Surface melt produces more mass loss than any other process on the Greenland Ice Sheet. In some regions of Greenland with high summer surface melt and high winter snow accumulation, the warm porous firn of the percolation zone can retain liquid meltwater through the winter. These regions of water-saturated firn, which may persist for longer than one year, are known as firn aquifers, commonly referred to as perennial firn aquifers. Here, we use airborne ice-penetrating radar data from the Center for Remote Sensing of Ice Sheets (CReSIS) to document the extent of four firn aquifers in the Helheim, Ikertivaq, and Køge Bugt glacier basins with more than six repeat radar flight lines from 1993 to 2018. All four firn aquifers first appear and/or show decadal-scale inland expansion during this time period. Through an idealized energy-balance calculation utilizing reanalysis data from Modèle Atmosphérique Régionale (MAR) regional climate model, we find that these aquifer expansions are driven by decreasing cold content in the firn since the late 1990s and recently increasing high-melt years, which has reduced the firn’s ability for refreezing local meltwater. High-melt years are projected to increase on the Greenland Ice Sheet and may contribute to the continued inland expansion of firn aquifers, impacting the ice sheet’s surface mass balance and hydrological controls on ice dynamics.

Surface melting can alter ice sheet sliding by supplying water to the bed, but subglacial processes driving ice accelerations are complex. We examine linkages between surface runoff, transient subglacial water storage, and short-term ice motion from 168 consecutive hourly measurements of meltwater discharge (i.e. moulin input) and GPS-derived ice surface motion for Rio Behar, a ~60 km2 moulin-terminating supraglacial river catchment the southwest Greenland ablation zone. Short-term accelerations in ice speed correlate strongly with lag-corrected measures of surface mass loss, specifically supraglacial river discharge (r= 0.9; p<0.001). Though our 7-day record cannot address seasonal-scale forcing, diurnal ice accelerations align with normalized differenced supraglacial and proglacial discharge, a proxy for subglacial storage change, better than GPS-derived ice surface uplift. These observations counter theoretical steady-state basal sliding laws and suggest that moulin- and proglacially induced fluctuations in subglacial water storage, rather than absolute subglacial water storage, drive short-term ice accelerations.