Optimal Integration of the Ensemble Forecasts from an Ensemble Quantitative Precipitation Forecast Experiment
Journal
Procedia Engineering
Journal Volume
154
Pages
1291-1297
Date Issued
2016
Author(s)
Abstract
For providing precipitation forecasts in Taiwan, Taiwan Typhoon and Flood Research Institute of the National Applied Research Laboratories executes a numerical weather model based quantitative precipitation forecast experiment. In this study, we first evaluate the performance of the ensemble precipitation forecasts during typhoons. Then, a strategy based on artificial intelligence is proposed to optimal integrate these ensemble forecasts for better forecasting performance. To demonstrate the potential of the proposed strategy, an application to 24-h precipitation forecasts during 5 typhoons is conducted. The results indicate that the skill scores of the forecasts provided by the proposed optimal integration strategy are higher as compared to those of the ensemble members and of the ensemble mean, especially for the extreme values. That is, an improved forecasting performance is obtained. The improved precipitation forecasts are expected to be helpful for the disaster decision making during typhoons.
SDGs
Type
conference paper