Seasonal Forecast of Antarctic Sea Ice Xiaojun Yuan

Seasonal Forecast of Antarctic Sea Ice Xiaojun Yuan and Dake Chen Lamont-Doherty Earth Observatory of Columbia University WCRP Workshop on Seasonal Prediction, 4 -7 June, 2007 Barcelona, Spain

Outline • What influence Antarctic Sea Ice Variability at Seasonal Time Scale? Ø Ø Remote influence Regional impact • The Statistical Forecast Model Ø Ø The Markov model Validation and Forecast • Summary • Outstanding Questions

~ ~ Temperature Differences between El Nino and La Nina Composites Between 1980 and 2002 SAT SST Liu, et al. 2002

Yuan 2004

Differences between El Nino and La Nina Composites SAT SLP El Nino Composite Sea Ice SLP La Nina Composite Yuan 2004

Yuan, 2004

Ice lags 2 -month Yuan, 2005

Percentage of grid points with correlation significant at 99% confidence level ice lag ice lead Yuan, 2005

Seasonal Forecast of Antarctic Sea Ice A linear Markov Model Sea ice Surface air temperature Sea level pressure Surface winds (1980 -2000 satellite and reanalysis data) Multivariate empirical orthogonal functions (MEOF) Eigenvectors of leading modes from MEOFs PCs of leading modes are used to train the seasonally dependent transition matrices The Markov model predicts the PCs from one month to the next. Chen and Yuan, 2004 Seasonal sea ice forecast

Chen and Yuan, 2004

Chen and Yuan, 2004

Chen and Yuan, 2004

Current Sea Ice Forecast http: //ldeo. columbia. edu/~xiaojun

Summary Ø Antarctic sea ice in the western hemisphere is predictable at season time scale with reasonable good skill. Ø The predictability comes from the spatially coherent large climate variability in the Ocean-Atmosphere-ice coupled system. Ø Outside of the Antarctic Dipole region, climate variability is not distinct or coherent, therefore no good prediction skill exists. Ø The best prediction skill is found in the large climate action centers associated with ENSO and dominant high latitude climate modes, and in winter.

Outstanding Questions Ø Can this predictability in southern high latitudes be useful to extrapolar seasonal forecast? The magnitude of the ADP anomaly is large, the spatial scale is half hemisphere, temporal scale of the anomaly is season and longer. Ø How do the polar processes feedback to extra-polar climate variability? Understanding the mechanisms of the feedback of polar processes to extrapolar climate variability leads to a better integration of cryosphere into the global climate system.