多目标遗传算法反演对流层大气温湿廓线研究

Research on Multi-Objective Genetic Algorithm Inversion of Tropospheric Atmospheric Temperature and Humidity Profile

针对地基微波辐射计反演对流层大气温湿廓线完全依赖于历史数据的问题,设计了一种新的反演算法和流程。在对历史探空数据的统计分析基础上,得到各层大气温湿参数经验范围。从经验库中随机构造一条大气温湿廓线作为初值,基于大气微波辐射传输模式MonoRTM模型,计算模拟亮温。比较计算值与实测亮温的接近程度,通过更新初值,不断迭代计算,逐步筛选出可行解,并通过设定大气温湿垂直递减率等约束条件,从可行解中约束出帕累托前沿。采用皮尔逊系数加权平均的方法,从帕累托前沿中得到全局满意解。研究结果表明,新建立的基于多目标遗传算法的不完全依赖于历史数据的对流层大气温湿廓线反演模型,有较好的自适应能力和鲁棒性,反演精度高。该模型可以满足微波辐射计在历史气象资料积累匮乏地区的使用需求。

A new inversion algorithm and process are designed to address the problem that the ground-based microwave radiometer inversion of tropospheric atmospheric temperature and humidity profiles is completely dependent on historical data.Based on the statistical analysis of historical sounding data,the empirical ranges of atmospheric temperature and humidity parameters for each layer are obtained.An atmospheric temperature and humidity profile is randomly constructed from the empirical library as the initial value,and the simulated bright temperature is calculated based on the atmospheric microwave radiation transport model MonoRTM model.Comparing the proximity between the computed value and the measured bright temperature,the feasible solutions are gradually screened out by updating the initial value and continuously iterating the computation,and the Pareto front is constrained from the feasible solutions by setting the constraints,such as the vertical rate of atmospheric temperature and humidity decreasing.The method of weighted average of Pearson coefficients is used to obtain a globally satisfactory solution from the Pareto front.The research results show that the newly established inversion model of tropospheric atmospheric temperature and humidity profiles based on multi-objective genetic algorithm that does not completely depend on historical data has good adaptive ability and robustness,and the inversion accuracy is high.The model can meet the usage requirements of microwave radiometers in areas where the accumulation of historical meteorological data is scarce.