区域历史气压数据预测方法研究  

Forecasting Methods of Regional Historical Atmospheric Pressure

作　　者：刘溅洪 罗来正[1] 王健坤 李茜 吴欣睿 孙有美 LIU Jianhong;LUO Laizheng;WANG Jiankun;LI Qian;WU Xinrui;SUN Youmei(Southwest Institute of Technology and Engineering,Chongqing 400039,China)

机构地区：[1]西南技术工程研究所,重庆400039

出　　处：《装备环境工程》2025年第2期167-172,共6页Equipment Environmental Engineering

摘　　要：目的 研究区域气压、温度、湿度等典型环境因素历史数据的精确预测方法。方法 通过分析气压的时空变化规律,采用样条函数拟合气压在时间维度上的变化规律,利用高斯分布刻画不同站点间空间上的相关性,构建区域历史气压数据时空分布模型。结果 模型对海南岛某站气压数据拟合值与实测值的最大绝对误差不超过0.8h Pa,平均绝对误差不超过0.2hPa。模型对海南岛某站气压的预测平均绝对误差在6h Pa以下(占比95.3%),对高原地区某站气压的预测平均绝对误差约为29.5 hPa,平均相对误差为4.6%,预测精度较高。结论 采用构建的时空分布模型能够为区域环境因素极值计算提供更多的基础数据,该方法还可以广泛应用于绝对湿度、温度等环境因素边界条件的预测。The work aims to research the accurate prediction method of historical data of regional typical environmental factors,such as pressure,temperature and humidity.Specifically,by analyzing the spatial and temporal variation of atmospheric pressure,spline function was used to fit the temporal variation,and Gaussian distribution was used to depict the spatial correla-tion between different stations.Then,a spatial and temporal distribution model of regional historical environmental factors was constructed.Results showed that the maximum absolute error and the average absolute error between the fitted and measured atmospheric pressure data of a station in Hainan Island were less than 0.8 hPa and 0.2 hPa respectively.In addition,the average absolute error of the model for a station in Hainan Island was less than 6 hPa(95.3%),which was about 29.5 hPa for a station in the plateau area.Meanwhile,the average relative error was about 4.6%,indicating high prediction accuracy of the model.Therefore,the constructed spatiotemporal distribution model provides more basic data for the calculation of the extremesdata of regional historical environmental factors,and it also provides new ideas for predicting boundary conditions of environmental factors such as absolute humidity and temperature.

关 键 词：历史环境因素 样条函数 时间序列模型 时空分布模型 边界条件 预测 

分 类 号：P426.6[天文地球—大气科学及气象学]