21st Conference on Climate Variability and Change


Relationship between sea Ice, land surface temperature, and vegetation in the Arctic coastal zone

Uma Bhatt, University of Alaska, Fairbanks, AK; and D. Walker, M. Raynolds, and J. Comiso

Recent dramatic reductions in sea ice and changes in Arctic vegetation have been well documented and are of growing concern because of how they may impact the ecosystem at high latitudes, which includes permafrost, soils, fauna, as well as humans. It is hypothesized that an earlier ice melt forces atmospheric and landsurface temperature changes, leading to increased summer warmth, higher NDVI and enhanced greenness of vegetation. To investigate the nature of these relationships, climate analysis techniques are applied to high-resolution passive microwave sea ice concentration and AVHRR land surface temperatures to evaluate the direct relationship between coastal ice and the adjacent land. The analysis employs 25 km resolution SSMI passive microwave Sea Ice Concentration (Comiso 1999) and AVHRR Surface Temperature (Comiso 2006, 2003) covering the 24-year period from January 1982 to December 2007. The spatial variations of the climate-vegetation relationships are examined by performing analysis for various regions as delineated by Treshnikov (1985).

Pan-arctic trends and variations in sea ice concentration, NDVI, and surface temperature were evaluated in contiguous 50-km strips of land and ocean. The 50km land surface zone represents 60% of the tundra and while the 50-km ocean region is the key oceanic domain that impacts the nearby land surface. At the pan-arctic scale, Summer Warmth Index (SWI, number of degree days over freezing during summer) and NDVI show statistically significant increases while sea ice show significant decreases. On the continental scale, NDVI displays larger positive trends in North American than in Eurasia, while sea ice decreases have been largest in the Laptev and E. Siberian Seas. The Arctic was divided into regions based on the Treshnikov divisions and there is substantial heterogeneity in trends and variability across the regions. In general SWI and integrated NDVI are significantly positively correlated, sea ice and SWI are moderately negatively correlated and sea ice is not correlated with integrated NDVI. There is notable co-variability between the ocean (through sea ice concentration) and nearby land (through SWI) but the exact causality is not currently understood. The largest trends for SWI and NDVI are in the Chukchi/Beaufort region, where the ocean has warmed the most. We have found a strong relationship between SWI and sea ice area during the preceding spring. Since both are strongly correlated with the large-scale circulation it cannot be concluded that the ice directly forces the warmer summers.


Comiso, J. C., 2006: Arctic warming signals from satellite observations. Weather, 61, 70- 76.

Comiso, J. 2003: Warming Trends in the Arctic from Clear Sky Satellite Observations, J. Climate, 16, 3498-3510.

Comiso, J. 1999, updated 2005: Bootstrap sea ice concentrations for NIMBUS-7 SMMR and DMSP SSM/I, June to September 2001. Boulder, CO, USA: National Snow and Ice Data Center. Digital media.

Treshnikov, A. F., Atlas of the Arctic, Moscow, 204 pp., 1985. (in Russian).

Session 7A, Global dynamics and processes - II
Tuesday, 13 January 2009, 3:30 PM-5:45 PM, Room 129A

Previous paper  Next paper

Browse or search entire meeting

AMS Home Page