Discrepancies between model‐calculated and observed shortwave atmospheric absorption in areas with high aerosol loadings

Abstract

Model calculations of absorption of solar radiation in the atmosphere are afflicted with large uncertainties. In many current general circulation models (GCMs) the atmosphere is too transparent to solar radiation. Collocated surface and satellite observations indicate that the underestimation of shortwave atmospheric absorption in these models is most pronounced in low‐latitude areas, typically of the order of 20–40 W m−2. The present study focuses on one of these areas (equatorial Africa) to investigate the biases in three GCMs and in a model in assimilation mode (European Centre for Medium‐Range Weather Forecasts (ECMWF) reanalysis). The underestimation of atmospheric shortwave absorption by these models is found to be particularly evident in areas and seasons with large aerosol loadings from extensive vegetation fires. This indicates that the crude aerosol climatologies typically used in current GCMs and reanalyses do not properly account for these effects. The lack of absorbing aerosols may be responsible for biases of up to 30 W m−2 regionally and seasonally in the GCM surface and atmospheric shortwave budgets. For a more realistic simulation of the radiation budget in these regions, an adequate spatial and temporal representation of aerosol absorption is therefore crucial. No indications for a significant underestimation of cloud absorption in the GCMs are found. The remaining underestimation of shortwave absorption in the tropical GCM atmosphere, not attributable to the lack of aerosol, may rather be due to underestimation of water vapor amount and deficiencies in the clear‐sky radiative transfer calculations. The ECMWF reanalysis is significantly improved in these latter respects and is shown to compare favorably with the observed estimates of atmospheric absorption on annual mean scales. However, compensational effects between underestimated aerosol absorption and spurious cloud absorption may favor this agreement. On seasonal scales the compensational effects no longer apply, and peak biases similar to the GCMs become apparent, due to the lack of an adequate, seasonally resolved aerosol climatology.