Enhancing Water Level Forecasting Performance in High-Variability Basins through Data Restructuring: A Case Study of the Yom River Basin, Thailand

The Yom river basin in one of the 22 main river basins of Thailand. This experiences perennial floods and droughts that heavily impact the agricultural sector. In order to reduce the impact, water management, including water level estimation. A considerable task of management is the quantitative forecasting of water levels. This study proposes appropriate forecasting models for time series of daily water level data from four water level measurement stations. The study period is from 2007 to 2022 on September. The efficiency of this forecasting model was determined from comparisons to three approaches, centered moving average model (CMA), additive decomposition model (DEC), Holt’s Winter additive model (WIN). Results indicated that: The forecasts of two years gave similar forecast patterns to the previously observed values. Mainly, (decomposition) was more accurate than the other approaches for all stations. The RMSEs of upstream was slightly greater than the downstream RMSEs for three approaches.