Skip to content

Module bmb

Brief summary

The bmb module computes the basal mass balance state.bmb (m ice eq. yr⁻¹, positive for a gain, like smb), the basal source term of the thickness equation \(\partial_t h + \nabla\cdot(h\,\bar{\mathbf{u}}) = \mathrm{smb} + \mathrm{bmb}\) solved by the thk module. Under grounded ice it is minus the thermodynamic basal melt of the enthalpy module; under floating ice exposed to the ocean it is minus the sub-shelf melt rate of the selected law (prescribed, quadratic, pico, picop or plume), which reads the ocean properties of the ocean module. The parameters of the module are described here.

State variables

Reads: —

Writes: bmb, shelf_melt_rate, grounded_fraction

The module is opt-in: a land-terminated glacier does not need it. With method: zero and include_grounded_melt: true it only forwards the enthalpy melt (bmb = -basal_melt_rate).

Melt laws

The sub-shelf melt rate \(m\) (m ice eq. yr⁻¹, positive for melt) is given by method:

  • zero — no ocean melt.
  • prescribed — prescribed.form selects a uniform rate (constant, prescribed.value), a 2-D state variable (field, named by prescribed.field), or the MISMIP+ Ice1 melt (mismip_plus, Asay-Davis et al., 2016): \(m = \Omega \tanh(H_c/H_0)\,\max(z_0 - z_d, 0)\), with \(H_c\) the water-column thickness and \(z_d\) the ice draft.
  • quadratic — the quadratic thermal-forcing law \(m = \gamma_0 \left(\frac{\rho_w c_o}{\rho_i L}\right)^2 (\mathrm{TF} + \delta T)\,|\langle \mathrm{TF}\rangle + \delta T|\), with the thermal forcing TF of the ocean module at the ice base and \(\langle \mathrm{TF}\rangle\) its averaging: local (TF itself), shelf (mean over each ice shelf) or basin (mean over the shelves of each basin of the basins state variable, the ISMIP6 non-local form of Jourdain et al., 2020). With water_equivalent: true (the default) \(\gamma_0\) gives a melt in m w.e., as in ISMIP6, converted to ice.
  • pico — the box model of Reese et al. (2018): each ice shelf is divided into up to n_boxes boxes by distance from the grounding line, and the ocean water of the continental shelf (bed above continental_shelf_depth) is carried box by box with the overturning strength \(C\) and the heat-exchange velocity \(\gamma_T\); the freezing point is the PICO potential-temperature form \(T_f = aS + b - cp\). reference_distance sets the grounding-line distance at which a shelf has n_boxes boxes (0: the largest distance in the domain), and grounded areas smaller than maximum_ice_rise_area are ice rises rather than grounding line.
  • picop — the box properties of PICO with the plume melt of Lazeroms et al. (2018) (Pelle et al., 2019).
  • plume — the buoyant-plume melt of Lazeroms et al. (2019), fed by the ocean properties at the grounding-line depth of each shelf node.

For picop and plume, the grounding-line depth feeding each shelf node is carried along the ice flow from the grounding line (grounding_line_depth, a diffusion-regularized advection solved by Jacobi sweeps), so these two laws need iceflow to have run.

The law is evaluated every update_freq years (0: every step); the geometry, the grounding-line weights and bmb are updated every step. Between two evaluations, a node that has become floating takes the melt of its shelf neighbours at the last evaluation. The result is multiplied by melt_enhancer; negative rates (refreezing, e.g. in the outer PICO boxes) are clipped at zero unless allow_refreezing: true. With ocean_connected_only: true (the default), ocean melt is applied only under floating ice connected, through water, to the open ocean (a domain side with a zero thickness boundary, or else the ice-free water), not over subglacial lakes.

Grounding line

Across the grounding line the grounded and ocean melts are weighted by the grounded fraction \(g\) of each node's cell, estimated from grounding_line.sub_samples samples of the bilinear flotation function (published as state.grounded_fraction):

\[ \mathrm{bmb} = -\left(g\, m_\mathrm{grounded} + w\, m_\mathrm{ocean}\right) \quad \text{where there is ice, 0 elsewhere,} \]

with the ocean weight \(w\) set by grounding_line.treatment: nmp no ocean melt in the cells crossed by the grounding line, fmp full ocean melt in every cell containing floating ice (both after Seroussi and Morlighem, 2018), pmp ocean melt weighted by the floating fraction (Leguy et al., 2021).

Module ordering

Run bmb after enthalpy (for the grounded melt), after ocean (for the ocean properties), after iceflow for picop and plume, and right before thk:

override /processes:
  - time
  - ocean
  - iceflow
  - bmb
  - thk

Example usage

ISMIP6-like quadratic melt from an ocean profile, with the melt scaled by the floating fraction near the grounding line:

# @package _global_

processes:
  ocean:
    method: profile
  bmb:
    method: quadratic
    quadratic:
      averaging: shelf
      gamma0: 1.4477e+4
    grounding_line:
      treatment: pmp
  thk:
    front:
      method: sub_grid

MISMIP+ Ice1 melt, no melt in the grounding-line cells:

processes:
  bmb:
    method: prescribed
    prescribed:
      form: mismip_plus
    grounding_line:
      treatment: nmp

Parameters

The complete default configuration file can be found here: bmb.yaml.

Structure of the parameters:

bmb
├── method
├── update_freq
├── include_grounded_melt
├── melt_enhancer
├── allow_refreezing
├── ocean_connected_only
├── grounding_line
│   └── ...
├── physics
│   └── ...
├── prescribed
│   └── ...
├── quadratic
│   └── ...
├── pico
│   └── ...
├── picop
│   └── ...
├── plume
│   └── ...
└── grounding_line_depth
    └── ...

Description of the parameters:

Name Description Default value Units
method Sub-shelf melt law: zero (no ocean melt; with include_grounded_melt, bmb = -basal_melt_rate), prescribed, quadratic, pico, picop or plume. prescribed —
update_freq Re-evaluate the sub-shelf melt law every X years; the geometry, the grounding-line weights and bmb are updated every step. 0 evaluates the law every step. 0.0 yr
include_grounded_melt Use the thermodynamic basal melt of the enthalpy process (basal_melt_rate) under grounded ice, when it is available. True —
melt_enhancer Multiplicative factor applied to the sub-shelf melt rate. 1.0 —
allow_refreezing Keep negative sub-shelf melt rates (marine-ice accretion) instead of clipping them at zero. PICO refreezes in its outer boxes under cold shelves: set it to true to keep that. False —
ocean_connected_only Apply ocean melt only under floating ice connected, through water, to a domain side with a zero thickness boundary (or, when no such side touches water, to ice-free water), not over subglacial lakes. Periodic sides are not wrapped around. True —
grounding_line
Name Description Default value Units
grounding_line.treatment Melt near the grounding line: nmp no ocean melt in cells crossed by the grounding line and fmp full ocean melt in every cell containing floating ice (Seroussi and Morlighem, 2018), pmp ocean melt weighted by the floating fraction (Leguy et al., 2021). The grounded part of each cell receives the thermodynamic melt. nmp —
grounding_line.sub_samples Samples per direction and per quarter cell of the bilinear flotation function, used to estimate the floating fraction of each node's cell. 4 —
physics
Name Description Default value Units
physics.c_ocean Specific heat capacity of sea water. 3974.0 J kg\( ^{-1} \) K\( ^{-1} \)
physics.L_ice Latent heat of fusion of ice. 334000.0 J kg\( ^{-1} \)
prescribed
Name Description Default value Units
prescribed.form Prescribed melt: constant, field (a 2-D state variable) or mismip_plus (MISMIP+ Ice1, Asay-Davis et al., 2016). constant —
prescribed.value Uniform melt rate of the constant form. 0.0 m yr\( ^{-1} \) ice eq.
prescribed.field Name of the state variable holding the melt rate of the field form (m ice eq. yr^{-1}, positive for melt). Do not use bmelt, an alias of the enthalpy basal_melt_rate. prescribed_melt_rate —
prescribed.mismip_plus.omega Melt-rate factor Omega of m = Omega tanh(H_c/H0) max(z0 - z_d, 0). 0.2 yr\( ^{-1} \)
prescribed.mismip_plus.H0 Water-column thickness scale H0. 75.0 m
prescribed.mismip_plus.z0 Depth z0 below which melting occurs, relative to the water level. -100.0 m
quadratic
Name Description Default value Units
quadratic.averaging Thermal forcing in m = gamma0 (rho_w c_o/(rho_i L))^2 (TF + dT)| + dT|: local (TF itself), shelf (mean over each ice shelf) or basin (mean over the ice shelves of each basin of the basins field, the ISMIP6 non-local form of Jourdain et al., 2020). shelf —
quadratic.gamma0 Heat-exchange velocity. The default, 14477 m yr^{-1}, is the value used by Kori-ULB for the ISMIP6 non-local MeanAnt calibration of Jourdain et al. (2020). 14477.0 m yr\( ^{-1} \)
quadratic.delta_T Offset added to the thermal forcing. 0.0 K
quadratic.water_equivalent Interpret the melt given by gamma0 in m water equivalent, as in Jourdain et al. (2020), and convert it to ice. True —
pico
Name Description Default value Units
pico.n_boxes Maximum number of boxes of an ice shelf. 5 —
pico.gamma_T Turbulent heat-exchange velocity. 2e-05 m s\( ^{-1} \)
pico.overturning Overturning strength C. 1000000.0 m\( ^{6} \) s\( ^{-1} \) kg\( ^{-1} \)
pico.rho_star Reference sea-water density of the linear equation of state. 1033.0 kg m\( ^{-3} \)
pico.alpha Thermal expansion coefficient. 7.5e-05 K\( ^{-1} \)
pico.beta Haline contraction coefficient. 0.00077 (g kg\( ^{-1} \))\( ^{-1} \)
pico.reference_distance Distance from the grounding line, measured along grid paths (edge-neighbour steps) as the PICO distances are, for which a shelf has n_boxes boxes. 0 uses the largest distance in the domain, as Reese et al. (2018) do over all Antarctic ice shelves; set it for regional domains, whose largest shelf would otherwise always get n_boxes boxes. 0.0 m
pico.continental_shelf_depth Deepest bed, relative to the water level, of the ice-free ocean whose temperature and salinity feed the ice shelves. -800.0 m
pico.maximum_ice_rise_area Grounded-ice areas smaller than this are ice rises (e.g. pinning points): they are not grounding line, the distances cross them, and shelves joined by one form one shelf, as in PISM (default 1e5 km^{2} there). 0 treats all grounded ice as grounding line, as in Reese et al. (2018). 0.0 km\( ^{2} \)
pico.freezing_point.a Salinity coefficient of the potential-temperature freezing point T_f = a S + b - c p of PICO (Reese et al., 2018). -0.0572 °C (g kg\( ^{-1} \))\( ^{-1} \)
pico.freezing_point.b Constant of the PICO freezing point. 0.0788 °C
pico.freezing_point.c Pressure coefficient of the PICO freezing point, with the pressure p = rho_w g |z| at the ice draft z. 7.77e-08 °C Pa\( ^{-1} \)
picop
Name Description Default value Units
picop.M0 Melt scale of the Lazeroms et al. (2018) plume. 10.0 m yr\( ^{-1} \) K\( ^{-2} \)
picop.gamma1 Coefficient of the effective heat-exchange (Stanton) number. 0.545 —
picop.gamma2 Coefficient of the effective heat-exchange (Stanton) number. 3.5e-05 m\( ^{-1} \)
picop.Cd_Gamma_T Product sqrt(C_d) Gamma_T. 0.0011 —
picop.Cd_Gamma_TS0 Product sqrt(C_d) Gamma_TS0. 0.0006 —
picop.E0 Entrainment coefficient. 0.036 —
picop.Cd Drag coefficient. 0.0025 —
picop.x0 Dimensionless position of the melt-rate transition. 0.56 —
plume
Name Description Default value Units
plume.gamma_T Effective Stanton number. 0.00059 —
plume.E0 Entrainment coefficient. 0.036 —
plume.Cd Drag coefficient. 0.0025 —
plume.C_eps Slope correction parameter. 0.6 —
plume.alpha Thermal expansion coefficient. 3.87e-05 K\( ^{-1} \)
plume.beta Haline contraction coefficient. 0.000786 (g kg\( ^{-1} \))\( ^{-1} \)
grounding_line_depth
Name Description Default value Units
grounding_line_depth.epsilon Diffusion regularising the advection of the grounding-line depth along the ice flow (picop and plume). 1e-05 m\( ^{2} \) yr\( ^{-1} \)
grounding_line_depth.tol Convergence tolerance of the grounding-line depth iteration. 0.01 m
grounding_line_depth.max_iter Maximum number of Jacobi sweeps of the grounding-line depth iteration. 10000 —

Contributors: Thomas Gregov.