Downscaled Rainfall Prediction Model (DRPM) using a Unit Disaggregation Curve (UDC)

Abstract. This study was undertaken to identify the process for generating finer time scaled rainfall from higher time scaled data. The Downscaled Rainfall Prediction Model (DRPM) using the technique of unit disaggregation curve (UDC) was developed under the concept of coupling the stochastic autoregressive (AR) model with a wavelet filter and disaggregation model. Sequences of the number of rainy days and monthly rainfall were simulated from 52-year rainfall records at 4 stations in the northeastern part of Thailand. Compared with actual rainfall sequences, the 30 year generated sequences provided R-square values of 0.47-0.60. The model was applied to forecast the number of rainy days and monthly rainfall for the year of 2002. When compared with actual records the prediction model provided R-square values of 0.50 to 0.79.