Quantifying Greenhouse Gas Emissions from Small Arctic Lakes Using High-Resolution Remote Sensing and Thermal-Biogeochemical Modeling in the Yukon Kuskokwim Delta, Alaska

Water bodies play a critical role in the carbon dynamics of the Northern high latitudes. The rapid warming of Arctic and sub-Arctic regions is driving significant changes in the hydrology of these ecosystems, which in turn influences greenhouse gas (GHG) feedbacks. Small lakes (less than 0.1 km²), which are hydrologically sensitive and strong emitters of methane, are particularly important but remain poorly quantified contributors to overall GHG emissions. This uncertainty is driven by the limited resolution of remote sensing imagery, the high spatiotemporal variability in GHG emissions, and the lack of extensive field data. This study addresses these gaps by using high-resolution (3.5 m) satellite-derived lake maps and a fully coupled thermal-biogeochemical lake model to estimate GHG emissions from lakes in the Yukon Kuskokwim Delta, Alaska. The water maps were produced using cubesat imagery combined with a convolutional neural network (U-Net) to inform a series of LAKE model simulations. The model was calibrated using three years of biogeochemical data, allowing for estimates of both discrete and cumulative CH4 and CO2 fluxes based on the high-resolution water body distribution. Additionally, our work incorporates variability in lake morphometry, biogeochemistry, and terrestrial inputs to refine GHG emission estimates. The findings offer constrained regional estimates of lake GHG emissions and identify key areas where additional data collection is essential.