Arctic climate simulations by coupled models an overview

Arctic climate simulations by coupled models - an overview Annette Rinke and Klaus Dethloff Alfred Wegener Institute for Polar and Marine Research, Research Department Potsdam, Germany DSAR konference 9. november 2000 1

Surface temperature anomalies in 1890 -2000 Observation Experiment 3 Large internal variability of the coupled atmosphere-ocean system Experiment 5 Experiment 4 To what extent is polar warming amplification attributed to real physical processes rather than to model imperfections? Delworth and Knutson, 2000 [K] Anomalies relative to 1961 -90 climatology

· Global Coupled Models (AOGCMs) - AOGCMs performance in the Arctic (seasonal cycle, interannual & decadal variability) · Regional Models (RCMs) - atmospheric RCMs performance in the Arctic (seasonal cycle, interannual variability) - coupled RCMs for the Arctic (case studies) · Outlook

(1) Annual cycle of surface air temperature models temperature observation poleward 70 o. N, excluding land 8 coupled models from IPCC/DDC; „control“ 1961 -90 Walsh et al. , 2002

(3) Decadal variability AO Pattern and its temporal variability Data (NCEP, 1948 -2001) AOGCM (ECHO-G, 1000 yrs) Dominant spatial pattern z 500, NH, DJF Variability of dominant pattern Handorf et al. , 2002

AOGCM summary § Reasonable representation of mean state and variability by the ensemble, but considerable across-model scatter § Biases in Arctic climate from an Arctic perspective: systematic differences in key variables (SLP, clouds, sea ice) influence of global climate on Arctic & vice versa development of Arctic specific parameterizations (PBL, clouds, permafrost, …) § Resolution (200 -300 km horiz. , few-tens of vertical levels) limits the ability to capture important aspects of climate (e. g. , topographic effects, storms, sea ice-atmosphereinteraction) higher resolution

Regional climate model (RCM) method GCM (or observation-based analyses) RCM Initial & time-dependent boundary conditions for the RCM provided by GCM

GCM (T 30, 3. 75 o) RCM (0. 5 o) Courtesy W. Dorn

(1) Annual cycle of surface air temperature observation Temperature [o. C] averaged over model domain model (Period: 1979 -93, RCM: HIRHAM)

[K] HIRHAM 1979 -93

Arctic Regional Climate Model Intercomparison Project (ARCMIP) Participating Models 1. ARCSy. M (USA) 2. COAMPS (S) 3. HIRHAM (D, DK) 4. NARCM (CAN) 5. RCA (S) 6. Reg. CM (N) 7. REMO (D) 8. Polar. MM 5 (USA) Experimental set-up § Same horizontal resolution & boundary conditions § Different dynamics & physics § Simulation during SHEBA year (Sept 1997 -Sept 1998) Same domain § Beaufort Sea & pan-Arctic http: //paos. colorado. edu/~currja/arcmip/index. html

Different domains allows elucidation of the interaction of the parameterized processes with the atmospheric dynamics influence of resolution Different boundary conditions separate errors associated with - lateral boundary advection - interaction with ice/ocean surface

ARCMIP- Results: 850 h. Pa temperature May 1998 Across-model std dev [o. C] [K]

ARCMIP- Results: Temporal development of the vertical atmospheric structure January 1998

Anomalous sea ice retreat in Siberian Seas during summer 1990 Observation Coupled Regional Models HIRHAM-MOM August 1990 Sea ice concentration ARCSy. M Maslanik et al. , 2000 Rinke et al. , 2003

Atmospheric circulation, August 1990 - Mean sea level pressure - Atmosphere-alone with satellite sst/ice Coupled regional models HIRHAM-MOM H ARCSy. M H H L L L HIRHAM L L H L L Models L Observation L Maslanik et al. , 2000 Rinke et al. , 2003

RCM summary § Added value due to downscaling compared with GCM output RCMs improve (should we expect to): § reduction of mean bias § better spatial variability § more realistic variance § better tail behaviour (i. e. , extremes) § Importance of synoptic-scale processes in simulating strong regional variability of sea ice cover RCM‘s problems: § large-scale errors of driving model § nesting technique

Outlook Model development § going to finer horizontal and vertical resolutions § Arctic specific parameterizations (surf. albedo, clouds, PBL) § extensive ensemble integrations § include more components of the climate system § combined use of AOGCMs and RCMs EU project “Global Implications of Arctic Climate Processes & Feedbacks” Understanding § natural climate variability on multiple scales in space & time § atmosphere-ocean-ice-land interactions on regional scale § interplay between Arctic regional climate feedbacks & global circulation patterns