基于GNSS数据的MODIS近红外水汽评估与无偏优化

MODIS Near-infrared Water Vapor Assessment and Unbiased Optimization Based on GNSS Data

针对MODIS近红外水汽产品精度不足的问题,提出了MODIS近红外水汽无偏优化方法。首先,在研究区域基于SuomiNet提供的GNSS PWV对MODIS近红外PWV进行了验证;然后,引入一种人工神经网络BPNN模型,构建MODIS PWV无偏优化模型,并对该模型在空间域和时间域进行全面验证;最后,将优化模型应用于MODIS近红外PWV影像对其进行校准。结果表明:原始MODIS近红外PWV产品精度为4.07 mm;BPNN优化后的MODIS近红外PWV精度为1.41 mm,精度改善率为65.36%。该模型可有效改善MODIS近红外PWV产品精度,对其在短临天气预报和对地观测误差改正的应用具有价值。

To solve the problem of insufficient accuracy of MODIS near infrared water vapor products,an unbiased optimization method for MODIS near infrared water vapor is proposed in this study.Firstly,MODIS NIR PWV with GNSS PWV from SuomiNet is validated in the study area.Then,an unbiased optimization model for MODIS PWV is constructed using an artificial neural network,BPNN model,which is fully validated in the spatial and temporal domains.Finally,the optimization model is used to calibrate the MODIS NIR PWV image.The results show that the original MODIS NIR PWV product has an accuracy of 4.07 mm,while the accuracy of MODIS NIR PWV after BPNN optimization is 1.41 mm,resulting in a 65.36%accuracy improvement rate.The model presented in this study effectively improves the accuracy of MODIS NIR PWV products,making it valuable for short-range weather forecasting and Earth observation error correction applications.