Dań Zhùr (Donjek) Glacier, located in the St. Elias Mountains, Yukon, is a surge-type glacier that undergoes cyclical periods of rapid advance over a period of ~1-2 years, followed by retreat for a period of ~10 years (Kochtitzky et al., 2019). Dań Zhùr Chù’ (Donjek River) runs perpendicular to the terminus of the glacier and past surges have, at times, caused the terminus to advance enough to block the river, leading to the formation of an ice-dammed lake (Kochtitzky et al., 2020). The glacier most recently surged between 2012 and 2014, and since then Dań Zhùr Lake has drained three times: in 2017, 2018 and 2019. When a glacier dam fails, the drainage of the lake can be catastrophic and cause flooding downstream. In the case of Dań Zhùr Lake, the most recent drainage event occurred on July 13th, 2019, when the ~2.45 km2 lake drained in less than 36 hours and created an ice canyon through the glacier terminus. Time-lapse cameras and pressure sensors were used to capture the drainage event, and air photos taken during the melt season (June and September) were used to construct digital elevation models (DEMs) of the glacier terminus, lake, and lake basin. The method of drainage for the 2019 event was determined to be flotation of the terminus, leading to rapid subglacial drainage of the lake. There were also noticeable changes in water extent downstream during the 2019 event, meaning that there is a potential risk to downstream recreational users. Because of the formation of a large ice canyon after the 2019 drainage, it is very unlikely that that the lake will reform until the next surge, which is anticipated to occur around 2024. Following that surge, the size of Dań Zhùr Lake is expected to increase during the next quiescent phase, as the continued glacier recession will expose a larger basin for the lake to form in, and flotation will continue to be a likely mechanism for drainage. However, in the long term it is unlikely that ice-dammed lakes will continue to form at Dań Zhùr Glacier, as there is a trend of the maximum terminus extent during each surge being smaller than the previous one, meaning that the glacier will no longer block the flow of Dań Zhùr Chù’.

Glacial-dominated areas pose unique challenges to downstream communities in adapting to recent and continuing global climate change, including increased threats of glacial lake outburst floods (GLOFs) that have substantial impacts on regional social, environmental and economic systems increasing risk due to flooding of downstream communities. In this dissertation, two lakes with potential to generate GLOFs were studied, Imja Lake in Nepal and Palcacocha Lake in Peru. At Imja Lake, basic data was generated that allowed the creation of a conceptual model of the lake. Ground penetrating radar and bathymetric surveys were performed. Also, an inundation model was developed in order to evaluate the effectiveness of a project that seeks to reduce flooding risk by lowering the lake at least 3 meters. In Peru, a GLOF inundation model was created. Also, the vulnerability of the people living downstream in the City of Huaraz was calculated, and the impacts of an early warning system were evaluated. The results at Imja indicated that the lake deepened from 98 m in 2002 to 116 m in 2012. Likewise, the lake volume increased from 35.8 to 61.6±1.8 million m3 over the past decade. The GPR survey at Imja and Lhotse-Shar glaciers shows that the glacier is over 200 m thick in the center of the glacier. The modeling work at Imja shows that the proposed project will not have major impacts downstream since the area inundated does not reduce considerably unless the lake is lowered by about 20 m. In Huaraz, the results indicate that approximately 40646 people live in the potentially inundated area. Using the flow simulation and the Peru Census 2007, a map of vulnerability was generated indicating that the most vulnerable areas are near the river. Finally, the potential number of fatalities in a worst case GLOF scenario from Lake Palcacocha was calculated to be 19773 with a standard deviation of 1191 if there is no early warning system and 7344 with a standard deviation of 1446 people if an early warning system is installed. Finally, if evacuation measures are improved the number reduces to 2865 with a standard deviation of 462. / text