Abstract: A data homogenization procedure was applied to adjust systematic errors in monthly station pressure data series of 825 stations in China for the period 1951-2014. Penalized maximum F-test was applied to examine the homogeneity of the data series and metadata was used to confirm the break points. Results showed that 400 of the 825 sites were confirmed to be homogenous while the other 425 were not. For the latter, firstly, the hydrostatic model was applied to adjust the errors caused by the use of incorrect station elevation values. Then statistical approaches (mean adjustment) were applied when a physical-based correction was in vain due to lack of correct metadata. As a result, pressure data for 255 of the 425 inhomogeneous sites were homogenized by the hydrostatic model, and for additional 170 sites were homogenized with the mean adjustment method. By comparing the trend of the raw and homogenized data, it was found that after homogenization the pattern exhibits a better spacial consistency, and the east coast and the northwest parts of China show a decrease trend while the middle part an increase one.

Key words: homogenization, pressure data, 825 stations in China, RHtest