Stratosphere‐Troposphere Exchanges of Air Mass and Ozone Concentrations From ERA5 and MERRA2: Annual‐Mean Climatology, Seasonal Cycle, and Interannual Variability

Abstract

AbstractStratosphere‐Troposphere exchange (STE) of air mass and ozone in ERA5 and Modern Era Retrospective analysis for Research and Application, version 2 (MERRA2) reanalyses from 1980 to 2022 are investigated on their seasonal cycle, annual‐mean climatology, and monthly anomalies smoothed using a 1‐year Lanczos low‐pass filter. We employ a lowermost stratosphere mass budget approach with dynamic isentropic surfaces fitted to tropical tropopause as the upper boundary of lowermost stratosphere. The annual‐mean ozone STEs over the NH extratropics, SH extratropics, tropics, extratropics, and globe in ERA5 are −342, −239, 201, −581, and −380 Tg year−1, respectively, versus −305, −224, 168, −529, −361 Tg year−1 from MERRA2. The annual‐mean global ozone STE difference between ERA5 and MERRA2 is dominated by the diabatic heating difference, partly compensated by the ozone concentration difference. There are about 40% (−40%) differences between ERA5 and MERRA2 in global ozone STEs in boreal summer (autumn), mainly due to the difference in seasonal breathing of the lowermost stratosphere ozone mass between reanalyses. The correlation coefficient between ERA5 and MERRA2 global ozone mass STE monthly anomalies is 0.57 and thus ERA5 and MERRA2 can only explain each other's variance by 33%. Multiple linear regression analysis shows that El Niño–Southern Oscillation, quasi‐biennial oscillation, and Brewer‐Dobson circulation explain the variance in the ERA5 (MERRA2) global ozone STE monthly anomalies by 17.3 (5.0), 5.4 (7.2), and 1.0 (3.1)%, respectively. The volcanic aerosol impacts on ozone STEs from ERA5 and MERRA2 have opposite signs and thus are inconclusive. Cautions are therefore needed when using ERA5 and MERRA2 to investigate the STE seasonal cycle and interannual variability.