Evaluation and quantile mapping-based correction of TIGGE multi-model precipitation forecasts in mountainous watersheds
Year: 2026
Type: article
Abstract: Accurate precipitation forecasts are essential for characterizing and predicting extreme rainfall events, particularly in mountainous watersheds where precipitation exhibits strong spatial and temporal variability. This study evaluated 6 h accumulated precipitation forecasts from three TIGGE (THORPEX Interactive Grand Global Ensemble) systems, CMA, ECMWF, and NCEP, during the April-October seasons of 2017–2022 over the Jinbao River Basin, Guangxi, China. Forecast performance was assessed using categorical and continuous verification metrics. The ECMWF ensemble maximum precipitation forecasts showed the highest skill for heavy rainfall and rainstorm events but exhibited systematic overestimation. To mitigate this bias, a segmented quantile mapping (QM) method was applied to post-process the ECMWF ensemble maximum forecasts. After correction, overprediction and false alarms were substantially reduced, while the TS (Threat Score) for rainstorm events remained stable at approximately 0.2 for lead times of 1–4 days. These results indicate that QM improves the reliability of deterministic extreme precipitation forecasts from the ECMWF ensemble maximum. Because only six years of observations were available, however, the study cannot fully characterize precipitation return periods of 50 years or longer; results beyond this level should be interpreted with caution and evaluated against longer observational records.
Source: Journal of Hydroinformatics
Authors: Fanshui Zeng, Meng Yuan, Zuxiang Situ, Dantong Zhu
Institution: South China University of Technology
Language: English
Subfield: Atmospheric Science
Field: Earth and Planetary Sciences
Domain: Physical Sciences
SDG: Climate action
Open Access status: diamond
OpenAlex ID: W7236704500
