An official website of the United States government
Abstract We describe elongate, wet, subglacial bedforms in the shear margins of the NE Greenland Ice Stream and place some constraints on their formation. Lateral shear margin moraines have been observed across the previously glaciated landscape, but little is known about the ice-flow conditions necessary to form these bedforms. Here we describe in situ sediment bedforms under the NE Greenland Ice Stream shear margins that are observed in active-source seismic and ground-penetrating radar surveys. We find bedforms in the shear margins that are ~500 m wide, ~50 m tall, and elongated nearly parallel to ice-flow, including what we believe to be the first subglacial observation of a shear margin moraine. Acoustic impedance analysis of the bedforms shows that they are composed of unconsolidated, deformable, water-saturated till. We use these geophysical observations to place constraints on the possible formation mechanism of these subglacial features.
more »
« less
Automatic identification of streamlined subglacial bedforms using machine learning: an open‐source Python approach
Subglacial processes exert a major control on ice streaming. Constraining subglacial conditions thus allows for more accurate predictions of ice mass loss. Due to the difficulty in observing large‐scale conditions of the modern subglacial environment, we turn to geological records of ice streaming in deglaciated environments. Morphometric values of streamlined subglacial bedforms provide valuable information about the relative speed, direction, and maturity of past ice streams and the relationship between ice streaming and subglacial erosion and deposition. However, manually identifying streamlined subglacial bedforms across deglaciated landscapes, sometimes in clusters of several thousand, is an arduous task with difficult‐to‐control sources of variability and human‐biased errors. This paper presents a new tool that utilizes a machine learning approach to automatically identify glacially derived streamlined features. Slope variations across a landscape, identified by topographic position index, undergo analysis from a series of supervised machine learning models trained from over 600 000 data points identified across the deglaciated Northern Hemisphere. A filtered data set produced through the combination of scientifically driven preprocessing and statistical downsampling improved the robustness of our approach. After cross‐validation, we found that Random Forest detected the most true positives, up to 94.5% on a withheld test set, and an ensemble average of machine learning models provided the highest stability when applied within the range of applicable data sets, performing at up to 79% identification of true positives on an out of distribution area of interest. We build these models into an open‐source Python package, bedfinder, and apply it to new data in the Green Bay Lobe region, USA, finding the general ice‐flow direction and average streamlined subglacial bedform elongation with minimal effort. This type of open, reproducible machine learning analysis is at the leading edge of glacial geomorphology research and will continue to improve with integration of newly acquired and previously collected data.
more »
« less
- Award ID(s):
- 2305317
- PAR ID:
- 10557149
- Publisher / Repository:
- Wiley-Blackwell
- Date Published:
- Journal Name:
- Boreas
- Volume:
- 54
- Issue:
- 4
- ISSN:
- 0300-9483
- Format(s):
- Medium: X Size: p. 488-505
- Size(s):
- p. 488-505
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Spatial variability in bed topography, characterized as bed roughness, impacts ice-sheet flow and organization and can be used to infer subglacial conditions and processes, yet is difficult to quantify due to sparse observations. Paleo-subglacial beds of formerly expanded glaciers found across the Antarctic continental shelf are well preserved, have relatively limited post-glacial sediment cover and contain glacial landforms that can be resolved at sub-meter vertical scales. We analyze high-resolution bathymetry offshore of Pine Island and Thwaites glaciers in the Amundsen Sea to explore spatial variability of bed roughness where streamlined subglacial landforms allow for the determination of ice-flow direction. We quantify bed roughness using std dev. and Fast Fourier Transform methods, each employed at local (100km) and regional (101–2km) scales and in along- and across-flow orientations to determine roughness expressions across spatial scales. We find that the magnitude of roughness is impacted by the parameters selected – which are often not sufficiently reported in studies – to quantify roughness. Important spatial patterns can be discerned from high-resolution bathymetry, highlighting both its usefulness in identifying patterns of streaming ice flow and underscores the need for a standardized way of characterizing topographic variability.more » « less
-
Ice sheets reshape Earth’s surface. Maps of the landscape formed by past ice sheets are our best tool for reconstructing historic ice sheet behavior. But models of glacier erosion and deposition that explain mapped features are relatively untested, and without observations of landforms developing in situ, postglacial landscapes can provide only qualitative insight into past ice sheet conditions. Here we present the first swath radar data collected in Antarctica, demonstrating the ability of swath radar technology to map the subglacial environment of Thwaites Glacier (West Antarctica) at comparable resolutions to digital elevation models of deglaciated terrain. Incompatibility between measured bedform orientation and predicted subglacial water pathways indicates that ice, not water, is the primary actor in initiating bedform development at Thwaites Glacier. These data show no clear relationship between morphology and glacier speed, a weak relationship between morphology and basal shear stress, and highlight a likely role for preexisting geology in glacial bedform shape.more » « less
-
Abstract Bedforms of Thwaites Glacier, West Antarctica both record and affect ice flow, as shown by geophysical data and simple models. Thwaites Glacier flows across the tectonic fabric of the West Antarctic rift system with its bedrock highs and sedimentary basins. Swath radar and seismic surveys of the glacier bed have revealed soft‐sediment flutes 100 m or more high extending 15 km or more across basins downglacier from bedrock highs. Flutes end at prominent hard‐bedded moats on stoss sides of the next topographic highs. We use simple models to show that ice flow against topography increases pressure between ice and till upglacier along the bed over a distance that scales with the topography. In this basal zone of high pressure, ice‐contact water would be excluded, thus increasing basal drag by increasing ice‐till coupling and till flux, removing till to allow bedrock erosion that creates moats. Till carried across highlands would then be deposited in lee‐side positions forming bedforms that prograde downglacier over time, and that remain soft on top through feedbacks that match till‐deformational fluxes from well upglacier of the topography. The bedforms of the part of Thwaites surveyed here are prominent because ice flow has persisted over a long time on this geological setting, not because ice flow is anomalous. Bedform development likely has caused evolution of ice flow over time as till and lubricating water were redistributed, moats were eroded and bedforms grew.more » « less
-
Ancient glaciated landscapes reveal interactions among ice dynamics, bed composition, and preglacial river networks. Subglacial landscapes, revealed in regions of recent ice-sheet retreat, provide a window into ice-sheet dynamics and interactions with evolving subglacial topography. Here, we document landscape evolution beneath the southern Laurentide Ice Sheet of North America since the end of the Pliocene, 2.6 million years (Ma) ago, by reconstructing the isostatically adjusted preglacial surface and modern bedrock topography at 250 m horizontal resolution. We use flow routing to reconstruct drainage networks and river longitudinal profiles, revealing the pattern and extent of their glacially forced reorganization. The overall mean Quaternary (2.6 Ma ago to present) erosion rate is 27 m/Ma, rising within ice-streaming corridors to 35 m/Ma (and locally reaching 400 m/Ma) and falling to 22 m/Ma in non–ice-streaming regions. Our results suggest that subglacial erosion was sufficient to lower the southern Laurentide Ice Sheet into warmer environments, thereby enhancing ablation and reducing ice-sheet extent over time.more » « less
