摘要
使用香港地区2016~2018年探空数据,基于随机森林算法构建了大气加权平均温度(RFT_(m))模型,结合北斗卫星系统预测了2018年7月香港地区大气可降水量。结果表明:单一北斗卫星系统获取的对流层延迟精度满足水汽反演精度,RFT_(m)模型较各类单因子模型有17%~44%的精度提升,RFT_(m)模型预测单月大气可降水量精度较单因子模型有20%以上提升。研究表明,北斗卫星系统结合随机森林算法能够更为精确地预测区域大气可降水量。
Using the sounding data of Hong Kong region from 2016 to 2018,an atmospheric weighted average temperature(RFTm)model based on random forest algorithm was constructed to predict the atmospheric precipitable water in Hong Kong region in July 2018 in combination with the BeiDou navigation satellite system.The results show that the accuracy of tropospheric delay obtained by a single BeiDou navigation satellite system meets the accuracy of water vapor inversion,and RFTmmodel has about 17%-44%accuracy improvement over various single-factor models,and RFTmmodel predicts single-month atmospheric precipitable precipitation accuracy has more than 20%improvement over single-factor models.The study shows that the Beidou navigation satellite system combined with the random forest algorithm can predict the regional atmospheric precipitable water more accurately.
作者
丁凯文
张晨晰
陈宇
DING Kai-wen;ZHANG Chen-xi;CHEN Yu(Shandong University of Science and Technology,School of Resources and Environment,TaianShandong 271019,China;China University of Mining and Technology,School of Environment Science and Spatial Informatics,Xuzhou Jiangsu221116,China)
出处
《现代测绘》
2022年第6期7-11,共5页
Modern Surveying and Mapping
关键词
对流层延迟
随机森林算法
大气加权平均温度
大气可降水量
tropospheric delay
random forest algorithm
atmospheric weighted average temperature
atmospheric precipitable water vapor
