基于星地链路的垂直降雨场反演方法

在分析星地链路几何结构及传播模型的基础上,研究基于联合代数重建技术的星地链路反演二维垂直降雨场的方法.利用实测降雨资料构建3类降雨场,并搭建3条17 GHz垂直极化星地链路进行数值仿真.实验结果表明:单星地链路无法实现二维垂直降雨场的重构,反演场与真实场的相关系数分别为0.556,0.504和0.364;基于双星地链路的反演结果和真实场的相关系数均高于0.98,平均偏差分别为0.122, 0.159和0.537 mm/h,欧式距离均低于0.9 mm/h,熵的相对误差均小于1.6%,基本实现了二维垂直降雨场的反演;三星地链路的应用进一步提升了反演精度,相关系数接近于1,能够精准重构降雨场.实验结果验证了基于星地链路的垂直降雨场反演方法的可行性、准确性和有效性.星地链路的架设和维护简单,探测降雨范围广,易于在高原、山区、海岛等传统地面观测资料缺失的地区使用,可以作为已有降水测量手段的补充.

Inversion of vertical rainfall field based on earth-space links

The accurate measurement of rain intensity and its distribution in vertical direction can not only help to understand the process of rainfall development, but also play an important role in human life such as agriculture, weather forecasting, water resources management, and natural disaster warning. According to the analysis of the geometric structure of earth-space link and propagation model of electromagnetic wave in atmosphere, in this paper we propose a method to reconstruct two-dimensional(2 D) vertical rainfall field by using earth-space links. Firstly, the measured data of micro rain radar(MRR) from Nanjing are used to generate three real vertical rainfall fields which are marked as I, II and III respectively. Secondly, based on the analysis of the earth-space link’ s geometry and the effect of signal attenuation from other factors such as scintillation, atmosphere gas and cloud, the vertical rainfall field inversion model is established. According to the power-law relationship between rain intensity and rain attenuation, which is given by International Telecommunication Union(ITU), the simultaneous algebraic reconstruction technique(SART) is used to inverse the vertical rainfall field. Then, one earth station which can receive a 17 GHz signal from satellite is employed to detect the vertical rainfall field. However, the simulation results show that it is difficult for one earth station to achieve the inversion of rainfall field, and that the correlation coefficients between rainfall fields and inversed fields are 0.556, 0.504 and 0.364 respectively. Based on the result, two earth stations are jointly used. In this simulation, the result shows that after 500 iterations the correlation coefficients all increase above0.98, and the average biases between rainfall field I, II, III and their inversed fields are 0.122, 0.159 and 0.537 mm/h, respectively. Meanwhile, the Euclidean distances decrease to 0.246, 0.235 and 0.812 mm/h, and the relative errors of entropy are both less than 2%. It can be seen from the inversion fields that the vertical distribution of rain rate is close to that of the real field, which suggests that the method proposed in this paper can basically achieve the inversion of vertical rainfall field by using earth-space links. In addition, with the combined detection of three earth stations the accuracy of the inversion results is significantly improved. The correlation coefficients are all close to 1 and the mean deviations are all on the order of 10–12 mm/h, indicating that the 2 D vertical rainfall fields are accurately reconstructed. In the near future, the satellite constellation system will be globally deployed, which can promote the applications of our method in areas, such as plateaus,mountains and islands, where there exist no traditional observation data, serving as a supplement to existing precipitation measurements.