Comparing the Degree of Land–Atmosphere Interaction in Four Atmospheric General Circulation Models

The strength of the coupling between the land and the atmosphere, which controls, for example, the degree to which precipitation-induced soil moisture anomalies affect the overlying atmosphere and thereby the subsequent generation of precipitation, has been examined and quantified with many atmospheric general circulation models (AGCMs). Generally missing from such studies, however, is an indication of the extent to which the simulated coupling strength is model dependent. Four modeling groups have recently performed a highly controlled numerical experiment that allows an objective intermodel comparison of land-atmosphere coupling strength, focusing on short (weekly down to subhourly) timescales. The experiment essentially consists of an ensemble of 1-month simulations in which each member simulation artificially maintains the same (model specific) time series of surface prognostic variables. Differences in atmospheric behavior between the ensemble members then indicate the degree to which the state of the land surface controls atmospheric processes in that model. A comparison of the four sets of experimental results shows that coupling strength does indeed vary significantly among the AGCMs.