Research on Quality Control Method of Surface Temperature Observations for Complex Physical Geography

Abstract

Abstract This article aims to improve the quality control (QC) of surface daily temperature observations over complex physical geography. A new QC method based on multiverse optimization algorithm, variational mode decomposition, and kernel extreme learning machine (MVO–VMD–KELM) was eomployed to identify potential outliers. For the selected six regions with complex physical geography, the inverse distance weighting (IDW), the spatial regression test (SRT), KELM, and the empirical mode decomposition improved KELM (EMD–KELM) methods were employed to test the proposed method. The results indicate that the MVO–VMD–KELM method outperformed other methods in all the cases. The MVO–VMD–KELM method yielded better mean absolute error (MAE), root-mean-square error (RMSE), index of agreement (IOA), and Nash–Sutcliffe model efficiency coefficient (NSC) values than others via the analysis of evaluation metrics for different cases. The comparison results led to the recommendation that the proposed method is an effective quality control method in identifying the seeded errors for the surface daily temperature observations.