Link to a PowerPoint presentation of this article


DRASTIC Model

Depth to Aquifer
Recharge
Aquifer Media
Soil Media
Topography
Intermediate Materials
Conductivity

(Figures)



Introduction
In the Black Hills of western South Dakota, there is an extreme scarcity of water.  This is not an unusual phenomenon in the western prairie states.  However, in light of a two-year drought and a population boom that is unprecedented, the issue of water contamination is a huge one in the Black Hills right now.   The DRASTIC model is one that is widely used by the U.S. Environmental Protection Agency (USEPA) to determine potential for water contamination.  In fact, neighboring states have already used this model to determine their own directions for agricultural and urban planning (Wyoming, 1998).  
    Geology in the Black Hills is unique.  A crystalline pluton caused the uplift and bending of what would normally be flat-lying rock units, exposing the granite core from which Mount Rushmore is carved.  Some of the previously flat-lying rocks are classified as aquifers.  That is, they are rock units with sufficient porosity and permeability to contain and transmit economically valuable quantities of water over large distances (Cherry and Freeze, 1988.)  These aquifers are tapped by cities and individual wells in this region as well as by communities as far as 300 miles east at the Missouri River.  Because the plutonic uplift encouraged weathering and exposure of the regional aquifers, the Black Hills has become an area of recharge and extreme vulnerability for the aquifers.  Strobbel, et al have produced a cross section displaying the exposure of these aquifer units.   

Model
Several models exist for determining groundwater sensitivity.  The USEPA has produced a national coverage of aquifer sensitivity using a model called “DRASTIC.”  This study applies the same model to a smaller region.  The DRASTIC model for assessing groundwater sensitivity requires seven criteria: Depth to water table, Recharge rates, Aquifer permeability, Soil type, Topography, Impact of the Vadose Zone, and Conductivity of the Vadose Zone (Aller, 1985).  DRASTIC evaluates pollution potential based on weighted combination of these hydrogeologic settings. Each factor is assigned a weight based on its relative significance in affecting the pollution potential. Each factor is further assigned a rating for different ranges of the values. The typical ratings range are from 1-10 and the weights are from 1-5. The DRASTIC Index, a measure of the pollution potential, is computed by summation of the products of rating and weights for each factor.

A mathematical representation similar to that used at other locations is this:
 (DI)  DRASTIC Index = D r D w + R r R w + A r A w + S r S w + T r T w + I r I w + C r C w  where:
Dr = Ratings to the depth to water table
Dw = Weights assigned to the depth to water table.
Rr = Ratings for ranges of aquifer recharge
Rw = Weights for the aquifer recharge
Ar = Ratings assigned to aquifer media
Aw = Weights assigned to aquifer media
Sr = Ratings for the soil media
Sw = Weights for soil media
Tr = Ratings for topography (slope)
Tw = Weights assigned to topography
Ir = Ratings assigned to vadose zone
Iw = Weights assigned to vadose zone
Cr = Ratings for rates of hydraulic conductivity
Cw = Weights given to hydraulic conductivity

(Wyoming, 1998).

The actual assigned ratings for this project are listed on the Figures page

Data
Data shall be discussed according to its specific part in the model. 




Depth to Aquifer

Data for the multiple regional aquifers  was acquired in UTM projection for zone 13 from the United States Geological Survey (USGS). It was created from point data taken in The Black Hills Hydrology Study. http://sd.water.usgs.gov/projects/bhhs/digitaldata.html   The themes are available as .00e files.  Resolution is about 50 meters.  Themes used for this project include depth to formation for three separate rock units :  the Deadwood aquifer, the Madison aquifer, and the Minnelusa aquifer. Because this data was first received in 50 meter resolution shapefiles, a single, more compact theme had to be created  to contain only the attributes of units where they are the most surficial aquifer.  This was accomplished in several steps. 
First, a new field was created for name, so that while in one theme, each quifer could still be identified individually.  Each aquifer was assigned a number, representing its name. Then, the upper two aquifers, Minnelusa and Madison, were connected through a union operation, connecting both themes while keeping attributes of each theme. Then, to “thin out” the heavy data, several manipulations were performed that would, eventually, enable faster processing.  The attribute table was queried to find polygons where the lower aquifer had a depth value, but the upper aquifer had no depth value.  These would be the locations where the upper aquifer had been weathered away.  The selection was then inverted so that the selection included all areas where both aquifers were present.  All of these records were deleted.  The resulting shapefile consisted only of the polygons of the lower aquifer that was exposed.  Another union was then created between the new theme and Minnelusa so that it only contained the parts of aquifers where they were superficial.  This process was repeated, then, with the Madison/Minnelusa theme and the Deadwood Aquifer.  The computer process for merging where the Deadwood aquifer was concerned, was especially cumbersome, taking two to three hours.  Once the Deadwood was minimized to only the exposed polygons, all processing moved much faster. 
The product of all this manipulation was one theme:  Regional Aquifers.  Depth to all aquifers was transferred into a common field, GRID_CODE, names were also placed in a single field, NAME.  Figure 2 displays the depth to aquifer.



Recharge


According to the Hydrology Study, recharge values are established for each of these aquifers.  A new field was created in the attribute table for recharge, and values were placed in this field according to each aquifer.  This was accomplished quite easily, by first performing a query on the NAME field for the individual aquifer, and then calculating the value for RECHARGE.  The Deadwood Aquifer has a recharge of only 4 inches per year, the Minnelusa and Madison aquifers have a recharge of 6 inches per year.  For the most space-efficient theme, the Regional Aquifer theme was dissolved according to recharge, as is shown in Figure 3.


Aquifer Media
Aquifer media values were assigned in much the same way as recharge.  The Black Hills Hydrology Study has described these units in detail.  The Deadwood Aquifer is a poorly cemented sandstone, the Madison is limestone with caves in the upper third, and the Minnelusa is a dense limestone.  Since this model is intended for regional comparisons, ratings were directly associated with these descriptions.  See Figure 4.


Conductivity
Although discussing conductivity at this point is not in accordance with the order of the DRASTIC model, it does to be the best place to discuss assignment of hydraulic conductivity to the layers because this was performed much the same way as the assignment of recharge.  Hydraulic conductivity is, generally, a function of the aquifer media.  It is tested wherever test wells are drilled.
The materials were assigned hydraulic conductivity according to standard values found in Groundwater by Freeze and Cherry, 1992.  Although the region has only 3 aquifers, there are 4 different ratings for hydraulic conductivity.  There is the general conductivity of the Deadwood Aquifer, the general conductivity of the Minnelusa Aquifer, but the Madison Aquifer is divided into two distinct parts:  Upper and lower.  The upper Madison Aquifer is a series of interconnected caves, through which water can flow as fast as any stream.  The lower part of the aquifer is much denser, with a more “normal” conductivity.  In order to separate the two parts, all areas where the Madison is completely buried were given the highest rating, thus assuming that the cavernous part is present.  However, where the Madison is exposed and weathered to less than two-thirds its original thickness, it has a lower hydraulic conductivity.  Generally, this aquifer is 450 ft thick.  So, where the Madison Aquifer is exposed and thickness is less than 300 ft down to the next unit -the Deadwood Aquifer-  the conductivity was assigned a lower rating.  The ratings were put into their own field, called CONDUCT.  These values were dissolved, then, to create a the Hydraulic Conductivity map as seen in Figure 5.

 
Soil Media
Soil data was downloaded from the United States Department of Agriculture as ArcInfo files.  Using Import 71, supplied with the ArcView Suite, this package of soil data was converted into one shapefile and several .dbf files, which are included for reference.  Attributes within the shapefile for individual soils are rather cryptic.  Most data fields hold numbers identified by other data files.  For instance, one datafield in the shapefile was titled “ID.”  Values in this field are in a letter-number format, such as SD001.  There is a .dbf file included in the package that contains all of the proper names for these ID numbers.  To reduce some time spent switching between tables, the soil.shp file was joined to the table with ID numbers.  This was not the end of manipulating soil data.  The USDA uses a complex method of soil description and names its soils based on their location, not their descriptive qualities.  In the Black Hills, one of the more popular soils is called “Pierre-Okaona-Blackfoot”.  Although this type of name is readily understood by soil scientists, it did not fit well into the proposed GIS plan. 
There were 41 specifically named soils in the region, all of which needed to fit into just 10 general descriptions.  In order to understand the characteristics of each soil name, one must go to the USDA Soil Description website www.usda.gov/soils/descrip.htm  and enter each name individually.  A description of each soil, then, can be read.  Based on the descriptions at the USDA website, all 41 soils were grouped into categories as shown in Figure 6.

 
Topography
Since there were problems manipulating the DEM for this region into a useful projection, the soil data was also used to extrapolate topography.  In the descriptions of soils, there was a general thickness for each type.  Assuming that thick soils accumulate in flat-lying areas and thin soils signify steep slopes, a general range of slope was also applied to this theme.  The range was very general with only four values:  Flat, Gentle Slope, Moderate Slope, and Steep Slope.  Figure 7 shows the result of this assumption. 
This is not the ideal way to interpret regional slope, but for the purposes of this project, it did serve its purpose.  When run without the topography layer, the model finds that sensitive areas within this region are different, demonstrating that topography has some control on aquifer sensitivity. 
At this point, the problems relating to the DEM are unclear.  Besides being able to properly project it, or to change other themes into an appropriate projection, attempts to derive slope values from this DEM seemed to produce unreasonable results.  Most likely, there was some conflict between the understood map units and elevation units.  After several attempts to make the DEM work for this project, it was discarded.  This was somewhat due to the time required for processing a DEM that covers a 100 mile by 40 mile region at 10 ft resolution.  If an effort were made to fix this problem, the first order of business would be to reclassify the DEM into 100 ft resolution, which was the resolution for the final output of this project. 


Intermediate Materials
     Determining the intermediate, or vadose zone materials, requires a knowledge of local geology.  These values reflect the material that is above the aquifer, but below the soil.   According to the DRASTIC model, vadose zone materials of least permeability determine the hydraulic conductivity values.  Vadose Zone materials were interpreted from depth to aquifer and compared to the known stratigraphy in this flat-lying area.  Above the Minnelusa Aquifer, is a clay confining layer.  Visible as a thin layer in Figure 1. 
     Consequently, where all aquifers are flat and buried to at least a depth of 30 ft , the intermediate materials do not transmit much water.  They have a very low rating.  At other places, there are varying degrees of weathered limestone and gravels.  Where the crystalling core rocks are exposed, the material between bedrock and soil is only gravels.  Gravel transmits water very well.  It is also assumed that some of this gravel has eroded from the top of the mountain a little ways outward to increase the permeablility of materials above the other aquifers.  In general, permeability decreases outward from the center of the Black Hills.  Such an assumption must be seen as just that.  In a regional overview, this is most likely not an over-generalization.  However, if the same model were applied to a smaller area, samples should be examined to determine more specifically the intermediate materials.  As for all other layers, the final theme was created by dissolving for the Intermediate Materials field.  Figure 8 shows that although there is not an aquifer present over the entire region, there are soils and intermediate materials covering the entire region. 

ModelBuilder

The specific equation for the index model was described above as:
(DI)  DRASTIC Index = D r D w + R r R w + A r A w + S r S w + T r T w + I r I w + C r C w.  Using ModelBuilder for the map addition is a simple process.  Each of the themes is easily converted into a grid and then ranked in the ModelBuilder Wizard.  The addition and weights are shown in the following figure.  Weights were manipulated and changed many times, until the resulting DI theme was reasonable to an educated eye.  This theme also needed to coincide with some conclusions approached by the USGS, that specific regions were likely to be quite sensitive.  In the final Drastic Index map, the regions they predicted were within the top 3 most sensitive areas. 
Results
After much tinkering and manipulation Figure 9 has become the final result.  Nine represents the areas of highest sensitivity and three the lowest.  The Drastic Index in the region is definitely a combination of all seven themes.  When the weighted model was run leaving out any one parameter, the end results were different.  There is no area that ranks a ten for the Drastic Index, but those areas with nine and eight values are worth noting as locations for potential contamination.  The inner circular region represents the exposure of the Madison Aquifer where it is cavernous.  The outer, thin circular region is largely controlled by its lack of slope and sensitive soils.  The small region ranking a 9, portrayed in red, is the most sensitive area and lies within a national forest region, so is generally protected from pollution related to development.  These small areas would be best revealed to land management areas with jurisdiction in the region, so that any kind of dumping or, even, camping could be directed away from these highly sensitive places. 


Conclusions and Discussion
This project accomplished its goal of poducing a single map representing areas of high potential or introduction of contaminants into regional groundwater.  In a model for actual use in the region, problems with the DEM should be resolved. In addition, it would be advisable to take samples to determine intermediate materials directly, rather than extrapolating them from general geological knowledge.  The final output was done in 100m resolution, rather the intended 50m resolution due largely to the time required for processin the aquifer layer.  It is important for anyone reading this report to recognize that the resulting sensitivity map represents relative sensitivity within the region.  One cannot take this Drastic Index and directly compare it to other regions, even if they are using the same ratings and processes.  One minor modification was made in interpretation of the model:  recharge values were assigned based on literature describing the region.  The model as applied in the Wyoming study used precipitation values for the recharge.  A regional precipitation theme may slightly affect the output of the Drastic Index, but it is hard to say which would be more accurate. 

Bibliography
Aller,L., T. Bennett, J.H. Lehr, and R.J. Petty. 1985. DRASTIC: A Standardized System for
Evaluation Groundwater Pollution Potential Using Hydrogeologic Settings. USEPA 600/2-
85/0108. Ada, Okla.: USEPA, Robert S. Kerr Environmental Research Laboratory.

Freeze and Cherry.  Groundwater. 

USGS.  South Dakota Hydrological Study, Open-file report

Wyoming Water Resource Center.  1998.  Wyoming Ground Water Vulnerability
Assessment Handbook, Volume 1.  SDVC Report 98-01.  University of Wyoming.