Large-Scale Atmospheric Transport in GEOS Replay Simulations

Offline chemical transport models (CTMs) have traditionally been used to perform studies of atmospheric chemistry in a fixed dynamical environment. An alternative to using CTMs is to constrain the flow in a general circulation model using winds from meteorological analyses. The Goddard Earth Observing System (GEOS) replay'' approach involves reading in analyzed fields every 6 h and recomputing the analysis increments, which are applied as a forcing to the meteorology at every model time step. Unlike in CTM, all of the subgrid-scale processes are recalculated online so that they are consistent with the large-scale analysis fields, similar in spirit to nudged'' simulations, in which the online meteorology is relaxed to the analysis. Here we compare the transport of idealized tracers in different replay simulations constrained with meteorological fields taken from The Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2). We show that there are substantial differences in their large-scale stratospheric transport, depending on whether analysis fields or assimilated fields are used. Replay simulations constrained with the instantaneous analysis fields produce stratospheric mean age values that are up to 30% too young relative to observations; by comparison, simulations constrained with the time-averaged assimilated fields produce more credible stratospheric transport. Our study indicates that care should be taken to correctly configure the model when the replay technique is used to simulate stratospheric composition.