四川泥石流频发区暴雨的可预报性研究

初始条件的不确定性是四川泥石流频发区暴雨预报不确定性的重要来源之一。本项目拟用条件非线性最优扰动（CNOP）方法和繁殖模（BV）方法考查初始条件的不确定性对四川泥石流频发区暴雨预报的影响，探讨该区域暴雨的可预报性问题。首先，建立基于WRF模式的非线性最优化系统，以较好地模拟出暴雨的强度和落区。接着，为了使研究结果更加有说服力，同时利用CNOP方法和BV方法考查初始条件的不确定性对暴雨强度和落区预报的影响，分析两种方法所得结果的异同。在此基础上，寻找对四川泥石流频发区暴雨预报结果影响最大的初始关键性区域和关键性物理量，并考查上述区域和物理量的初始场的改善对提高泥石流频发区暴雨预报技巧的有效性。这些研究不仅有助于理解四川泥石流频发区暴雨预报不确定性的机理以及非线性影响机制，而且可以为提高该区域暴雨的预报技巧奠定理论基础，为该区域暴雨的外场加密观测试验的开展提供科学依据。

The initial uncertainty is one of the main sources of heavy rainfall in the debris-flow area in Sichuan province. The program will use the methods of conditional nonlinear optimal perturbation (CNOP) and the breeding vector (BV) to investigate the impact of initial uncertainty on the forecast of heavy rainfall in the Sichuan debris-flow area, and discuss the predictability of the heavy rainfall. First, we will establish a nonlinear optimal system to better simulate the rainfall's amount and area, then in order to make the results more convictive, the CNOP and BV will be used to investigate the impact of initial uncertainty on the forecasts of rainfall amounts and area, and their results will be compared. Based on this, the initial key areas and key variables that would have the largest impacts on the forecasts will be identified. The initials in the key areas and of the key variables will be improved, and their impacts on the forecast skills will be examined. The above studies would help us understand the uncertainty of the heavy rainfall forecasts in Sichuan debris-flow area and the impact of nonlinearity on the heavy rainfalls. Also, they could help us to improve the forecast skill of heavy rainfall and give us scientific references on the performance of the additional observations in field experiment for the heavy rainfalls in the Sichuan debris-flow area.