2025-06-05

Results

Reanalysis Intercomparison

general features - high-winds band 40-60S, strongest over indian ocean. (Fig 1)
How remaining (R1, JRA, MEARRA) reanalyses compare to ERA5 in means, tails, interannual variability (Table 2, ST1, ST2)

JRA and ERA5 behave similarly, have stat. sig. trend mostly in austral summer and autumn (Fig 2, Fig 3, Table 3)
R1 and MERRA2 stronger trends than JRA/ERA5, stat. sig. year round, and significant over larger spatial area
for all reanalyses, trends 1980-1999 basically double those over full 1980-2019 period

jet position roughly similar in products (Fig 4, ST3), strong seasonal cycle (southerly in autumn, northerly in spring), seasonal cycle amplitude similarly represented, as is interannual variability of zonal mean jet position and zonally-varying jet position.
strongest trend in jet position occurs in the austral summer and autumn, much stronger over 1980-1999 than over full period, not always statistically significant

obs sam index could be used to constrain reanalyses, in general reanalyses pretty synchronous with observational sam index (Fig 5)
trend in sam strongest in summer/autumn, again strongest during period of max ozone loss.
trend in SAM correlates tightly with trend in wind speed; since they are mechanistically linked, reanalyses that represent SAM well could represent winds well.
R1/MERRA2 austral winter/spring wind trends could be an overestimate as they are accompanied by strong sam inddex trends that we don’t see in the obs.
In autumn, ERA5 & JRA3Q underestimate 1980-1999 SAM trend, R1 and MERRA2 overestimate it 1980-2019.
Less clear correlation between SAM index and wind in summer