An Investigation Of The Forecast Bust In GRAPES_GFS Medium-range Forecast In East Asia

The GRAPES＿GFS model is a global medium-range weather system self-developed by the China Meteorological Administration,version 2.0 of which was put into service in 2016.The quality of the medium-range weather forecast is constantly improving,mainly through the improvement of the physical processes,the 4D-variational assimilation system,the application of satellite data assimilation,and the dynamic framework,individual predictions still have very large errors.we focus on East Asia and define a “forecast bust” to be an occasion when the day-6 high-resolution forecast of East Asia Z500 has an ACC(Anomaly Correlation Coefficient,ACC)less than 0.4 and an RMSE greater than 60 gpm(Root Mean Square Error,RMSE).In this paper,the source and development of model error are studied from two aspects of average characteristics and case analysis,and the internal mechanism and influencing factors are further studied to provide scientific basis for improving the results of GRAPES＿GFS model.The main results are as follows:(1)In this paper,the business forecast data of GRAPES＿GFS model from April 1,2016 to December 31,2018 were used for inspection and evaluation.It was found that the ACC of GRAPES＿GFS model in East Asia decreased with the increase of lead time,while the RMSE increased with the increase of lead time,and the growth rate was significantly lower than the linear growth rate.The forecast performance of the height field is better than that of the temperature field and the full wind speed field.The forecast performance of the 0-3 wave is better than that of the 10-20 wave.The RMSE large value area of the height field was concentrated in the middle and upper layer,the RMSE large value area of the temperature field was concentrated in the near surface at 250 h Pa,and the RMSE large value area of the full wind speed field was mainly concentrated in the 300-250 h Pa jet stream area.Furthermore,this paper studied the global error distribution characteristics of the GRAPES＿GFS model through diagnostic analysis,and found that the forecast errors of the GRAPES＿GFS model presented the difference characteristics of sea and land in the horizontal distribution and the symmetry characteristics of the northern and southern hemispheres in the vertical distribution,and the characteristics were significantly deepened with the increase of the lead time.(2)The medium-range forecast busts usually occur in spring and autumn,when the circulation situation changes.The frequency of the busts in the ECMWF model is relatively low.The frequency of this event in the GRAPES＿GFS and NCEP models is relatively high,and the time points are close to or even same.The forevast busts of GRAPES＿GFS model’s medium-range forecast in East Asia is usually manifested in the failure of the forecast of the East Asian trough,the Siberian trough and the subtropical high.The initial conditions are an important factor affecting the busts.When the forecasting skills for short-range weather forecast are poor,the forecasting skills for medium-range weather forecast are also relatively poor.The negative deviation of geopotential height in the initial conditions tends to deepen baroclinic systems,such as weather-scale troughs.The upward movement in front of the trough and the positive deviation of CAPE are conducive to the occurrence and development of small and medium-scale weather systems in the area.At the same time,convective activities It will support further development by changing its circulation situation.The interaction between changes in the large-scale circulation situation and meso-and small-scale disturbances is the source of extreme decline events in medium-term forecasting techniques.(3)The sources of error for forecast busts in the medium-range forecast in East Asia are more complicated,mainly from the polar regions north of East Asia,the Atlantic Ocean to the west,and the Pacific Ocean to the southeast.In winter,the initial error of the polar region is transferred to the downstream east Asia with the strong jet stream.In spring,the strong baroclinic development process of the cyclone on the Atlantic Ocean and the downstream effect of Rossby waves are important factors affecting the medium-term weather forecast in East Asia;In late summer and early autumn,the intensity and position changes of typhoons in the western Pacific Ocean will have a close influence on the adjustment of the circulation situation in East Asia,especially the influence of strong typhoons or even double typhoons.