Assessing the potential of passive microwave radiometers for continuous temperature profile retrieval using a three year data set from Payerne
Abstract. The motivation of this study is to verify theoretical expectations placed on ground-based radiometer techniques and to confirm whether they are suitable for supporting key missions of national weather services, such as timely and accurate weather advisories and warnings. We evaluate reliability and accuracy of atmospheric temperature profiles retrieved continuously by a HATPRO (Humidity And Temperature PROfiler) system operated at the aerological station of Payerne (MeteoSwiss) in the time period August 2006–December 2009. Assessment is performed by comparing temperatures from the radiometer against temperature measurements from a radiosonde accounting for a total of 2088 quality-controlled all-season cases. In the evaluated time period, HATPRO delivered reliable temperature profiles in 88% of all-weather conditions with a temporal resolution of 15 min. Random differences between HATPRO and radiosonde are down to 0.5 K in the lower boundary layer and rise up to 1.7 K at 4 km height. The differences observed between HATPRO and radiosonde in the lower boundary layer are similar to the differences observed between the radiosonde and another in-situ sensor located on a close-by 30 m tower. Temperature retrievals from above 4 km contain less than 5% of the total information content of the measurements, which makes clear that this technique is mainly suited for continuous observations in the boundary layer. Systematic temperature differences are also observed throughout the retrieved profile and can account for up to ±0.5 K. These errors are due to offsets in the measurements of the microwave radiances that have been corrected for in data post-processing and lead to nearly bias-free overall temperature retrievals. Different reasons for the radiance offsets are discussed, but cannot be unambiguously determined retrospectively. Monitoring and, if necessary, corrections for radiance offsets as well as a real-time rigorous automated data quality control are mandatory for microwave profiler systems that are designated for operational temperature profiling. In the analysis of day/night differences, it is shown that systematic differences between radiosonde and HATPRO decrease throughout the boundary layer if 2 m surface temperature measurements are included in the retrieval.