Analysis and correction of the observation error of the microwave radiometer in dongying city

Three-year observation errors of microwave radiometer in Dongying city are analyzed and corrected by layers referring the EC-interim reanalysis data as the true. Quality controls are carried out in time and space, which can effectively eliminate the outliers in observations. The correlation coefficients are 0.98 and 0.89 for the sampled temperature profiles and water vapor density profiles after classifications. Under non-precipitation conditions, the correlation coefficient for water vapor density profiles is 0.89, while that is only 0.86 under precipitation conditions. Precipitation has larger impact on water vapor density than temperature. The average deviation of the temperature profiles below 6.5 km is around 0-2K with high precision. The root mean square error (RMSE) of the temperature profiles is about 2K from 0-2 km in the boundary layer. The average bias for water vapor density is 0-1.5gm-3. The maximum bias is larger than 2gm-3 with the existence of precipitations at the altitude of 1-4.5 km. The bias and RMSE profiles of the water vapor density show spinning conical structures. The RMS error is less than 3.5gm-3 when there's no precipitation and is greater than 4gm-3 under precipitation conditions in 1-4.5 km. The bias of the temperature profile is almost 0 after corrections with RMSE of 0.5K and those corrected for water vapor density profiles is 0-0.2 gm-3.