顾及GNSS水汽线性和非线性特征的ERA5 PWV产品校正方法  

The ERA5 PWV correction method considering linear and nonlinear characteristics of GNSS PWV

作　　者：郭宏武 李浩杰 杜芬玲 李祖锋[5] 马智 赵庆志[3] 翟园 杜玉柱[6] 岳延兵[6] GUO Hongwu;LI Haojie;DU Fenling;LI Zufeng;MA Zhi;ZHAO Qingzhi;ZHAI Yuan;DU Yuzhu;YUE Yanbing(Xi'an Meteorological Bureau,Xi'an 710032,China;Key Laboratory of Transportation Meteorology of China Meteorological Administration,Nanjing Joint Institute for Atmospheric Sciences,Nanjing 210041,China;College of Geomatics,Xi'an University of Science and Technology,Xi'an 710054,China;Xianyang Institute of Surveying and Mapping,Xianyang,Shaanxi 712000,China;Power China Northwest Engineering Corporation Limited,Xi'an 710065,China;Shanxi Conservancy Technical Institute,Yuncheng,Shanxi 044000,China)

机构地区：[1]西安市气象局,西安710032 [2]南京气象科技创新研究院中国气象局交通气象重点开放实验室,南京210041 [3]西安科技大学测绘科学与技术学院,西安710054 [4]咸阳市勘察测绘院,陕西咸阳712000 [5]中国电建集团西北勘测设计研究院有限公司,西安710065 [6]山西水利职业技术学院,山西运城044000

出　　处：《导航定位与授时》2025年第1期97-110,共14页Navigation Positioning and Timing

基　　金：国家自然科学基金(42274039);陕西省气象局秦岭和黄土高原生态环境气象重点实验室开放研究基金(2023K-1);山西省水利科学技术研究推广项目(2024GM10,2024GM11);南京气象科技创新研究院北极阁开放研究基金(BJG202411)。

摘　　要：高精度大气可降水量(PWV)对数值天气预报和短临极端天气研究等具有重要意义。第五代欧洲中尺度天气预报中心再分析数据集(ERA5)能够提供高时空分辨率的PWV产品,但在局部区域其精度并不理想,无法满足区域精细化天气预报预警的现实需求。为提高ERA5 PWV产品的局部区域精度,提出了一种顾及GNSS水汽线性和非线性特征的EAR5 PWV校正方法。该方法考虑了GNSS PWV与ERA5 PWV之间的系统偏差,利用Lomb-Scargle周期图方法分析了PWV偏差周期项,基于最小二乘原理建立了PWV偏差的线性周期校正模型。其次,综合考虑了线性校正后PWV、不同气象参数和时空因子对水汽残差的影响,基于反向传播神经网络(BP-NN)构建了不同季节ERA5 PWV非线性校正模型,以优化ERA5 PWV局部区域精度。选取中国大陆构造环境监测网络2021-2023年GNSS,ERA5和无线电探空站的PWV,以及气象站实测数据进行实验。结果表明,该方法在不同时空水汽对比上均具有较好的精度,与ERA5 PWV产品相比其均方根(RMS)平均改善率为32.15%。该方法能够有效改善局地ERA5 PWV精度,为区域精细化天气预警预报研究等提供高精度的水汽信息。Highly accurate precipitable water vapor(PWV)is of great importance for numerical weather prediction,short-range extreme weather prediction studies,etc.The fifth generation European centre for medium-range weather forecasts reanalysis(ERA5)is capable of providing PWV products with high spatiotemporal resolution.However,its accuracy is not ideal in local regions and cannot meet the practical needs for regionally refined weather forecasts and warnings.In order to effectively improve the accuracy of ERA5 PWV products in the local region,an ERA5 PWV correction method considering linear and nonlinear characteristics of GNSS PWV is proposed.This method considers the systematic deviation between GNSS PWV and ERA5 PWV,analyses the period term of PWV deviation by using the Lomb-Scargle(LS)periodogram method,and develops a linear period correction model for PWV deviation based on the least squares principle.In addition,the influence of linearly corrected PWVs,various meteorological parameters and spatiotemporal factors on the PWV residuals is comprehensively considered.The nonlinear correction model of ERA5 PWV in different seasons is developed based on the back propagation neural network(BP-NN),and the accuracy of ERA5 PWV in local regions is optimized.The PWV derived from the crustal movement observation network of China(CMONOC)GNSS stations,ERA5,and radiosondes stations,as well as the meteorological data measured by meteorological stations from 2021 to 2023 are selected for the experiment.The results show that the proposed method has good accuracy in different spatiotemporal water vapor contrasts,and its root mean square(RMS)average improvement rate is 32.15%compared with the ERA5 PWV product.The proposed method can effectively improve the accuracy of local ERA5 PWV and provide high-precision water vapor information for regional refined weather warning and forecasting research.

关 键 词：GNSS ERA5 PWV Lomb-Scargle周期图 反向传播神经网络 

分 类 号：P228.4[天文地球—大地测量学与测量工程]