Glacial lakes are expanding globally due to climate change, particularly in Patagonia, Alaska, the Himalayas, and Greenland. These lakes influence glacier dynamics and mass loss, with freshwater calving glaciers experiencing more rapid retreat than land-terminating glaciers. The expansion of glacial lakes also increases the risk of glacial lake outburst floods (GLOFs), which can cause significant damage and loss of life. GLOFs occur when a dam (moraine, glacier ice, or bedrock) is breached, often triggered by events like calving, landslides, or rockfalls. Understanding outburst mechanisms and monitoring glacial lakes are crucial for hazard management. Satellite remote sensing offers a powerful tool for monitoring glacial lakes, enabling mapping, area change measurement, and lake level assessment using altimetry and DEMs. While studies have used GRACE/GRACE-FO data to quantify water mass changes in ordinary lakes, its application to GLOFs has been limited. This study focuses on the abrupt drainage of Lago Greve, a large proglacial lake in Patagonia, using multiple satellite datasets to analyze the event's characteristics and triggering mechanism.

The increase in glacial lake number and size due to glacier retreat has been well documented globally (Shugar et al., 2020). Studies have focused on the impact of glacial lakes on ice dynamics (Truffer & Motyka, 2016; Tsutaki et al., 2019; Sugiyama et al., 2019; Carrivick et al., 2020) and the accelerated retreat of calving glaciers (King et al., 2019; Minowa et al., 2021; Abdel Jaber et al., 2019; Dussaillant et al., 2018; Sakakibara & Sugiyama, 2014; Zemp et al., 2019). The influence of glacial meltwater on lake ecosystems has also been examined (Sommaruga, 2015; Tiberti et al., 2020; Sugiyama et al., 2016, 2021; Carrivick & Tweed, 2013). GLOFs are a significant hazard (Meerhoff et al., 2019; Glasser et al., 2016; Thorndycraft et al., 2019), with various triggering mechanisms identified (Neupene et al., 2019; Hubbard et al., 2005; Westby et al., 2014; Sugiyama et al., 2008; Emmer & Cochachin, 2013; Emmer et al., 2022; Veh et al., 2020; Carrivick & Tweed, 2016). Satellite remote sensing has been employed to map lakes and quantify changes (Nie et al., 2017; Xu et al., 2021; Jiang et al., 2017; Grinsted et al., 2017; Carabajai & Boy, 2021; Pereira et al., 2021; Yi et al., 2017), but comprehensive analysis of GLOFs using GRACE data is still lacking.

The study used multiple satellite datasets to analyze the Lago Greve outburst. Lake area was measured from Sentinel-2 MSI and Landsat 8 OLI images (19 images total, from September 2016 to November 2020), with the lake margin manually delineated using false-color composite images and NDWI to handle shaded areas. Uncertainty was assessed by randomly shifting the lake polygon nodes within the image resolution. Glacier area changes were also measured. Sentinel-1 SAR images were used to constrain the event's onset. Lake water level was measured from ICESat-2/ATLAS L3A Inland Water Surface Height products (from November 2018 to July 2020) and DEMs (SRTM-DEM and a photogrammetrically derived 5-m resolution WV-DEM from July 2020). Vertical offsets between DEMs and altimetry data were corrected. The onset of drainage was further investigated using Sentinel-1 backscatter images. GRACE/GRACE-FO mascon solutions from CSR were analyzed to detect the gravity field changes associated with the outburst. A model incorporating linear, seasonal, and decadal variations was fitted to the mascon time series, and additional mass loss due to the outburst was estimated. The spatial distribution of mascon changes was also analyzed to identify the location of mass loss.

The Lago Greve outburst occurred between April 9th and 19th, 2020, resulting in significant lake area reduction (14.5 ± 0.02 km²) and water level drop (18.3 ± 1.2 m). The total water discharge was 3.7 ± 0.2 km³, ranking among the largest GLOFs ever recorded. Satellite images revealed that a bump near the lake outlet collapsed, altering the outlet stream's flow path and causing erosion of the northern bank and stream bed. This deepened the valley and facilitated lake drainage. The erosion likely affected unconsolidated glacial deposits. While the trigger for the bump's collapse remains unclear, it is speculated to be a relatively rapid event unrelated to immediate climate change, but possibly a delayed response to the warming climate and retreat of Glaciar Occidental around 1870. GRACE data detected a signal consistent with the outburst, although the estimated mass loss was significantly larger than the actual water discharge due to the limited spatial resolution of GRACE gravimetry. The study highlighted the importance of higher resolution gravity data for accurate quantification of GLOF events.

The findings demonstrate the power of multi-sensor satellite observations in detecting and characterizing large GLOFs. The abrupt drainage of Lago Greve, triggered by a seemingly minor event (the bump collapse), highlights the vulnerability of large glacial lakes. The discrepancy between the GRACE-estimated mass loss and the actual water discharge emphasizes the limitations of using GRACE data for quantifying smaller-scale events like GLOFs from individual lakes. The impact of the lake level drop on the glaciers feeding the lake, such as influencing ice speed and calving frequency, should be further investigated. This event highlights the necessity for continued monitoring of glacial lakes in vulnerable regions.

This study documents a significant GLOF event at Lago Greve using multiple satellite datasets, providing insights into the event's magnitude, triggering mechanism, and detection via GRACE. Future work should focus on identifying the cause of the bump collapse, further analyzing the impact on the glaciers, and investigating the use of high-resolution gravity data for precise quantification of such events. Continued monitoring of similar glacial lakes is crucial for improved GLOF hazard assessment and risk management.

The triggering mechanism of the bump collapse remains speculative. While the study utilizes various datasets, a ground-based field survey would provide crucial corroborating information and resolve ambiguities. The use of GRACE data, although detecting the mass loss, was limited by its spatial resolution, causing an overestimation of the water discharge volume. More precise methods are needed to effectively quantify water volume changes for smaller scale events.