Long‐term impact of the proglacial lake Jökulsárlón on the flow velocity and stability of Breiðamerkurjökull glacier, Iceland

Abstract

Proglacial lakes are becoming ubiquitous at the termini of many glaciers worldwide due to continued climate warming and glacier retreat, and such lakes have important consequences for the dynamics and future stability of these glaciers. In light of this, we quantified decadal changes in glacier velocity since 1991 using satellite remote sensing for Breiðamerkurjökull, a large lake‐terminating glacier in Iceland. We investigated its frontal retreat, lake area change and ice surface elevation change, combined with bed topography data, to understand its recent rapid retreat and future stability. We observed highly spatially variable velocity change from 1991 to 2015, with a substantial increase in peak velocity observed at the terminus of the lake‐terminating eastern arm from ~1.00 ± 0.36 m day−1 in 1991 to 3.50 ± 0.25 m day−1 in 2015, with mean velocities remaining elevated from 2008 onwards. This is in stark comparison to the predominately land‐terminating arms, which saw no discernible change in their velocity over the same period. We also observed a substantial increase in the area of the main proglacial lake (Jökulsárlón) since 1982 of ~20 km2, equating to an annual growth rate of 0.55 km2 year−1. Over the same period, the eastern arm retreated by ~3.50 km, which is significantly greater than the other arms. Such discrepancies between the different arms are due to the growth and, importantly, depth increase of Jökulsárlón, as the eastern arm has retreated into its ~300 m‐deep reverse‐sloping subglacial trough. We suggest that this growth in lake area, forced initially by rising air temperatures, combined with the increase in lake depth, triggered an increase in flow acceleration, leading to further rapid retreat and the initiation of a positive feedback mechanism. These findings may have important implications for how increased melt and calving forced by climate change will affect the future stability of large soft‐bedded, reverse‐sloped, subaqueous‐terminating glaciers elsewhere. © 2020 The Authors. Earth Surface Processes and Landforms published by John Wiley & Sons Ltd

Toward Numerical Modeling of Interactions Between Ice-Marginal Proglacial Lakes and Glaciers

“…Dynamic water levels are technically difficult to achieve and computationally expensive to consider. Similar effects of ice-marginal lakes affecting longitudinal flow extending many kilometres up-ice have been interpreted recently for large lobate outlet glaciers in southern Iceland (e.g., Storrar et al, 2017;Dell et al, 2019;Baurley et al, 2020) and numerically modeled for an ice sheet margin in west Greenland (Price et al, 2008).…”

Ice‐Marginal Proglacial Lakes Across Greenland: Present Status and a Possible Future

“…Ice velocities on the edge of the GrIS are locally enhanced by ice-marginal lakes compared to those from land terminating margins (Figure 2b, Figure SI_1b in Supporting Information S1) typically by a factor of ∼+25%. The association of faster-flowing ice with ice-marginal lakes compared to glaciers terminating on land is relatively well-documented elsewhere (e.g., Baurley et al, 2020;Dell et al, 2019;King et al, 2018;Liu et al, 2020;Pronk et al, 2021;Sato et al, 2021;Sutherland et al, 2020;Tsutaki et al, 2013Tsutaki et al, , 2019Tsutaki et al, , 2011. However, for the GrIS, whilst ice velocities might be higher in the vicinity of lakes than at land-terminating ice-margins (Figure 2b, Figure SI_1b in Supporting Information S1), this does not currently translate into receipt of very much ice (discharge) because; (a) whilst acknowledging the large uncertainty in the ice thickness data set especially on the margins of the GrIS the ice in those areas is thin, at least in comparison to the thickness of marine terminating outlet glaciers (Figure SI_1b and 1d in Supporting Information S1) and, (b) at present ice-marginal lakes on the GrIS are predominantly situated on the lateral margins of glaciers, that is, oblique to the main flow direction, rather than at termini (e.g., Figure SI_1a in Supporting Information S1).…”

Glacier Surface Speed Variations on the Kenai Peninsula, Alaska, 2014–2019

“…With the automatic iterative refinements, the final coregistration was on the order of 1/1,000 of a 20 m azimuth pixel for S1 IW SLC data (1/10 of a 3 m pixel for TerraSAR‐X images) and no phase jumps were visible. Offset tracking: The azimuth phase ramp for each TOPS‐mode data was deramped with an oversampling factor of two. Based on the expected maximum ice displacements and results from previous studies (Baurley et al., 2020; Dall et al., 2015; Sánchez‐Gámez & Navarro, 2017), for the Sentinel‐1 data, we applied a patch window size of 256 × 128 single‐look pixels in the range and azimuth directions and a step size of 40 × 8 pixels in the offset tracking procedure, while for TerraSAR‐X we used a window size of 512 × 512 pixels and 32 × 32 pixels for the step size. We discard pixels with unreliable offset values by the signal‐to‐noise ratio threshold of 6.0, which is defined as the ratio of the cross‐correlation function peak to the average correlation on the tracking window (Strozzi et al., 2002).…”