Comparison of Radiosonde and RUC Model Mean Radiating Temperature of the Atmosphere

COMPARISON OF RADIOSONDE AND RUC MODEL MEAN RADIATING TEMPERATURE OF THE ATMOSPHERE

Zlatko R. Vukovic and Walter J. Strapp Cloud Physics and Severe Weather Research Section, Environment Canada, Toronto, Canada

The main objective of this study is to analyze the accuracy of using the mean radiating temperature of the atmosphere (Tmr) obtained from the RUC model, instead of the seasonal averaged climatological Tmr values from radiosonde measurements. In our application, Tmr values are used in the calculation of precipitable water vapour and precipitable liquid water from microwave brightness temperatures. From a one year data set of Maniwaki radiosonde and RUC data the Tmr values for 23.8 and 31.4 GHz were calculated. The data were linearly correlated, but with the slope in a range of 1.2 – 1.3, and rms values in a range of 8 – 13 oK. After the adjustment of the RUC Tmr values based on this regression, the new data set naturally had a 1:1 linear correlation. For the following five periods, the entire year of 2003, Dec-Feb, Mar-May, Jun-Aug, and Sep-Nov, the rms difference between the radiometer and the RUC Tmr values were in the range of 2.5 to 5.2 oK, which was much better than the radiosonde difference from it's averaged Tmr values of 3.5 to 10.0 oK for the same periods. This suggests that using averaged seasonal radiosonde values of Tmr in the microwave radiometer statistical retrievals generates larger errors than if adjusted RUC forecast Tmr values were used. As the forecast duration is shortened, the error is reduced. The three main points can be concluded from this study: a) The Tmr calculated from the RUC model forecasts can successfully substitute for the Tmr derived from the historical radiosonde data. In this study the coefficient of linear correlation increased from 0.92 for a forecast of 12 h (23.8 GHz), to 0.97 for a 3 h forecast (31.4 GHz). b) The RUC model can often provide improved accuracy in the estimation of Tmr. In this study, in about ¼ of comparison points the improvement was more than 5%, and up to a maximum of 8%. c) The use of the RUC model forecast data can also provide Tmr estimates for locations where historical radiosonde data are not available or are at unacceptable distances.