Statistical Downscaling Model Based-on Support Vector Regression to Predict Monthly Rainfall: A Case Study in Indramayu District

Abstract

Knowledge of weather and etimate, especially rainfall is significantly necded in agricultural sector. Accurate information of raintall can be used to determine the planting pattern and time appropriately, 50 that farrners can avoid crop failure caused by floods due to high rainfall and drought due to low rainfall. Techniques of statistical downscaling (SD) using a global circulation model output (GCM) are cornrnonly used as a primary tool to learn and understand the etimate system. The airn ofthis research was to develop an SD model using support vector rcgression (SVR) with GCM as input to prcdict monthly rainfall in the district of Indramayu. The research results showcd that GCM can be used to prcdict the average value of monthly rainfall. The best rcsult of prcdiction is at the Bondan Station having an average correlation of 0.766.