Different Types of Observations Show Robust Tropospheric Warming and Stratospheric Cooling
Scientists at the Wegener Center for Climate and Global Change, in collaboration with B.D. Santer at Lawrence Livermore National Laboratory and with colleagues from 15 other institutions, performed a detailed intercomparison of observed atmospheric temperature trends estimated from satellites, radiosondes, lidar systems, and GPS radio occultation (RO). This analysis provides compelling observational evidence of global-scale tropospheric warming and stratospheric cooling over the satellite era (1979 to 2018).
This is the most comprehensive comparison to date of atmospheric temperature trends inferred from different measurement systems. It reveals that over the past four decades, there has been robust cooling of the stratosphere (by ca. 1–3°C) and robust warming of the troposphere (by ca. 0.6–0.8°C). Over the shorter post-2001 RO period, satellite data exhibit consistently smaller tropospheric warming than RO and radiosondes. This discrepancy requires further investigation and points towards the possibility of a systematic low bias in satellite tropospheric temperature trends.
This large international team included experts in all currently available systems used to monitor multi-decadal changes in atmospheric temperature. The key results from this study are the following. First, there is considerable scientific value in rigorous intercomparison of temperature trends obtained with different measurement systems. Second, intercomparison helps to identify robust features of the observations, such as the large, global-scale tropospheric warming and stratospheric cooling signals. Third, formal intercomparison reveals that over the post-2001 RO period, the tropical upper tropospheric warming inferred from RO and radiosondes (0.25–0.35°C per decade) is systematically larger than in satellite data. Results from RO and radiosondes show amplification of surface warming over 2002 to 2018, consistent with a moist adiabatic lapse rate. One possible interpretation of these differences between trends in satellite data and in RO/radiosondes is that satellites may have residual inhomogeneities – a conclusion in accord with a recently published paper by Santer et al. (2021).