Simulated inundation extent and depth in Harriman Fjord and Barry Arm, western Prince William Sound, Alaska, resulting from the hypothetical rapid motion of landslides into Barry Arm Fjord, Prince William Sound, Alaska

Summary

This data release contains postprocessed model output from a simulation of hypothetical rapid motion of landslides, subsequent wave generation, and wave propagation. A simulated displacement wave was generated by rapid motion of unstable material into Barry Arm fjord. We consider the wave propagation in Harriman Fjord and Barry Arm, western Prince William Sound (area of interest and place names depicted in Figure 1). We consider only the largest wave-generating scenario presented by Barnhart and others (2021a, 2021b). As in Barnhart and others (2021c), we used a simulation setup similar to Barnhart and others (2021a, 2021b), but our results differ because we used different topography and bathymetry datasets. We used datasets that supported refinement of the model to a 5-meter resolution along the shorelines. This data release contains GIS files and geographically registered figures representing the maximum wave height and extent of simulated inundation along shorelines. This data release does not include simulation results related to other related phenomena (for example, current strength or the interaction of the wave with glaciers).

Model Description

The simulation was generated by the D-Claw model (George and Iverson, 2014; Iverson and George, 2014). D-Claw simulates the coupled evolution of fluid and solid phases while satisfying mass and momentum conservation. D-Claw is a single layer, depth-averaged model that conceptualizes landslide material as fully saturated granular material. The model is capable of simulating motion of landslide material, interaction of that material with water, displacement wave generation, and wave propagation. Because the mobile material in D-Claw may be water, landslide material, or a mixture of the two, we will use the term “wave height” to refer to the altitude of the surface of mobile material, regardless of its composition.

Reference frame

The horizontal reference frame for all files is North American Datum of 1983 (NAD 83) Universal Transverse Mercator (UTM) Zone 6 N (European Petroleum Survey Group Code 26906). The vertical reference frame is mean higher high water (MHHW) at Whittier, Alaska (National Oceanic and Atmospheric Administration, NOAA Station 9454949). At this station, MHHW is defined as 3.395 m above the North American Vertical Datum of 1988 (NAVD88). Elevation, altitude, and height, as used in this data release, refer to distance above the MHHW vertical datum.

Considered Scenario and Model Implementation

We present results from a simulation of a single scenario (Table 1). This scenario used the landslide source characteristics of the larger (689 million m3), contractive, more mobile scenario C-689 from Barnhart and others (2021a, 2021b), which generated the largest wave near Whittier, Alaska. In this scenario, three landslides (Coe and others, 2021) on an unstable slope northwest of the northern portion of Barry Arm fjord concurrently and rapidly move into the fjord, generating a displacement wave.
The spatial extent of the computational domain (full extent of Figure 1) and the area of interest (AOI, depicted as the thick black line in Figure 1) are smaller than what was considered by Barnhart and others (2021a, 2021b). The AOI is limited to Harriman Fjord and Barry Arm and does not cover all of western Prince William Sound. The duration of simulated time was 30 minutes, reflecting the arrival of the largest wave heights and associated inundation extents in the AOI within the first half hour of simulated time (Barnhart and others, 2021a; their Figures 7 and 8).

The D-Claw model supports adaptive mesh refinement, and like Barnhart and others (2021a,b,c), we used a computational grid cell size of 50 m around the landslide and along the wave propagation path. Our implementation differs from our previous work in that we permitted grid refinement to a finer resolution of 5 m along the shoreline in the AOI. Simulations used the Denali high performance computer (Falgout and others, undated).

To control adaptive mesh refinement in the simulation, we used the presence of mobile material (“wet” grid cells), the presence of a propagating wave, and the location of shallow (less than 50 m deep) or deep (less than or equal to 50 m deep) water. Wet grid cells were defined as water or landslide material with a thickness greater than 0.001 m. A wave was present when a grid cell was wet and the difference in altitude between the mobile material surface and sea level exceeded 0.05 m. Over the entire computational domain, locations with no wave present used the coarsest grid cell size of 1,000 m, whereas locations with a wave present used a maximum grid cell size of 50 m. When a propagating wave was present within the AOI, a 5 m grid cell size was used where water depths were less than 50 m and 10 m grid cells were used everywhere else. This model specification focused refinement along the shoreline during propagation of the wave.

Topographic and Bathymetric Data Sources

This work relied on integrating multiple topographic and bathymetric data sources (Figure 1). We converted all data sources to NAD 83 UTM Zone 6 N and MHHW before compiling.
In Barry Arm fjord north of Port Wells, we used a digital terrain model (DTM) derived from subaerial light detection and ranging (lidar) data collected on June 26, 2020, (Daanen and others, 2021) and submarine bathymetric data collected between August 12 and 23 (NOAA, 2020). These data were combined and used at 5 m horizontal resolution.

For the subaerial portions of the computational domain outside of Barry Arm, we used a 5 m interferometric synthetic aperture radar (IFSAR)-derived DTM for Alaska (U.S. Geological Survey, 2018, accessed through Alaska Division of Geological and Geophysical Surveys, 2013). Below the MHHW waterline we used one of two existing topo-bathymetric sources. In Passage Canal, we used an 8/15 arc-second dataset (~12 m grid cells) for Whittier and Passage Canal (NOAA, 2009b). Elsewhere, we used an 8/3 arc-second dataset (~59 m grid cells) for Prince William Sound (NOAA, 2009a). These two topo-bathymetric datasets were themselves derived from multiple data sources, including, but not limited to: National Ocean Service hydrographic surveys, National Elevation Dataset topography, and digital coastlines datasets. The source data for both topo-bathymetric datasets was sparse in both deep water and near shore (up to 1.5 km spacing between observations), which necessitated interpolating over those areas. This process, which is detailed by Caldwell and others (2011), gave substantial weight to the shoreline topography in the assignment of interpolated depths in the nearshore zone. Because our results use the more recent and higher resolution IFSAR-derived topography, which has a different shoreline, we re-interpolated a narrow band of nearshore grid cells using a similar methodology.

We defined the near-shore re-interpolation zone based on a constant horizontal distance from the edge of valid IFSAR observation. We used a distance of 83 m because it results in the re-interpolation of at least one but no more than two of the 8/3 arc-second grid cells in the coarser of the two topo-bathymetric datasets Any grid cell of either the Prince William Sound topo-bathymetric dataset (NOAA, 2009a) or the Whittier and Passage Canal topo-bathymetric dataset (NOAA, 2009b) at its original resolution, that overlapped the near-shore re-interpolation zone was not used. After removing the grid cells in the near-shore re-interpolation zone from these two topo-bathymetric datasets, we bilinearly interpolated and resampled both datasets from their original resolution to match the 5 m resolution of the IFSAR DTM. We then merged the two topo-bathymetry datasets with the IFSAR DTM. This yielded a 5 m dataset with missing values only in the near-shore re-interpolation zone. Finally, we interpolated across these missing values using bilinear interpolation, thereby generating a single, continuous 5 m topo-bathymetric raster.

Simulation results and data files

Herein, we provide two result files of spatially distributed model output in GeoTiff format and one polygon shapefile for scenario C-689 (Table 1). The shapefile delineates the boundary between model grid cells that were and were not inundated during the simulation. The two GeoTiff files contain the following variables: maximum wave height (maximum_wave_height_meters.tif) and the maximum inundation depth (inundated_depth_meters.tif) over the extent of the computational domain (full extent of Figure 1).
In addition, we provide two figures depicting the simulation results. Figure 2 depicts the maximum wave height throughout the AOI. Figure 3 depicts the extent of inundation relative to the MHHW shoreline for Harriman Fjord and Barry Arm. All figures are provided as GeoPDFs.

Inundation extent
The file “C689_5m_inund_extent.shp” is an ESRI Shapefile containing a polygon demarcating the boundary between areas that were inundated and areas that were not inundated. It was constructed by delineating all areas where values of “inundated_depth_meters.tif” were greater than zero.
Inundation depth
The file “inundated_depth_meters.tif” contains the maximum inundation depth for grid cells that were inundated by the propagating wave at some point in the simulation. The maximum inundation depth was calculated by analyzing model output at 15 second increments and identifying the maximum inundation depth across all output timesteps. Model grid cells that were never inundated are indicated with “no data.”

Maximum wave height

The file “maximum_wave_height_meters.tif” contains the maximum wave height for cells that were inundated by the propagating wave. The wave height is given in meters relative to the vertical reference frame datum. The maximum wave height was calculated by analyzing the model output at 15 second increments and identifying the maximum wave height for all output timesteps. Model grid cells that were never inundated are indicated with “no data.” The maximum wave height reflects the sum of the inundation depth and the grid cell elevation. Note that in grid cells that were initially dry but were inundated later, this value does not reflect the inundation depth.

Figure Captions

Figure 1. Extent of simulation domain, area of interest, topo-bathymetric data sources used, and place names.
Figure 2. Map of maximum wave height for Harriman Fjord and Barry Arm.
Figure 3. Map of inundation extent for Harriman Fjord and Barry Arm.

References Cited
Alaska Division of Geological & Geophysical Surveys [DGGS], 2013, Elevation Datasets of Alaska: Alaska Division of Geological & Geophysical Surveys Digital Data Series 4, https://elevation.alaska.gov/, accessed May 6, 2022, at https://doi.org/10.14509/25239.
Barnhart, K.R., Jones, R.P., George, D.L., Coe, J.A., and Staley, D.M., 2021a, Preliminary assessment of the wave generating potential from landslides at Barry Arm, Prince William Sound, Alaska: U.S. Geological Survey Open-File Report 2021–1071, 28 p., accessed July 22, 2021, at https://doi.org/10.3133/ofr20211071.
Barnhart, K.R., Jones, R.P., George, D.L., Coe, J.A., Staley, D.A., 2021b, Select model results from simulations of hypothetical rapid failures of landslides into Barry Arm Fjord, Prince William Sound, Alaska: U.S. Geological Survey data release, accessed July 22, 2021, at https://doi.org/10.5066/P9XVJDNP.
Barnhart, K.R., Jones, R.P., George, D.L., Coe, J.A., Staley, D.M., Haeussler, P. J. and Labay, K. 2021c, Simulated inundation extent and depth at Whittier, Alaska resulting from the hypothetical rapid motion of landslides into Barry Arm Fjord, Prince William Sound, Alaska: U.S. Geological Survey data release, accessed May 6, 2022, at https://doi.org/10.5066/P9IAPCZ5.
Caldwell, R. J., Eakins, B. W., and Lim, E., 2011, Digital Elevation Models of Prince William Sound, Alaska: Procedures, Data Sources, and Analysis: NOAA Technical Memorandum NESDIS NGDC-40, accessed June 16, 2022, at https://www.ngdc.noaa.gov/mgg/dat/dems/regional_tr/prince_william_sound….
Coe, J.A., Wolken, G.J., Daanen, R.P., and Schmitt, R.G., 2021, Map of landslide structures and kinematic elements at Barry Arm, Alaska in the summer of 2020: U.S. Geological Survey data release, accessed May 25, 2022, at https://doi.org/10.5066/P9EUCGJQ.
Daanen, R.P., Wolken, G.J., Wikstrom Jones, K., and Herbst, A.M., 2021, High resolution lidar-derived elevation data for Barry Arm landslide, southcentral Alaska, June 26, 2020: Alaska Division of Geological & Geophysical Surveys Raw Data File 2021–3, 9 p., accessed June 17, 2021, at https://doi.org/10.14509/30593.
Falgout, J.T., Gordon, J., Williams, B., Davis, M. J., [undated], U.S. Geological Survey Advanced Research Computing—U.S. Geological Survey Denali Supercomputer: U.S. Geological Survey webpage, accessed May 25, 2022, at https://doi.org/10.5066/P9PSW367.
George, D.L., and Iverson, R.M., 2014, A depth-averaged debris-flow model that includes the effects of evolving dilatancy—II. Numerical predictions and experimental tests: Proceedings of the Royal Society of London. Series A, v. 470, no. 2170, p. 20130820, accessed June 17, 2021, at https://doi.org/ 10.1098/rspa.2013.0820.
Iverson, R.M., and George, D.L., 2014, A depth-averaged debris-flow model that includes the effects of evolving dilatancy—I. Physical basis: Proceedings of the Royal Society of London. Series A, v. 470, no. 2170, p. 20130819, accessed June 17, 2021, at https://doi.org/ 10.1098/rspa.2013.0819.
Haeussler, P. J., Parsons, T., Finlayson, D. P., Hart, P., Chaytor, J. D., Ryan, H., Lee, H., Labay, K., Peterson, A., and Liberty, L., 2013, New Imaging of Submarine Landslides from the 1964 Earthquake Near Whittier, Alaska, and a Comparison to Failures in Other Alaskan Fjords: Advances in Natural and Technological Hazards Research, p. 361–370, accessed July 22, 2021, at https://doi.org/10.1007/978-3-319-00972-8_32
National Oceanic and Atmospheric Administration [NOAA], 1995, Report for H10655: National Oceanic and Atmospheric Administration [NOAA] web page, accessed July 22, 2021, at https://www.ngdc.noaa.gov/nos/H10001-H12000/H10655.html.
National Oceanic and Atmospheric Administration [NOAA], 2020, Report for H13396: National Oceanic and Atmospheric Administration [NOAA] web page, accessed April 5, 2021, at https://www.ngdc.noaa.gov/nos/ H12001- H14000/ H13396.html.
National Oceanic and Atmospheric Administration [NOAA] National Geophysical Data Center, 2009a, Prince William Sound, Alaska 8/3 arc-second MHHW coastal digital elevation model: National Oceanic and Atmospheric Administration [NOAA], National Centers for Environmental Information web page, accessed June 16, 2022, at https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.n….
National Oceanic and Atmospheric Administration [NOAA] National Geophysical Data Center, 2009b, Whittier, Alaska 8/15 arc-second MHHW coastal digital elevation model: National Oceanic and Atmospheric Administration [NOAA], National Centers for Environmental Information web page, accessed April 5, 2021, at https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.n… gg.dem:530.
U.S. Geological Survey, 2018, USGS EROS Archive – Digital Elevation – Interferometric Synthetic Aperture Radar (IFSAR) – Alaska, Accessed May 6, 2022, at https://doi.org/10.5066/P9C064CO.
Table 1. Summary of considered scenario including key simulation input parameter values.
Simulation input parameters Scenario name and description
C-689
Symbol Units Description Larger, contractive, more mobile
α degrees Headscarp angle 60
C - Logarithmic spiral coefficient (defined in Equation 2 of Barnhart and others, 2021a) 0.5
V m3 Volume of all three landslides 689,000,000
m0 - Initial solid volume fraction 0.62
mcrit - Critical state solid volume fraction 0.64
φ degrees Basal friction angle 36
φΔ degrees Basal friction angle offset 0
k0 m2 Hydraulic permeability 10-10