Probabilistic Precipitation Forecast in (Indonesia) Using NMME Models: Case Study on Dry Climate Region

Abstract

Reliable precipitation forecast is one of the key inputs to generate accurate and reliable hydrological forecast. This paper uses the North American Multimodel Ensemble (NMME) models to generate seasonal precipitation forecasts in Indonesia. The NMME models are verified against observed precipitation, and the analysis shows that they are biased and underdispersive. The Bayesian Model Averaging (BMA) approach was applied to calibrate the forecast for reliable prediction. East Nusa Tenggara (NTT) is chosen as the pilot study since the region has been well recognized as a dry region with the highest degree of vulnerability toward drought. The results show that the BMA improves the forecast reliability. Moreover, the Canadian Meteorological Center (CMC) models outperform the others. The map of the forecasted Standardized Precipitation Index (SPI) is validated with the observation and shows a high prediction accuracy.