TITLE:
Evaluation of Simulated Summer Weather Patterns over the Yangtze River Basin Using Multiple Climate Models
AUTHORS:
Shengnan Wu, Pei Wu
KEYWORDS:
CMIP6, Self-Organizing Map (SOM), Multi-Model Ensemble, Synoptic Weather, Precipitation Bias
JOURNAL NAME:
Journal of Geoscience and Environment Protection,
Vol.14 No.8,
August
31,
2026
ABSTRACT: Based on daily ERA5 reanalysis data from the European Centre for Medium-Range Weather Forecasts covering 2005-2014, this study adopts the Self-Organizing Map (SOM) method to evaluate and rank sea level pressure fields under historical forcing scenarios from 16 climate models of the Coupled Model Intercomparison Project Phase 6 (CMIP6). Optimal models are selected for multi-model ensemble, and the precipitation discrepancies between ensemble simulations and observations are further analyzed. The main conclusions are as follows: 1) Most CMIP6 climate models can reasonably reproduce the observed frequency characteristics of summer weather patterns, yet their simulation performance varies significantly across individual months. The top five models in the comprehensive ranking are MPI-ESM1-2-HR, IPSL-CM6A-MR1, IPSL-CM6A-LR, MPI-ESM1-2-HAM and FGOALS-f3-L. Multi-model ensemble effectively improves the overall simulation capacity for summer weather patterns, but more ensemble members do not guarantee better performance. An ensemble of the top three best-performing models (denoted MME3) can greatly boost model simulation skill. 2) Three ensemble schemes (MME3, MME5 and MME16) improve consistency with observations by reducing biases in simulating the frequency of noisy weather patterns, among which MME3 presents the smallest simulation errors. In terms of precipitation biases, MME3 can basically reproduce the spatial distribution of observed precipitation, with precipitation deviations ranging from 0 to 2 mm per day. Model performance varies widely across different weather patterns; overall, multi-model ensemble greatly optimizes the simulation of heavy rainfall magnitudes.