Abstract
The purpose of this study is to forecast rainfall and river water levels for the Kahayan River in Indonesia and to provide the forecast information to mitigate peat fires. Specifically, rainfall prediction was done first by using the rainfall values observed on the land surface, the dataset provided by the Global Precipitation Climatology Project (GPCP) and the Nearest-Neighbor Method (NNM). Then, water level prediction was done by inputting the predicted rainfall values into the hydrological model. In forecasting the water level, climate changes such as those related to El Niño were taken into consideration by incorporating the Sea Surface Temperature (SST) into the forecast factors. The accuracy of water level forecasts at times of low water in the dry season with scant rainfall was improved by using the lower limit value of predicted rainfall based on the error evaluation. It was demonstrated that long-term water level forecasts with 1- to 3-month lead times were able to be done with reasonable accuracy.
