2026-10-11T14:18:43Z https://www.scidb.cn/oai

10.57760/sciencedb.27886 2025-09-09T10:11:08Z

2025-09-09 Source Code for: An Image Recognition Framework for Detecting Large-Scale Persistent Extreme Precipitation Events (LPEPEs) doi:10.57760/sciencedb.27886 en This dataset contains the source code for the Image Recognition Algorithm (IRA) developed in the manuscript "An Image Recognition Framework for Detecting Large-Scale Persistent Extreme Precipitation Events (LPEPEs) and Its Revelation of Southward Displacement and Dynamic Mechanisms".The algorithm utilizes a connected-component labeling technique from image processing to achieve rapid and accurate automatic detection of Large-scale Persistent Extreme Precipitation Events (LPEPEs). Compared to the Traditional Iterative Algorithm (TIA), the IRA improves computational efficiency by two orders of magnitude (~100x) when processing high-resolution gridded precipitation data while maintaining identical detection accuracy.The code is implemented in Python, with its core functionality leveraging the scipy.ndimage.label library. This repository includes the core algorithm scripts, sample data, and comprehensive documentation. Researchers can use this code to efficiently process various reanalysis and satellite precipitation products (e.g., CMA DPDC, ERA5, GPM IMERG) for the detection, analysis, and study of extreme weather and climate events.Key finding of the associated paper: Employing this algorithm, we provide the first unified evidence across three independent high-resolution precipitation datasets revealing a pronounced southward displacement preference of non-typhoon LPEPEs over China during summer (especially the Meiyu period), and investigate the underlying dynamic-thermodynamic mechanisms. Extreme Precipitation Events; Image Recognition; Connected-Component Labeling; Computer Vision; Climatology; Precipitation; Algorithm; Python Jiao Ma PUBLIC https://api.github.com/licenses/mit dataset http://www.doi.org/10.1007/s00382-022-06539-z Science Data Bank