A Development of a Prediction Model for Ungauged Catchment in the North of Thailand

Abstract

Flow data are essential for hydrological study, planning, and management to prevent drought and flood in a region. In catchments where flow data are not recorded or of poor quality, hydrological indices could be an alternative for predicting flow in ungauged catchments. This study demonstrates the methodology for predicting flow in ungauged catchments through the case study of 37 sub-catchments of the upper Ping catchment in northwest Thailand from 2006-2014. The regression method was applied to investigate the relationship between three flow indices including runoff coefficient, base flow index, and 95th percentile of flow, and catchment properties. The prediction interval of the regression relationship was used to condition rainfall-runoff model parameters. The model performance was tested by NSE* and reliability. The 95th percentile of flow was found to be the most informative index to regionalize flow followed by RC. The BFI had least contribution to the prediction of flow with poor NSE* and large uncertainty. The 95th percentile of flow and RC generally worked well for small sub-catchments.