基于先进的微波扫描辐射计AMSR-E/2观测的辐射值,利用-维变分算法(1D-Var)反演各类水成物(云水、雨水和云冰)的垂直廓线,并对其反演结果进行检验。以2014年8月台风"夏浪"为例,分两步对变分反演的云微物理参数进行了检验。首先,将反演的各类水成物含量补充到辐射传输模式的输入场,观测算子模拟的AMSR-2各通道亮温与实况观测相比非常接近,可以很好地模拟出台风外形、强度及螺旋结构。其次,将反演的水成物廓线与载在CloudSat上的云雷达CPR同时段观测的雷达反射率因子进行对比,发现反演出的云水、雨水含量大值区与毫米波云雷达观测的雷达反射率因子高值区一一对应,进一步说明1D-Var反演的水成物参数精度很高。然而,由于星载AMSR-E/2观测通道少且空间分辨率低,对尺度较小、较薄的云不敏感,同时对云层较厚的密闭云区和多层云区反演能力也有限。

The amount of hydrometeors in clouds plays an important role in the Earth's radiation balance. It is also an important parameter in representing clouds in global circulation models used for climate study and weather forecasting. Satellite data have been widely used to estimate global atmospheric parameters. In this study, we introduce in detail a 1D-Var retrieval algorithm and assess the quality of the devived hydrometeor products. This algorithm could provide an estimate of the geophysical state, especially hydrometeor profiles, which are used as first guess and/or background before starting data assimilation. This algorithm is beneficial to the assimilation of satellite measurements under cloudy and rainy conditions. A one-dimensional variational retrieval algorithm is developed to retrieve hydrometeor parameters (profiles of liquid cloud, liquid precipitation, and ice cloud) from spaceborne microwave AMSR-E/2 measurements. The algorithm is an iterative physical inversion system that finds a consistent geophysical solution to fit all radiometric measurements simultaneously. It inverts the radiative transfer equation by finding radiometrically appropriate profiles of geophysical parameters. In addition, the retrieved parameters include a set of derived products that are a simple vertical integration of fundamental profiles, such as total precipitable water, cloud liquid water, ice water path, and rainfall rate. AMSR-2 measurements from Halong Typhoon in 2014 were used as examples, and all of the retrieved products were assessed These hydrometeor profiles were integrated into the radiative transform model (observation operators), in which cloud absorption and scattering effect were measured. The simulated and observed brightness temperatures were consistent in all microwave channels. The retrieved hydrometeor profiles were validated using the observed reflectivities of the Cloud Profiling Radar uploaded on CloudSat satellite. Comparison results showed that areas with high radar reflectivity matched the cloud water content and liquid precipitation regions at high amounts, proving the high precision of hydrometeor retrievals from the 1D-Var algorithm. However, the AMSR-E/2 observations were not sensitive to small-scale shallow clouds because of its few channels and poor spatial resolution. In addition, the inversion ability of satellite microwave measurments was limited to overcast or layered clouds with a high optical thickness. These hydrometeor parameters are extremely difficult to assess because of the lack of effective ways to measure these quantities (either from ground-based or satellite sensors). Mutual validation of these hydrometeor products from different sensors for long periods of time is still needed.