POLAR CLIMATE SIMULATION OF THE NCAR CCM3

Present-day Arctic and Antarctic climate of the National Center for At mospheric Research (NCAR) Community Climate Model version 3 (CCM3) is presented. The CCM3 simulation is from a prescribed and interannually varying sea surface temperature integration from January 1979 through August 1993. Observations from a variety of sources, including the Eur opean Centre for Medium-Range Weather Forecasts analyses, rawinsonde, and surface station data, are used for validation of CCM3's polar clim ate during this period. Overall, CCM3 can simulate many important pola r climatic features and in general is an incremental improvement over CCM2. The 500-hPa polar vortex minima are too deep by 50-100 m and too zonally symmetric. The Arctic sea level pressure maximum is displaced poleward, while the Icelandic region minimum is extended toward Europ e, and the Aleutian region minimum is extended toward Asia. The Antarc tic circumpolar trough of low sea level pressure is slightly north of the observed position and is 2-3 hPa too low. Antarctic katabatic wind s are similar to observations in magnitude and regional variation. The Antarctic surface wind stress is estimated to be 30%-50% too strong i n some regions. Polar tropospheric temperatures are 2 degrees-4 degree s C: colder than observations, mostly in the summer season. Low-level winter inversions over the Arctic Ocean are only 3 degrees-4 degrees C , rather than the observed 10 degrees C. In the Antarctic midcontinent they are around 25 degrees-30 degrees C (about 5 degrees stronger tha n observed) and continue to be stronger than observed along the coast. Although water vapor column is uniformly low by 10%-20% compared to a nalyses in both polar regions, the regional patterns of minima over Gr eenland and the East Antarctic plateau are well represented. Annual 70 degrees to pole CCM3 values are 5.8 kg m(-2) for the Arctic and 1.7 k g m(-2) for the Antarctic. The regional distribution of precipitation minus evaporation compares reasonably with analyses. The annual 70 deg rees to pole values are 18.1 cm yr(-1), which are close to the most re cent observational estimates of 16 to 18 cm yr(-1) in the Arctic and 1 8.4 +/- 3.7 cm yr(-1) in the Antarctic. In both polar regions, summer surface energy budgets are estimated to be low by roughly 20 W m(-1). Suggestions as to causes of simulation deficiencies Bye 1) polar heat sinks that are too strong; 2) inadequate representation of sea-ice-atm osphere hear exchange, due to lack of fractional coverage of sea ice o f variable thickness 3) effects of low horizontal resolution; and 4) b iased extrapolar influence.