Using Machine Learning to Estimate Near-Surface Permafrost Extent at NGEE Arctic Sites on the Seward Peninsula in Alaska
Authors
Evan Thaler1* (thaler@lanl.gov), Sebastian Uhlemann2, Joel Rowland1, J. Schwenk1, Chen Wang2, Baptiste Dafflon2, Katrina Bennett1, Colleen Iversen3
Institutions
1Los Alamos National Laboratory, Los Alamos, NM; 2Lawrence Berkeley National Laboratory, Berkeley, CA; 3Oak Ridge National Laboratory, Oak Ridge, TN
URLs
Abstract
Permafrost soils are a critical component of the global carbon cycle and are locally important regulators of hydrologic fluxes from uplands to rivers. Furthermore, degradation of permafrost soils can lead to land surface subsidence, damaging crucial infrastructure for local communities. Regional and hemispherical maps of permafrost are too coarse to resolve distributions at a scale relevant to assessments of infrastructure stability or to illuminate geomorphic impacts of permafrost thaw. Here, researchers train 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 data set 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. However, transferrable machine learning models for estimating near-surface permafrost extent across ecosystems and disturbances histories will require additional ground truth datasets, which are currently underdeveloped.