卫星微波观测资料的全天候同化关键技术研究

Most severe weather processes are related to clouds and precipitation and the observation error of cloud- and precipitation-affected observations is very important for assimilating the microwave observations，which directly impacts the data usage and the quality of the analysis in the numerical weather prediction (NWP). There are two main methods in estimation of satellite radiance observation errors which are statistical estimation and physical simulation, and the former one is mainly applied in clear-sky conditions, namely the constant observation error. However, in all-sky conditions (clear, cloudy and precipitating), this approach would overestimate the observation error in the clear-sky while underestimate the observation error in the cloudy and rainy areas. Consequently, one way is to assume that the first guess departure (O-B) is situation-dependent in the current data assimilation framework to improve the data usage of the microwave observations in the cloudy and rainy areas and the predictability of the severe weather processes. Based on this, this project is to study the impacts of the cloud amount variability on the observation error of the microwave satellite radiance in all-sky conditions, and develop the all-sky model of observation error of the microwave satellite radiance, which is varies as a function of cloud amount. Furthermore, the assimilating and forecasting impacts by using the all-sky observation error model are to be discussed, and it will provide the important references for improving the used data rate of the microwave satellite observations of the Chinese satellites of the Fengyun-3 series in NWP.

绝大部分重大天气过程均与云雨有关，而云雨区的卫星微波辐射率资料的观测误差在卫星资料同化中非常重要，会直接影响分析场质量。卫星资料的观测误差的给定主要基于统计估计和物理模拟两种方法，而前者主要应用于在晴空条件下给定常量观测误差。然而在晴空、多云、雨区等全天候条件下，采用给定常量观测误差法，就会高估晴空条件下的观测误差而低估云雨区的观测误差。为了提高云雨区卫星微波观测资料的利用率以及对重大天气过程的预报能力，方法之一就是在现有同化框架中假设观测背景误差（O-B）是状态依赖的。基于此，本项目将研究云量的变化对卫星微波资料观测误差的影响，进一步发展随云量变化而变化的全天候条件下卫星微波资料观测误差模型，并探讨在启用全天候观测误差模型的情况下产生的同化及预报影响，为有效提高我国风云三号卫星微波观测数据在数值天气预报中的使用率提供重要理论依据。

为了充分且准确利用风云三号系列卫星微波观测资料在云雨区的观测，在针对风云三号C/D星微波湿度计和微波成像仪观测资料进行基本的质量评估的基础上，利用Geer and Bauer提出的对称误差模型（symmetric error model），选用RTTOV-SCAT做为观测算子，建立了适用于FY-3C/D微波湿度计和微波成像仪数据的随云量变化而变化的观测误差模型。同时进行了全天候条件下动态地表发射率的反演研究，完成了以动态地表发射率为主，气候态发射率为辅的地表发射率分布计算。并选取台风、西南涡、梅雨等多个重大天气过程的典型个例，利用WRFDA系统和WRF模式开展了较为全面的全天候同化预报与晴空同化预报对比试验。研究结果表明，全天候同化算法的应用，有效增加了进入同化系统的微波湿度计MWHS-2的数据量，所增加的观测均位于深厚云区内，充分利用了云雨区的观测；全天候同化通过改善温度，湿度，风场等物理量场的预报，对于重大天气过程的总体预报结果有轻微正面的影响；而全天候条件下的动态地表发射率的反演可为卫星资料同化提供更加接近实况的地表发射率，在增加进入同化系统的风云三号C/D星微波观测资料的同时，亦可对陆地上的强降水预报产生中性到轻微正面的影响。这些算法均适用于风云三号系列卫星微波辐射计及同类卫星观测仪器，也可进一步应用于GRAPES业务预报模式，使得风云三号系列卫星的微波观测资料在气象业务上得到更加充分有效地利用，为国家和人民提供更加准确的天气预报服务。