An Assimilating Model Using Broad Learning System for Incorporating Multi‐Source Precipitation Data With Environmental Factors Over Southeast China

Abstract

Accurate estimation of precipitation is of critical importance for hydrometeorological applications and hazard prevention (Gao et al., 2019;Stephens & Kummerow, 2007). In general, precipitation can be derived from ground-based (e.g., rain gauges and radars) observations and satellite-based products, as well as atmospheric reanalysis products (e.g., ERA5;Tarek et al., 2020). Rain gauges and weather radars could measure precipitation with comparatively high credibility. However, rain gauges and radars are often distributed unevenly and sparsely, especially in mountainous areas (Ma et al., 2015), which leads to a deficiency of precipitation data in some regions (Gottschalck et al., 2005). Moreover, in serious natural environments, some gauges may be out of operation resulting in discontinuous rainfall measurements. Errors in the weather radar measurements can hardly be avoided, and the accuracy is influenced by weather conditions, precipitation process, and the spatial distribution of precipitation particles (Joss et al., 1998;Sokol et al., 2021). Remote sensing techniques have been applied to detect precipitation data with near-global coverage and decadal records (Ma et al., 2018;Sunilkumar et al., 2016), which could overcome the spatial and temporal restrictions to some extent. Vast satellite-based precipitation products have been released recently, such as the latest Version-7 Tropical Rainfall Measuring Mission (TRMM) Multi-satellite Precipitation Analysis (TMPA) near-real-time (3B42RT) & post-real-time (3B42V7) products (Huffman et al., 2010), precipitation products using Climate Prediction Center morphing (CMORPH) method

The evaluation and downscaling‐calibration of IMERG precipitation products at sub‐daily scales over a metropolitan region

“…It is less accurate to drive hydrological and meteorological models at a local or urban basin using IMERG. However, some studies have shown that the spatial distribution of extreme storms can be captured well by IMERG (Tang et al, 2020; Zhou et al, 2022). Benefitting from its finer resolution than other satellite‐based products, the IMERG data are still a promising data source for flood risk analysis.…”

Fast Initial Model Design for Electrical Resistivity Inversion by Using Broad Learning Framework

“…The training time for the deep learning network on the same training set may be hundreds of times longer than that of the BL network [44]. Recently, The BL network has been successfully applied in geophysical inversions, achieving noteworthy results [46][47][48][49]. These applications demonstrate that BL possesses the advantages of excellent efficiency and robust mapping capability.…”