https://doi.org/10.5061/dryad.9kd51c5qv

Author Information:

A. Principal Investigator Contact Information

Name: Shivani Ehrenfeucht;
Institution: University of California- Irvine;
Email: sehrenfe@uci.edu

B. Co-Author

Name: Eric Rignot;
Institution: University of California- Irvine;
Email: erignot@uci.edu

C. Co-Author

Name: Mathieu Morlighem;
Institution: Dartmouth College;
Email: Mathieu.Morlighem@dartmouth.edu

Project Description:

Recent observations suggest that seawater rushes between the glacier base and the bedrock at high tide. The presence of ocean water at this boundary, referred to as seawater intrusion, has the potential to increase grounded ice melt. We test this hypothesis on Petermann Glacier, Greenland, using an ice sheet model. We run the model to reconstruct the ice velocity and grounding line position from 2010 to 2022 with and without seawater intrusions. We compare the results with satellite observations of velocity, grounding line position, and ice thinning. When we use enhanced ice melt rates from seawater intrusions, the model matches the observed retreat, speed up, and thinning. When we do not, the model fails to replicate observations. Seawater intrusions therefore play a critical role in glacier evolution. Adding them in ice flow models will increase their sensitivity to ocean warming and the projections of mass loss and sea level rise.

Here we provide all output results from 45 model simulations of ice flow generated using the Ice-sheet and Sea-level System Model (ISSM). We vary two parameters: maximum basal melt rate and distance of seawater intrusion. In addition to model output, we include the surface mass balance and melt water runoff data generated by the Regional Climate Model, MAR, that we used to force the model during simulations. We also include the files and code required to generate the manuscript figures.

Code/Software

This work was done using the Ice-sheet and Sea-level System Model (ISSM) and the Glacier Drainage System Model (GlaDS). GlaDS is implemented within ISSM, which is open access and available for download at https://issm.jpl.nasa.gov/. These scripts were run using ISSM version 4.23 and MATLAB version R2021b.

Description of the data and file structure

The data provided here is organized into 3 folders: Model_Forcing, Results_v2, and Figures. A model, PGIS_GL_GlaDS.mat, is provided as well, which contains the initial conditions for ice velocity (m/yr) and geometry (surface, thickness, base, bed, all of which are in units of meters), and the results from our steady state subglacial hydrology simulation:

• step: timestep of model simulation (unitless)
• time: time at each timestep (years)
• EffectivePressure: water pressure - ice pressure at the bed (Pa)
• HydraulicPotential: the hydraulic potential of the subglacial hydrology system (Pa)
• HydrologySheetThickness: height of the water layer in the distributed hydrology system (m)
• ChannelArea: cross-sectional area of channels (m2)
• ChannelDischarge: water discharge through the channelized hydrology system (m3/s)
• HydrologyWaterVx: x component of the water velocity in the distributed hydrology system (m/yr)
• HydrologyWaterVy: y component of the water velocity in the distributed hydrology system (m/yr)

To load the model in MATLAB on a computer with ISSM installed, and then plot the initial conditions for ice thickness using ISSM's plotting function plotmodel():

>> md = loadmodel('PGIS_GL_GlaDS.mat');

>> plotmodel(md, 'data', md.geometry.thickness);

MODEL FORCING:

Spatially varying, depth-dependent basal melt rates for each model simulation and all relevant model output is contained in the Model_Forcing folder in MATLAB files. Basal melt rate files each contain 5 vectors that provide the basal melt rate applied at each node of the model domain for each of the 5 seawater intrusion distances tested. They are labeled based on the basal melt rate used in the model simulation. For example, BMR_10.mat corresponds to the 5 model simulations that used a basal melt rate of 10 m/yr. The contents of this file can be viewed by loading and calling it:

>> load BMR_10.mat;

>> BMR_10

BMR_10 =

struct with fields:

   L_0: [12505x1 double]

   L_1500: [12505x1 double]

   L_3000: [12505x1 double]

   L_4500: [12505x1 double]

   L_6000: [12505x1 double]

Here L indicates the parameter seawater intrusion distance, and the underscored number (0, 1500, 3000, 4500, or 6000) is the distance of intrusion in meters. So the file BMR_10 contains the basal melt rates used in all five model simulations that had a maximum basal melt rate of 10 m/yr. A map of the spatially variable basal melt rate model forcing for the simulation with a 10 m/yr maximum basal melt rate and a seawater intrusion distance of 1500 m can be called by:

>> BMR_10.L_1500

This filed can be plotted using ISSM's plotting function plotmodel() when the model, md, is loaded:

>> plotmodel(md, 'data', BMR_10.L_1500);

There are 9 BMR_XX.mat files contained in this folder, corresponding to each of the 9 maximum basal melt rates used in model simulations for this work, ranging from 10 m/yr to 90 m/yr in increments of 10 m/yr. Each BMR_XX.mat file is structured the same way as the example above, containing a 5 vectors of basal melt rates corresponding to the 5 simulations of varying seawater intrusion distance with a maximum basal melt rate specified by the number in the file name. A figure of all 45 basal melt rate maps, corresponding to the 45 model simulations run in this work, can be seen in supplemental documentation published with the main manuscript.

Daily surface mass balance generated from the Regional Climate Model, MAR, in m/yr from January 1st 2010 to August 1st 2022 is used to force the transient model simulations and is contained here in the MATLAB file MAR_SMB_daily.mat.

MODEL RESULTS:

All model results are avaliable in the Results_v2 folder. Individual .mat files containing model results are named according to the variable that they contain and the basal melt rate used during the model simulations. For example, H_DDmeltrate_10.mat contains the ice thickness results, H, in meters for the 5 model simulations that used a maximum basal melt rate of 10 m/yr and the time, in years, of model output:

>> H_DDmeltrate_10

H_DDmeltrate_10 =

struct with fields:

   time: [1x399 double]

   L_0: [12505x399 double]

   L_1500: [12505x399 double]

   L_3000: [12505x399 double]

   L_4500: [12505x399 double]

   L_6000: [12505x399 double]

In this folder, we provide model output results for 4 variables: ice thickness: H (m), ice surface elvation: Surf (m), ice surface velocity: Vel (m/yr), and grounding line location: GL (m). These 4 variables are provided for each of the 45 simulations at each timestep during the simulation. We use the same organization as we did for the basal melt rate model forcing, where there are 9 .mat files per variable and each .mat file contains data corresponding to 5 model simulations. Note that the grounding line location data is a map showing the distance away from the GL- negative values correspond to floating ice and positive values correspond to grounded ice. The zero contour of this data set represents the spatial location of the grounding line at that time.

Folder contents:

• Variable: Ice thickness
  • Abbreviation: H
  • Units: meters
  • Files: H_DDmeltrate_10.mat, H_DDmeltrate_20.mat, H_DDmeltrate_30.mat, H_DDmeltrate_40.mat, H_DDmeltrate_50.mat, H_DDmeltrate_60.mat, H_DDmeltrate_70.mat, H_DDmeltrate_80.mat, H_DDmeltrate_90.mat
• Variable: Surface elevation
  • Abbreviation: Surf
  • Units: meters *Files: Surf_DDmeltrate_10.mat, Surf_DDmeltrate_20.mat, Surf_DDmeltrate_30.mat, Surf_DDmeltrate_40.mat, Surf_DDmeltrate_50.mat, Surf_DDmeltrate_60.mat, Surf_DDmeltrate_70.mat, Surf_DDmeltrate_80.mat, Surf_DDmeltrate_90.mat
• Variable: Ice surface velocity
  • Abbreviation: Vel
  • Units: meters/year
  • Files: Vel_DDmeltrate_10.mat, Vel_DDmeltrate_20.mat, Vel_DDmeltrate_30.mat, Vel_DDmeltrate_40.mat, Vel_DDmeltrate_50.mat, Vel_DDmeltrate_60.mat, Vel_DDmeltrate_70.mat, Vel_DDmeltrate_80.mat, Vel_DDmeltrate_90.mat
• Variable: Grounding line location
  • Abbreviation: GL
  • Units: meters
  • Files: GL_DDmeltrate_10.mat, GL_DDmeltrate_20.mat, GL_DDmeltrate_30.mat, GL_DDmeltrate_40.mat, GL_DDmeltrate_50.mat, GL_DDmeltrate_60.mat, GL_DDmeltrate_70.mat, GL_DDmeltrate_80.mat, GL_DDmeltrate_90.mat

PLOTTING:

All files and scripts to generate the three figures in the main manuscript are provided in the Figures folder. There are three figure scripts which are used to generate the main three figures:

• Figure1.m: Script to generate Figure 1 in manuscript.
• Figure2.m: Script to generate Figure 2 in manuscript.
• Figure3.m: Script to generate Figure 3 in manuscript.

The file labeled Landsat_PG.png is the background image used in each of the panels in the three figures.

There are three subfolders in this directory that provide additional files used in the figure scripts.

• Exp: Folder containing Greenland outline and Petermann outline used in Figure 1. These are .exp files that provide the x-y coordinates of the two outline shapes.
• glResults: Folder containing individual .png files illustrating grounding line migration results from all 45 transient model simulations. Contents of glResults are used to generate Figure 2. The results shown in these images are the same as the results provided in GL_DDmeltrate_**.mat files in the Results_v2 folder.
• Petermann_GL: Folder containing shape files of grounding line position generated from satellite data and used to validate model output.

The MATLAB files included here to generated figures require an installation of ISSM on the user's computer to run and for the model, PGIS_GL_GlaDS.mat, to be loaded in MATLAB. When the model is loaded, execute the desired file and the corresponding figure will be generated. To make Figure 1 from the manuscript, for example, run the following commands in MATLAB:

>> md = loadmodel('PGIS_GL_GlaDS.mat');

>> Figure1;

Figure2.m and Figure3.m generate the second and third figures from the manuscript respectively, and will run with the model PGIS_GL_GlaDS.mat loaded using the data contained in the glResults and Results_v2 folders.