Research on rainfall and temperature forecasting and correction technology in mountainous microterrain

This study aims to improve the accuracy of weather forecasting, particularly rainfall and temperature predictions, in mountainous micro-terrain conditions. A combined approach using the WRF model and LightGBM algorithm is proposed to enhance the precision of meteorological forecasts. Initially, weather forecasts are generated using the WRF model, and the LightGBM algorithm is applied to correct temperature, precipitation, and other meteorological variables. The study analyzes weather data from December 2, 2023, to March 2, 2024, to construct training and testing datasets for weather stations in mountainous regions. The results show that the WRF-LightGBM correction method significantly improves the accuracy of precipitation and temperature forecasts, with rainfall forecast accuracy reaching 90% and temperature forecast accuracy exceeding 70%. Specifically, corrected precipitation and temperature forecasts exhibit minimal error, especially in regions such as the Yunnan-Guizhou Plateau and southwestern Guangdong. Through comparative analysis, this study demonstrates the practical application value of the proposed method, providing a new technical approach for improving weather forecasting accuracy in mountainous areas.