Unit Affiliation: Ocean and Climate Physics, Lamont-Doherty Earth Observatory (LDEO)
Modeling C fluxes in the Arctic and predicting them into a warmer future has been one of the greatest challenges to understanding the Arctic carbon cycle and is a large focus of recent research efforts, including NASA Arctic-Boreal Vulnerability Experiment (ABoVE). Predicting whether C becomes CH4 or CO2 is critical to understanding feedbacks to climate warming, as CH4 has 28-times the global warming potential of CO2. Top-down scaling using remote sensing and airborne CO2 campaigns and bottom-up scaling of field-based CO2 flux measurements both show the Arctic is becoming a C source, though most process-based models show the Arctic yet remains a C sink. Biogeochemical hotspots—areas of abnormally high activity—tend to be found at boundaries between contrasting terrestrial and aquatic ecosystems where resources mix. Wetlands and small ponds are often hotspots of C fluxes, driven by high dissolved CO2 and CH4 concentrations, and are recognized as the dominant sources of CH4 emissions in the Arctic. Uncertainties in the total area of wetlands, as well as the double counting of small lake area emissions again as wetland area emissions, have led to a discrepancy between bottom-up and top-down estimates of CH4 budgets for the Arctic. Small lakes and wetlands can also be significant sources of CO2; however, current interpretations of airborne CO2 fluxes have ignored lake and wetland contributions. The low resolution of most commonly-used satellite products and long-term records cannot detect small lakes and wetlands. The subarctic tundra of the Yukon-Kuskokwim Delta (YKD) of Alaska is a model ecosystem for testing scaling relationships from landscape-level processes to regional emissions. This study will investigate patterns of C fluxes, permafrost thaw, and hydrologic connectivity. Evaluating the impacts of climate change on C emissions is particularly relevant in the YKD given its preeminent role in Alaska’s CH4 budget and the vulnerability of permafrost in the YKD to thawing. This study will demonstrate the utility of high-resolution imagery, landscape classification, and indices of connectivity to capture the changes in surface and subsurface hydrology in permafrost ecosystems, which may be applied in other permafrost terrains in Alaska. This research is relevant to NASA ABoVE, and addresses the SMD focus to “Detect and predict changes in Earth’s ecosystems and biogeochemical cycles, including land cover, biodiversity, and the global carbon cycle”. The objectives are to use multi-scale satellite imagery to produce landscape classifications of wetland vegetation types and surface water for the YKD at high resolutions; quantify and scale functional relationships of wetland and inland aquatic C that incorporate biogeochemical mechanisms, hydrologic connectivity in contributing watersheds, and water-body shape and size; produce a timeseries for the YKD from 2012-2018 using NASA airborne C flux data; identify the relevant scale needed to accurately quantify C fluxes from wetlands and terrestrial-aquatic interfaces. Wetland, vegetation, and surface water maps will be produced from unsupervised classifications using the k-means algorithm and input layers from the 5m-ArcticDEM, WorldView-2, Sentinel-2, and Landsat imagery. Inland aquatic C will be modeled using machine learning techniques from environmental, biogeochemical, geospatial data. I will use the WRF-STILT model to determine spatially explicit airborne C fluxes for the YKD and seasonal and annual budgets. I will partition the sources of airborne C fluxes using scaled-up eddy tower fluxes to compare the contributions of inland waters, vegetated wetlands, and terrestrial components. This research will enable us to identify hot spots of C emissions, and better anticipate changes to the permafrost C feedback in a warming climate.