基于多源数据的地面降水质量控制方法研究北大核心

本文在研究地面小时降水量与自动站其他观测要素的关联关系和雷达定量估测降水的基础上,提出从自动站关联要素和雷达估测降水两个角度对地面降水进行综合质量控制的多源质量控制方法。利用该方法对2019年5至6月江苏省国家级自动站小时降水量进行质量控制,结果表明:1 h相对湿度变化值、1 h变温,相对湿度与小时降水量关联关系较好;基于SFLA-BP的雷达定量估测降水结果与Z-I关系法相比精度更高,提升了天气雷达对降水质量控制的有效性;多源降水质量控制方法比MDOS准确率提高1.45%,产生可疑数据量下降67.16%,该方法能有效提升现有降水质量控制方法的准确性。

Study on quality control method for surface rainfall based on multi-source data

Based on the study of the correlation between surface hourly rainfall and other observation factors of automatic stations and the quantitative estimation of rainfall by radar based on neural network, this paper presents a multi-source quality control method for ground rainfall from two aspects of automatic station correlation and radar estimation.Resultsshow that the change value of the relative humidity in one hour, the change temperature in one hour, the correlation between the relative humidity and the hourly rainfall is good; compared with the Z-I relation method, the radar based on SFLA-BP has higher precision, which improves the effectiveness of radar in rainfall quality control. The accuracy of multi-source rainfall quality control method is 1.45% higher than that of MDOS, the quantity of generating suspicious data decreased by 67.16%. This method can effectively improve the accuracy of existing rainfall quality control methods.