High-Resolution Mapping of Near-Surface Permafrost

The Science

Permafrost soils are a key component of Arctic and sub-Arctic ecosystems and the global carbon cycle. As Arctic climates warm, permafrost thaw has the potential to release large quantities of carbon into the atmosphere, further increasing warming. Moreover, permafrost thawing can cause rapid ground surface subsidence, which can severely damage infrastructure.

Current maps of predicted permafrost extent are too coarse to adequately evaluate either the potential contribution of thaw to the atmospheric carbon or infrastructure vulnerability. A team of researchers generated new high-resolution maps of permafrost extent using machine learning. These maps and algorithms provide a future direction for generating policy-relevant maps of permafrost extent.

The Impact

Permafrost extent is heterogeneous over spatial scales too fine to be accurately predicted by the available coarse-resolution map products. Researchers developed new high-resolution maps of permafrost extent for areas on the Alaskan Seward Peninsula using machine learning algorithms that incorporated geophysical data and remote sensing. These new maps indicate the potential for using machine learning and high-resolution field and remote sensing data to generate spatial predictions of permafrost at scales relevant to land managers and policy-makers.

Summary

Permafrost soils are a critical component of the global carbon cycle and are locally important because they regulate the hydrologic flux from uplands to rivers. Furthermore, degradation of permafrost soils causes land surface subsidence, damaging crucial infrastructure for local communities. Regional and hemispherical permafrost maps are too coarse to resolve distributions at a scale relevant to assessments of infrastructure stability or to illuminate geomorphic impacts of permafrost thaw.

A team of researchers trained machine learning models to generate meter-scale maps of near-surface permafrost for three watersheds in the discontinuous permafrost region. The models were trained using ground truth determinations of near-surface permafrost presence from measurements of soil temperature and electrical resistivity.

The team trained three classifiers: extremely randomized trees (ERTr), support vector machines (SVM), and an artificial neural network (ANN). Model uncertainty was determined using k-fold cross-validation, and the modeled extents of near-surface permafrost were compared to the observed extents at each site.

At-a-site near-surface permafrost distributions predicted by the ERTr produced the highest accuracy (70% to 90%). However, the transferability of the ERTr to sites outside of the training dataset was poor, with accuracies ranging from 50% to 77%. The SVM and ANN models had lower accuracies for at-a-site prediction (70% to 83%), yet they had greater accuracy when transferred to the nontraining site (62% to 78%).

These models demonstrate the potential for integrating high-resolution spatial data and machine learning models to develop maps of near-surface permafrost extent at resolutions fine enough to assess infrastructure vulnerability and landscape morphology influenced by permafrost thaw.

Principal Investigator

Katrina Bennett
Los Alamos National Laboratory
[email protected]

Program Manager

Daniel Stover
U.S. Department of Energy, Biological and Environmental Research (SC-33)
Environmental System Science
[email protected]

Funding

This research was supported by the Biological and Environmental Research program in the U.S. Department of Energy’s Office of Science, as a contribution to the Next-Generation Ecosystem Experiments Arctic project.

Related Links

• ESS-DIVE Dataset

References