Remote Sensing–Based Mapping and Scaling of the Low Arctic Tundra Plant Functional Types and Classes from Field Plots to Peninsula
Authors
Jitendra Kumar1* ([email protected]), Venkata Shashank Konduri1,2, Amy L. Breen3, William W. Hargrove4, Forrest M. Hoffman5, Verity G. Salmon1, Colleen Iversen1, Auroop R. Ganguly6
Institutions
1Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN; 2National Ecological Observatory Network, Boulder, CO; 3International Arctic Research Center, University of Alaska–Fairbanks, AK; 4Eastern Forest Environmental Threat Assessment Center, U.S. Department of Agriculture Forest Service, Asheville, NC; 5Computational Sciences and Engineering Division, Oak Ridge National Laboratory, Oak Ridge, TN; 6Department of Civil and Environmental Engineering, Northeastern University–Boston, MA
URLs
Abstract
High-resolution maps of vegetation distribution are crucial for studying the impacts of climate change on Arctic tundra ecosystems and their associated feedbacks. Bridging scales and integrating field-scale observations, airborne remote sensing and environmental datasets, researchers studied the current distribution and potential future changes in vegetation cover class and plant functional type covers across the low Arctic region of northwestern Alaska. Leveraging airborne hyperspectral imagery collected by NASA’s Arctic-Boreal Vulnerability Experiment (ABOVE), the research team developed deep neural network-based models, trained using field vegetation surveys conducted at NGEE-Arctic intensive sites at Seward Peninsula, to develop a high resolution (5 m) landscape scale vegetation class maps with accuracy exceeding 93%. Analysis shows that the low- and mid-elevation areas on the landscape are dominated by shrubs and graminoid, while the high elevation areas are often dominated by dryas-lichen and nonvegetated classes. The team also developed random forest-based multivariate environmental niche models to determine habitat suitability of vegetation classes and further upscale the vegetation class mapping to regional scale across southern Seward Peninsula. Using future projections from five Earth system models (RCP8.5, CMIP5), researchers developed projections of vegetation cover classes in future decades up to 2050. Results show a substantial increase in shrub-dominated vegetation classes, especially along hillslope and high-resource environments like floodplains and stream corridors. Using plot observations of plant functional type cover, researchers have developed model-ready data of fractional plant functional types at various resolutions to parameterize regional scale simulations of the ELM-Arctic model.