Impact Evaluation for Replacement of Temperature-humidity Sensor of Automatic Weather Station

To discuss the differences between automatic and manual observations for relative humidity and water vapor pressure, some comparative researches are implemented, analyzing the correlation coefficient, mean and variance between 38 long-time automatic observation meteorological stations and their adjacent stations in Hubei Province. Furthermore, based on the automatic and manual observations of four national reference climatological stations where the temperature-humidity sensors have been replaced 28 times from 2003 to 2011, impacts on relative humidity and water vapor pressure measurements are studied.The correlation coefficients for relative humidity, water vapor pressure between the checked station and its adjacent stations show a decreasing trend when mannual observation is replaced by automatic observation, the mean and variance for relative humidity and water vapour pressure are also obviously different, and these differences may be caused by the change of observation method.The relative humidity and water vapour pressure records jump caused by replacement of the temperature-humidity sensor appears in the ratio of 64%. The average relative humidity record jump is 3.4%, and the maximum reaches 8.5%, while the average for water vapor pressure is 0.74 hPa, and the maximum is up to 1.93 hPa.The main cause for such difference appearing at the same station as described above is that calibration error of indication, for old temperature-humidity sensor is much different from the new one when replacing. So improving the method of observation and perfecting the verification procedure of automatic weather station are vary important to ensure the homogeneity of data.Therefore, it suggests choosing the temperature-humidity sensor whose calibration error is smaller, and an error calibration module should be added to the data processing software or operational software of automatic weather stations.