Pierre St-Laurent, pst-laurent@vims.edu, 7 January 2023 ******************************************************************************** Abstract ******************************************************************************** Dataset: A numerical simulation of the ocean, sea ice and ice shelves in the Amundsen Sea (Antarctica) over the period 2006-2022 and its associated code and input files Abstract: A three-dimensional numerical model of the Amundsen Sea (Antarctica) was used to simulate the period Jan.2006-Mar.2022 under consistent atmospheric/oceanic forcings, bathymetry/ice shelf topography, and model equations/parameters. The model is an implementation of the Regional Ocean Modeling System (ROMS, https://www.myroms.org/) with extensions for sea ice (Budgell 2005) and ice shelves (Dinniman et al. 2011). It simulates the ocean hydrography and circulation, sea ice thermodynamics and dynamics, and the basal melt of the ice shelves, with a uniform horizontal mesh of 1.5km and 20 topography-following vertical levels. Forcings include the ERA5 reanalysis (3-hourly), 10 tidal constituents from CATS 2008, and ocean/sea ice conditions at the edges of the model domain taken from the 5km-resolution circumpolar model of Dinniman et al. 2020 and from daily SSM/I satellite images. The model outputs are divided into nine directories each containing two years worth of model results (run661-669) in the NetCDF format. Each directory contains: daily-averaged model fields (roms_avg_xxxx.nc), instantaneous snapshots every 3 hours for select fields (roms_qck_xxxx.nc), and instantaneous snapshots every 30 days (roms_his_xxxx.nc). All the metadata information necessary for the interpretation of the model outputs (dimensions, units, etc) is included inside the NetCDF files. The NetCDF files follow the CF conventions and can be opened with various software that are open source and freely available over the Internet. In addition to the model outputs, this archive includes the computer code as well as the input files necessary for reproducing the model outputs of this archive. ******************************************************************************** Information about the code ******************************************************************************** The tarball code_v20230107.zip contains the computer code used for the simulation of the Amundsen Sea (Antarctica) over the period 2006-2022. It includes the following directories/files: (a) amicus_roms/ Files that are specific to the Amundsen Sea implementation of ROMS. (b) roms_kshedstrom_git/ Generic ROMS files. (c) depths.m A Matlab/Octave script computing the vertical position of the model grid boxes at a given time. The file originates from the Rutgers Matlab toolbox (see myroms.org), but it includes customizations for the presence of ice shelves (their presence causes a re-definition of the vertical depths inside ROMS.) The script is fairly short and can be rewritten easily for other computer languages. The script is essential to obtain the depths associated with the three-dimensional model outputs. The generic ROMS code (directory roms_kshedstrom_git) originates from the Git repository of Kate Hedstrom (Univ.Alaska Fairbanks, kshedstrom@alaska.edu): https://github.com/kshedstrom/roms.git GIT Revision: kate_svn commit 0baf6674f10514306a8ab81ee42af5a9a3da87cc accessed (cloned) on 2021-06-29 ...that combines the Rutgers branch of ROMS/TOMS (its version 3.9 dated Apr.6, 2020 (according to roms_kshedstrom_git/ROMS/Version)), the sea ice module of Budgell 2005, and the ice shelf module of Dinniman et al. 2011 (see full references below). The present code archive is similar in scope to that of Mack et al. 2019 (see https://github.com/mnemoniko) except that the present archive is based on a more recent version of ROMS/TOMS. Note that Kate's code includes a very large number of customizations on top of the original Rutgers code, with these customizations being activated/deactivated via C Pre-Processing (CPP) directives. The Amundsen Sea configuration (defined by the file amicus_roms/amicus.h) activates only a very small fraction of these customizations. The best way to read and understand the code in this archive is to first compile the Amundsen Sea application, and then to study the preprocessed files (that will have the extension .f90) stored inside the directory amicus_roms/Build_roms/. Since the .f90 files have underwent preprocessing, they are infinitely smaller and more legible than their .F (non-preprocessed) counterpart. The Git ROMS repository of Kate Hedstrom is no longer maintained and a few of its files were not in a fully functional state at the time when her repository was accessed. Additional modifications were therefore applied by St-Laurent in order to obtain a functional configuration that produced results consistent with an earlier version of this codebase (Dinniman's SOGLOBEC3.3). These additional modifications can be tracked with a recursive "grep" command over the alias "psl" (Pierre St-Laurent). The key changes were: (a) use a non-zero value for the minimum sea ice concentration (see parameter min_a inside amicus_roms/ice.in); (b) use a non-zero value for the minimum sea ice thickness (see parameter min_h inside amicus_roms/ice.in); (c) complete the implementation of some sea ice boundary conditions that had been partially implemented (see roms_kshedstrom_git/ROMS/SeaIce/tibc.F); (d) un-comment the calculation of the wao term (see Mellor & Kantha 1989) so that it can contribute to the thermodynamics (see roms_kshedstrom_git/ROMS/SeaIce/ice_mk.h); (e) fix a typo concerning idBvar(isTice) inside roms_kshedstrom_git/ROMS/Modules/mod_ncparam.F; (f) fix the net evaporation term in the particular case where SCORRECTION is un-defined (see roms_kshedstrom_git/ROMS/Nonlinear/set_vbc.F). The code inside this archive is targeted toward users who have at least a basic experience with the Rutgers branch of ROMS (see myroms.org, which includes extensive documentation as well as introductory test-cases). Because the present code is coupled to sea ice and to ice-shelves, it is definitely not the right way to learn how to use ROMS. Minimal instructions on how to compile the code of this archive: (a) Open amicus_roms/build_roms.bash with your favorite text editor. (b) Edit the environment variable MY_ROMS_SRC so that it points toward your local copy of the directory roms_kshedstrom_git. (c) Edit the variables USE_MPI,which_MPI,FORT to reflect your computer configuration. If necessary, edit the files Linux-gfortran.mk or Linux-ifort.mk to reflect your computer configuration. (d) ./build_roms.bash (i.e. compile the code). (e) Edit the file roms_amicus.in to update the paths to the input files and to set the parameters of your calculation. The health of the code has been verified by compiling with ROMS' USE_DEBUG mode and then running the resulting executable under the GDB debugger. The code has been successfully used with both the GNU Fortran compiler (gfortran) and the Intel Fortran compiler (ifort). The original Rutgers ROMS branch is licensed under the MIT/X license. I'm quoting below an extract from this license in order to remind the user of the key passage (the full license text is available at http://www.opensource.org/licenses/mit-license.php ): THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESSED OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIMS, DAMAGES OR OTHER LIABILITIES, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE If you experience issues or unexpected behavior, you can contact me at the email address above and we'll see what could be done. The message board of myroms.org (which is maintained by Rutgers) isn't the best resource for new questions because Kate Hedstrom's extensive customizations represent a substantial departure from the official Rutgers code hosted at myroms.org. Nevertheless, the myroms.org message board can be helpful by searching its database for answers to past questions. ******************************************************************************** Information about the input files ******************************************************************************** Abundant documentation is available in the metadata of each input file; the basic information is repeated here for the reader's convenience. (a) grd_amicus1076x516_2010_v20220703.nc Created by Pierre St-Laurent (pst-laurent@vims.edu) on 03-Jul-2022 14:04:02 for project AMICUS (NASA 20-CRYO2020-0034). The grid has a uniform horizontal mesh spacing of 1.5km; see variables xprj,yprj and the global metadata `projection` for additional information. The bathymetry and ice shelf draft are based on IBCSO_v2 (Dorschel et al. 2022, https://doi.org/10.1038/s41597-022-01366-7 ), Jordan et al. 2020 (Dotson, Crosson, Thwaites ice shelves; https://doi.org/10.5194/tc-14-2869-2020 ), and BedMachineAntarctica-2020-10-08.nc (M.Morlighem, Dartmouth, pers.comm.). Areas of the grid where BedMachine`s water-column thickness was <15m were considered as `dry land` (i.e., set to mask_rho=0). The bathymetry and ice shelf draft were slightly smoothed to meet the criteria rx0 <= 0.8 and rx1 <= 30. The original (i.e., prior to smoothing) bathymetry, ice shelf draft and land/ice shelf/water mask are available from variables h__0, zi_0, and mask_rho_0 (respectively). Although ROMS overwrites the following parameters at runtime based on the values provided inside roms_my_app.in, the grid was designed assuming: N=20, Vtransform=1, Vstretching=1, THETA_S=2.5, THETA_B=0, TCLINE=15. The ice shelf front of Dotson ice shelf and eastern Getz ice shelf (originally from BedMachine) was pushed further inland to match the ship track of NPB202202 (TARSAN-ARTEMIS-THOR expedition) during their sampling of these two ice shelves. Icebergs representative of year 2010 were implemented either as a static top-to-bottom obstacle (mask_rho=0, for large tabular icebergs such as B-22), or as static shallow ice shelves (to inhibit sea ice drift while allowing for an oceanic flow) in locations such as the Thwaites ice tongue; see St-Laurent et al. 2017 for more information ( https://doi.org/10.1002/2017JC013162 ). The position and geographical extent of the icebergs is based on Fig.13 of Mazur et al. 2019 ( https://doi.org/10.33265/polar.v38.3313 ) and on year-specific satellite images ( Scambos et al. 2022, https://doi.org/10.7265/N5NC5Z4N ). Additional notes: The geographical extent of this Amundsen implementation differs from the one described in St-Laurent et al. 2017; all the components of the present archive are new and have been created from scratch. Although the configuration Vtransform=Vstretching=1 is considered to be "obsolete" in non-Antarctic implementations of ROMS, the more recent Vtransform/Vstretching configurations were designed under the assumption of a quasi-horizontal free surface, which is not at all the case in presence of ice shelves. This is the rationale for favoring the traditional Vtransform=Vstretching=1 configuration. ROMS does not explicitly represent icebergs or fast ice and therefore these must be parameterized. Additionally, the goal of this model experiment was to simulate the Amundsen Sea under a consistent (i.e. time-invariant) iceberg/fast ice/ice shelf thickness configuration. The configuration described above is representative of year 2010, which is (more or less) at the center of the simulation period (2006-2022). The role of changes in the iceberg/fast ice/ice shelf thickness configuration will be covered in future experiments (not part of this dataset.) (b) roms_ini.nc Initial condition for 1 January 2006, 00:00 UTC. (c) tides_amicus1076x516_v20211128.nc Tidal harmonics for 10 constituents: M2, S2, N2, K2, K1, O1, P1, Q1, MF, MM. Based on CATS 2008 ( https://doi.org/10.15784/601235 ), accessed 2021-08-04. References: Padman et al. 2002 ( https://doi.org/10.3189/172756402781817752 ), 2008 ( https://doi.org/10.1029/2008GL035592 ). The tidal reference date is recorded in the variable `zero_phase_date` (see https://www.myroms.org/projects/src/ticket/896 ). The 18.6 year cycle and the `nodal corrections` are ignored (i.e., neglected). (d) bry_amicus1076x516x20_ssmiyyyy_v20220928.nc (2006 <= yyyy <= 2022) Created by Pierre St-Laurent (pst-laurent@vims.edu) on 28-Sep-2022 14:34:23 for project AMICUS (NASA 20-CRYO2020-0034). The majority of the fields are a monthly climatology assembled from three consecutive years of simulation using the model setup of Dinniman et al. 2020 ( https://doi.org/10.1029/2019JC015736 ). The monthly fields are complemented by year-specific, observed daily sea ice concentrations (variable `aice`) from satellites. The satellite dataset for the period <=2020 is: ``Comiso, J. C. 2017. Bootstrap Sea Ice Concentrations from Nimbus-7 SMMR and DMSP SSM/I-SSMIS, Version 3. Subset used: year 2006. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. doi: https://doi.org/10.5067/7Q8HCCWS4I0R ``. For the period >=2021, the satellite dataset is: ``Brodzik, M. J. and J. S. Stewart. 2016. Near-Real-Time SSM/I-SSMIS EASE-Grid Daily Global Ice Concentration and Snow Extent, Version 5. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. https://doi.org/10.5067/3KB2JPLFPK3R``. The present BRY file is consistent with the ROMS grid file `/home/pierre/projects/amicus/amicus_roms/grd_amicus1076x516_2010_v20220703.nc`. These "bry" fields are only applied at the immediate edge of the model domain, following the passive/active approach of Marchesiello et al. 2001 ( https://doi.org/10.1016/S1463-5003(00)00013-5 ). Note that the use of a monthly climatology for oceanic fields at the edges of the model domain imply that year-to-year changes in the position/characteristics of large-scale features such as the Antarctic Circumpolar Current or Ross Gyre are not represented in the present simulation. All year-to-year changes visible in the model results must be the result of the atmospheric forcing and/or from the daily sea ice concentration (from SSM/I) prescribed at the edges of the model domain (see St-Laurent et al. 2022 on this topic.) (e) nud_amicus1076x516x20_v20220915.nc Created by Pierre St-Laurent (pst-laurent@vims.edu) on 15-Sep-2022 17:45:40 for project AMICUS (NASA 20-CRYO2020-0034). temp_NudgeCoef has a maximum value of 1/(14days) at the open boundaries. temp_NudgeCoef decreases linearly away from the along-shelf boundaries over 200 grid points, and the nudging is only active in the off-shelf areas where the bottom depth is 3000m or greater (and no nudging in shallower water). The same nudging is applied on salinity (salt_NudgeCoef) and on the three-dimensional horizontal velocities (M3_NudgeCoef). The "nud" file works in combination with the "clm" files. Their purpose is to complement the approach of Marchesiello et al. 2001 (which is only active at the immediate edge of the model domain) by providing additional relaxation for three model fields (salinity,temperature,M3 velocities) over the off-shelf area of the domain. The nudging is inactive (i.e. zero) outside of the off-shelf area. (f) clm_amicus1076x516x20_yyyy_v20220919.nc (2006 <= yyyy <= 2022) Created by Pierre St-Laurent (pst-laurent@vims.edu) on 19-Sep-2022 17:14:49 for project AMICUS (NASA 20-CRYO2020-0034). The fields (salinity, temperature, three-dimensional horizontal velocities) are a monthly climatology assembled from three consecutive years of simulation using the model setup of Dinniman et al. 2020 ( https://doi.org/10.1029/2019JC015736 ). These "clm" fields work in combination with the "nud" file. (g) frc_amicus_era5_yyyy_swrad0.71_v20211111.nc (2006 <= yyyy <= 2022) ROMS atmospheric forcing file for the years 2006-2006, created by pst-laurent@vims.edu on 11-Nov-2021 08:36:25. This forcing file has a temporal resolution of 3 hours, a spatial resolution of 0.25 degrees, and it covers the domain 86-142deg.W, 67-77deg.S. The fields originate from the ERA5 reanalysis (Hersbach et al. 2020, https://doi.org/10.1002/qj.3803 ). The ERA5`s downwelling shortwave radiation (swrad) is scaled by a factor of 0.71 to match real-life in situ observations (ASPIRE cruise, Amundsen Sea, austral summer 2010/2011.) ******************************************************************************** Information about the Amundsen Sea implementation of ROMS ******************************************************************************** The implementation is specifically focused on the *continental shelf* of the Amundsen Sea. Although the model domain includes a generous portion of the continental slope and deep sea, the realism of the model in deep areas is limited by the coarse vertical resolution of the implementation (only 20 vertical levels). This is what motivated the use of nudging (see above) over areas of bottom depth >3000m in order to maintain a realistic stratification there at all times. No other form of relaxation is used in the simulation (e.g. the CPP option SCORRECTION of Kate Hedstrom is set to #undef.) The coarse vertical resolution over the abyssal plain (and the poorly-resolved surface mixed layer there) also motivated a modification to the drag at the sea ice/ocean interface. Over the continental slope and abyssal plain, this drag gradually becomes a function of sea ice velocity alone (CPP option LIMIT_WATER_STRESS_IN_UNRESOLVED_SURFACE_LAYER). Over the rest of the model domain, the drag at the sea ice/ocean interface follows the usual function of sea ice velocity and surface ocean velocity. The Amundsen implementation uses the Large, McWilliams and Doney (1994, "LMD") sub-grid scale diffusivity along the vertical with the background values and limiters specified inside the file amicus_roms/roms_amicus.in. Note that vertical diffusivities for temperature and salinity are limited to a maximum of 10^(-2) m2/s (AKT_LIMIT). The third-order upstream-biased advection scheme described in Shchepetkin & McWilliams (2005) is used for horizontal advection of horizontal momentum, salinity and temperature. This advection scheme is supplemented by a very weak Laplacian diffusivity of 0.1 m2/s applied on horizontal momentum (along the topography-following surfaces, MIX_S_UV), salinity and temperature (along geopotential surfaces, MIX_GEO_TS). Parameters relative to sea ice are found inside the file amicus_roms/ice.in. The basal melt rates of ice shelves (see Dinniman et al. 2011 for a description of their computation) relies on transfer coefficients that vary in time/space according to the local friction velocity (see Schmidt et al. 2004). This approach works appropriately for Abbot, Cosgrove, Pine Island, Thwaites ice shelves and leads to basal melt rates that are consistent with remote sensing estimates. However, in the case of Dotson-Crosson and Getz ice shelves, the parameterization of Schmidt et al. leads to an overestimation of the basal melt rates relative to estimates from remote sensing. Following Nakayama et al. 2017 ( https://doi.org/10.1002/2016JC012538 ), an ad hoc adjustment is made to the transfer coefficients computed under Dotson-Crosson and Getz. Specifically, the transfer coefficients under Dotson-Crosson are scaled down to 67% of the Schmidt et al. values (CPP option REDUCE_GAMMA_TS_DOTSON_CROSSON), and for Getz to 25% of the Schmidt et al. values (CPP option REDUCE_GAMMA_TS_GETZ). The results of the simulation have been evaluated against: (a) historical sea ice concentrations from SSM/I (Comiso 2017), (b) historical hydrographic profiles from the ASPIRE (austral summer 2010/2011) and ARTEMIS (summer 2022) cruises, (c) basal melt rates estimates from remote sensing (Adusumilli et al. 2020, Rignot et al. 2013, Depoorter et al. 2013), (d) tidal harmonic constituents from the pressure gauge of the mooring described in Wahlin et al. 2019. ******************************************************************************** Using the model outputs: Extracting a 3-D modeled field ******************************************************************************** To read the modeled practical salinity on day 5 of year 2006, under Matlab/Octave: salt = ncread( 'run661/roms_avg_0001.nc', 'salt', [1, 1, 1, 5], [inf, inf, inf, 1] ); (type `help ncread' to learn the syntax of command `ncread'.) ...which returns a three-dimensional array of size 1076x516x20 corresponding to axes xi_rho, eta_rho, and s_rho (respectively). s_rho is the topography-following vertical axis, and it is ordered from sea floor to sea surface. ******************************************************************************** Using the model outputs: Horizontal position of the model grid boxes ******************************************************************************** The latitude/longitude of the 1076x516 grid boxes defining the horizontal plane of the model grid are stored inside lon_rho and lat_rho: lon_rho = ncread( 'run661/roms_avg_0001.nc', 'lon_rho' ); lat_rho = ncread( 'run661/roms_avg_0001.nc', 'lat_rho' ); These values are constant over time and available in any of the output files. ******************************************************************************** Using the model outputs: Deriving the vertical position of the model grid boxes (and their thickness) ******************************************************************************** To obtain the depth of the 1076x516x20 grid boxes of the example above (again on day 5 of 2006): z__r = depths( 'run661/roms_avg_0005.nc', 'run661/roms_avg_0001.nc', 1, 0, 5 ); ...which returns a three-dimensional array with the same size as the `salt' array of the example above. z__r is in meters relative to the geoid, and z__r is defined negative for a grid box positioned below the geoid. The Amundsen configuration has 20 topography-following levels. To obtain the vertical position of the interfaces (i.e. the 21 surfaces that separate the 20 vertical levels), type: z__w = depths( 'run661/roms_avg_0001.nc', 'run661/roms_avg_0001.nc', 5, 0, 5 ); ...which returns a three-dimensional array with 21 values along the third dimension. The first value corresponds to the vertical position of the sea floor. The 21st value corresponds to the vertical position of the sea surface. All these values change over time (in response to the movement of the sea surface) except the sea floor (whose vertical position is fixed.) To obtain the thickness of the 20 topography-following levels: dz = z__w(:, :, 2 : end) - z__w(:, :, 1 : end - 1); ******************************************************************************** Using the model outputs: Deriving horizontal velocities u,v ******************************************************************************** The archive includes daily-averaged three-dimensional horizontal volume fluxes, Huon (aligned along the xi_rho axis of the model grid) and Hvom (aligned along the eta_rho axis of the model grid), with units of m^3/s. The benefit of saving daily-averaged Huon,Hvom over saving the usual u,v velocities is that correlations between thickness and horizontal velocities are preserved. An example of how to derive u,v from Huon,Hvom (again on day 5 of year 2006): dx = 1.5e3; % Horizontal mesh size (meters; constant throughout the grid). % Interpolate dz horizontally at the location of the `u',`v' points: dz_u = ( dz(1 : end - 1, :, :) + dz(2 : end, :, :) ) * 0.5; dz_v = ( dz(:, 1 : end - 1, :) + dz(:, 2 : end, :) ) * 0.5; clear dz; huon = ncread( 'run661/roms_avg_0001.nc', 'Huon', [1 1 1 5], [inf inf inf 1] ); hvom = ncread( 'run661/roms_avg_0001.nc', 'Hvom', [1 1 1 5], [inf inf inf 1] ); u = huon ./ dz_u / dx; % Units of m/s. v = hvom ./ dz_v / dx; % Units of m/s. (The "w" velocities are directly available in the files roms_avg_xxxx.nc.) ******************************************************************************** Using the model outputs: Deriving basal melt rates under ice shelves ******************************************************************************** Basal melt under an ice shelf corresponds to a sink of oceanic heat, i.e. a negative value for the heat flux at the ice shelf/ocean interface. The interfacial heat flux is stored inside the field shflux and has units of Watts per square meter: shflux = ncread( 'run661/roms_avg_0001.nc', 'shflux' ); % W/m2. A shflux value from under an ice shelf can be converted into a mass of ice melted using Eq.3 of Holland & Jenkins 1999: mass_ice = - shflux / 334.e3, ...where 334.e3 J/kg is the latent heat fusion and mass_ice has units of kg per second per square meter. ******************************************************************************** Acknowledgements ******************************************************************************** This research was supported by NASA award 80NSSC21K0746 (Antarctic sea ice, fast ice and icebergs: Modulators of ocean-ice shelf interactions (AMICUS)) and by NSF award 1941292 (NSFGEO-NERC: Collaborative Research: Accelerating Thwaites Ecosystem Impacts for the Southern Ocean (ARTEMIS)). The authors acknowledge William & Mary Research Computing (https://www.wm.edu/it/rc) for providing computational resources and/or technical support that have contributed to the results reported within this dataset. ******************************************************************************** References ******************************************************************************** Adusumilli, S. H.A. Fricker, B. Medley, M.R. Siegfried, 2020, Interannual variations in meltwater input to the Southern Ocean from Antarctic ice shelves, Nature Geoscience, https://doi.org/10.1038/s41561-020-0616-z Brodzik, M. J. and J. S. Stewart. 2016. Near-Real-Time SSM/I-SSMIS EASE-Grid Daily Global Ice Concentration and Snow Extent, Version 5. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. https://doi.org/10.5067/3KB2JPLFPK3R Budgell, W.P., 2005, Numerical simulation of ice-ocean variability in the Barents Sea region: Towards dynamical downscaling, Ocean Dynamics, 55:370-387, https://doi.org/10.1007/s10236-005-0008-3 Comiso, J. C. 2017. Bootstrap Sea Ice Concentrations from Nimbus-7 SMMR and DMSP SSM/I-SSMIS, Version 3. Subset used: year 2006-2022. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center, https://doi.org/10.5067/7Q8HCCWS4I0R Depoorter, M.A., J.L.Bamber, J.A.Griggs, J.T.M.Lenaerts, S.R.M.Ligtenberg, M.R.van den Broeke, G.Mohold, 2013, Calving fluxes and basal melt rates of Antarctic ice shelves, Nature, https://doi.org/10.1038/nature12567 Dinniman, M.S., J.M. Klinck, W.O. Smith Jr., 2011, A model study of Circumpolar Deep Water on the West Antarctic Peninsula and Ross Sea continental shelves, Deep-Sea Research II, 58:1508-1523, https://doi.org/10.1016/j.dsr2.2010.11.013 Dinniman, M.S., P. St-Laurent, K.R. Arrigo, E.E. Hofmann, G.L. van Dijken, 2020, Analysis of Iron Sources in Antarctic Continental Shelf Waters, J. Geophys. Res.: Oceans, 125, e2019JC015736, https://doi.org/10.1029/2019JC015736 Dorschel, B., et al., 2022, The International Bathymetric Chart of the Southern Ocean Version 2, Scientific Data, 9, article 275, https://doi.org/10.1038/s41597-022-01366-7 Hersbach, H., et al., 2020, The ERA5 global reanalysis, Quarterly Journal of the Royal Meteorological Society, 146, 1999-2049, https://doi.org/10.1002/qj.3803 Holland, D.M. and A. Jenkins, 1999, Modeling Thermodynamic Ice-Ocean Interactions at the Base of an Ice Shelf, J. Phys. Oceanogr., 29, 1787-1800, https://doi.org/10.1175/1520-0485(1999)029<1787:MTIOIA>2.0.CO;2 Jordan, T.A., D. Porter, K. Tinto, R. Millan, A. Muto, K. Hogan, R.D. Larter, A.G.C. Graham, J.D. Paden, 2020, New gravity-derived bathymetry for the Thwaites, Crosson, and Dotson ice shelves revealing two ice shelf populations, The Cryosphere, 14, https://doi.org/10.5194/tc-14-2869-2020 Large, W.G., J.C. McWilliams, S.C. Doney, 1994, Oceanic vertical mixing: A review and a model with a nonlocal boundary layer parameterization, Reviews of Geophysics, 32(4), 363-403, https://doi.org/10.1029/94RG01872 Mack, S.L, M.S. Dinniman, J.M. Klinck, D.J. McGillicuddy, L. Padman, 2019, Modeling Ocean Eddies on Antarctica's Cold Water Continental Shelves and Their Effects on Ice Shelf Basal Melting, J. Geophys. Res.: Oceans, 124, 5067-5084. https://doi.org/10.1029/2018JC014688 Marchesiello, P., J.C. McWilliams, A. Shchepetkin, 2001, Open boundary conditions for long-term integration of regional oceanic models, Ocean Modelling, 3, 1-20, https://doi.org/10.1016/S1463-5003(00)00013-5 Mazur, A.K., A.K. Wahlin, O. Kalen, 2019, The life cycle of small- to medium-sized icebergs in the Amundsen Sea Embayment, Polar Research, 38, https://doi.org/10.33265/polar.v38.3313 Mellor, G.L., L. Kantha, 1989, An ice-ocean coupled model, J. Geophys. Res., 94(C8), 10,937-10,954, https://doi.org/10.1029/JC094iC08p10937 Nakayama, Y., D. Menemenlis, M. Schodlok, E. Rignot, 2017, Amundsen and Bellingshausen Seas simulation with optimized ocean, sea ice, and thermodynamic ice shelf model parameters, J. Geophys. Res.: Oceans, 122(8), https://doi.org/10.1002/2016JC012538 Padman, L., H.A. Fricker, R. Coleman, S. Howard, L. Erofeeva, 2002, A new tide model for the Antarctic ice shelves and seas, Annals of glaciology, 34, 247-254, https://doi.org/10.3189/172756402781817752 Padman, L., S.Y. Erofeeva, H.A. Fricker, 2008, Improving Antarctic tide models by assimilation of ICESat laser altimetry over ice shelves, The Cryosphere, 35(22), https://doi.org/10.1029/2008GL035592 Rignot, E., S.Jacobs, J.Mouginot, B.Scheuchl, 2013, Ice-shelf melting around Antarctica, Science 341, 266, https://doi.org/10.1126/science.1235798 Rutgers branch of ROMS: https://www.myroms.org/ Scambos, T., B. Wallin, J. Bohlander, 2022, Images of Antarctic Ice Shelves, Version 2 [Data Set]. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. https://doi.org/10.5067/W87VCY3CW0MJ Schmidt, G.A., C.M. Bitz, U. Mikolajewicz L.B. Tremblay, 2004, Ice-ocean boundary conditions for coupled models, Ocean Modelling, 759-74, https://doi.org/10.1016/S1463-5003(03)00030-1 Shchepetkin, A.F., J.C. McWilliams, 2005, The regional oceanic modeling system (ROMS): a split-explicit, free-surface, topography-following-coordinate oceanic model, Ocean Modelling, 9, 347-404, https://doi.org/10.1016/j.ocemod.2004.08.002 St-Laurent, P., P. L. Yager, R. M. Sherrell, S. E. Stammerjohn, M. S. Dinniman, 2017, Pathways and supply of dissolved iron in the Amundsen Sea (Antarctica), J. Geophys. Res.: Oceans, 122(9), 7135-7162, https://doi.org/10.1002/2017JC013162 St-Laurent, P., S.E. Stammerjohn, T. Maksym, R.M. Sherrell, 2022, On the relative importance of offshelf/onshelf drivers of variability in mCDW inventory on the Amundsen Shelf, Antarctica, paper C15D-0621 presented at 2022 Fall Meeting, AGU, 12-16 Dec, https://www.dropbox.com/s/dfxi1tiih7vnv59/amundsen_fall_agu20221212.pdf?dl=0 Wahlin, A.K., J. Rolandsson, K.M. Assmann, 2019, Water temperature, salinity, oxygen, and other oceanographic data collected by CTD from a mooring deployed and recovered by the research vessel ice breaker Araon in the Amundsen Sea from 2016-01-28 to 2018-01-18 (NCEI Accession 0206672), NOAA National Centers for Environmental Information. Dataset. https://doi.org/10.25921/6pwp-1791