Anisotropic Flow in the Red River Alluvium

Caddo Parish, Louisiana

By: John Hubbard

Department of Geography

University of Texas at Austin 

Introduction

      An anisotropic aquifer is defined as an aquifer which permeability varies with the direction of flow (Hornberger et al., 1998). Mapping flow in anisotropic aquifers is particularly problematic; requiring complicated modeling techniques.  Can a GIS be used to model these systems? Can ArcView be used to map the potentiometric surface in an anisotropic aquifer? Can spatial analyst be used to map flow direction?  The Red River alluvial aquifer is used as an example of an anisotropic aquifer in this study.   

Motivation 

    The alluvium is used by local farmers and ranchers as an agricultural water source. A project very similar to this one, using the same data set, was begun my senior year of undergraduate study.  That study was to ascertain if the river or precipitation is the major driving force with-in the alluvium.  The anisotropy of the flow, however, caused some complications.  This class caused me to wonder if ArcView could be used to model this complex flow.   

Location

     The  Red River alluvium covers 2,120 square-miles of Louisiana (Figure 1).

Figure 1: The Aquifers of Louisiana

The alluvium occupies the Red River Valley from the Arkansas border in the northwest until it joins the Mississippi River.  Caddo Parish (Figure 2) is located in the northwest corner of Louisiana. It is bounded on the west by Texas and on the east by the Red River. Shreveport is home to the majority of the residents       

Figure 2: Caddo Parish, Louisiana

Data

    Data were obtained from the The Red River Waterways Project for 64 USGS observation wells and imported into excel.  The data were sorted and the year 1980 was selected as an average year with the most data available prior to the construction of the Red River Lock-and-Dam system. The annual high and low  water elevation was found for each well, and the annual average was calculated. These data were then complied and saved as 1980.dbf (Table 1).

Table 1: 1980.dbf

Further well information was obtained from the Louisiana Department of Transportation and Development  (DOTD) Public and Water Resources Division. These data includes the latitude and longitude of each well in degrees, minutes, and seconds. This table was saved in excel as wells.dbf (Table 2) and the latitude and longitudes were converted to decimal degrees.

Table 2: Wells.dbf

Aquifer.shp (Figure 1) and Parish.shp (Figure 2) were also obtained from DOTD.  The National Hydrography Dataset (NHD) file for Cataloging Unit (CU) 1140202 was downloaded as 11140202.shp (Figure 4).  The EPA BASINS core data for the same CU were obtained, but were only used in the generation of reference maps and not used in the actual analysis. 

Procedure

            The 1980.dbf and Wells.dbf were added to ArcView and joined by well number. The newly joined elevation table mapped using "Add Event

            Theme." The new theme was saved as wells.shp. (Figure 3).

Figure 3: Wells.shp

    Aquifer.shp was then add to the view. The Alluvial aquifer was selected, and a "New Theme" was created, Alluvium.shp (Figure 4).

Figure 5: Alluvium.shp in brown.

11140202.shp was also added. The eastern boundary of Alluvium.shp was traced by hand using the line tool and copied. The western bank of the river (11140202.shp) was then selected and a "New Theme" was created, Boundry.shp. Boundry.shp was then edited and the line was pasted in. (Figure 6)

Figure 6: Boundry.shp in yellow.

Spatial Analyst was then loaded, and "Interpolate grid" was run (Figure 7).

Figure 7

The Z value Field was set as the Average (or mean) well elevation.  Boundry.shp was set as the Barrier. After two hours of processing a grid of the potentiometric surface was produced (Figure 8). 

Figure 8: Potentiometric Surface of average well elevation. 

An attempt was made to use spatial analyst to calculate the slop and aspect of the grid. However, the program then produced an "Error in reading STA file message."  The data were then twice reconstructed, and the same error was given.

The BUILDSTA function in Workstation ArcInfo Grid, was used on all three grids, but to no avail. 

Figure 9 shows what the aspect for such a grid might look like, but note that it has no boundaries. 

Figure 9: An aspect grid

Problems and Future Work

    The first and most obvious problem is the STA error.  All the data would need to be revaluated to ascertain the source of the error. Secondly, was the lack of data points. The well data for Caddo was no regularly taken, and therefore there are large data gaps in the set. 

    Another issue for the future is the fact that the river is not included in the creation of potentiometric surface. There is undoubtedly an interaction between the ground water and the river.  This is very important when modeling groundwater flow. River stage should be incorporated into future models. 

    Lastly, the hydraulic conductivity was not accounted for in this model. Workstation ArcInfo Grid has an application called Darcyflow that does take into account hydraulic conductivity.  However, hydraulic conductivity in anisotropic aquifers is relative to flow direction. Darcyflow assumes the aquifer has one hydraulic conductivity, and is therefore isotropic.  An application that uses a more complicated model would be needed for an accurate analysis. 

Acknowledgements 

Dr. David Bieler, Centenary College of Louisiana 

Dr. David Maidment, University of Texas at Austin

Damon Scott

Works Cited

Hornberger, G.M., J.P. Raffensperger, P. L. Wieberg, and K.N. Eshleman. 1998. Elements of Physical Hydrology. Johns 

     Hopkins   University Press, Baltimore, Maryland, USA.