Uncertainties in recent satellite ozone profile trend assessments (SI2N, WMO 2014) : A network-based assessment of fourteen contributing limb and occultation data records

Numerous vertical ozone profile data records collected over the past decades from space-based platforms have the potential to allow the ozone and climate communities to tackle a variety of research questions. A prime topic is the study and documentation of long-term changes in the vertical distribution of atmospheric ozone, as targeted by the recent SPARC/IO3C/IGACO-O3/NDACC Initiative (SI2N) and WMO’s ozone assessment. Such studies typically require data records with documented mutual consistency in terms of bias and long-term stability. Ground-based networks play a pivotal role in evaluating which satellite records comply with end-user requirements and are fit for their purpose. They provide high-quality, independent measurements on a pseudo- global scale from the ground up to the stratosphere. Here, we present an assessment of the long-term stability and mutual consistency of fourteen limb/occultation ozone profile data records, using NDACC/GAW/SHADOZ ozonesonde and NDACC lidar network data as reference standards. We show how a harmonized analysis framework and robust statistical methods allow us to derive reliable estimates of the drift, bias, and short-term variability of each satellite data record. We examine the dependence of these parameters on altitude and, whenever feasible, on latitude and season. The analysis is furthermore performed in four different ozone profile representations, as it turns out that auxiliary data used for unit and representation conversions can impact data quality. We discuss the mutual consistency and compliance of satellite data sets with respect to specific user requirements from GCOS and from climate research groups. We conclude by reflecting on the implication of our results for trend assessments on recently merged ozone profile records (Ozone_CCI, GOZCARDS, SWOOSH, ...)