Inventory map of submarine and subaerial-to-submarine landslides in Glacier Bay, Glacier Bay National Park and Preserve, Alaska

Mass-wasting events that displace water, whether they initiate from underwater sources (submarine landslides) or subaerial sources (subaerial-to-submarine landslides), have the potential to cause tsunami waves that can pose a significant threat to human life and infrastructure in coastal areas (for example towns, cruise ships, bridges, oil platforms, and communication lines). Sheltered inlets and narrow bays can be locations of especially high risk as they often have higher human populations, and the effects of water displacement from moving sediment can be amplified as compared to the effects from similarly sized mass movements in open water. In landscapes undergoing deglaciation, such as the fjords and mountain slopes adjacent to tidewater glaciers found in Southeast Alaska, glacial retreat and permafrost decay can destabilize rock slopes and increase landslide potential. Establishing and maintaining inventories of subaerial and submarine landslides in such environments is critical for identifying the magnitude and frequency of past events, as well as for assessing areas that may be susceptible to failures in the future. To maintain landslide inventories, multi-temporal surveys are needed. High-resolution digital elevation models (DEM) and aerial imagery can be used to establish and maintain subaerial landslide inventories, but repeat bathymetric surveys to detect submarine landslides are generally less available than their terrestrial counterparts. However, existing bathymetry can be used to establish a spatial inventory of landslides on the seafloor to provide a baseline for understanding the magnitude of past events and for locating areas of high submarine landslide susceptibility. These data can then be used to address how future failures and the tsunamis that they could trigger could impact surrounding areas. Here, we present an inventory of mapped landslide features in Glacier Bay, Alaska that includes landslide source areas, deposits, and scarps. This data release contains geographic information system (GIS) polygons and polylines for these mapped features; the underlying DEM raster compiled from available bathymetry from the National Oceanic and Atmospheric Administration (NOAA) and the U.S. Geological Survey (USGS); a slope map created from the compiled DEM; ­and a derivative topographic openness map used to help identify the landslide features. Bathymetric DEMs used in the compilation cover 1012.5 sq. km, which represents approximately 80% of the total area of Glacier Bay. The DEMs were collected in 2001 and 2009 for the southern and northern parts of the bay, respectively. To minimize resolution bias and maximize mapping consistency while maintaining visual fidelity, we re-sampled all the original bathymetry (resolution ranging from 1 to 16 m) to 5 m, which represents the minimum resolution for the majority of mapped areas; the lower resolution areas generally covered deeper and flatter portions of the bay where fewer landslides were present. For mapping, we used a topographic openness map (Yokoyama and others, 2002) in combination with a traditional slope map (see Red Relief Image Map in Chiba and others, 2008), which allows for good discernment of subtle concavities and convexities in the bathymetry and is well-suited for identifying landslide scars and deposits. We classified mapped landslides based on their source area type and used two primary classification categories of "slide" and "debris flow". We used a third category, "mixed", to classify landslides that showed evidence of both types of source area contributing to the deposit. For each landslide classified as slide or mixed, we mapped the source area and deposit as separate polygons. For landslides classified as debris flow, we mapped only deposits. Since debris flow source areas are subaerial drainage basins, delineating them should be part of larger future subaerial landslide mapping efforts in Glacier Bay National Park and Preserve. Similarly, for mixed landslides, we delineated source areas as the slide contribution area and not the larger debris-flow drainage basin component. For any source areas (for mixed and slide polygons) or deposits that included a subaerial portion, we used 2012 5-m IFSAR data, and Landsat and DigitalGlobe imagery to map subaerial parts of the polygons. IFSAR and Landsat data are available from Earth Explorer (https://earthexplorer.usgs.gov/) and DigitalGlobe imagery is available from DigitalGlobe (https://www.digitalglobe.com/). These data and images are not included in this data release. Thirty-five of the forty-four slide and mixed features initiated as subaerial landslides. However, in all cases, we only mapped landslides if we could identify a submarine deposit. For example, we did not map the subaerial Tidal Inlet landslide (Wieczorek and others, 2007) because we could not identify a submarine deposit associated with it. Additionally, we did not map subaerial and submarine deposits that appeared to be deposited by water-dominated flows (for example, alluvial fans and fan deltas), or large submarine fans that likely resulted from turbidite flows, such as the one at the junction of Queen Inlet and Glacier Bay. Because we could not observe mapped submarine landslides in the field, we assigned a level of moderate (77 landslides) or high (31 landslides) confidence based on our certainty that the mapped features represented actual slope failures. We omitted low confidence landslides from the map. In total, we mapped 108 landslides, with 22, 64, and 22 classified as slide, debris flow, and mixed, respectively. The total area (source and deposit) for slide and mixed type landslides ranged from 0.026 to 2.35 sq. km. Debris-flow deposits ranged from 0.012 to 0.61 sq. km. Finally, we mapped a total of 7,097 individual landslide scarps where we could not identify any clear associated deposits, and where the distance between lateral flanks was approximately 50 m or more. Though we did our best to map only arcuate-shaped scarps typically formed by landslides (that is, single-mass failures), as opposed to geomorphic features formed by gradual glacial or submarine-current-related erosion (for example, submarine canyon walls), we acknowledge that some mapped scarps may have been formed by processes other than landsliding. Thus, for purposes of landslide susceptibility mapping, these scarp data are intended to be used in conjunction with other data, such as slope angle, geologic substrate, or geomorphic units. Ultimately, the full dataset is meant to serve as a qualitative component to inform future submarine and subaerial landslide susceptibility assessments in Glacier Bay National Park and Preserve. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. References used: Chiba, T., Kaneta, S., and Suzuki, Y., 2008, Red relief image map: new visualization method for three dimensional data: The international archives of the photogrammetry, remote sensing and spatial information sciences, v. 37, no. B2, p. 1071-1076. Wieczorek, G.F., Geist, E.L., Motyka, R.J., Jakob, M., 2007, Hazard assessment of the tidal inlet landslide and potential subsequent tsunami, Glacier Bay National Park, Alaska: Landslides, v. 4 p. 205-215. Yokoyama, R., Shirasawa, M., and Pike, R.J., 2002, Visualizing topography by openness: a new application of image processing to digital elevation models: Photogrammetric engineering and remote sensing, v. 68, no. 3, p. 257-266.