction of meteorological drought in Khuzestan province was investigated with individual Galerkin and MARS models and the combined Galerkin-MARS model during a 30-year statistical period (1990-2020). To assess drought conditions, the Standardized Precipitation Index (SPI) obtained from data from eight synoptic stations was used. In the next step, the modeling results were compared with the aforementioned models using goodness-of-fit indices. The results indicated that the combined Galerkin-MARS model is highly efficient for estimating SPI in Khuzestan province. Also, long-term SPI time windows had higher accuracy than short-term time windows in the study area. In general, it can be said that combining numerical models with machine learning in Khuzestan Province leads to increased accuracy in SPI modeling.

Keywords

Main Subjects