Forecast Assessment of CMA-GFS on the Meteorological Elements in the Northeast Hemisphere

Abstract

Based on the daily forecast data initialized at 20:00 （Beijing time） of the CMA global forecast system （CMA-GFS） during the period from October， 2019 to December， 2020， the forecasting quality of the upper air meteorological elements in the Northeast Hemisphere are verified.The verified meteorological elements include geopotential height and wind fields at 500 hPa， temperature field at 850 hPa and relative humidity field at 700 hPa.The results show that： （1） From the characteristics of time variability， there were obviously seasonal variation of the forecasting performance.The similarity between the forecast field and analysis field of 500 hPa geopotential height and 850 hPa temperature is the lowest in summer， and the similarity of 700 hPa relative humidity field is the lowest in winter.With the increase of forecasting time， the root mean square error of each field is the smallest in summer.（2） From the characteristics of spatial variability， 500 hPa geopotential height has high similarity between the forecast field and analysis field in most areas of the Northeast Hemisphere， and the similarity of mid-high latitudes are more similar than the low latitudes.The root mean square error of geopotential height is relatively small in the low latitudes and large in the high latitudes， and the mean error is mainly a large-scale negative deviation in the Northeast Hemisphere.In contrast， the model has higher forecasting skill in East Asia for the geopotential height and its gradient.The distribution of forecasting similarity and root mean square error of 850 hPa temperature field are like that of 500 hPa geopotential height， and the mean error is lower in the mid-high latitudes and higher in the low latitudes.In former forecast period （for the 24~48 hours）， the similarity of temperature field is higher and the root mean square error is basically within 2 ℃ in the north of 20°N， which means the forecast effect here is preferable.The distribution of forecasting similarity and root mean square error of 500 hPa zonal wind and meridional wind fields also increase with increasing latitude.However， for zonal wind， there is an obvious high value belt of forecasting skill between 20°N -30°N， and for meridional wind， the forecast effect is poor in the south of 20°N.For 700 hPa relative humidity field， the model has a certain forecast effect for the 24 hours.But with the increase of forecasting time， the forecast of humidity field only has some reference significance in North Africa to South Asia.

Keywords

• cma-gfs
• verification
• anomaly correlation coefficient
• root mean square error