Abstract：The article selected surface observational data from October 2019 to September 2020 in the Everest Nature Reserve to test and analyze precipitation products from three numerical forecasting models:ECMWF,GRAPES-GFS,and GRAPES-MESO.The results showed that the three models had significantly better predictive abilities for small-scale precipitation in the Mount Everest Nature Reserve than other levels,with ECMWF and GRAPES-MESO having better predictive abilities than GRAPES-GFS.For different start times,the numerical model forecasting ability from 20:00 is higher than that from 08:00,especially for larger precipitation forecasts.The TS scores of the three numerical models in the winter half year from January to April are much higher than those from October to December.For severe weather with blizzards of more than 10 mm,the ECMWF model has the highest score and the best forecasting ability,especially for precipitation of ≥50 mm,which has a much higher score than GRAPES.MESO.Overall,ECMWF precipitation products have high forecasting ability and reference value for catastrophic weather in the Mount Everest Nature Reserve.
2026, Vol. 43 
