Greenland’s snowline exhibits large fluctuations and is a primary amplifier of ice sheet surface melt and runoff.

Recent deployments of CubeSat imagers by companies such as Planet may advance hydrological remote sensing by providing an unprecedented combination of high temporal and high spatial resolution imagery at the global scale. With approximately 170 CubeSats orbiting at full operational capacity, the Planet CubeSat constellation currently offers an average revisit time of <1 day for the Arctic and near-daily revisit time globally at 3 m spatial resolution. Such data have numerous potential applications for water resource monitoring, hydrologic modeling and hydrologic research. Here we evaluate Planet CubeSat imaging capabilities and potential scientific utility for surface water studies in the Yukon Flats, a large sub-Arctic wetland in north central Alaska. We find that surface water areas delineated from Planet imagery have a normalized root mean square error (NRMSE) of <11% and geolocation accuracy of <10 m as compared with manual delineations from high resolution (0.3–0.5 m) WorldView-2/3 panchromatic satellite imagery. For a 625 km² subarea of the Yukon Flats, our time series analysis reveals that roughly one quarter of 268 lakes analyzed responded to changes in Yukon River discharge over the period 23 June–1 October 2016, one half steadily contracted, and one quarter remained unchanged. The spatial pattern of observed lake changes is heterogeneous. While connections to Yukon River control the hydrologically connected lakes, the behavior of other lakes is complex, likely driven by a combination of intricate flow paths, underlying geology and permafrost. Limitations of Planet CubeSat imagery include a lack of an automated cloud mask, geolocation inaccuracies, and inconsistent radiometric calibration across multiple platforms. Although these challenges must be addressed before Planet CubeSat imagery can achieve its full potential for large-scale hydrologic research, we conclude that CubeSat imagery offers a powerful new tool for the study and monitoring of dynamic surface water bodies.

Fine‐scale, subseasonal fluctuations in Arctic‐Boreal surface water reflect regional water balance and modulate trace gas emissions to the atmosphere but have eluded detection using traditional satellite remote sensing. We use high‐resolution (~3–5 m), high‐frequency CubeSat sensors to measure near‐daily changes in lake surface area through an object‐based tracking method that incorporates machine learning to overcome notable limitations of CubeSat imagery. From ~76,000 images we obtain >2.2 million individual observations of changing surface areas for 85,358 lakes in Northern Canada and Alaska between 1 May and 1 October 2017. We find broad‐scale lake area declines across diverse climatic, hydrologic, and physiographic terrains. Localized exceptions reveal lowland flooding and aquatic vegetation phenology cycles. Cumulative small shoreline changes of abundant lakes on the Canadian Shield exceed total inundation variations of better‐studied lowland environments, revealing a surprisingly dynamic landscape with respect to subseasonal variations in surface water extent and trace gas emissions.

Small water bodies (i.e., ponds; 0.01 km ² ). We leverage high‐resolution (3 m) optical satellite imagery from Planet Labs and deep learning methods to map seasonal changes in pond and lake areal extent across four regions in Alaska. Our water maps indicate that changes in open water extent over the snow‐free season are especially pronounced in ponds. To investigate potential impacts of seasonal changes in pond area on carbon emissions, we provide a case study of open water methane emission budgets using the new water maps. Our approach has widespread applications for water resources, habitat and land cover change assessments, wildlife management, risk assessments, and other biogeochemical modeling efforts.