文献类型：期刊文献

英文题名：Deep Learning-Based Feature Importance for Rainfall Nowcast Driven by GNSS PWV and CAPE

作者：He, Lin Xu, Chaoqian Wang, Hong Zhang, Hui Cao, Juncai Li, Qingsong Zhao, Jian Liu, Yang Zhang, Tian

第一作者：He, Lin

通信作者：Liu, Y[1]

机构：[1]Wuhan Univ, Guiyang 550000, Peoples R China;[2]Guizhou Panjiang Coal & Elect Grp Technol Res Inst, Guiyang 550000, Peoples R China;[3]Wuhan Univ, Wuhan 430000, Peoples R China;[4]Northeastern Univ, Shenyang 110000, Peoples R China;[5]Guizhou Inst Technol, Guiyang 550000, Peoples R China;[6]Guizhou Coal Mine Design Res Inst Co Ltd, Guiyang 550025, Peoples R China;[7]Guizhou Elect Power Design & Res Inst Co Ltd, Guiyang, Guizhou, Peoples R China;[8]Xian Univ Sci & Technol, Xian 710054, Peoples R China;[9]Shaanxi Prov Architectural Design & Res Inst, Electromech Design Inst, Xian 710000, Peoples R China

第一机构：Wuhan Univ, Guiyang 550000, Peoples R China

通信机构：corresponding author), Xian Univ Sci & Technol, Xian 710054, Peoples R China.

年份：2025

卷号：18

起止页码：26688-26698

外文期刊名：IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING

收录：;EI(收录号：20254319389793);Scopus(收录号：2-s2.0-105019563551);WOS:【SCI-EXPANDED(收录号:WOS:001604856100003)】；

基金：This work was supported by the National Key Research and Development Program of China under Grant 2023YFB3907100.

语种：英文

外文关键词：Rain; Global navigation satellite system; Weather forecasting; Long short term memory; Predictive models; Forecasting; Accuracy; Wind speed; Ocean temperature; Data models; Convective available potential energy (CAPE); global navigation satellite system (GNSS); long short-term memory (LSTM); precipitable water vapor (PWV)

摘要：Accurate rainfall nowcasting remains one the most challenging tasks in weather forecasting. Previous studies mainly developed rainfall forecast models by combining global navigation satellite system (GNSS) derived precipitable water vapor (PWV) with shallow neural networks or simple machine learning algorithms. However, these shallow models suffer from capturing time-dependent phenomena, difficulties in achieving stable solutions, filtering insignificant inputs, and overfitting. In addition, the quantitative contributions from predictors for rainfall forecasting are still unclear. Therefore, a deep learning rainfall nowcast (DLRN) model based on a Long Short-Term Memory network and Decision Tree Regression, driven by GNSS PWV, convective available potential energy (CAPE), and multiple meteorological data, is proposed in this study. Hourly PWV, CAPE, temperature, relative humidity, pressure, wind speed, wind direction, and rainfall data from 20 GNSS stations in Taiwan Province, recorded over a period of five years, were selected to evaluate the performance of the proposed DLRN model. The root mean square error (RMSE), mean absolute error (MAE), and correlation coefficients for the DLRN model are 1.25 mm, 0.37 mm, and 0.75, respectively. Compared to existing quantitative rainfall forecast studies, the DLRN model achieved a 34% improvement in RMSE, and a 43% improvement in MAE. In addition, a wavelet coherence analysis indicated that most predictors contribute to rainfall occurrence. Feature importance experiments illustrated that CAPE and PWV are the top two contributing factors to rainfall nowcasting. Experimental results confirmed satisfactory performance of the proposed DLRN model, and that it possesses the capability of deployment in practical scenarios for high-precision rainfall nowcasting.