tcalving.rst - pism - [fork] customized build of PISM, the parallel ice sheet model (tillflux branch)
 (HTM) git clone git://src.adamsgaard.dk/pism
 (DIR) Log
 (DIR) Files
 (DIR) Refs
 (DIR) LICENSE
       ---
       tcalving.rst (8865B)
       ---
            1 .. include:: ../../../global.txt
            2 
            3 .. _sec-calving:
            4 
            5 Calving and front retreat
            6 -------------------------
            7 
            8 .. contents::
            9 
           10 The :numref:`tab-calving` summarizes options controlling calving parameterizations
           11 implemented in PISM.
           12 
           13 .. list-table:: Options for the calving models in PISM.
           14    :name: tab-calving
           15    :header-rows: 1
           16    :widths: 1,1
           17 
           18    * - Option
           19      - Description
           20 
           21    * - :opt:`-front_retreat_cfl`
           22      - Apply CFL-type criterion to reduce (limit) PISM's time step using the horizontal
           23        calving rate computed by ``eigen_calving`` or ``vonmises_calving``.
           24     
           25    * - :opt:`-calving eigen_calving`
           26      - Physically-based calving parameterization :cite:`Levermannetal2012`,
           27        :cite:`Winkelmannetal2011`. Whereever the product of principal strain rates is positive,
           28        the calving rate is proportional to this product.
           29 
           30    * - :opt:`-eigen_calving_K` (`m s`)
           31      - Sets the proportionality parameter `K` in `\text{m}\,\text{s}`.
           32 
           33    * - :opt:`-calving vonmises_calving`
           34      - Physically-based calving parameterization :cite:`Morlighem2016` that uses the tensile
           35        von Mises stresses.
           36 
           37    * - :opt:`-vonmises_calving_sigma_max` (`Pa`)
           38      - Sets the maximum tensile stress `\tilde{\sigma}` in `\text{Pa}`.
           39 
           40    * - :opt:`-calving thickness_calving`
           41      - Calve all near-terminus ice which is thinner than ice threshold thickness
           42        `H_{\textrm{cr}}`.
           43 
           44    * - :opt:`-thickness_calving_threshold` (m)
           45      - Sets the thickness threshold `H_{\textrm{cr}}` in meters.
           46 
           47    * - :opt:`-thickness_calving_threshold_file`
           48      - Specifies the file containing the variable :opt:`calving_threshold` to be used as
           49        the spetially-variable thickness threshold
           50 
           51    * - :opt:`-calving float_kill`
           52      - All floating ice is calved off immediately.
           53 
           54    * - :opt:`-float_kill_margin_only`
           55      - At each time step, calve cells at the ice margin only instead of removing all
           56        floating ice.
           57 
           58    * - :opt:`-float_kill_calve_near_grounding_line`
           59      - Calve floating ice near the grounding line (this is the default). Disable using
           60        :opt:`-float_kill_calve_near_grounding_line off`.
           61 
           62 To select several calving mechanisms, use a comma-separated list of keywords mentioned in
           63 :numref:`tab-calving`:
           64 
           65 .. code-block:: none
           66 
           67    -calving eigen_calving,thickness_calving,ocean_kill,vonmises_calving
           68 
           69 .. _sec-eigen-calving:
           70 
           71 Eigen calving
           72 ^^^^^^^^^^^^^
           73 
           74 PISM-PIK introduced a physically-based 2D-calving parameterization
           75 :cite:`Levermannetal2012`. This calving parameterization is turned on in PISM by option
           76 :opt:`-calving eigen_calving`. Average calving rates, `c`, are proportional to the product
           77 of principal components of the horizontal strain rates, `\dot{\epsilon}_{_\pm}`, derived
           78 from SSA-velocities
           79 
           80 .. math::
           81    :label: eq-calv2
           82 
           83    c = K\; \dot{\epsilon}_{_+}\; \dot{\epsilon}_{_-}\quad\text{and}\quad\dot{\epsilon}_{_\pm}>0\:.
           84 
           85 The rate `c` is in `\text{m}\,\text{s}^{-1}`, and the principal strain rates
           86 `\dot\epsilon_\pm` have units `\text{s}^{-1}`, so `K` has units `\text{m}\,\text{s}`. The
           87 constant `K` incorporates material properties of the ice at the front. It can be set using
           88 the :opt:`-eigen_calving_K` option or a configuration parameter
           89 (:config:`calving.eigen_calving.K` in |config-cdl|).
           90 
           91 The actual strain rate pattern strongly depends on the geometry and boundary conditions
           92 along the confinements of an ice shelf (coast, ice rises, front position). The strain rate
           93 pattern provides information in which regions preexisting fractures are likely to
           94 propagate, forming rifts (in two directions). These rifts may ultimately intersect,
           95 leading to the release of icebergs. This (and other) ice shelf calving models are not
           96 intended to resolve individual rifts or calving events, but it produces
           97 structurally-stable calving front positions which agree well with observations. Calving
           98 rates balance calving-front ice flow velocities on average.
           99 
          100 The partially-filled grid cell formulation (section :ref:`sec-part-grid`) provides a
          101 framework suitable to relate the calving rate produced by ``eigen_calving`` to the mass
          102 transport scheme at the ice shelf terminus. Ice shelf front advance and retreat due to
          103 calving are limited to a maximum of one grid cell length per (adaptive) time step. The
          104 calving rate (velocity) from ``eigen_calving`` can be used to limit the overall timestep
          105 of PISM--thus slowing down all of PISM--by using :opt:`-front_retreat_cfl`. This "CFL"-type
          106 time-step limitation is definitely recommended in high-resolution runs which attempt to
          107 model calving position accurately. Without this option, under certain conditions where
          108 PISM's adaptive time step happens to be long enough, dendritic structures can appear at
          109 the calving front because the calving mechanism cannot "keep up" with the computed calving
          110 rate.
          111 
          112 .. _sec-stress-calving:
          113 
          114 Von Mises stress calving
          115 ^^^^^^^^^^^^^^^^^^^^^^^^
          116 
          117 .. warning::
          118 
          119    This code is experimental and has not yet been thoroughly tested, use at your own risk.
          120    
          121 While eigen-calving (section :ref:`sec-eigen-calving`) is appropriate for Antartic ice
          122 shelves, it does not work for outlet glaciers that flow in narrow fjords. Along valleys
          123 with nearly parallel walls, the transverse component of the velocity is close to zero, and
          124 the transversal strain rate is therefore also close to zero and noisy.
          125 
          126 Instead of the product of the eigen strain rates, :cite:`Morlighem2016` proposes a calving
          127 law where the calving rate `c` is a functionally related to tensile stresses:
          128 
          129 .. math::
          130    :label: eq-calv3
          131 
          132    c = |\mathbf{u}| \frac{\tilde{\sigma}}{\sigma_{max}},
          133 
          134 where `\tilde{\sigma}` is the tensile von Mises stress and `\sigma_{max}` is a threshold
          135 that has units `Pa`. It can be set as a configuration parameter
          136 (:config:`calving.vonmises.sigma_max` in |config-cdl|). As the tensile fracture strength
          137 is much smaller than the compressive fracture strenth, the effective tensile strain rate
          138 is defined as
          139 
          140 .. math::
          141    :label: eq-calv4
          142 
          143    \tilde{\dot{\epsilon}}_e = \left(\frac{1}{2}\left(\max(0,\dot{\epsilon}_{_+})^2 +
          144    \max(0,\dot{\epsilon}_{_-})^2\right)\right)^{1/2}.
          145 
          146 Following :cite:`Morlighem2016` `\tilde{\sigma}` is given by
          147 
          148 .. math::
          149    :label: eq-calv5
          150 
          151    \tilde{\sigma} = \sqrt{3} B \tilde{\dot{{\epsilon}}}_e^{1/n},
          152 
          153 where `B` is the ice hardness.
          154 
          155 Similar to ``eigen_calving``, the calving rate from ``vonmises_calving`` can be used to
          156 limit the overall timestep of PISM --- thus slowing down all of PISM --- by using
          157 :opt:`-front_retreat_cfl`.
          158 
          159 .. _sec-additional-calving:
          160 
          161 Additional calving methods
          162 ^^^^^^^^^^^^^^^^^^^^^^^^^^
          163 
          164 PISM also includes three more basic calving mechanisms (:numref:`tab-calving`). The
          165 option :opt:`-calving thickness_calving` is based on the observation that ice shelf
          166 calving fronts are commonly thicker than about 150--250 m (even though the physical
          167 reasons are not clear yet). Accordingly, any floating ice thinner than `H_{\textrm{cr}}`
          168 is removed along the front, at a rate at most one grid cell per time step. The value of
          169 `H_{\mathrm{cr}}` can be set using the :opt:`-thickness_calving_threshold` option or the
          170 :config:`calving.thickness_calving.threshold` configuration parameter.
          171 
          172 To set a spatially-variable ice thickness threshold, use the option
          173 :opt:`-thickness_calving_threshold_file` or the parameter
          174 :config:`calving.thickness_calving.threshold_file`. This file should contain the variable
          175 :var:`calving_threshold` in meters (or other compatible units).
          176 
          177 Option :opt:`-calving float_kill` removes (calves), at each time step of the run, any ice
          178 that satisfies the flotation criterion. Use of this option implies that there are no ice
          179 shelves in the model at all.
          180 
          181 Use the option :opt:`-float_kill_margin_only` to restrict this to cells at the ice margin.
          182 
          183 Sometimes it is useful to preserve a one-cell-wide shelf near the grounding line. To do
          184 this, set :config:`calving.float_kill.calve_near_grounding_line` to false.
          185 
          186 .. _sec-prescribed-retreat:
          187 
          188 Prescribed front retreat
          189 ^^^^^^^^^^^^^^^^^^^^^^^^
          190 
          191 Option :opt:`-front_retreat_file` allows prescribing retreat of the ice front. The forcing
          192 file specified using this option should contain :var:`land_ice_area_fraction_retreat` ---
          193 a 2D field, possibly time-dependent, that contains ones in areas that may be covered by
          194 ice and zeros in areas that have to be ice-free. Values between `0` and `1` allow for a
          195 "partial" retreat on coarser grids.
          196 
          197 More precisely, :var:`land_ice_area_fraction_retreat` is a mask prescribing the *maximum
          198 ice extent* at a given time throughout a simulation; a certain rate of retreat can be
          199 prescribed by creating a field with an appropriately decreasing maximum extent.
          200 
          201 Changes in ice mass resulting from using this mechanism are reported as a part of the
          202 *discharge* (:var:`tendency_of_ice_mass_due_to_discharge`).
          203 
          204 .. note::
          205 
          206    This replaces the :literal:`ocean_kill` mechanism available in previous PISM versions.