Permafrost thaw study investigates landscape changes on Alaska’s Seward Peninsula

Feb 04, 2015
Researchers measure the spectral reflectance of tundra vegetation near Nome, Alaska, on the Seward Peninsula as part of larger mapping efforts to characterize climate-driven changes in Arctic landscapes. Left to right: Jennifer Liebig of the University of Tennessee, Santonu Goswami of Oak Ridge National Laboratory, and Guido Grosse of the Alfred Wegener Institute in Germany. 

Now in its third year, the Oak Ridge National Laboratory Climate Change Science Institute’s (CCSI’s) Model-Data Fusion project (as it is called for short) consolidates data from satellite imaging, remote sensing, and fieldwork in Arctic Alaska to assess the thawing of permafrost, or frozen soil, and its impact on carbon release and uptake. After 2 years of surveying thaw across the pan-Arctic, in 2014, the team focused on a particular region undergoing a lot of change. The Seward Peninsula on the Alaskan west coast has previously been known for its large, crater-like maar lakes created from hot magma erupting through cold groundwater, as well as its history as part of the Bering land bridge that once connected Asia to North America. Today, CCSI researchers are interested in the peninsula for its vulnerability to climate change.

Using remote sensors to determine the density of vegetation, the team observed that some of the Seward Peninsula’s trees, shrubs, and grasses are creeping northward and new lakes are appearing as permafrost thaws due to warming. Team members traveled to the peninsula several times in 2014 to corroborate satellite and sensing data with field observations. As they continue to monitor changes in the region’s landscape ecology, they are also using the information they gain from the peninsula and other Arctic locations to improve the representation of ecological processes that drive the carbon cycle in climate models. In particular, they are developing specialized parameters for different types of Arctic vegetation to avoid “green sponges” in climate models, or the uniform modeling of diverse plant processes and their unique contributions to the carbon cycle, which can lead to over- or underestimating future greenhouse gas concentrations in the atmosphere.

The official title of the Model-Data Fusion project is “Model-data Fusion Approaches for Retrospective and Predictive Assessment of the Pan-Arctic Scale Permafrost Carbon Feedback in Global Climate.”

Research contact: Daniel Hayes,