A Comparison of Groundwater Modeling Approaches: Conceptual vs Numerical

January 24, 2014

With the release of Visual MODFLOW Flex, we introduced a new streamlined approach to building groundwater models. Along with the traditional numerical modeling approach, we provided groundwater professionals with the option to develop MODFLOW groundwater models from a 3D conceptual model. By incorporating raw GIS data that is completely independent of a numerical grid, the conceptual modeling approach offers many benefits over traditional numerical modeling methods. Despite the benefits, we realize that adopting new workflows can be a daunting task – especially with time constraints and project deadlines looming. To help ease the transition, we have prepared a short guide that compares both modeling workflows side-by-side for each step in the modeling process.

Step 1

Visual MODFLOW Classic – Traditional Numerical ModelingVisual MODFLOW Flex – Conceptual Modeling
Define the gridImport your data
Define the grid (rows and columns), import basemaps to reference boundaries or properties and digitize inactive cells.Import relevant data files from a variety of sources such as shapefiles, raster grids, images, points, and well locations. Create surfaces and determine the area for your conceptual model.

Step 2

Visual MODFLOW ClassicVisual MODFLOW Flex
Define the Grid Layer ElevationsDefine Conceptual Model Structure
Assign grid layer elevations by importing surfaces, interpolating XYZ points or specifying a thickness or elevationDefine the model area and surfaces using polygon and surface data objects. Convert surfaces to horizons, by specifying a horizon type for each surface (Erosional, Conformable, or Discontinuous)

Step 3

Visual MODFLOW ClassicVisual MODFLOW Flex
Assign Property Parameters to Grid CellsDefine Property Zones
Assign hydraulic parameters to selected grid cells by digitizing polygons or polylines, importing and interpolating points or importing surfaces.Assign hydraulic parameter to 3D geological formations by defining a constant value, linking to a surface or using polygon shapefiles with attributes.

Step 4

Visual MODFLOW ClassicVisual MODFLOW Flex
Select Grid Cells and Define Boundary Condition ParametersSelect Shapes and Define Boundary Condition Parameters
Assign boundary condition parameters to selected grid cells by importing: shapefiles with attributes, gridded data, MODFLOW packages, transient data from external files; or digitizing polygons or polylines.Assign boundary condition parameters to selected shapes by defining the geometry (using user created shapes or importing DXF or shapefiles with attributes) and defining the parameters (time schedules for transient data, constant values, shapefile attributes or raster grids).

Step 5

Visual MODFLOW ClassicVisual MODFLOW Flex
Translate and RunGenerate Numerical Grid
Translate the numerical inputs into MODFLOW packages and run MODFLOW.Define the grid sizing and the layering type (Deformed or Non-conforming). Refine the grid where needed or add local (child) grids around areas of interest.

Step 6

Visual MODFLOW Flex
With Visual MODFLOW Flex, you’re almost there! Once you have the grid, you can generate a numerical model from the conceptual model objects (property zones and boundary conditions). This is done automatically through a conversion routine. The results are shown below.

Step 7

Visual MODFLOW Flex
Translate and Run
Translate the numerical inputs into MODFLOW packages and run MODFLOW.

Step 8

After the model has been run you may need adjust the grid to account for new features or to improve model accuracy and stability. This involves refining or coarsening the horizontal grid or the vertical layering.

Visual MODFLOW ClassicVisual MODFLOW Flex
Adjust the grid and manually correct the properties and boundary conditions.Adjust the grid and automatically re-generate the properties and boundary conditions.
With numerical modeling, changing the grid is easy, but the difficult part comes afterwards when you need to adjust the property zones and boundary conditions that are assigned to those grid cells. This is a time-consuming task, and often it is just simpler to delete the previous properties and boundaries, and re-assign to the new grid cells.With conceptual modeling, the property and boundary condition objects are assigned to shapes, not grid cells. This gives you the freedom to make any change to the numerical grid and automatically re-generate the numerical model with the appropriate properties and boundary conditions.