中亚地区常规观测站点稀少,需借助星载高光谱AIRS资料分析出该地区数值预报最优初始场。以CRTM中输入探空模拟出的AIRS辐射亮温为参考值,分析了AIRS观测亮温偏差,并评估了 AIRS卫星资料在中亚数值天气预报业务系统中的适用性。结果表明：（1） 各通道模拟所选站点上空亮温最大正偏差的平均值约为3.3 K,最大负偏差的绝对值约为2.6 K。（2） 多个站点平均的AIRS观测辐射亮温整体略高于模拟亮温,其概率密度分布比单个站点更加接近正态分布曲线。（3） AIRS的同化改善了RMAPS-CA对位势高度、温度、比湿等高空要素的预报效果,并未改善高空风速的预报。对各个要素,AIRS的同化改善幅度在低层较高层大。同化后,位势高度、温度、比湿和风速的预报RMSE分别小于20 gpm、2 K、8×10^-4 kg·kg^-1以及5 m·s^-1。

Due to a scarcity of observation sites, only a small amount of conventional observation data on the temporal and spatial distribution of temperature and humidity in Central Asia can be obtained, which makes analysis difficult. High-resolution air infrared detector (AIRS) data can effectively fill the gap. In this paper, the radiance temperature of AIRS simulated by the input radiosonde in the Community Radiative Transfer Model was utilized as the reference value, deviations in the brightness temperature of the AIRS observation were analyzed, and the applicability of AIRS satellite data in the Central Asia numerical weather forecast operation system was evaluated. First, it shows that the average value of the maximum positive deviation of brightness temperature over the selected stations was approximately 3.3 K, and the absolute value of the maximum negative deviation was approximately 2.6 K. Second, the average brightness temperature of the AIRS observation in multiple stations was slightly higher than the overall simulated brightness temperature, and its probability density distribution was closer to the normal distribution curve than that of a single station. Finally, the assimilation of AIRS improved the prediction effect of RMAPS-CA on the geopotential height, temperature, specific humidity, and other high-altitude elements, but did not improve the prediction of high-altitude wind speed. For each factor, the assimilation improvement range of AIRS was larger at the lower and higher levels. After assimilation, the root mean square error of the geopotential height, temperature, specific humidity, and wind speed were less than 20 gpm, 2 K, 8×10^-4 kg·kg^-1 and 5 m·s^-1, respectively. It should be noted that, due to the time limitations imposed on the encrypted sounding experiment, only one full month in July 2016 was selected as the research time period in the deviation analysis of this study. Due to the relatively less vegetation coverage on the underlying surface in Central Asia, the surface radiation was large in summer, while the area without snow cover in winter was relatively small. Therefore, the conclusion of the deviation analysis in this paper is not necessarily applicable to other seasons. In addition, considering the possibility of a business transformation of the research results, this study was based on Rapid-Refresh Multiscale Analysis and Prediction System-Central Asia (RMAPS-CA) to carry out assimilation analysis when evaluating the applicability of AIRS. The system uses the three-dimensional variational data assimilation (3DVAR) method. If a more advanced four-dimensional variational data assimilation (4DVAR) assimilation method is adopted, the assimilation effect may improve.