Comment on gmd-2022-10

Abstract

A previous study on the use of nudging in EAMv1 had an unresolved issue, namely a simulation nudged to EAMv1’s own meteorology showed non-negligible deviations from the free-running baseline simulation over some of the subtropical marine stratocumulus and trade cumulus regions. Here, we demonstrate the deviations can be substantially reduced by two changes in the nudging implementation: first, revising the sequence of calculations in a nudged EAM simulation to improve consistency with the free-running baseline; second, increasing the frequency of constraining data from 6-hourly to 3-hourly to better capture strong sub-diurnal variations. The resulting improvements in climate representativeness motivate an investigation on the potential benefits of using newer reanalysis products with higher data frequency in nudged simulations that aim at capturing observed weather events. Simulations nudged towards 6-hourly ERA-Interim reanalysis or 6-hourly, 3-hourly, and hourly ERA5 reanalysis are evaluated against EAMv1’s own climatology, global-scale satellite retrievals of outgoing longwave radiation and precipitation, as well as in-situ measurements of air temperature, humidity, and winds from the Atmospheric Radiation Measurement (ARM) user facility. Our overall recommendation is to use the revised sequence of calculation and 3-hourly data from ERA5 for the nudged simulations. The anthropogenic aerosol effects in various EAMv1 simulations are evaluated. For estimating the global mean effect, the source of constraining data has a relatively small impact. Consistent with conclusions from previous studies, in order to obtain estimates of global mean aerosol effects that are consistent with results from the free-running baseline, nudging the horizontal winds but not air temperature is recommended.