基于CALMET与LSTM的风场预报降尺度研究

中图分类号：TP183 文献标志码：A 文章编号：2095-2945（2025）28-0095-04

Abstract：Thispaperconductsa5kmCALMETdownscale technologyresearchbasedontheECfinegrid forecastfieldfrom 2017to2019，providinganewobjectivebackgroundfieldforintellgentgridstylefieldforecasting.TheWRF_CALMET downscale forecasting system was completed， and the 10m wind speed of the WRF CALMET downscale forecasting product was - inspected and evaluated.The downscale forecasting system had a good forecasting effect on the 10m wind speed; the 10m wind speed test results of the WRF_CALMET downscale forecasting and the 10m wind speed test results of the EC-thin mode were found to be that the 10m wind speed root mean square error of the WRF_CALMET downscale forecasting was significantly lower than thatof the EC-thin mode，that is，the downscale forecasting effect was better than the EC-thin mode.

Keywords：wind field; forecast; downscaling; long-term memory neural network；WRF_CALMET

开展精细化预报[1-2是天气预报发展的一个重要性战略目标，随着社会科学的进步和经济社会的发展，专业预报用户及公众对天气预报的准确率要求越来越高，制作“定时、定点、定量”的客观精细化预报已成为天气预报发展的必然。（剩余6170字）