Coastline Alteration

The mpas_tools.ocean.coastline_alteration module contains several functions for modifying the coastline in MPAS cells. This is accomplished by modifying the land mask that indicates which MPAS cells are ocean vs. land (or grounded ice).

Adding Land Blockages

The function mpas_tools.ocean.coastline_alteration.add_critical_land_blockages() add the masks associated with one or more transects to the land mask. This is useful if there are land features (typically peninsulas or narrow land bridges) that must block the ocean flow even if they are too narrow to be resolved in the MPAS mesh.

An example of the typical workflow that uses this function would be:

import xarray

from geometric_features import GeometricFeatures
from mpas_tools.mesh import conversion
from mpas_tools.ocean.coastline_alteration import add_critical_land_blockages


# an object used to read feature collections from the geometric_features
# package
gf = GeometricFeatures()

# get geometry of the land coverage from Natural Earth
fcLandCoverage = gf.read(componentName='natural_earth', objectType='region',
                         featureNames=['Land Coverage'])

# read in the base mesh
dsBaseMesh = xarray.open_dataset('base_mesh.nc')
# create a mask on the base mesh that is ones for land or grounded ice and
# zeros for ocean
dsLandMask = conversion.mask(dsBaseMesh, fcMask=fcLandCoverage)

# get a collection of features from geometric_features that are meant for
# use as critical land blockages
fcCritBlockages = gf.read(componentName='ocean', objectType='transect',
                          tags=['Critical_Land_Blockage'])

# make masks from the transects
dsCritBlockMask = conversion.mask(dsBaseMesh, fcMask=fcCritBlockages)

# add the masks to the "land" mask
dsLandMask = add_critical_land_blockages(dsLandMask, dsCritBlockMask)

Masking Land-locked Cells

By default, the land mask produced by the MPAS Mask Creator can produce ocean cells with vertices that are all on the land-ocean boundary. MPAS-Seaice uses a so-called Arakawa B-grid, meaning that velocities are located at vertices of the MPAS mesh. This means that sea-ice flow into or out of a given cell is not possible if all vertices of that cell are boundary vertices (where the velocity is, by definition, zero). This problem is alleviated by calling mpas_tools.ocean.coastline_alteration.add_land_locked_cells_to_mask(). Any “land-locked” cells with only boundary vertices are removed from the mesh. Land-locked cells are only removed poleward of a threshold latitude (43 degrees by default). The user can specify the number of iterations (nSweeps) of land-locked cell removal, since removing land-locked cells can produce new land-locked cells.

Here is an example workflow that removes land-locked cells:

import xarray

from geometric_features import GeometricFeatures
from mpas_tools.mesh import conversion
from mpas_tools.ocean.coastline_alteration import \
    add_land_locked_cells_to_mask


# an object used to read feature collections from the geometric_features
# package
gf = GeometricFeatures()

# get geometry of the land coverage from Natural Earth
fcLandCoverage = gf.read(componentName='natural_earth', objectType='region',
                         featureNames=['Land Coverage'])

# read in the base mesh
dsBaseMesh = xarray.open_dataset('base_mesh.nc')
# create a mask on the base mesh that is ones for land or grounded ice and
# zeros for ocean
dsLandMask = conversion.mask(dsBaseMesh, fcMask=fcLandCoverage)

# Find ocean cells that are land-locked, and alter the cell mask so that
# they are counted as land cells
dsLandMask = add_land_locked_cells_to_mask(dsLandMask, dsBaseMesh,
                                           latitude_threshold=43.0,
                                           nSweeps=20)

Widening Transects

Similarly to Masking Land-locked Cells, if critical passages in polar regions are too narrow, they can become blocked by sea ice that cannot be advected. Sea-ice flow is not possible unless channels are at least 2 cells wide. This widening is accomplished with mpas_tools.ocean.coastline_alteration.widen_transect_edge_masks(). Channels are only widened poleward of a threshold latitude (43 degrees by default).

An example workflow that includes transect-widening is:

import xarray

from geometric_features import GeometricFeatures
from mpas_tools.mesh import conversion
from mpas_tools.ocean.coastline_alteration import widen_transect_edge_masks

# an object used to read feature collections from the geometric_features
# package
gf = GeometricFeatures()

# read in the base mesh
dsBaseMesh = xarray.open_dataset('base_mesh.nc')

# merge transects for critical passages into critical_passages.geojson
fcCritPassages = gf.read(componentName='ocean', objectType='transect',
                         tags=['Critical_Passage'])

# create masks from the transects
dsCritPassMask = conversion.mask(dsBaseMesh, fcMask=fcCritPassages)

# Alter critical passages to be at least two cells wide, to avoid sea ice
# blockage.
dsCritPassMask = widen_transect_edge_masks(dsCritPassMask, dsBaseMesh,
                                           latitude_threshold=43.0)