W. M. Connolley, British Antarctic Survey, Cambridge, UK
C. Genthon, LGGE/CNRS, Grenoble, France.
There is good agreement between "observations" (actually NWP analyses from CEP) and the UKMO model (UM) that the area-averaged annually-averaged heat balance at the surface of Antarctica is between longwave radiative loss (about 40+) and gain by visible (about 28) and sensible (less agreement, about 14-19) fluxes. Latent heat is negligible. The Arpege models suffer from too-high surface temperatures and a non-heat-conserving deep-soil model and produce less plausible results. However they have roughly the same balance although the exact numbers are different and the budget does not balance.
Verifying this against true observations is very hard due to the sparsity of observations which is caused by the difficulty of making accurate long-term measurements of any of the components of the budget. A proper comparison awaits a thorough compilation of observations. This preliminary study uses only some net radiation observations compiled by John King, a few observations from the South Pole and some AWSs from the Ross Ice Shelf. Both of the latter are good locations for comparison because they are horizontally uniform in the model and reality, at least in terms of orography. The Ross sites, however, appear to show considerable variability possibly due to the influence of katabatic winds from the Transantarctic Mountains.