Printer Friendly

Using GIS to investigate septic system sites and nitrate pollution potential.

Introduction

According to a soil survey conducted by the Soil Conservation Service (now the Natural Resource Conservation Service), many on-site wastewater treatment systems in Summit County, Colorado, are located in soils for which the use of standard septic-system design is "severely limited" (1). The limitations are primarily due to slope, depth to bedrock, slow percolation, large stones, and/or extremely wet or flood conditions. In combination with the physical characteristics of the sites, the relative location of septic systems determines the impact those systems will have on groundwater and surface water quality The relationship between on-site wastewater disposal and groundwater quality cannot be properly evaluated without knowledge of factors that vary geographically, such as distance to residential water wells, land surface slopes, and soil conditions.

GIS for Environmental Health Departments

Because a geographic information system (GIS) can spatially represent site characteristics and hydrological responses with mapped features, such a system can help describe the suitability of a septic-system site and the potential for nitrate pollution of groundwater. Hydrological responses, which vary geographically, determine groundwater quality A GIS can incorporate spatial interactions to determine the extent of potential pollution loading on the basis of area and land types (2, 3).

GlS is very useful for county governments, which need to integrate multiple land uses into planning and which deal with issues that have spatial components (e.g., property boundaries, costs associated with distance to utilities, environmental factors). The integration of models and spatial data provides managers and planners with a new evaluation tool. A GIS thus provides resources for evaluating regulatory policies and management practices, economic feasibility, suitability of specific practices, and long-term impacts at a site (2, 4-7). These evaluations are often accomplished through examination of "before" and "after" conditions at a site or through examination of relationships between geographically related features.

This cooperative study, conducted by the Summit County Environmental Health Department, the Summit Water Quality Committee, and the Environmental Health Advanced Systems Laboratory at Colorado State University, resulted in the creation of a dynamic, multiuse GIS that Summit County can use for environmental health and land use planning operations. The geographic database can be updated with water quality data, maintenance records, altered property records, or septic-system placement and removal information. lt is anticipated that the GIS will be used to prioritize water quality improvement activities, integrate results, and measure improvements as water quality projects continue. Confirmation of problem areas should reduce unnecessary pollution control expenditures. The GIS will also enable the identification of sites suitable for septic systems.

Study Site and Background

The study area is confined to the 124 square miles of the upper Blue River watershed, [TABULAR DATA FOR TABLE 1 OMITTED] defined as the Blue River and its tributaries south of the Dillon Reservoir [ILLUSTRATION FOR FIGURE 1 OMITTED]. From its highest elevation at Quandry Peak (14,265 feet above sea level), the study basin drops almost 5,300 vertical feet to Dillon Reservoir at 9,000 feet above sea level.

On-site wastewater disposal systems are used by approximately 12 percent of the residences in southern Summit County, a mountain area that encompasses three major ski resorts and has a growing population. Agriculture is not a significant land use in this area, but increasing residential development in the basin threatens to increase the load of nutrients that flow into Dillon Reservoir from many other sources. Accelerated septic-system placement in recent years is the motivation for understanding the contribution of septic systems to the nutrient loading of the reservoir. Conclusions from a study of eutrophication in Dillon Reservoir suggest that on-site wastewater treatment accounts for the majority of the anthropogenic nonpoint-source nitrogen load (8). Little evidence is available, however, about the extent to which septic systems in various locations contribute to nitrogen in groundwater, or about where and how this contribution is reflected in surface water quality.

This study evaluated the potential for nitrate loading to groundwater primarily through classification of soil and geology. The subsurface hydrologic environment, which is the primary influence on groundwater movement, is interconnected with the surface-water regime in the study area and thus affects surface water quality. The location of high-permeability sediments, faults, fractures, and rocks directly influences the distribution of groundwater resources and the direction of groundwater flow in Summit County (9). Groundwater traveling through bedrock fractures may travel large distances very quickly with little natural purification (10).

Because the topography of the upper Blue River basin is extremely diverse, there is wide variation in snow accumulation and snowmelt regimes, and therefore in the magnitude of groundwater recharge. ln this moderately sized watershed, the annual snowfall averages from 160 inches near Dillon (elevation 9,000 feet) to 255 inches at the Breckenridge Ski Area (elevation 11,000 feet) (11). Precipitation and snowpack generally increase with elevation; however, wind, slope, aspect, and vegetation control snow accumulation (12).

GIS Development

The Summit County GIS was constructed on a SUN SPARCstation platform (from SUN Microsystems) with Arc/Info 7.0 software (from Environmental Systems Research Institute [ESRI]). Database manipulation that did not occur in the Arc/Info database module was performed in Microsoft Excel 7.0a (from Microsoft Corporation), Borland C++ 4.02 (from Borland International, Inc.), and dBASE III Plus 1.1 (from Ashton-Tate) on a Pentium computer (from Gateway 2000). For statistics operations and display purposes, SAS 6.11 (from SAS Institute, Inc.) and ArcView 3.0 (from ESRI) were used in a Unix environment in the final stages of the project.

Three coverages (data layers) were obtained in Arc/Info format: a topographic contour coverage from the Summit County Planning Department (SCPD), land parcel information and boundary coverages from SCPD, and geologic delineation coverages from the U.S. Geological Survey. Databases containing domestic well-sampling results from the Summit County Environmental Health Department (SCEHD), SCEHD septic-system permitting data, and well log information from the Colorado Department of Natural Resources (CDNR) each had a tax schedule number field. Each schedule number in these databases could be directly associated with a unique, county-issued personal property identification number (PPI) used by the Summit County Assessor, so that the assessor's data could then be joined to the spatial SCPD parcel coverage. The parcel coverage gives accurate property boundaries (represented as polygons), with a unique PPI for each property.

The 1980 Soil Conservation Service maps were registered and hand-digitized into the GIS with Universal Transverse Mercador control points at the corners of the four map plates (1). The digital data sets were all converted to the state plane coordinate system. Table 1 summarizes the raw data that were available for this study and indicates their sources.

The following fields were imported into the GIS from the SCEHD septic-system permitting database:

* PPI,

* activity (install, repair, or replace),

* inspection area,

* date,

* lot,

* block,

* tract,

* filing,

* public water (yes or no),

* lot size (determined at site assessment),

* occupancy (commercial or residential),

* number of bedrooms,

* percolation rate,

* soil type (identified at site assessment),

* presence of a French (interceptor) drain (yes or no),

* design code,

* variance (yes or no),

* tank capacity,

* easement (yes or no),

* dosing (pump or siphon), and

* tank material.

Arc/Info automatically generated lot-size values for the SCPD parcel coverage. These values, which were determined to be a more reliable measure than the values in the "lot size" field of the SCEHD septic-system permitting database, were used as lot area measurements in this study.

[TABULAR DATA FOR TABLE 2 OMITTED]

In 1994, SCEHD initiated a drinking-well monitoring program to measure nitrate concentrations in drinking water. After the assessor's database was used to find a unique PPI for each tax schedule number, the information in the database of samples from residential drinking wells was joined to the appropriate property in the GIS parcel coverage. The dates of the samples were included, as were the results of the three samples taken on each date. Among the 211 samples taken in the study area, the median nitrate concentration was 0.4 milligrams per liter (mg/L). Concentrations ranged from nondetectable to 8.95 mg/L, with an average of 0.83 mg/L and a standard deviation of 1.15 mg/L. In the present study these results were used for comparison with the nitrate pollution potential in the hydrogeologic area described in the following analysis.

Once the GIS coverages were developed for the data, analysis was performed to identify locations of nonpermitted septic systems and to characterize the vulnerability of septic-system sites on the basis of system design and location. The SCEHD septic-system permitting database was developed in 1973. It was assumed, therefore, that not all septic systems had been recorded in the permitting database. Information in the assessor's database about tax code areas and sewer fields was used to identify those parcels in the study area that were not on public sewer. Information from the local sanitation districts and personal input from the environmental health department served as checks in the process of determining septic-system locations.

Methods of Analysis

A digital geologic map of Colorado with a 1:500,000 scale was used to distinguish five geologic groups in the upper Blue River basin according to their ability to transmit water (13). The actual transmissivities of the formations are not known, so hydraulic conductivity ranges, groundwater supply data, and texture descriptions were used to classify the groups (9, 14, 15). From highest to lowest transmissivity, the geologic groups were as follows:

* Quaternary age alluvium deposits,

* Tertiary age intrusive crystalline rocks,

* shales from the Cretaceous age,

* sedimentary bedrock, and

* metamorphic-igneous rocks of Tertiary age.

For each polygon associated with the 1: 24,000-scale soil survey maps, the following attributes were recorded: permeability, depth to high water table, hydrologic soil group, and depth to bedrock. These characteristics were used to manually distinguish 11 unique soil groups.

The intersection of the soil and geologic groups resulted in 39 unique combinations that were termed "hydrogeologic units," the first unit of analysis. Subunits of the hydrogeologic unit (constituting groups of contiguous parcels that used septic systems and that belonged to the same type of hydrogeologic unit) were called "clusters," and served as a second unit of analysis. More specifically, cluster units comprised adjacent parcels of less than 10 acres each, each containing a septic-system, and all located in the same hydrogeologic unit. The flowchart in Figure 2 illustrates the process by which these units were delineated.

The DRASTIC Model

After the available data were integrated into a GIS, pollutant source areas were defined. The nitrate source areas (both hydrogeologic units and septic-system cluster units) were distinguished by their relative potential influence on groundwater quality and by the hydrogeologic response area in which they were located. The vulnerability of groundwater at each site was quantified with the DRASTlC model developed by the U.S. Environmental Protection Agency (EPA) and the National Water Well Association (NWWA) (16).

The DRASTIC model rates factors that affect pollution potential on a scale of 1 to 10. Ratings are based on the values associated with factors at a given site. Thus the rating for the depth-to-water-table factor is based on the number of feet from soil surface to water table; all possible values for the factor are assigned to 10 classes of depth, and the rating is the number of the class within which the depth-to-water-table factor falls at a given site. The classes and their ratings are defined for each factor in the documentation that accompanies the model (16). Each factor is also assigned a weight (ranging from 1 to 5), which represents the relative importance of that factor. The weights vary according to hydrogeologic setting, as defined in the model. From site to site within the Summit County study area, weights were uniform. To derive the pollution potential index for a site, the DRASTIC model multiplies rating by weight for each factor and adds all the results. Thus, for the study area,

INDEX = 5(D) +4(R) + 3(A) + 2(5) + 1(T) + 5(I) + 3(C) where

D = rating for depth to water table,

R = rating for net recharge,

A = rating for aquifer media,

S = rating for soil media,

T = rating for topography,

I = rating for impact to vadose zone, and

C = rating for hydraulic conductivity.

A higher DRASTIC factor rating entails a higher index and therefore greater vulnerability and greater potential for nitrate pollution.

The DRASTIC index was first calculated for each hydrogeologic unit to represent relative trends of groundwater vulnerability among the units. The local characteristics of depth to water table (the static water level in the CDNR well log data was used) and slope (derived from digital elevation coverage) were substituted to calculate the DRASTIC index for the cluster units.

Residential Well Nitrate Samples

The well nitrate data were stratified into four categories and regressed as independent variables against four septic-system site factors. These categories were

* nitrate concentration (mg/L) at each well,

* nitrate concentration (mg/L) at each well by quartile of DRASTIC results,

* median nitrate concentration for each hydrogeologic unit, and

* median nitrate concentration for each cluster.

The four septic-system site factors were

* septic-system age,

* the mean of the average slope of the parcels,

* the static water level measured at the well, and

* the DRASTIC groundwater vulnerability index.

Not all of the nitrate samples could be matched with parcels for which the SCEHD septic-system permitting database contained data or with the CDNR well log database; some PPIs used to link these databases were missing. Table 2 identifies the number of variables that matched for each category. As the categories became more specific, the number of observations in each group became very small. The median nitrate levels for each hydrogeologic unit and for each cluster were regressed with averages of each site factor. In all cases, correlations were evaluated with an alpha value of 0.05. Alpha was set at 0.05 to test for significance of the parameter estimation in the regressions. The fit of each regression was measured with an [R.sup.2] value, which is the square of the Pearson correlation coefficient (17).

A multiple regression that used all site factors was run for the outcome variable of nitrate level in individual wells. Complete information on which to perform this analysis was available in the records for 74 parcels (or observations).

Results

Thirty-nine unique hydrogeologic groups were created from combinations of five geologic groups and 11 soil groups, which were based on hydraulic conductivity and soil characteristics, respectively. Normalized to a scale of 100, the DRASTIC nitrate pollution potential values calculated for each unit ranged from 37 to 88 for the hydrogeologic units and from 35 to 84 for the clusters. The average rating for the hydrogeologic units was 58. For the clusters, it was 55. The wide range of values successfully distinguishes potential nitrate pollution in the county according to properties of soil and geology. Figure 3 illustrates the hydrogeologic unit coverage and the DRASTIC ratings given for each unit. The areas of alluvial fill or cumulic cryaquoll soil (poorly drained loam or clay surface over a shallow water table) stand out as being most vulnerable to nitrate pollution because the depth-to-water-table and impact-to-vadose-zone ratings were high for these groups. Figure 4 shows an example of the cluster coverage coded with DRASTIC index results. Because of the magnitude of the data set, only a small area is shown. Many clusters can have the same DRASTIC index.

The results of the regression analyses are presented in Table 2. None of the regressions resulted in an [R.sup.2] value greater than 0.5. Regression analyses were duplicated with base 10 logs of the nitrate concentrations in an attempt to normalize the population distribution of the nitrate measurements. Water quality data is often transformed with a log function because of the numerous small values (18). This procedure did not significantly change the results of the analyses.

Discussion

This paper describes a GIS that was developed for use by SCEHD in assessing the impact of septic-tank wastewater disposal systems on potential nitrate contamination of groundwater. With this GIS, spatial relationships between property locations can be queried and analyzed in terms of geophysical factors that influence the fate and transport of nitrates in groundwater systems. To demonstrate the analytical capabilities of the GIS, the authors coupled the GlS database with a groundwater vulnerability index model called DRASTIC. The GIS was used to determine DRASTIC values at three levels of resolution: individual parcels with septic systems, geographic regions with similar soil and geologic hydrologic response variables (termed hydrogeologic units), and subareas within the hydrogeologic units consisting of aggregated parcels that contain septic systems (termed septic-system cluster units). This study used GIS-generated data to conduct a statistical analysis of the relationship between groundwater vulnerability, well water quality, and physical site characteristics.

A GIS was essential in querying the database and extracting boundaries for each unit of analysis described above. Using overlay analyses, the authors were able to extract associated DRASTIC variables for each unit of analysis. The output from these queries resulted in additional coverages that could be used in further analysis. Without a GIS, the development of these spatially dependent databases would have been very tedious. The capability of mapping locations of specific vulnerability to groundwater pollution - and of linking these locations to parcel-level data on septic-tank locations and characteristics - greatly enhances the water quality protection processes of planning and infrastructure implementation.

This study was severely limited by the need to rely on existing data. The authors were unable to input data into the DRASTIC model that were specific to the analytical units used in this study The water quality samples that were used to assess the model were not collected specifically for the purpose of this study, which was to determine if septic-tank leachate is migrating to groundwater resources. Instead, the samples were taken to assess water quality in wells used as drinking water; these wells are located to minimize contamination to the well. Information was not available concerning the depth from which the water sample was taken, the location of the septic system on the property, or the potential for a hydraulic connection between the septic tank site and the water well. In fact, the nature of the hydrologic regime in the study area makes it possible for the groundwater recharge area to be quite distant from the well location. As a result, correlation statistics describing relationships between septic-system site characteristics and nitrate levels in the groundwater systems were very, weak ([R.sup.2] was less than 0.5 at p = .05). The authors recommend that to duplicate this approach future studies include, at a minimum, water quality sampling at locations that best represent the hydrologic regimes of the study areas.

The DRASTIC model has historically been used to characterize landscape vulnerability in study units and areas of much greater geographic extent than the ones used in this study. The limitation on its use at the spatial scale of this study is, again, availability of data. The model requires variables that are difficult to attain, particularly at the parcel level of resolution. This study did, however, demonstrate that, with GIS technology, data describing these variables at a coarser resolution could be used to analyze a landscape for groundwater vulnerability and locations of potential nitrate pollution sources.

Additional data enhancements are recommended. A geologic map at a finer scale (1:24,000) should be incorporated to allow the degree and density of fracturing to be considered. The 1:500,000-scale surficial geology map generalizes geologic contact zones and minimizes fault definitions (13). When a map of this scale is used, the degree of faulting for a particular formation may be misjudged. The interpretation of any results derived from these data must take these limitations into consideration.

The addition of a water table elevation map, including magnitudes of seasonal fluctuation, would greatly aid in the delineation of hydrogeologic units in the upper Blue River basin. At a scale of finer resolution, snow accumulation may be as important as hydrogeologic setting for the delineation of subsurface flow regimes. In mountainous areas, stream flow (and therefore nitrate loading to the stream) is sustained primarily by snowmelt recharge to shallow mountainous groundwater systems (12,19). The identification of water table fluctuations, along with hydraulic gradient, would provide information about the relative magnitude of recharge that aids nitrate transport through groundwater. The importance of tile recharge factor and differing interactions between snowmelt and groundwater in this basin may have been underestimated in this study, which simply chose a recharge value based on an annual recharge rate. Seasonal impacts of recharge could be addressed with finer-resolution data. Additional limitations of the DRASTIC model have been discussed by Aller et al. (16).

Conclusion

This study demonstrated the effective use of a GIS to represent septic-system site characteristics and associated potential for nitrate contamination. Potential nitrate contamination was quantified with the DRASTIC groundwater vulnerability model. The validity of this kind of model depends on the quality of the input data: The spatial scale should be appropriate, and the data should be representative of the hydrologic regime of the area. Although the data used in this study clearly were limited in these respects, the study did establish the utility of a GIS for managing and extracting the factors necessary for the DRASTIC model. Even at a coarse scale, the mapped index for the hydrogeologic units and the "clusters" serves as a tool with which variations in the vulnerability of groundwater to nitrate pollution can be distinguished when the environmental health manager is considering septic-system placement or is targeting areas for septic-system removal. A GIS is a comprehensive planning tool. Coupled with the DRASTIC model, this tool will be used in Summit County, Colorado, to recognize potentially contaminated areas, help prioritize water quality improvement projects, and support septic-system placement and removal decisions.

REFERENCES

1. Soil Conservation Service (1980), Soil Survey of Summit County, CO, Lakewood, Colo.: U.S. Department of Agriculture.

2. Kim, K., and S. Ventura (1993), "Large-Scale Modeling of Urban Nonpoint Source Pollution Using a Geographic Information System," Photogrammetric Engineering and Remote Sensing, 59(10):1539-1544.

3. Tim, U.S., D. Jain, and H. Liao (1996), "Interactive Modeling of Groundwater Vulnerability Within a Geographic Information System Environment," Groundwater, 24(4):618-626.

4. Flockhart, D.E., C.H. Sham, and Y. Xiao (1993), "Maximizing the Value of Information for Ground Water Protection: Three Test Cases," Water Resources Bulletin, 29(6):957-964.

5. Lemme, G., C.G. Carlson, R. Dean, and B. Khakural (1990), "Contamination Vulnerability Indexes: A Water Quality Planning Tool,"Journal of Soil and Water Conservation, March/April: 349-351.

6. Pringle-Baker, C., and E.C. Panciera (1990), "A Geographic Information System for Groundwater Protection Planning," Journal of Soil and Water Conservation, March/April:246-248.

7. Morse. G., A. Eatherall, and A. Jenkins (1994), "Managing Agricultural Pollution Using a Linked Geographical Information System and Non-Point Source Pollution Model," Journal of the Institute of Water and Environmental Management, 8:277-286.

8. Lewis. W.M., C. Brendecke, D. Crumpacker, and J.F. Saunders (1984), Eutrophication and Land Use, New York: Springer-Verlag, Inc.

9. Voegeli, P.T. (1965), Ground-Water Resources of North Park and Middle Park, Colorado - A Reconnaissance, Water-Supply Paper 1809-G, Denver, Colo.: U.S. Geological Survey.

10. Waltz, J.P. (1975), A System for Geologic Evaluation of Pollution Potential at Mountain Dwelling Sites, Completion Report Series, No. 59, Fort Collins, Colo.: Environmental Resources Center, Colorado State University.

11. National Climatic Data Center (1996), Summary of the Day, West 1 [compact disk]. Available from Earth Info [CD.sup.2], Boulder, Colo.

12. Chaney, T., T.H. Brooks, G. Kuhn, and others (1987), Hydrology of Area 58, Northern Great Plains and Rocky Mountain Coal Provinces, Colorado and Utah, Water-Resources Investigations, Open-File Report 85-479, Lakewood, Colo.: U.S. Geological Survey.

13. Green, G.N. (1992), The Digital Geologic Map of Colorado in ARC/INFO Format, Denver, Colo.: U.S. Department of the Interior Geological Survey.

14. Tweto, O., R.H. Moench, and J.C. Reed (1978), Geologic Map of the Leadville 1 x 2 Degree Quadrangle, Northwestern Colorado, Miscellaneous Investigations Series Map I-999, Denver, Colo.: U.S. Geological Survey

15. Domenico, P.A., and F.W. Schwartz (1990), Physical and Chemical Hydrogeology, New York: John Wiley and Sons, Inc.

16. Aller, L., T. Bennett, J.H. Lehr, and R.J. Petty (1987), "DRASTIC: A Standardized System for Evaluating Ground Water Pollution Potential Using Hydrogeologic Settings," EPA/600/2-85/018, Ada, Okla.: R.S. Kerr Environmental Research Laboratory, U.S. Environmental Protection Agency.

17. Graybill, F.A., and H.K. Iyer (1994), Regression Analysis: Concepts and Applications, Belmont, Calif.: Wadsworth, Inc.

18. McCuen, R.H. (1989), Hydrologic Analysis and Design, Englewood Cliffs, N.J.: Prentice Hall.

19. Flerchinger, G.N., K.R. Cooley, and D.R. Ralston (1992), "Groundwater Response to Snowmelt in a Mountainous Watershed," Journal of Hydrology, 133(1992):293-311.

Corresponding Author: John R. Nuckols, Associate Professor of Environmental Health, Environmental Health Advanced Systems Lab, Department of Environmental Health, College of Veterinary Medicine and Biomedical Science, Colorado State University, Fort Collins, CO 80523.
COPYRIGHT 1999 National Environmental Health Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1999, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:geographic information systems
Author:Rada, Jim
Publication:Journal of Environmental Health
Date:Apr 1, 1999
Words:4168
Previous Article:Hazard analysis critical control point approach to food safety: past, present, and future.
Next Article:Environmental pollution in India: heavy metals and radiogenic elements in Nacharam Lake.
Topics:


Related Articles
Geomatics on the cutting edge of information-age technology.
Geographic information systems: their use in environmental epidemiological research.
Using a spatial filter and a geographic information system to improve rabies surveillance data.
More on GIS at ATSDR: 2000 and beyond. (Products & Services).
High-tech maps: how to get here from there. (Social Studies & Science).
GIS improvements allow prediction of disaster risks, both natural and manmade.
GIS--a proven tool for public health analysis.
Savvy schools.
The introduction and expansion of GIS into a small local health department drinking-water program.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters