搜索

x

基于时空独立的随机森林模型对海南热带气温数值预报的订正

Tropical Temperature Correction for Numerical Forecast in Hainan Based on Spatiotemporal Independence Random Forest Model

  • 摘要: 面向海南省所特有的海岛以及热带特点,结合海南岛独特的地理地貌,本研究设计了基于时空独立的随机森林模型,并利用站点的实测数据以及欧洲中期天气预报中心(ECMWF)的模式数据,实现了对每个站点未来7天预报时效为3小时的气温精准订正;同时采用小于2℃的准确率、小于1℃的准确率及均方根误差等指标,对ECMWF模式的预报温度和本文模型的订正气温进行了评估,结果表明,本文所提的订正模型结果要显著优于ECMWF模式的结果,前者更接近真实的温度值,它对ECMWF的模式数据进行了较好的订正.

     

    Abstract: In our report, aimed at the island and tropical characteristics and the unique geographical features of Hainan Island, a random forest model based on spatiotemporal independence was designed, the observed data and the forecast data from ECMWF were used for the accurate correction for each station of 7 days forecast, which has 3 h precision; the proposed method was evaluated by the index of <2 ℃ accuracy, <1 ℃ accuracy and RMSE. The result showed that the proposed method is obviously better than the model of ECMWF, which is much closer to the observed temperature value, and which also revised the data form ECMWF very well.

     

/

返回文章
返回