Land use affects regional weather and climate. The use of land-use data that is updated and closer to the observation can provide strong support for numerical prediction. The high-temperature processes in August 2019 in Sichuan Basin are simulated by using the mesoscale numerical model WRF. Updated land-use datasets extracted from MODIS data are compared with original WRF datasets, and the influences of updated land-use data on WRF simulation are analyzed. The results show that there are significant differences between two sets of land use data, the updated data is more detailed, and the land use types of the simulated area are more abundant. The 2 m temperature is very sensitive to land use data, with update of land-use data, the accuracy ratio (absolute errors less than 2 ℃) of simulated 2 m temperature increased by 6.2%, and that of daily maximum temperature increased by 31.3%; the accuracy of simulated daily minimum temperature decreased by 2.1 %. Simulated temperature increases significantly in the Sichuan Basin where the land-use changed obviously, the increase in some areas is more than 4 ℃, and the negative bias in simulated daily maximum temperature by using original land-use datasets significantly reduces. The Hechuan station is selected as a typical station in the Sichuan Basin for analysis, where land-use type is changed from croplands to urban and built-up. It is found that after the change of land use type, the emissivity decreases, the roughness increases, and the stomata resistance of vegetation increases, and as a result, the upward sensible heat flux increases, surface evaporation and atmospheric water supply reduces, 2 m temperature increases, and planet boundary layer height increases. Updated land-use data results in strong locality of the simulation difference in the initial stage, as the lead time increases it gradually affects the upper atmosphere and surrounding areas. Land-use data lead to differences of simulations by affecting the selection of land-surface parameters. The accuracy of simulated 2 m temperature and maximum temperature are improved with update of land-use data.

 Based on ECMWF ERA-Interim reanalysis data, 8 factors (T2 m, T1000, T975, T950, T925, T850, H700-850 and H850-1000) for identifying precipitation phases were obtained through analyzing the temperature and geopotential thickness of precipitation phases (rain, snow, sleet) in winter half year from 2008 to 2017 in Shandong Province, and the threshold indicators of the 8 factors were provided. The discriminant equation for precipitation phase identification was established and the deep learning DNN model was trained using the 8 factors and their threshold values, and the forecast accuracy of rain, snow and sleet increased by 1.9%, 0.2% and 21.6% using DNN method through randomization test, respectively. The inspection using ECMWF fine grid model products indicated that among a total of 106 stations of rain, snow and sleet, the discriminant equation and DNN method carried out wrong identifications for 29 and 14 stations, respectively. The results show that the DNN method performed better than the discriminant equation, and in particular, it significantly improved the identification ability of sleet.