测绘科学技术学报2025，Vol.41Issue(6)：557-565,582,10.DOI:10.3969/j.issn.1673-6338.2025.06.002
CYGNSS土壤湿度时空特征增强的多尺度SPEI预测模型
Multi-scale SPEI Prediction Model Enhanced by Spatiotemporal Characteristics of CYGNSS Soil Moisture
摘要
Abstract
Drought and flood disasters pose significant threats to regional agricultural and water security.Soil mois-ture(SM),a crucial indicator of drought and flood conditions,plays a vital role in accurately predicting disasters through analysis of its spatiotemporal characteristics and dynamic monitoring.This study develops a multi-scale standardized precipitation evapotranspiration index(SPEI)prediction model,which integrates seasonal autoregres-sive integrated moving average with eXogenous regressors-long short-term memory extended(SARIMAX-LSTME)model.The model is based on cyclone global navigation satellite system(CYGNSS)SM remote sensing data and ground meteorological observations.Using Guangdong Province as a case study,the fast fourier transform-wavelet transform(FFT-WT)method is applied to analyze the periodic oscillation characteristics of CYGNSS SM during normal,drought,and flood conditions,and their correlation with the SPEI.Performance evaluation experiments of the SARIMAX-LSTME model are conducted at typical sites.The results indicate that the SM oscillation characteris-tics exhibit spatial heterogeneity under different climatic conditions.Compared to random forest(RF)and Trans-former models,the SARIMAX-LSTME model demonstrates superior performance in recognizing the spatial hetero-geneity of CYGNSS SM across various climatic regions and enhances the prediction accuracy of SPEI.The R2 val-ues for the 3-month,6-month,and 12-month prediction windows are 0.921,0.898,and 0.886,respectively.关键词
旋风全球导航卫星系统/旱涝灾害/土壤湿度/标准化降水蒸散指数/SARIMAX模型/LSTME网络Key words
cyclone global navigation satellite system/drought and flood disasters/soil moisture/standardized precipitation evapotranspiration index/SARIMAX/LSTME分类
天文与地球科学引用本文复制引用
王尔申,顾轶楠,于腾丽,徐嵩,曲萍萍,娜拉..CYGNSS土壤湿度时空特征增强的多尺度SPEI预测模型[J].测绘科学技术学报,2025,41(6):557-565,582,10.基金项目
国家自然科学基金项目(62173237) （62173237）
智慧中原地理信息技术河南省协同创新中心联合基金项目(232203). （232203）
