Research On The Revision Of Precipitation Probabilistic Forecast Considering Topographic Effects

This study is based on the daily 24–168h ensemble precipitation forecast datasets from September 2014 to September 2015,which are derived from the European Centre for MediumRange Weather Forecasts extracted from the TIGGE(The Interactive Grand Global Ensemble)dataset.Considering the influence of topographical factors,the precipitation probability forecast in China is revised.First,the error of the ECMWF ensemble forecast under the influence of different topographic factors is analyzed via different evaluation methods.Then considering the influence of topography,seasonal changes,location,and other factors,the GAMLSS model is used to simulate the spatial distribution of the daily precipitation probability parameters in China.The results indicate that available spatial probability of the daily precipitation climatic distribution parametric model is obtained.Finally,based on the CNLR method,the correction effects of the numerical forecast results are compared.Numerical weather forecasting model is unable to accurately and quantitatively describe the dynamical and thermal effects of terrain on the atmosphere due to its own design reasons and the complex interaction between atmospheric motion and terrain,resulting in model forecast errors.The study found that the forecast errors of ECMWF ensemble forecasts tend to increase as the complexity of the terrain increases.The degree of terrain complexity and the error between the model and the terrain mainly affect the precipitation forecast error of the light rainfall level.It shows that as the terrain complexity and the model terrain height error increase,the model's forecast error for light rain increases.Considering the influence of various factors on the probability distribution of daily precipitation,the left-censored Logistic distribution of the daily precipitation probability distribution parameters of the GAMLSS model are used.The results indicate that the spatial model of the daily precipitation climatic probability distribution can well reflect the actual precipitation distribution characteristics in the area,and can be well applied to the simulation of the climatic probability distribution of precipitation in the whole space under complex terrain.Using the left-censored non-homogeneous logistic regression method(CNLR)and standardized model post-processing method(SAMOS)to calibrate the precipitation forecasts in the southeast of China.The results show that the CNLR method can effectively improve the mean absolute error(MAE)and continuous ranked probability score(CRPS)of the raw ensemble forecast,and improve the forecasting skills of quantitative and probabilistic precipitation forecasts.Using SAMOS method to preprocess the data and considering the impact of topography and other factors,the forecast error caused by the terrain influence can be further revised on the basis of CNLR method,and a more accurate probabilistic forecast of precipitation is thus obtained.