Supraglacial lakes are an important component of the Greenland Ice Sheet's mass balance and hydrology, with their drainage affecting ice dynamics. This study uses imagery from the recently launched Sentinel-1A Synthetic Aperture Radar (SAR) satellite to investigate supraglacial lakes in West Greenland. A semi-automated algorithm is developed to detect surface lakes from Sentinel-1 images during the 2015 summer. A combined Landsat-8 and Sentinel-1 dataset, which has a comparable temporal resolution to MODIS (3 days vs. daily) but a higher spatial resolution (25-40 vs. 250-500 m), is then used together with a fully automated lake drainage detection algorithm. Rapid (<4 days) and slow (>4 days) drainages are investigated for both small (<0.125 km 2 , the minimum size detectable by MODIS) and large (≥0.125 km 2 ) lakes through the summer. Drainage events of small lakes occur at lower elevations (mean 159 m), and slightly earlier (mean 4.5 days) in the melt season than those of large lakes. The analysis is extended manually into the early winter to calculate the dates and elevations of lake freeze-through more precisely than is possible with optical imagery (mean 30 August; 1,270 m mean elevation). Finally, the Sentinel-1 imagery is used to detect subsurface lakes and, for the first time, their dates of appearance and freeze-through (mean 9 August and 7 October, respectively). These subsurface lakes occur at higher elevations than the surface lakes detected in this study (mean 1,593 and 1,185 m, respectively). Sentinel-1 imagery therefore provides great potential for tracking melting, water movement and freezing within both the firn zone and ablation area of the Greenland Ice Sheet.

Abstract. Surface lakes on the Greenland Ice Sheet play a key role in its surface mass balance, hydrology and biogeochemistry. They often drain rapidly in the summer via hydrofracture, which delivers lake water to the ice sheet base over timescales of hours to days and then can allow meltwater to reach the base for the rest of the summer. Rapid lake drainage, therefore, influences subglacial drainage evolution; water pressures; ice flow; biogeochemical activity; and ultimately the delivery of water, sediments and nutrients to the ocean. It has generally been assumed that rapid lake drainage events are confined to the summer, as this is typically when observations are made using satellite optical imagery. Here we develop a method to quantify backscatter changes in satellite radar imagery, which we use to document the drainage of six different lakes during three winters (2014/15, 2015/16 and 2016/17) in fast-flowing parts of the Greenland Ice Sheet. Analysis of optical imagery from before and after the three winters supports the radar-based evidence for winter lake drainage events and also provides estimates of lake drainage volumes, which range between 0.000046 ± 0.000017 and 0.0200 ± 0.002817 km3. For three of the events, optical imagery allows repeat photoclinometry (shape from shading) calculations to be made showing mean vertical collapse of the lake surfaces ranging between 1.21 ± 1.61 and 7.25 ± 1.61 m and drainage volumes of 0.002 ± 0.002968 to 0.044 ± 0.009858 km3. For one of these three, time-stamped ArcticDEM strips allow for DEM differencing, which demonstrates a mean collapse depth of 2.17 ± 0.28 m across the lake area. The findings show that lake drainage can occur in the winter in the absence of active surface melt and notable ice flow acceleration, which may have important implications for subglacial hydrology and biogeochemical processes.

Abstract. Surface lakes on the Greenland Ice Sheet play a key role in its surface mass balance, hydrology, and biogeochemistry. They often drain rapidly in the summer via hydrofracture, which immediately delivers lake water to the ice sheet base over timescales of hours to days and then allows meltwater to reach the base for the rest of the summer. Rapid lake drainage, therefore, influences subglacial drainage evolution, water pressures, ice flow, biogeochemical activity, and ultimately the delivery of water, sediments and nutrients to the ocean. It is assumed that rapid lake drainage events are confined to the summer, as this is when all observations to date have been made. Here we develop a method to quantify backscatter changes in satellite radar imagery, which we use to document the drainage of six different lakes during three winters in fast flowing parts of the Greenland Ice Sheet. Analysis of optical imagery from before and after the three winters supports the radar-based evidence for winter lake drainage events and also provides estimates of lake drainage volumes, which range between 0.000046 and 0.0202 km3. For three of the events, optical imagery allows photoclinometry (shape from shading) calculations to be made showing mean vertical collapse of the lake surfaces ranging between 4.04 m and 7.25 m, and drainage volumes of 0.004 km3 to 0.049 km3. The findings show that background winter ice motion can trigger rapid lake drainage, which may have important implications for subglacial hydrology and biogeochemical processes.

We present a newly developed 1-D numerical energy-balance and phase transition supraglacial lake model: GlacierLake. GlacierLake incorporates snowfall, in situ snow and ice melt, incoming water from the surrounding catchment, ice lid formation, basal freeze-up and thermal stratification. Snow cover and temperature are varied to test lake development through winter and the maximum lid thickness is recorded. Average wintertime temperatures of −2 to $-30^{\circ }{\rm C}$ and total snowfall of 0 to 3.45 m lead to a range of the maximum lid thickness from 1.2 to 2.8 m after ${\sim }250$ days, with snow cover exerting the dominant control. An initial ice temperature of $-15^{\circ }{\rm C}$ with simulated advection of cold ice from upstream results in 0.6 m of basal freeze-up. This suggests that lakes with water depths above 1.3 to 3.4 m (dependent on winter snowfall and temperature) upon lid formation will persist through winter. These buried lakes can provide a sizeable water store at the start of the melt season, expedite future lake formation and warm underlying ice even in winter.