Publications

Liu, X; Wang, Y; Zhan, W; Yu, TL (2023). Improving MODIS Precipitable water vapour in mainland China based on the LSF model. ADVANCES IN SPACE RESEARCH, 72(8), 3133-3149.

Abstract
Precipitable Water Vapour (PWV) measurements derived from the Moderate-resolution Imaging Spectroradiometer (MODIS) exhibit considerable promise for global high-resolution PWV monitoring but with limited accuracy. This study proposes a correction model for MODIS PWV, leveraging the least squares method (LS) and fast Fourier transform (FFT). The study used Global Navigation Satellite System (GNSS) PWV products (the accuracy is equivalent to Radiosonde-derived PWV) from 230 stations provided by the Crustal Movement Observation Network of China (CMONOC) for 2016 to 2019, serving as the ground truth for correcting MODIS PWV. Subsequently, the model undergoes an external coincidence test utilising the data from 2020. Initially, a linear or quadratic expression (LS model) was established through the application of the least squares method, effectively capturing the trends between the two PWV datasets. Then the periodic characteristics of residuals from the LS model were analysed by FFT. The outcomes reveal that within regions and seasons abundant in PWV, the quadratic expression is more consistent with the trend. The primary variation period of residuals exhibits variability across regions, with the annual period emerging as the primary fluctuation pattern in the majority of regions. Following this, we constructed the correction model (LSF model) that considered the cycle characteristics of residuals with the model type adopted for the LS model as the main part. After correction by the LSF model, the root-mean-square error (RMSE) of MODIS PWV decreased by 33.1%, 55.8%, 33.0%, and 19.4% in four seasons, respectively. Compared to the LS model, the LSF model showcased relative enhancements of the coefficient of determinations (R2) by 119.8%, 99.6%, 165.9%, and 39.6% in the four seasons, respectively, with an overall further decrease of 35.7% in RMSE. The LSF model demonstrates a superior degree of fitting and yields improved correction (c) 2023 COSPAR. Published by Elsevier B.V. All rights reserved.

DOI:
10.1016/j.asr.2023.06.041

ISSN:
1879-1948