Toward Advanced Understanding and Prediction of Arctic Climate Change Wieslaw Maslowski Naval Postgraduate School 34th Annual Climate Diagnostics and Prediction Workshop, Monterey, CA, October, 2009

Outline Model predictions of Arctic ice extent Trends of melt rates based on Ice extent vice ice thickness and volume Atmospheric vice Oceanic forcing of local sea ice melt Future projections of ice melt Regional Arctic Climate System Model

Observed Rate of Loss Faster Than GCM Predicted SSM/I period Obs Mean = 7.0 Mln km 2 Adapted from Stroeve et al., 2007 Wang & Overland, 2009: Ice-free summer Arctic Ocean by 2037 "A linear increase in heat in the Arctic Ocean will result in a non-linear and accelerating loss of sea ice.“ - Norbert Untersteiner, Prof. Emeritus, Univ. of Washington, July 2006

Observed Arctic sea ice extent (a,b) and modeled sea ice thickness (c,d) during September 1979 (a,c) and 2002 (b,c) SSM/I – 2D MODEL - 3D 09/7909/7909/0209/02 Significant decrease in observed sea ice extent (17-20%; top) and in modeled ice thickness (up to m or ~35%; bottom) in the 2000s. Note that largest changes are downstream of Pacific / Atlantic water inflow into the Arctic Ocean. (Maslowski et al., 2007)

Observed MY Ice Fraction

Forcing of Arctic sea ice melt “Atmospheric circulation trends are weak over the record as a whole, suggesting that the long-term retreat of Arctic sea ice since 1979 in all seasons is due to factors other than wind-driven atmospheric thermal advection.” - Deser and Teng, J. Clim Oceanic Forcing can locally play critical role in melting sea ice via: –horizontal advection of warm Pacific / Atlantic water into/under the sea ice cover (e.g. Stroeve and Maslowski, 2007) –Locally induced (upwelling, topographically controlled flow, eddies) upward heat flux into the mixed layer (Maslowski and Clement Kinney, in review, 2009)

Mean Oceanic Heat Convergence: m; T ref = T freezing Modeling Challenges: Inflow of Pacific / Atlantic Water into the Arctic Ocean Pacific Water entering via Pacific Water entering via narrow (~60mi) Bering Strait narrow (~60mi) Bering Strait outflow through Fram outflow through Fram Strait vs. Atlantic Water Strait vs. Atlantic Water inflow (FSBW) inflow (FSBW) Atlantic (BSBW) and Pacific Atlantic (BSBW) and Pacific Water each losses majority Water each losses majority of heat to the atmosphere of heat to the atmosphere before entering Arctic Basin before entering Arctic Basin Arctic ocean-ice-atm feedbacks – not represented realistically in climate models Modeling Challenges: Inflow of Pacific / Atlantic Water into the Arctic Ocean Pacific Water entering via Pacific Water entering via narrow (~60mi) Bering Strait narrow (~60mi) Bering Strait outflow through Fram outflow through Fram Strait vs. Atlantic Water Strait vs. Atlantic Water inflow (FSBW) inflow (FSBW) Atlantic (BSBW) and Pacific Atlantic (BSBW) and Pacific Water each losses majority Water each losses majority of heat to the atmosphere of heat to the atmosphere before entering Arctic Basin before entering Arctic Basin Arctic ocean-ice-atm feedbacks – not represented realistically in climate models Heat Loss High resolution is one of the top requirements for advanced modeling of High resolution is one of the top requirements for advanced modeling of Arctic climate FSBW BSBW Heat Loss (Maslowski and Clement Kinney, 2009; Maslowski et al., 2008, Clement et al., 2005)

EW Modeled Oceanic heat flux exiting the Chukchi Shelf Heat Flux via Alaska Coastal Current accounts for ~67% of the Total Heat Flux across Chukchi Shelf Line Sept Sept. 2002

Modeled Upper Ocean Heat Content and Ice Thickness Anomalies ( Mean Annual Cycle Removed) Heat content accumulated in the sub-surface ocean since mid-1990s explains at least 60% of total sea ice thickness change

Intrusion of oceanic heat into the mixed layer due to anti-cyclonic eddies Temperature above freezing ( o C) and velocity (cm/s) is shown in the top four panels. Potential density (σ θ ) and velocity (cm/s) profiles are shown in the lower four panels. The vertical section locations are shown by the black and red lines, which are across the Chukchi Rise and Eddy 2.

Volume (x104km3) Area (x107km2) Mean Thickness (m) Between : - annual mean sea ice concentration has concentration has decreased by ~17% decreased by ~17% - mean ice thickness has decreased by ~0.9 m decreased by ~0.9 m or ~36% or ~36% - ice volume decreased by 40%, which is ~2.5x the 40%, which is ~2.5x the rate of ice area decrease rate of ice area decrease If this trend persists the Arctic Ocean will become ice-free by ~2013! time series of Ice Volume, Area, Mean Thickness

Need for Regional Arctic Climate System Model There are large errors in global climate system model simulations of the Arctic climate system Observed rapid changes in Arctic climate system –Sea ice decline –Greenland ice sheet –Temperature Arctic change has global consequences –Sea ice change can alter the global energy balance and thermohaline circulation

Regional Arctic Climate System Model (RAMC) Participants: Wieslaw Maslowski- Naval Postgraduate School John Cassano- University of Colorado William Gutowski - Iowa State University Dennis Lettenmeier- University of Washington Other collaborators: David Bromwich - OSU Greg Newby, Andrew Roberts, Juaxion He-UAF/IARC/ARSC Primary science objective: to synthesize understanding of past and present states and thus improve decadal to centennial prediction of future Arctic climate and its influence on global climate. DOE Climate Change Prediction Program Funded Project

Science Objectives Perform multi-decadal simulations to: –Gain improved understanding of coupled Arctic climate system processes responsible for changes in Arctic sea ice cover hydrologic cycle freshwater export –Improve predictions of Arctic climate change –Identify limitations and physical and numerical requirements of global climate system model simulations of Arctic

RACM components and resolution Atmosphere - Polar WRF (gridcell ≤50km) Land Hydrology – VIC(same as WRF) Ocean - LANL/POP (gridcell ≤10km) Sea Ice - LANL/CICE (same as POP) Flux Coupler – NCAR CPL7 Use NCAR CCSM4 framework for developing RACM Higher component resolutions to be evaluated subject to availability of computer resources

RACM model domain and elevations Pan-Arctic region to include: - all sea ice covered ocean in the northern hemisphere - Arctic river drainage - critical inter-ocean exchange and transport - large-scale atmospheric weather patterns (AO, NAO, PDO)

Accomplishments to Date 2 nd year of 4 year DOE funded project Coupled individual model components to CPL7 and full RACM testing Model component evaluation studies –Polar WRF development and climatology –Polar WRF and fractional sea ice –Simulation of sea ice loss with POP/CICE –Oceanic heat transport

Next Steps Finalize component model / CPL7 coupling Fully coupled simulations Evaluation of fully coupled model Multi-decadal retrospective and future climate simulations Long-term goals –Regional simulations for next IPCC report –Additional climate system components Ice sheets Biogeochemistry

1.Oceanic heat advection / storage has contributed significant forcing (>60%) to local sea ice melt during the last decade 2.Ice-edge & shelf/slope upwelling, eddies and other mesoscale circulation features in the Canada Basin provide a mechanism for horizontal heat distribution throughout the basin and up into the mixed layer 3.Oceanic heat accumulating in the western Arctic is potentially a critical initial factor in reducing ice concentration / thickness before & during the melt season 4.The rate of melt of sea ice volume possibly much greater than that of sea ice extent 5.A regional high-resolution Arctic Climate System Model can address these deficiencies and improve predictive skill of climate models Conclusions