What is it?
The groundwater flow packages are used to define the aquifer properties and conductance schemes used in formulating the finite difference problem to be solved. Depending on the MODFLOW Version, different flow packages with different parameters are available.- The Node Property Flow package (NPF), is the formulation of aquifer inputs for MODFLOW-6.
- In the Layer Property Flow package (LPF), aquifer properties are defined directly or through the use of parameters.
- In the Block-Centered Flow package (BCF6), aquifer properties are defined for all cells directly. This is the oldest and simplest formulation.
- The Upstream Weighting Flow package (UPW), is a formulation that was developed for MODFLOW-NWT and weights cell conductance based on upstream head (if MODFLOW-USG is used with the Newton formulation an upstream weighted version of the LPF or BCF package is automatically applied).
In general the parameters controlled in this tab include:
- Horizontal conductance
- Cell wetting and drying
- Stranded cell handling (for the NWT and PCG solvers)
- Dry cell smoothing thickness
- Vertical Flow and Conductance
- XT3D Formulation (for MODFLOW-6)
Horizontal Conductance
- Cell Face Averaging Scheme:- Harmonic mean (This is most appropriate for confined and unconfined aquifers with abrupt boundaries in transmissivity at the cell boundaries or for confined aquifers with uniform hydraulic conductivity.)
- Logarithmic mean (This is most appropriate for confined aquifers with gradually varying transmissivities.)
- Arithmetic mean (This is most appropriate for unconfined aquifers with uniform hydraulic conductivities.)
- Arithmetic mean of saturated thickness and harmonic-mean hydraulic conductivity (This method is used in MODFLOW-USG for the Newton-Raphson formulation.)
- Arithmetic mean of saturated thickness and logarithmic-mean hydraulic conductivity. (This is most appropriate for unconfined aquifers with gradually varying transmissivities.
Note: If the Newton-Raphson formulation is used, horizontal conductance is calculated from cell thicknesses and then weighted by the saturated fraction of the upstream cell.
Cell Wetting and Drying
- Wetting Enabled [checkbox; expandable]: This indicates that dry cells can be rewetted during iteration [default enabled]. Note: the wetting capability does not work well with the SSOR solver.- Wetting threshold [decimal; meter]: [default 0.01; range 0.001-0.5]
- Threshold for initial head calc. [checkbox]: A flag (IHDWET) that determines which equation is used to define the initial head at cells that become wet.
Default (unchecked; IHDWET=0): wetted_head = cell_bottom + wetting_factor(hn - cell_bottom); where hn is the head in the neighboring cell that is causing the dry cell to become wet.
Checked (IHDWET!=0): wetted_head = cell_bottom + wetting_factor(wetting_threshold); where wetting_threshold is the same value that converts the cell to wet.
- Wetting Factor [decimal]: A factor for the calculation of the head at a cell when it is converted from dry to wet [default 0.1].
- Scan Dry/Wet [integer]: The iteration interval for attempting to wet cells [default 1; range 1-5]
- Wetting Option: Wet cells from below only (default), or from any adjacent cell.
Stranded Cell Handling
- Stranded cell options:- PCG solver ihcofadd: Always set dry (default; ihcofadd=0) sets an active cell that is surrounded by dry cells to dry.
Only set dry if no nonlinear terms (ihcofadd!=0) sets an active cell surrounded by dry cells to dry only if there are no head-dependent stresses or storage terms (HCOF=0).
- NWT solver ibotav: Do not apply correction (default; ibotav=0) allows cell heads in the deepest layer to fall below the cell-bottom altitude.
Apply correction (ibotav=1) so that head for cells in the deepest layer, and for single-layer models, cannot fall below the cell-bottom altitude.
Note: while ibotav does not change the effective solution, it can affect convergence behavior for cells with thin saturated thickness.
Dry Cell Smoothing Thickness Ω
When the Newton-Raphson formulation is used, the abrupt transitions between dry, saturated, and unsaturated thickness must be smoothed to allow calculation of partial derivatives. This is handled by using a smoothed, saturated cell fraction that is calculated using Ω, a small length, over which smoothing occurs. In MODFLOW-6 and MODFLOW-USG this value is fixed at 1e-6 meter.
- Dry cell smoothing thickness [decimal; meter]: This sets Ω, the smoothing thickness (THICKFACT) for the NWT solver [default 1e-5; range 1e-1 to 1e-10].
Vertical Flow and Conductance
- Vertical Scheme for Unconfined Conditions [expandable]: There are several options for computing the intercell vertical flow and conductance between overlying to underlying cells in adjacent layers.- Conductance: From Saturated Thickness [default see 'Variable' below] employs variable conductance based on saturated thickness, from Cell Thickness (see 'Constant' below) holds the vertical conductance constant and under the assumption that cells above and below are fully saturated. Constant conductance is potentially more stable but less accurate and is always used with the Newton Method.
- Perched Water Table: These corrections can but turned on or off if a perched aquifer (see 'Perched' figure below) is being modeled.
- Conductance Correction: Without the
- Flow Calc. Correction: If the correction for a cell overlying a dewatered convertible cell is applied, the head difference used in Darcy’s Law is equal to the head in the overlying cell minus the bottom elevation of the overlying cell. Otherwise, the head difference between the two cells is used.
XT3D Formulation
The XT3D option extends the capabilities of MODFLOW by enabling simulation of fully three-dimensional anisotropy with the full, three-dimensional conductivity tensor. It can also improve the accuracy of groundwater-flow simulations in cases in which the model grid violates certain geometric requirements. The XT3D formulation includes “neighbors of neighbors” of the two cells for which the flow is being calculated and is thus typically more computationally intensive than the default conductance-based NPF formulation. The user should consider whether the conductance-based formulation alone can provide acceptable accuracy for the problem being solved before using the XT3D option. Documentation for the “XT3D” Option in the Node Property Flow (NPF) Package of MODFLOW 6 can be found at Provost, Langevin & Hughes (2017)- Use XT3D FLOW Formulation [checkbox]: This options turns on the XT3D option. By default the XT3D terms will be put into the coefficient matrix.
- Move extra XT3D terms from coefficient matrix to right hand side [checkbox]: If checked ('RHS'), XT3D terms will be added to the right-hand side of the head equation [default unchecked]