翁源县乡镇尺度日最高最低气温预报方法构建

Construction of the Method of Daily Maximum and Minimum Temperature Forecast at Township Scale in Wengyuan County

摘要: 本研究旨在提升粤北岭南山区翁源县所辖乡镇的精细化温度预测精度。采用线性回归分析方法，构建了乡镇温度预报模型，并运用统计手段深入解析了乡镇区域自动站温度与翁源国家基本站温度的差异性。本文采用线性回归分析方法进行了优化与完善，最终确立了适用于翁源县8个乡镇的日最高及最低温度预测方程。结果表明：翁源国家基本站的日最高、最低温度与各乡镇区域自动站温度的相关系数均高于99%；回归分析法预测结果具有较高准确度，整体显著优于传统差值法；多数乡镇的预测准确率超过90%，对于预测精度较低的个别乡镇，建议采用差值法预测结果予以校正。

Abstract: This study aims to enhance the precision of refined temperature forecasting for townships in Wengyuan County, located in the Lingnan mountainous region of northern Guangdong. By employing linear regression analysis, a township-level temperature forecasting model was established, and statistical methods were utilized to analyze the differences between temperatures recorded by township regional automatic stations and those from the Wengyuan National Basic Station. Through optimization and refinement using linear regression analysis, the study ultimately developed prediction equations for daily maximum and minimum temperatures applicable to eight townships in Wengyuan County. Results indicate that the correlation coefficients between daily maximum and minimum temperatures at the Wengyuan National Basic Station and those at township regional automatic stations all exceed 99%. The regression analysis method demonstrates high accuracy, significantly outperforming the traditional difference method. Most townships achieved prediction accuracy exceeding 90%, while for those with lower precision, it is recommended to use the difference method for correction.