A Quality Control Algorithm for Surface Temperature Observations Based on Improved Kriging Method

A method called Ordinary Kriging (OK) in Geostatistics was introduced for quality control of surface temperature observations based on spatial correlation of temperature. Due to the continuity of temperature, Gaussian model was chosen as the semi-variogram model. Because of some possible disadvantages in applying Gaussian model, it is necessary to improve Gaussian model. A quality control method based on Improved Ordinary Kriging (IOK) was developed in this study. In order to assess the effectiveness and applicability of the proposed method, daily mean surface temperature observations collected at 67 stations in Jiangsu Province were used for quality control, and the results were compared with that from OK and Inverse Distance Weighted (IDW) methods. It was found that IOK performs better than IDW and OK in error checking, and the proposed method can effectively identify errors in temperature data. In addition, IOK also has higher stability and applicability.