Time‐Domain Reflectometry Measurements and Modeling of Firn Meltwater Infiltration at DYE‐2, Greenland

Abstract

Surface meltwater can be retained in an ice sheet if it infiltrates the firn and refreezes. This is an important mass balance process for the Greenland Ice Sheet, reducing meltwater runoff and associated sea-level rise. The processes of meltwater infiltration and refreezing are not fully understood, however, and remain difficult to monitor remotely. We deployed vertical arrays of thermistors and time-domain reflectometry (TDR) probes to 4-m depth in the firn to continuously monitor meltwater infiltration and refreezing processes at DYE-2, Greenland. The observations provide a detailed picture of the coupled thermal and hydrological evolution of the firn through the 2016 melt season, including estimates of firn water content. The thaw and wetting fronts reached a maximum depth of 1.8 m, with meltwater infiltration concentrated in four main pulses of melting and subsurface warming that reached progressively deeper into the firn. The observations were used to constrain a coupled model of firn thermodynamics and hydrology, which was then run over the period 1950-2020, driven by meteorological forcing from GC-Net station data and ERA5 climate reanalyses. Model results suggest that decadal-scale firn evolution at DYE-2 is strongly influenced by extreme melt seasons such as those of 1968, 2012, and 2019, when meltwater infiltration reached depths of 6-7 m. Extreme melt years drive increases in firn temperature, ice content, and density, reducing firn meltwater retention capacity. Such processes are likely to govern future meltwater retention as the percolation zone extends to higher elevations in Greenland in the coming decades.Plain Language Summary On polar ice sheets, the vast majority of surface meltwater either runs off to the ocean or is refrozen in porous layers of snow and firn. These processes are important to understand across the Greenland Ice Sheet because they influence how much meltwater contributes to sea level rise versus being retained within the ice sheet in any given year. However, the subsurface nature of meltwater percolation and refreezing makes these processes difficult to accurately monitor, and the conditions that control meltwater retention are also evolving with the changing climate. In the spring of 2016, we deployed a novel array of buried temperature and wetness sensors in the snow and firn at the DYE-2 site in Greenland to monitor the infiltration and refreezing of meltwater in firn, as well as a meteorological station at the surface to collect weather and radiation data. The station measured the evolution of firn structure and temperature over a full annual cycle of melting and refreezing. We used the field measurements to constrain a computer model that simulates firn density, melt, and refreezing processes at the same location in Greenland from 1950 through 2020, based upon historical climate reconstructions. We find that, while large summer melt events appear periodically in the long-term record, including a notable melt summer in 1968, recent extreme melt events at DYE-2 in 2012 and 2...

Expansion of Firn Aquifers in Southeast Greenland

“…This confirms our hypothesis: that upstream expansion of firn aquifers in southeast Greenland will occur in response to high‐melt years if the firn warms and the cold content is insufficient for refreezing meltwater. Recent coupled firn‐thermodynamic and hydrology modeling of meltwater infiltration at DYE‐2 shows that the firn is strongly impacted by high‐melt years such as those in 2012 and 2019 through the increase of firn temperature, ice content, and firn density (Samimi et al., 2021), which supports our results.…”

Ice Sheet Surface and Subsurface Melt Water Discrimination Using Multi‐Frequency Microwave Radiometry

“…Similarly, 40, 70 (between the 40 and 80 cm maxima), 100, and 180 cm are used for the “top‐down” y 2 to y 5 thicknesses representative of 18.7, 10.7, 6.9, and 1.4 GHz sensing depths; the latter value is selected to capture the sensitivity to melt observed at 1.4 GHz at 1.8 m. $$LW{C}_{n}^{max}$$ for each layer was roughly estimated from 0.035 to 0.020 (from layer 1–5) based on the modeled vertical profile of the fraction of irreducible water presented in Samimi et al. (2021).…”

“…In order to estimate $${\mathit{LWC}}_{n}$$ from the TB measurements, $$S{F}_{f(n)}$$ is used to scale $${\mathit{LWA}}_{TDn}$$ between zero and $${\mathit{LWA}}_{TDn}^{max}$$ as: $$${\mathit{LWA}}_{TDn}=S{F}_{f(n)}{\mathit{LWA}}_{TDn}^{max}=S{F}_{f(n)}{y}_{n}{\mathit{LWC}}_{TDn}^{max}$$$ where $${\mathit{LWC}}_{TDn}^{max}$$ represents the fraction of irreducible water (e.g., Samimi et al., 2021). While Equation implies wetness retrieval through each top‐down layer, in reality melt water in the top part of the surface may block emissions coming from the deeper parts of the layer.…”

“…where represents the fraction of irreducible water (e.g., Samimi et al, 2021). While Equation 4 implies wetness retrieval through each top-down layer, in reality melt water in the top part of the surface may block emissions coming from the deeper parts of the layer.…”

“…𝐴𝐴 𝐴𝐴𝐴𝐴 𝐴𝐴 𝑚𝑚𝑚𝑚𝑚𝑚 𝑛𝑛 for each layer was roughly estimated from 0.035 to 0.020 (from layer 1-5) based on the modeled vertical profile of the fraction of irreducible water presented in Samimi et al (2021).…”