Analysis of the downward shortwave surface flux biases in the coupled GFDL General Circulation Model

A state-of-the-art coupled General Circulation Model developed at the Geophysical Fluid Dynamics Laboratory (GFDL) (version CM2.1) is used to examine the difference with observations in the downward shortwave flux at the surface. The important objectives here are to assess the geographical distribution of the CM2.1 flux biases that arise (both for clear-sky and all-sky), how they relate to deficiencies in the derived cloud amount, aerosol optical depth and surface albedo, and how they impact the simulation of surface temperature in the model. To assess the flux biases, two ground-based observational datasets, the Baseline Surface Radiation Network (BSRN) and the Global Energy Balance Archive (GEBA), and the satellite-based International Satellite Cloud Climatology Project (ISCCP-FD) dataset are utilized.

A monthly-mean climatology of BSRN flux values for both the inferred “clear-sky” and for all-sky atmospheres, and for the direct, diffuse and total beam components have been obtained. The clear-sky flux climatology is used to assess the roles of aerosol optical depth and surface albedo in the resultant bias. When aerosols are excluded, CM2.1 overestimates the direct and underestimates the diffuse monthly mean surface flux values for all the site locations. When aerosols are included, the European, Asian and North American sites all display a notable reversal in the sign of these differences, along with a significant overestimate in the observed (AERONET) aerosol optical depth. Contrastingly, these flux differences are of similar sign but reduced at the Darwin, Florianopolis and Nauru Island sites, and occur in conjunction with an underestimate in the observed aerosol optical depth. For polar site locations, an underestimate in the surface albedo accompanies a similar difference with respect to the observed diffuse flux.

Similarities in the CM2.1's all-sky flux biases are determined among the BSRN, GEBA and ISCCP-FD datasets. Notable flux underestimates occur over North America, Europe, eastern Asia, and southern Africa, while notable overestimates appear over the Amazon, equatorial Africa, the western equatorial Pacific, and off the west coast of the Americas. These flux biases show an association with overestimates and underestimates in cloud amount, respectively, with respect to ECMWF ERA40 climatology. A moist bias present in CM2.1 contributes towards an overestimate in aerosol optical depth over Europe, eastern Asia and North America by hygroscopic sulfate, further enhancing the flux underestimates seen in these regions due to too much cloudiness. CM2.1 also underestimates the optical depth of carbonaceous aerosols in regions of biomass burning, further contributing to the flux overestimates seen over the Amazon and southern Africa. In response to these flux biases, CM2.1 surface temperatures are generally too low over North America and most of the Eurasian continent and too high over the Amazon region. Thus, the various observational datasets have been useful in providing a firmer basis for deciphering improvements needed in modeling cloud and aerosol absorption.