提出了一种利用支持向量机(SVM)模拟前向辐射传输模式的方法.利用欧洲中期天气预报中心(ECMWF)的RTFOV_8_7前向辐射传输模式和60L-SD廓线集牛成了AMSU-A模拟亮温资料,用模拟亮温和相应的廓线集资料组成训练样本和检验样本,采用SVM方法进行训练.对检验样本的模拟显示,SVM可以用于描写前向辐射传输模式中的非线性映射关系,较好地由大气廓线集资料模拟出与其相关的AMSU-A仪器5-14通道亮温,其中通道6-14的均方根误差在0.1K以内,平均误差算术平均值在0.01K以内.通过多元线件同归方法对温度廓线进行反演试验,发现用SVM模拟的亮温可以用于温度廓线反演,其反演精度可以达到甚至高于RTTOV_8_7计算的亮温.

Based on SupportVectorMachine(SVM), a technique simulating radiative transfer forwardmodel is presented. Using EuropeanCentre forMedium-Range Weather Forecasts (ECMWF) RTTOV_8_7radiative transfer forward model and60L_SD profile database, we simulate the brightness temperature received in AMSU-A instrumen.t Combine this brightness temperature datasets and correspondence profile datasets as training and validation database. After training the SVM network, the simulating technique is validated. The results show that SVM network describes the nonlinear projection relationship between input space and output space very wel,l and the simulated brightness temperatureofchannel5—14isprecise. TheRMS errorofchannel6—14is less than0.1K and themean standard deviation is less than0.01K. In order to find whether SVM simulated brightness temperature is appropriate for temperature retrieva,l muti-regression retrievalmethod isused to retrieve temperatureprofile. Experimentresultshows that the SVM simulate brightness temperature is appropriate for the purpose, and the retrievalprecision is notonly equally butalso a littlemore precise than theRTTOV_8_7simulated brightness temperature.