A One-Dimensional Variational Precipitation Retrieval Algorithm Considering Cloud Types for Western North Pacific Tropical Cyclones Using FengYun-3E Microwave Sounders

Abstract

Accurate precipitation retrieval from passive microwave observations is essential for global meteorological and climatological studies particularly in tropical cyclone (TC) environments. Building upon the Global Scene-Dependent Atmospheric Retrieval Testbed (GSDART), this study develops an enhanced one-dimensional variational (1DVAR) precipitation retrieval algorithm specifically tailored for monitoring TCs based on Chinese FengYun-3E (FY-3E) passive microwave observations over the northwestern Pacific. The algorithm introduces a precipitation-type differentiation module into the 1DVAR framework (1DVARDPT), employs the Advanced Radiative Transfer Modeling System (ARMS) as the forward operator, and utilizes observations from the FY-3E microwave humidity and temperature sounders (MWTHS). A comprehensive comparative analysis from 19 TCs reveals that considering precipitation types significantly improves retrieval performances—reducing relative bias from −9.89% to 2.02% and mean absolute error (MAE) from 0.38 mm/hr to 0.32 mm/hr—while enhancing the detection of both light/moderate and heavy precipitation. Furthermore, using ARMS instead of the Community Radiative Transfer Model (CRTM) as the forward operator markedly reduces the systematic underestimation seen in conventional retrievals (bias improved from −23.50% to −9.89%), demonstrating ARMS's superiority in precipitation-sensitive radiative transfer modeling. The findings from this study underscore the importance of accounting for precipitation-type variability in 1DVAR retrievals and highlight the strong potential of FY-3E observations and ARMS for advancing microwave-based precipitation estimation.