Abstract:In this paper,the inconsistency index distribution and variation characteristics of the ECMWF surface air temperature forecasts over China were studied based on the surface 2 m air temperature forecast data of the ECMWF operational forecast model,during the period of December 1,2015 to November 30,2016,along with the calculation of the Jumpiness index of the forecast.The results revealed that the annual mean forecast inconsistency index of the surface air temperature forecasts over China decreased from north to south,with two areas of maximum values over Tibetan Plateau and eastern Inner Mongolia,respectively,thus indicating that the forecast jumpiness over these two areas is quite high.Among the four seasons of a year,the forecast inconsistency index of the surface air temperature is the greatest during the boreal winter,and smallest in the boreal summer,while there is little difference between the forecast inconsistency indexes in spring and autumn,the inconsistency indexes thereof lying between those in winter and summer.The inconsistency index of the surface air temperature forecast on a given day during boreal summer is slightly different from the summer average value of the inconsistency index,while that on a day during the boreal winter is significantly different from the winter average value of the inconsistency index.