Tutorial 16: Data-intensive analysis for predicting groundwater flow directions

This video illustrates the use of the Michigan Groundwater Management Tool (MGMT) to delineate groundwater flow directions using extensive hydro-geologic data from Michigan’s massive statewide groundwater database. MGMT uses more than half a million water well records in Michigan and other hydro-geologic data available to quickly and efficiently create a map of hydraulic head contours in any part of Michigan. This example uses this approach to map the flow directions and thereby use particle tracking to delineate a contaminant plume in Antrim County in north-western Michigan. The plume delineated using this approach is compared to that delineated by traditional methods, i.e., by collecting extensive site-specific data. The plume delineated by the data-intensive approach of MGMT compares favorably to the traditional approach. MGMT also allows reverse particle tracking, which can be used to identify the source of contamination.

The steps to create the model are outlined briefly:

(1)    Upon opening MGMT, we see the MGMT interface with the map of Michigan shown in the display window. The data layers, shown to the left, lists all the different layers of data that are available in Michigan. The top of the window shows the various tools available to process, visualize and analyze the data.

(2)    The map is zoomed in to focus on the area of interest, Antrim County.

(3)    In the data layers, expand and check the “Potential Contaminants” group and check the “Plume” layer. This displays a polygon in Antrim County that represents an existing contaminant plume delineated by traditional data collection.

(4)     Zoom into the map to focus on the area around the plume.

(5)    Click on “Create polygon raster” and while holding on to the Alt button, draw a rectangle around the area containing the plume.

(6)    A “Model Options” window opens up which has a number of options, including:

·         Number of grids, NX and NY, to discretize the area into and the corresponding grid cell size

·         Effective porosity

·         Molecular diffusion coefficients and local dispersivities.

·         Hydraulic conductivity used in the model

·         the data used to interpolate the hydraulic head contours in this area

·         the interpolation method used and the resolution of the interpolation, and;

·         the number o smoothing iterations.

(7)    Under scale, select the 540 m resolution

(8)    Two methods are available for interpolation, namely Inverse Distance Weighted (IDW) or Non-stationary Kriging (NSK). Both methods can use either 30 or 50 nearest neighbors to perform interpolation. Under method, select IDW with 30 neighbors.

(9)    Under Source, select “All Rivers Big Lakes, Well with outliers removed” – this data combination uses all rivers, big lakes and a combination of water wells in which, those wells whose water levels are outliers have been removed.

(10)Under Smoothing, select “Pass010”, i.e., the interpolated heads have been smoothed through 10 passes of a low-pass filter.

(11)Click “Apply”

(12) A flow field showing hydraulic head contours appears – the colors red and blue refer to high and low heads respectively. Water flows from high to low in a direction perpendicular to the head contours.

(13) Click on “Place particle by drawing a polygon” and draw a polygon near the source of the plume. After drawing the polygon, check the “Display Particle Path” on the right hand side.

(14) Click “Start Running Model” – this starts advecting the particles along the flow directions. Notice that the plume delineated by the particles is very similar to the existing plume delineation.

(15) Click “Stop running model”

(16) Click “Remove all particles”

(17) Click “Refresh” to refresh the image.

(18)Click “Place single particle on the map” and click to create multiple particles at the downstream end of the plume.

(19)Click “Reverse running model”. Notice that the particles all track back to the source of the contamination.