Abstract:

Using the PMTred (penalized maximum T test) and PMFT (penalized maximum F test) developed by theEnvironment Canada, inhomogeneity test of monthly surface relative humidity of national station in China during 1951-2014 has been launched. The homogenous nearby stations with high correlation coefficients with the area averaged annual relative humidity are chosen to construct reference series by correlation coefficient weighted average method. Combined with the detailed metadata of each station, the main reasons causing the homogenization problem of the surface relative humidity in China have been analyzed. The result shows that the surface relative humidity data have serious problems of inhomogeneity, 68% stations existing break points. Automation, relocation and change of observe times are the main causes of the data discontinuity. For the whole dataset, the negative correction values have the high proportion, mainly between -5% and 0, which is linked tightly to the humidity decreasing after automation. The averaged surface humidity is decreasing in 1951-2014 before adjustment, and there is no obvious trend after adjustment.

Key words: relative humidity, homogenization, observation automation, PMTred and PMFT