The Surface Energy Balance at Local and Regional Scales-A Comparison of General Circulation Model Results with Observations

Abstract

Aspects of the mean monthly energy balance at continental surfaces are examined by appeal to the results of general circulation model (GCM) simulations, climatological maps of surface fluxes, and direct observations. Emphasis is placed on net radiation and evaporation for (i) five continental regions (each approximately 20°?150°) within Africa, Australia, Eurasia, South America, and the United States; (ii) a number of continental sites in both hemispheres. Both the mean monthly values of the local and regional fluxes and the mean monthly diurnal cycles of the local fluxes are described. Mostly, GCMs tend to overestimate the mean monthly levels of net radiation by about 15% ?20% on an annual basis, for observed annual values in the range 50 to 100 Wm?2. This is probably the result of several deficiencies, including (i) continental surface albedos being undervalued in a number of the models, resulting in overestimates of the net shortwave flux at the surface (though this deficiency is steadily being addressed by modelers); (ii) incoming shortwave fluxes being overestimated due to uncertainties in cloud schemes and clear-sky absorption; (iii) land-surface temperatures being under-estimated resulting in an underestimate of the outgoing longwave flux. In contrast, and even allowing for the poor observational base for evaporation, there is no obvious overall bias in mean monthly levels of evaporation determined in GCMS, with one or two exceptions. Rather, and far more so than with net radiation, there is a wide range in values of evaporation for all regions investigated. For continental regions and at times of the year of low to moderate rainfall, there is a tendency for the simulated evaporation to be closely related to the precipitation-this is not surprising. In contrast, for regions where there is sufficient or excessive rainfall, the evaporation tends to follow the behavior of the net radiation. Again, this is not surprising given the close relation between potential evaporation and net radiation, as discussed by Priestley and Taylor. Finally, the introduction into GCMs of an ?improved? surface scheme (incorporating more realistic representations of soil and canopy processes and revised albedos) does tend to improve the calculations of both regional net radiation and evaporation.