Local comparisons of tropospheric ozone: vertical soundings at two neighbouring stations in southern Bavaria

Abstract

In this study ozone profiles of the differential-absorption lidar at Garmisch-Partenkirchen are compared with those of ozone sondes of the Forschungszentrum Jülich and of the Meteorological Observatory Hohenpeißenberg (German Weather Service). The lidar measurements are quality assured by the highly accurate nearby in situ ozone measurements at the Wank (1780 m a.s.l.) and Zugspitze (2962 m a.s.l.) summits and at the Global Atmosphere Watch station Schneefernerhaus (UFS, 2670 m a.s.l.), at distances of 9 km or less from the lidar. The mixing ratios of the lidar agree with those of the monitoring stations, with a standard deviation (SD) of 1.5 ppb, and feature a slight positive offset of 0.6 ± 0.6 ppb (SD) conforming to the known −1.8 % calibration bias of the in situ instruments. Side-by-side soundings of the lidar and electrochemical (ECC) sonde measurements in February 2019 by a team of the Forschungszentrum Jülich shows small positive ozone offsets for the sonde with respect to the lidar and the mountain stations (0.5 to 3.4 ppb). After applying an altitude-independent bias correction to the sonde data an agreement to within just ±2.5 ppb in the troposphere was found, which we regard as the wintertime uncertainty of the lidar. We conclude that the recently published uncertainties of the lidar in the final configuration since 2012 are realistic and rather small for low to moderate ozone concentrations. Comparisons of the lidar with the Hohenpeißenberg routine measurements with Brewer-Mast sondes are more demanding because of the distance of 38 km between the two sites implying significant ozone differences in some layers, particularly in summer. Our comparisons cover the 3 years September 2000 to August 2001, 2009, and 2018. A slight negative average offset (−3.64 ± 3.72 ppb (SD)) of the sondes with respect to the lidar was found. We conclude that most Hohenpeißenberg sonde data could be improved in the troposphere by recalibration with the Zugspitze station data (1978 to 2011 summit, afterwards UFS). This would not only remove the average offset but also greatly reduce the variability of the individual offsets. The comparison for 2009 suggests a careful partial re-evaluation of the lidar measurements between 2007 and 2011 for altitudes above 6 km, where occasionally a negative bias occurred.