Using High-resolution imaging satellite constellations to understand glacier mass balance and dynamics in High Mountain Asia

Abstract

Meltwater from glaciers in High Mountain Asia (HMA) is an essential source of dischargefor major Asian rivers, which sustain one of the most densely populated regions on Earth. HMA glaciers are in a state of overall decline, with models projecting that this decline will continue in the coming decades as the Earth warms. An improved understanding of past, present, and future glacier change in HMA is crucial for informing ongoing and future policy decisions. While recent efforts have documented the long-term to interannual evolution of HMA glacier mass balance and ice flow, these records do not capture local, fine-scale processes operating on shorter timescales (e.g., surface debris-cover evolution, basal sliding), and their effect on glacier mass balance and dynamics. Precise geodetic measurements of 3D surface displacement from high-resolution satelliteimagery can be used to study glacier motion and associated volume changes over time. The dissertation outlines several new methods to derive geodetic products from commercial satellite imaging constellations (Planet Labs, Maxar). We use these methods to generate new data products that document HMA glacier change with unprecedented spatial and temporal resolution. In the second chapter, I developed automated workflows for multi-view stereo cameramodel refinement and image geolocation corrections for the Planet Labs SkySat-C constel- lation. The refined camera models are used to generate accurate and self-consistent DEM products at 2 m resolution and orthoimages at the native 0.7-0.9 m resolution. I analyzed sample DEM products for a triplet stereo collection over Mt. Rainier, USA, and two video collections over Mt. St. Helens, USA. The output DEM products display ≤ 1 m relative and ≤ 2 to 3 m absolute vertical accuracy, which easily captures the 5 to 15 m of elevation change signals over Mt. St. Helens due to the melting of seasonal snow and glacier flow. In the third chapter, I used very-high-resolution images acquired by the Maxar constella-tion to prepare a time series of contemporaneous DEM and surface velocity products for six debris-covered glaciers in Nepal. I combined these measurements with publicly available ice thickness estimates to produce flow-corrected Lagrangian surface mass balance (SMB) maps and used these maps to study local surface melt signals. Our results show melt reduction under thick debris cover and enhanced ablation over ice cliffs. These ice cliffs were responsible for 17 to 43% of the total ablation over debris-covered areas, even though they occupied ≤ 11% of the entire area. Our seasonal SMB products reveal the timing and patterns of summer accumulation and ablation, underscoring the importance of snow avalanches for low elevation debris-covered glaciers in the region. Finally, the fourth chapter outlines new techniques to derive monthly to seasonal surfacevelocity time series for HMA glaciers using near-daily, high-resolution (∼ 3 m) optical imagery from the PlanetScope constellation. To address the geolocation issues associated with Planet data, I combined 1000s of noisy velocity maps derived from individual image pairs using robust spatiotemporal statistics to obtain systematic, analysis-ready monthly velocity measurements. Our results offer detailed documentation of the surge front propagation and seasonal variability for the Gando Glacier during the 2018-2021 surge event. We also observe previously undetected seasonal velocity variations over the slow-moving, heavily debris- covered Ngozumpa and Khumbu Glaciers, including the Khumbu Icefall. Our observations indicate that peak glacier velocities occur during the summer and monsoon months, with a summer speedup of ~67% observed over the main, fast-flowing western tributary of Ngozumpa Glacier. These methods and datasets offer new insights on the seasonal evolution of subglacial hydrology, basal sliding, meltwater storage, and glacier sensitivity to future change.