Based on conventional ground observational data,forecast products from European Center (EC),Ameri-ca GFS and T639 numerical models,the maximum air temperature prediction errors were analyzed.The result shows that the temperature was obviously overestimated during the Spring Festival from January 30 to February 2, 2014 in Huaibei of Jiangsu province.It has been found that during the Spring Festival of 2014,the strong warm temperature advection at middle and lower troposphere is favorable to increase temperature greatly.However,abun-dant low-level cloud,dense mist and serious pollution weaken the incoming solar radiation.The surface with low air temperature absorbs the heat from the air,thus counteracting the warming effects of warm temperature advection and inhibiting temperature from rising over Huaibei of Jiangsu province,even decreasing the local temperature. Temperature inversion layer tends to be presented vertically.Meanwhile,the thermal inversion layer is favorable to maintenance of stable stratification at boundary layer and weak wind condition at the ground layer,which are not favorable to diffusion of pollution and water vapor.As a result,the fog and haze are enhanced.All these form a cooling feedback.Low cloud,fog and haze as well as temperature inversion layer have great impact on variation of daily air temperature.Main reasons for temperature prediction error are as follows:it has difference between ob-served and simulated circulation by a numerical weather prediction model;relative humidity near surface layer is underestimated;forecast total cloud cover is less than the observed.Cooling effect of fog and haze is not predicted. Strong warm temperature advection raises air temperature at 850 hPa.However,when surface air temperature is low,the surface absorbs heat from the air,thus leading to weak warming of surface air.The relationships of air temperature at 850 hPa and surface air temperature is weakened.Therefore,the warming of strong warm tempera-ture advection is overestimated.Forecast for 2 m air temperature is excessively depended on daily forecast products from EC and GFS models and so on.Circulation and parameters of the numerical forecast products has not been tested.%利用常规地面气象观测资料及欧洲 EC、美国 GFS 和 T639数值预报产品分析了2014年1月30日至2月2日（春节期间）江苏淮北地区日最高气温预报明显高于实况的可能原因。结果表明：2014年江苏淮北地区春节期间对流层中低层强暖平流有利于大幅升温，期间低云较多、雾较浓且空气污染较重，减弱了到达地面的太阳辐射，地面气温较低，吸收大气热量，对暖平流的升温有明显的抵消作用，不利于淮北地区的大幅升温，甚至造成局地降温，垂直方向的温度层结上易出现逆温。当逆温层维持时，层结较稳定，地面风力较小，不利于空气中污染物和水汽的扩散，雾霾加重，形成一个降温正反馈机制。低云和雾霾及逆温层对地面气温的变化有重要影响。造成此次最高气温预报失误的主要原因为，模式预报的形势场与实况存在较大差异，未充分考虑近地层的相对湿度条件，预报的云量少于实况；对白天雾霾的降温效应估算过低；强暖平流增强850 hPa 气温时，当地面气温较低时，地面吸收大气热量，升温不明显，850 hPa 温度与地面气温变化的对应关系减弱，过高估算了强暖平流的升温作用；对 EC 和 GFS 等模式2 m 气温数值预报产品过度依赖，未对形势和要素的数值模式预报结果进行检验。
展开▼
