Intercomparisons are performed of simulated Arctic and Antarctic climate for 1979-1988 by 17 currently available international climate models for selected surface and vertically integrated monthly averaged quantities, such as sea-level pressure, precipitable water, precipitation, and evaporation. The accuracy of modeled sea-level pressure depends on horizontal resolution. It seems that a critical resolution exists at about 4 x 4 degrees, i.e., simulated results are significantly improved when the model resolution is better than the critical resolution. There is a maximum discrepancy in simulated precipitation over the Arctic and Antarctica. The cause of the excess simulated precipitation in high latitudes must be factors other than moisture availability, and is most likely due to the simulated general circulations of the respective models. Most of the models are able to capture the evaporation/sublimation rate quite well over the polar regions. Thus the large errors in the net precipitation over high latitudes are probably caused by the anomalously large simulated precipitation rates.
Although the ensembles of model simulations for both higher resolution
(HR) and lower resolution (LR) are generally able to capture strength and
location of the semiannual variation of the Southern Hemisphere circumpolar
trough, biases in the HR models are less than those in the LR models. A
noticeable one-month phase lag exists from February to August. The dominant
interannual variability (QBO) in the southern circumpolar trough is not
well simulated by both HR and LR models. The simulations have much smaller
amplitude and longer period variations. The time-latitude Hovmoeller diagram
for the Pacific sector (120 W-180) zonal wind for five T-42 models reveals
that the prominent interannual variation signals of ENSO and the double
jetstream strength are quite well simulated in the Southern Hemisphere.
However, the relatively weak interannual variability in the Northern Hemisphere
is not well captured.