Abstract:Convective temperature (Tc) can be utilized to estimate the likelihood of local thermal convection and convective cloud (CC). However, its application in operational cloud forecasting is restricted to some extent due to the limited sounding time and certain preconditions. Addressing this problem, we propose a numerical computation scheme of Tc, as well as an idea for CC forecast on the basis of model-forecasted sounding. Firstly, Tc at each forecasting time can be calculated by using dew point temperature at 2 m height, surface pressure, and temperature on pressure levels from numerical weather prediction (NWP) model. Then, an index of thermal convection (Icv), which is defined as the difference between temperature at 2 m height (T2 m) and Tc, can be easily got as T2 m is usually provided by NWP operational models. If Icv meets certain threshold value, CC will be predicted to occur. In this study, this idea is successfully used to explain the false negative prediction of CC in Shandong on 27 April 2020. Besides, Icv has been quasi-operationally applied in the forecast of CC since May 2020, and it performs quite well in predicting both the thermal cumulus on the land and the wintertime cold airflow-induced low-level cloud over the sea. In order to forecast the occurrence of thermal convective cloud, we suggest that emphasis be paid on the condition analysis of thermal convection, rather than that of dynamical lifting or water vapor. In addition, the influences of a couple of special atmospheric profile scenarios, both of which are characterized by the existence of temperature inversion, on the computing of Tc are discussed, and the corresponding solutions are given.