20 October 2022 in Earth and Space Science
Eliminating the systematic errors of member models is a key step before the variant‐weight multi‐model ensemble forecasting. However, how to reasonably calculate model systematic errors is a problem worthy study. Taking surface temperature as the subject investigated, this study explores this problem through comparative analyses of a series of variant‐weight multi‐model ensemble forecasting experiments. The results showed that eliminating the systematic errors of member models dramatically improves the ensemble forecasting outcomes. The calculation of model systematic errors based on the optimal time lengths regarding the spatiotemporal distribution characteristics of model forecasting errors is a reasonable and effective approach. The surface temperature forecasts with model systematic errors calculated with the optimal time lengths and those beyond the optimal time lengths exhibit higher accuracies. Owing to the approximately uniform distribution of temperature, no significant difference happens between the accuracies of the surface temperature forecasts with model systematic errors calculated using the average and quantile methods. In addition, the updating of the models participating into the ensemble forecasting would weaken the representativeness of the optimal time lengths for model systematic errors calculation that are obtained based on historical period data, and thus, weaken the forecasting skill. The research results provide some train of thought for enhancing the variant‐weight multi‐model ensemble forecasting accuracy.
5 August 2026 in SSRN Electronic Journal
Rainstorms occur frequently with pronounced forecast biases over the complex-terrain region of southwest China—especially the key area on the steep eastern slope of the Tibetan Plateau (TP)—often causing severe losses and impacts. Based on hourly multi-source merged precipitation analysis observations from June to September 2019-2024, this study systematically evaluates the forecast performance of operational high-resolution global and regional models over this key precipitation area, using the Threat Score (TS)、 BIAS score、decomposed intensity and frequency metrics, and analyzes the bias patterns separately for frontal versus warm-sector rainstorms. The results show that the TS for ≥10 mm precipitation is below 0.1 for all models: ECMWF exhibits systematic underestimation, whereas the regional models tend to overforecast. The diurnal precipitation peak over the key area is generally too early, with the magnitude of the phase bias increasing with elevation—ECMWF advances the peak by about 3-4 h, while CMA-3KM advances it by 3 h on the plateau side of the eastern slope, but controls the bias within 1 h at the low-elevation center near Ya'an, thus better capturing the nocturnal rainfall peak. Spatially, the ECMWF rainfall area is shifted southward and is too weak; CMA-SH9 severely overforecasts at the steep peaks; CMA-3KM provides the closest spatial distribution to the observations, yet exhibits a bias of excessive intensity and insufficient frequency. For the two typical rainstorm types, ECMWF shows almost no discrimination between frontal and warm-sector events, indicating a single physical response; CMA-3KM captures their spatial differences, but under the weak forcing of warm-sector conditions it is prone to excessive convection triggering, resulting in overly strong precipitation intensity. The comprehensive evaluation indicates that CMA-3KM is relatively the best; however, all models exhibit systematic biases arising from physical parameterizations such as terrain resolution and convection triggering. The findings suggest that elevation- and event-type-specific bias correction, together with multi-model adaptive ensemble forecasting, represents a promising pathway toward improving fine-scale precipitation forecasts over the eastern slope of the Tibetan Plateau.