Of course atmospheric-ocean-coupled global climate models are, by their nature, very strong on ocean-atmosphere interaction, but what about large lakes? Lake Victoria in East Africa, the N American Great Lakes, and Lake Nasser, and maybe some others are of comparable size to the average grid cell size of the AO-GCM. Hence wind, humidity and evaporation should affect adjacent grid cells. Is this taken into account?
The resolution of AO-GCM are not high enough to include lake model in their code, at least as far as I can remember (looking at google scholar, I have not found anything related to lake model in AOGCM).
However, lake model are implemented (and operational) in regional climate model (RCM), such as CRCM5 (Canadian Regional Climate Model v5). See for example Martynov et al. (2012), where they describe the results of implementing lake model in a 45km resolution simulations. On that note, I think that until AOGCM (and now the Earth System Models, which implement carbone cycle) goes to resolution close to 50km, they won't adopt massively such lake models.
From Martynov et al. (2012)
"The presence of lakes influences mostly the lake-rich region of the Canadian Shield by altering the low-level air temperature, humidity and precipitation. The influence of lakes on the precipitation is most significant over the Laurentian Great Lakes and within the distance of 200-300 km downstream of these lakes, especially with open water on lakes in wintertime. The patterns of enhanced precipitation correspond to regions of the well-known lake effect snow on downwind coasts of the Great Lakes."
Since the effect is rather regional, I suspect that other climate processes are more important for the AOGCM modellers.
Martynov, Andrey, Laxmi Sushama, René Laprise, Katja Winger, and Bernard Dugas. 2012. “Interactive Lakes in the Canadian Regional Climate Model, Version 5: The Role of Lakes in the Regional Climate of North America.” Tellus A; Vol 64 (2012), February. http://www.tellusa.net/index.php/tellusa/article/view/16226.