solver_nb.args_array

solver_nb.args_array(cell, dt, dz, LW_in, SW_in, T_air, p_air, T_dp, wind, Tsfc=273.15, lid=False, Sfrac_lid=np.array([np.nan]), k_lid=np.nan, fixed_sfc=True, N=50)

Convert the variables from an instance of the IceShelf class into the format required to run NumbaMinpack.hybrd, a root finder written in Fortran compatible with Numba jitted functions. This should result in some speedup compared to using scipy’s equivalent function fsolve.

We need to convert any non-arrays into 1d arrays with one element to concatenate them into this vector, hence the “np.array([var])” syntax. This is because Numba is compiled rather than interpreted; the code expects to see arrays rather than floats and cannot infer types like regular Python.

Called whenever the various heat equation implemenations are used, notably in heateqn, heateqn_lid and heateqn_fixedsfc, which are in turn called by the relevant solver functions (see get_firn_heateqn_solver, get_lid_solvers).

Parameters:

  • cell (core.iceshelf_class.IceShelf) – IceShelf object containing the relevant information on the temperature, etc. of the firn column

  • dt (int) – timestep in seconds [s]

  • dz (float) – size of each vertical grid cell [m]

  • LW_in (array_like, float, dimension(cell.vert_grid)) – Surface downwelling longwave radiation [W m^-2]

  • SW_in (array_like, float, dimension(cell.vert_grid)) – Surface downwelling shortwave radiation [W m^-2]

  • T_air (array_like, float, dimension(cell.vert_grid)) – Surface air temperature [K]

  • p_air (array_like, float, dimension(cell.vert_grid)) – Surface air pressure [Pa]

  • T_dp (array_like, float, dimension(cell.vert_grid)) – Dewpoint temperature [K]

  • wind (array_like, float, dimension(cell.vert_grid)) – Wind speed [m s^-1]

  • fixed_sfc (bool, optional) – Boolean flag to determine whether to use the fixed surface heat equation.

  • Tsfc (float, optional) – Surface temperature to fix the surface of the firn column to, if applicable (i.e. if using fixed_sfc=True)

  • lid (bool, optional) – Boolean flag determining whether we want the lid heat equation solver.

  • Sfrac_lid (array_like, float, optional, dimension(cell.vert_grid)) – Solid fraction of the frozen lid, if applicable. Default to np.array([np.nan]), as array type, and not used in non-lid cases.

  • k_lid (float, optional) – thermal conducitvity of the frozen lid, if applicable. Default np.nan, as not used in non-lid cases.

Returns:

args – Numpy array containing the arguments we want to pass into the heat equation solver.

Return type:

array_like