空间面板数据模型在地面气温观测资料质量控制中的应用

由于地面气温观测资料的时空分布符合空间面板数据结构特征,提出一种基于改进空间面板数据模型的地面气温观测资料质量控制算法(ST-RH算法)。该算法在利用邻近站地面气温观测资料时空相关性信息对目标站气温观测资料进行质量控制的基础上,兼顾了地面气温与相对湿度之间的强耦合关系,将相对湿度作为解释变量融入算法,增加了算法的内部复杂度,提高了算法的动态稳定性。为了检验该算法的有效性及适应性,利用ST-RH算法对多个场景地面气温观测资料进行质量控制,并与反距离加权算法(IDW算法)和空间回归算法(SRT算法)进行比较分析。试验结果表明,ST-RH算法相对于IDW算法和SRT算法更能有效地标记出地面气温观测资料中的存疑数据,同时多组独立案例的分析结果说明ST-RH算法具有更好的稳定性和适用性。

Application of spatial panel data model to the quality control for surface temperature observations

Considering the spatial and temporal distribution of surface temperature observations which meet the spatial panel data structure, a new quality control method (ST-RH) based on spatial panel data was proposed. The ST-RH method employs the information from the neighboring stations to estimate the value of target station. As the strong-coupling relation between surface temperature and relative humidity, the relative humidity was introduced into ST-RH as an explanatory variable, which made the ST-RH more complicate and more stable. In order to test the new method, the IDW method and SRT method was used to evaluate the ST-RH method on different cases. The results of the comparison led to the recommendation that the ST-RH method is an effective quality method in identifying the seeded errors for the surface temperature with the best robustness and adaptability.