Study of Eleven Tropical Cyclones Simulated by Sample Optimization of an Ensemble Forecast Based on the Observed Track

Abstract

ABSTRACT The quality of ensemble forecasts is significantly affected by sample quality. In this paper, the influence of sample quality on simulation results is analyzed by optimizing the distribution of ensemble members. As part of our research, simulated and observed tracks are compared; samples with smaller track errors are retained, and samples with larger track errors are eliminated in order to improve the overall quality of the ensemble forecast. The Weather and Research Forecasting model was used to simulate 11 tropical cyclones that occurred in the northwest Pacific to test the ability of our scheme to improve the forecast track of these cyclones. The results show that, in most cases, sample optimization effectively reduces the track error of tropical cyclones. Overall, the 12-hour, 24-hour, and 36-hour errors in the forecast track are reduced by 10.95 km (20.35%), 10.26 km (16.95%), and 10.1 km (14.71%), respectively. In addition, the forecast of tropical cyclone intensity was improved to a certain extent. Thus, it was confirmed through quantitative measurements that sample optimization based on the observed track improves the track prediction of tropical cyclones.