Application of soft-computing techniques for forecasting meteorological drought

Abstract

Drought is amongst the most hazardous climatic hazard, having a tremendous impact on the environment and humansociety. Accurate and reliable drought prediction is critical for effective planning and decision-making in drought-prone areas around the world. Data-driven techniques have been commonly employed for drought forecasting, but choosing an appropriate forecasting model remains challenging due to a lack of information on model performance. The purpose of this study is to compare three machine learning (ML) techniques commonly used for meteorological drought forecasting. The overall results of this study indicated that based on performance criteria and computation time, the ML-techniques are in order SVM > ANN > ANFIS.