Presentation on theme: "Daniela Flocco, Daniel Feltham, David Schr eder Centre for Polar Observation and Modelling University College London."— Presentation transcript:
Daniela Flocco, Daniel Feltham, David Schr eder Centre for Polar Observation and Modelling University College London
MODELS VS OBSERVATIONS Stroeve et al., 2007
ALBEDO AND ICE-ALBEDO FEEDBACK Global warming is intensified in polar regions due to the albedo feedback mechanism [IPCC, 2007]: since ice and snow are highly reflective compared to other surface types, a reduction in the ice or snow coverage due to melting results in enhanced absorption of solar radiation, which leads to further melt and local warming ( snow = 0.8, ocean = 0.1, grass = 0.3).
Using ERS satellite data, strong correlation is found between ice thickness and length of previous melt season (Laxon et al, Nature, 2003) The fraction of first year ice is increasing in time Pond area is larger over thin undeformed ice SEA ICE MASS BALANCE
MELT PONDS Photo Credit: ICESCAPE project - NASA/Kathryn Hansen July 4, 2010: Arctic sea ice and melt ponds in the Chukchi Sea.
MELT PONDS FORMATION Melt ponds form on Arctic sea ice during summer (rarely seen in Antarctic) when the snow/ice surface melts due to absorbed solar, short wave radiation and freeze in September Pond coverage ranges from 5% to 50% of the sea ice cover Albedo of pond-covered ice ( ) is lower than the albedo of bare sea ice or snow covered ice ( ) In terms of the heat budget of the Arctic, a change in mean albedo of 0.1 (e.g. a change in pond fraction of 20%) is equivalent to a change in sea ice extent of 10%. Underice refrozen ponds are insulated by an ice layer and take up to 2-3 months to freeze completely
GCM-COMPATIBLE MELT POND MODEL Requirements of constructing a melt pond model for use in existing GCMs places strong constraints on the form the model can take: main difficulty is that GCMs do not determine the sea ice topography. Modern GCMs contain a thickness distribution function g(h). Fractional Area Thickness g1g1 g2g2 g3g3 g4g4 g5g5 Flocco and Feltham, 2007
HEIGHT AND DEPTH DISTRIBUTION FUNCTIONS NEED: information about the surface height to redistribute surface water. ANSWER: We introduce surface height α(h) and basal depth β(h) distributions, which give the relative area of ice of a given surface height or basal depth. α β h sl NOTE: α(h) and β(h) are derived from the thickness distribution g(h) but do not describe the topography. Reference height
DELTA EDDINGTON and “TOPO” POND’S SCHEME “Retained” melt water fraction: a i for h i > 1 cm Refreezing: V p = V p e −0.005(−2−Tsfc) for T sfc < −2 ◦ Areaa p = (V p /0.8) 0.5 < 1 Depth h p = 0.8a p < 0.9h i Volume Vp = a p h p Snow a p =(1−h s /0.03)a p Rransport V p, a p Retained melt water fraction: a i for h i > 1 cm Reduce V p by fraction of ice area that ridged Melt water fills thinnest categories first, may overflow Snow occupies space according to ρ s. If the ice is permeable, pond can drain to sea level V ip = refrozen “lid” on pond newly forming: use F sfc thickening: use Stefan solution for melting: use meltt (Elizabeth Hunke talk at NCAR)
MELT POND AREA AND DEPTH RETRIEVAL snow pond A snow category i category i-1 category i-2 A pond (i-1) A bare ice (i+1) h full h surfl Flocco and Feltham (JGR, 2007)
DIFFERENTIAL ICE GROWTH Sea ice growth difference between the two pond’s schemes runs Cm/month RED: Sea ice growth greater in new scheme
POND AREA AND ICE FRACTION Ice thickness category Pond area IIIIIIVIV Pond area is larger over thin undeformed ice Sea ice concentration %
POND AREA JUN-SEPTEMBER 2007
SEA ICE EXTENT: SEPTEMBER 2007
CONCLUSIONS We study melt ponds to quantify their importance for the increased melting rate due to global warming Global Circulation Models are sensitive to the presence of a melt pond routine Initial runs suggest that incorporation of melt ponds might allow simulations of the 2007 minimum sea ice extent The melt pond routine is being incorporated into the latest version of CICE, which is the sea ice model used also by the UK Met Office