基于实测资料的京津冀地区地表总辐射最优估算模型研究

Study on Optimal Estimation Model of Total Surface Radiation in the Beijing-Tianjin-Hebei Region based on Measured Data

地表总辐射(solar radiation, R_(s))是一种对人类发展及其重要的新能源,同时也是影响区域环境变化的主要因素,R_(s)的准确估算对区域新能源发展及环境保护政策制定意义重大。故找出适用于京津冀地区辐射R_(s)模拟的标准方法是十分重要的,基于机器学习模型、辐射法和温度法共9种不同模型,分别计算了R_(s)在日值和月值尺度的精度,并计算了不同精度指标,结果表明:3种类型方法精度由高到低依次为机器学习模型>辐射法>温度法,其中极限学习机模型ELM在日值和月值尺度精度最高。

Rs is an important new energy source for human development and also a main factor affecting regional environmental change. Accurate estimation of Rs is of great significance for regional new energy development and environmental protection policyformulation. Therefore, it is very important to find out the standard method suitable for radiation Rs simulation inthe Beijing-Tianjin-Hebei area. In this paper, based on nine different models including machine learning model, radiation method,and temperature method, the precision of Rs at daily and monthly scales and different precision indexes were calculated respectively. The results showed that: in order of accuracy, the three types of methods were machine learning model > radiation method and > temperature method. Extreme learning machine model ELM has the highest accuracy in daily and monthly scales.