Mine Reclamation Using a Collaborative GIS-based Approach
Abstract
The repercussions of metal mining during the late 1800's currently threaten the Tenmile Creek watershed. Tetra Tech EM Inc. (TtEMI), was tasked by the EPA to determine the nature and extent of mine waste at over 100 sites and to prevent additional mine wastes from entering Helena's water supply. To meet the demands of this task, TtEMI implemented a GIS program which required the collaboration of two TtEMI offices, state and federal agencies, mining companies, and several consultants to adequately characterize the site and prevent further deterioration of Helena's water supply. TtEMI implemented an innovative ArcInfo and Arc View GIS solution on a watershed basis to organize thousands of analytical results, process imagery, present model results, and monitor clean up. As the project continues, the EPA will require all data to be posted to a GIS-enabled web site where regulatory personnel, data users and the public will have 24 hour access to reclamation status and data. The enterprise GIS solution was the logical approach due to the distribution of sites throughout the watershed and project requirements for sharing data.
 
I. Introduction
TtEM Inc. has developed a Geographic Information System (GIS) for the Region VIII Environmental Protection Agency (EPA) as a means to address threats to human health and the environment within a municipal watershed. Under the Comprehensive Environmental Response Compensation and Liability Act (CERCLA), also known as Superfund, the Helena municipal watershed has reached critical levels of heavy metals resulting from over a century of mining. The primary water source for the City of Helena, Montana is in the Upper Tenmile Creek watershed, located almost entirely in Lewis and Clark County, approximately 10 miles southwest of Helena (Figure 1). The extent of the municipal watershed is approximately 51 square miles. Approximately 150 abandoned or inactive mine sites are present in the Upper Tenmile Creek watershed within or near the historic Rimini Mining District.
 
II. Background
 
Most historic mining activity in the watershed took place within the Rimini Mining District and included hard rock mining for gold, lead, zinc, and copper. Active hard rock mining began in the 1870's and continued through the 1930's. Limited intermittent mining activities were conducted during and after World War II. The last active commercial mining in the Rimini Mining District ended in 1953. The site also includes the properties of the defunct Basin Creek Mine, an open pit gold mine located at the headwaters of Tenmile Creek, which had operated under a Montana Department of Environmental Quality (DEQ) permit until the mid-1990's. The Basin Creek Mine is being reclaimed by DEQ using money from a reclamation bond forfeited by the operator upon bankruptcy. Helena's Upper Tenmile Creek water supply system includes Chessman Reservoir and Scott Reservoir with storage capacities of 350 and 196 million gallons of water, respectively. In addition, tributaries throughout the watershed flow into Tenmile Creek where diversion structures convey water via underground pipe to the Tenmile Creek water treatment plant (Figure 2). Degradation of water quality in upper Tenmile Creek is attributed to the residual waste rock/tailings piles and discharging mine adits at the approximately 150 abandoned hard rock mine sites located in this watershed. Environmental data gathered from these mine sites indicate a release of high concentrations of arsenic, cadmium, copper, lead, zinc, and other hazardous substances that may pose unacceptable risks to human health and the environment. These substances, which are present in the watershed groundwater, surface water, stream sediments, mine waste material, and residential soils are located in close proximity to the creek and often exceed established regulatory standards for drinking water and aquatic life. Since these mine wastes may potentially contaminate drinking water supplies for residences of both the city of Helena and the community of Rimini, they pose a potential health threat. Tissue samples from Tenmile Creek fish have also shown elevated levels of arsenic.
III. A Collaborative Approach to Reclamation
 

Cleaning up Tenmile Creek began as a grass roots effort made up of local citizens concerned for the health of the community of Rimini. The group has been very active at the site and is now composed of agency representatives and private citizens. This group, The Upper Tenmile Watershed Steering Committee, has identified three goals in their effort address contaminant problems in the Upper Tenmile Creek Watershed. Their goals include to 1) discuss and offer suggestions on improving water quality in the basin; 2) educate the general pubic about environmental problems and the potential to improve water resources within the basin, especially drinking water for residents of Rimini and the city of Helena; and 3) guide development of land in the watershed in order to preserve its amenities for the enjoyment of all, now and in the future. The group is also concerned about identifying a potable water supply and ensuring fire protection for residents of Rimini, controlling noxious weeds, improving the quality of the fishery, abating road dust, and considering measures to specifically address adit discharges to Tenmile Creek to improve water quality. From discussions among committee members and various government agencies, the EPA has gained a broader perspective on the concerns of people directly affected by the legacy of historic mining and have been able to identify issues and potential cleanup approaches that may meet the common goals and objectives of many of the affected parties. Numerous individuals and local, state, and federal governmental representatives have been working for many years to address contaminant problems in the upper Tenmile Creek watershed. They include the following:

Upper Tenmile Watershed Steering Committee
Upper Tenmile Technical Subcommittee
Lewis and Clark County
City of Helena
Rimini Caucus
Montana Department of Fish, Wildlife and Parks (MDFWP)
Montana Department of Natural Resources and Conservation (DNRC)
Montana Department of Environmental Quality (MDEQ)
Montana Bureau of Mines and Geology (MBMG)
U.S. Geological Survey (USGS)
U.S. Forest Service (USFS)
U.S. Fish and Wildlife Service (USFWS)
U.S. Environmental Protection Agency (EPA)
 
The now defunct Basin Creek Mine, located at the headwaters of Tenmile Creek and directly on the continental divide, was selected as the repository site for mine waste material. Using state of the art technology, one of the historic gold mine pits, known as the Luttrell Pit, will be lined and store nearly 1.5 million cubic yards of contaminated soil. The mining operation has installed several deep wells surrounding the gold mine pits as a means of monitoring water quality.
IV. The GIS Tool
 
With input from so many sources, project purpose, scope and objectives were the focus of initial organizational meetings. The purpose of the project is to identify the major sources of mine waste contamination in the Upper Tenmile basin. Once the major mine waste contamination sites were identified, remediation plans were designed to address the sources of pollution. Focusing mainly on how this pollution causes risk to human and environmental health required a group of scientists to assess the situation from multi-disciplines, such as hydrology, soil scientology, engineering, chemistry, and geology. The issue of data retrieval, display and analysis was one that crossed over to each discipline. The result was the obvious need for a GIS. What wasn't obvious was the method to approach the myriad of needs for each discipline. Following is an outline of TtEM Inc.'s approach to the establishment of a GIS that would serve as a retrieval, storage, display and analysis system for all tabular and graphical data.
 
Step 1: Gather Existing Data Nearly every group that was involved with the project from its inception had gathered some form of data. For example, the USFS identified mine locations, collected soil and water samples, and prioritizing high-risk mine reclamation sites. The USGS collected surface and groundwater samples. The Upper Tenmile Watershed Steering Committee not only utilized local Rimini citizens to collect water samples at strategic locations along the creek, the group was actively seeking a potable water supply for residents of Rimini and looking at measures to specifically address adit discharges to Tenmile Creek to improve water quality. The USFWS had conducted fishery studies, and the Basin Creek Mine had installed deep wells at strategic locations surrounding the mine to monitor groundwater contamination. Collection of this historical data was vital to the success of the assessment. Most scientists involved with the project were eager obtain existing data for their disciplines so research and acquisition of data required a thorough yet quick result.
 
Step 2: Plan for future data needs. This step required answers to unknown questions. Each discipline was lacking the details that would steer their study. It was important to know what each group was trying to accomplish, what data exists to support this endeavor, what data will need to be collected, and what are the priorities. Some issues would have to be dealt with at a later date, such as, what depth of soil samples were required to assess soil contamination. What became obvious after reviewing the existing data was the large extent of contamination. Questions arose such as 'how many mine sites are we talking about' and 'has each stream within the watershed been affected' and even more basically, 'does every stream have a quantifiable flow?'. Some of the groups could provide large-scale, site-specific maps of their study areas but a visual depiction of the entire area required small-scale maps that dropped out details. The need for a large-scale visual depiction of the study area came to the forefront. Team scientists identified present needs as base maps, roads, streams, ownership, physical features, and mine locations.
 
Step 3: Establish data formats, Integrate existing data TtEM Inc. designed a standardized framework for the tabular and graphical data prior to any data integration. This framework gave GIS specialists guidelines for integration and consistency for all output. The data management of the Upper Tenmile Creek project consisted of a relational database management system (RDBMS) and a geographic information system (GIS). Spatial information such as roads, streams, and topographical contours, along with sampling locations of monitoring wells, surface water, groundwater, sediments and waste rock are stored in the GIS. The spatial data is then linked to the relational database, which stores attribute information associated with the spatial data. This attribute information includes any nonspatial component to a sample. When spatial data are linked to attribute data, a clearer and more complete picture of site was developed. Tetra Tech developed standards for both the database design and the GIS that would accommodate the different data sources while maintaining integrity for the life of the project.
 
Relational Database Design
The database was designed in MS Access and was linked to the GIS data through standard query language (SQL) connects. The design was an adaptation of the Montana Department of Health & Environmental Sciences Clark Fork Data System. Using this project as a framework reduced the work needed for initial database development and design phase. In its initial phase, the database included information supplied by the Mine Waste Clean-up Bureau (MWCB, a division of DEQ) and other sources as part of available historic data. The Montana State Library, Natural Resource Information System (NRIS), supplied Montana statewide data. All available analytical data from past investigations, available cultural and resource data, and other GIS data, was also entered into the database. The Access attribute database contains 10 primary tables and three supporting tables, as illustrated in (Figure 3). Each table is analogous to a data file and is a means of modeling or logically organizing data. The 10 tables presented in Figure 3 are considered primary tables because they contain the principal field observation, analytical, and survey data. The 10 primary tables are as follows: PT_KEY, WEL, STN, ST, SU, SC, SLab, RESULTS, FOBs, and ResultsCFO. The remaining three supporting tables in the attribute database promote standardization of essential data. These tables contribute standard parameter names, analytical group information, and investigation grouping information. This section explains each of the 10 primary database tables in further detail and briefly discusses the three supporting tables.
PT_KEY: Table TtEMI modified the Clark Fork Data System by adding a PT_KEY, or point key table, to the State of Montana RDBMS structure. This table links the GIS portion to the relational database. The table consists of a unique number and a point name. The GIS can link to the number rather than the point name and, in so doing, will eliminate redundant or duplicate point names in the GIS system. The PT_KEY table will be linked to the ST table.
SU Table: The SU, or survey, table will contain information from each survey event. Data will include survey identification, survey name, date, purpose, use, and import status. These fields will be the basis for the survey. Other tables will use the survey identification and survey name as references. This table is the starting point for data entry, as all information placed in the database will be associated with the survey event. Each survey will be maintained in a unique fashion so that data may be queried based on individual surveys.
ST Table: The ST, or station, table will hold information collected during each individual sampling event. Details include: survey location, station name, and location information. The spatial information contained in this table will be used by the GIS to generate maps. This table will also include spatial information, such as coordinate or "xy" method, "xyz" error, latitude, longitude, township, range, and section. Other location-specific details are included such as floodplain and surrounding geology.
WEL Table: The WEL, or well, table will be used to store information about monitoring wells. Columns are provided for recording all pertinent well?related information. The WEL table is linked under a one?to?one relationship with the ST table, as only one well can occupy a single location at a given time.
STN Table: The STN, or substation, table will hold any additional sample information for a given station. This table is used so that multiple sample types, for example, soil and water, can use the same location information. The ST table is linked to the STN table in a one?to?many fashion since it is possible that there can be many substations at each station.
FOBs Table: Field observations will be made at each station. A list of field observations is included in the FOBs, or field observations, table. Site-specific information such as time, temperature, water flow, and water elevation are stored here. This table is linked to the STN table and the ResultsCFO table.
ResultsCFO Table: The measured values for field observations will be stored in the ResultsCFO, or field observation results table. This table will resemble the results table from laboratory analysis but will hold only field observation data. It will be linked to the FOBs table.
SC Table: The SC, or sample collection, table contains data regarding the actual sampling event. Data fields include date, time, sample type, depth, and collection equipment used. This table will provide a link between the substation and the laboratory information.
Slab Table: Laboratory details are stored in the Slab, or sample laboratory, table. This table is the final link between the sample collection event and the actual analytical laboratory results. This table stores laboratory information such as basis, method, laboratory identification, batch, and sample preparation. The analytic results, however, are held in a separate table.
RESULTS Table: Analytical results are stored in the RESULTS table. This table contains sample collection identification, analyte, dilution, result, units and qualifier. The RESULTS table is linked to the SC table and then to the STN, ST, and SU tables through cascading relationships. A lookup table is also used in conjunction with other tables to provide consistent chemical analyte names and units of measurement.
ChemGroups Table: The ChemGroups, or chemical groups, table is a lookup table that holds information used throughout the database. As data types are added, they will be made available to the other tables automatically. The ChemGroups table contains the chemical analyte groups such as volatile organic compounds, semivolatile organic compounds, pesticides and metals, as well as polychlorinated biphenyls, petroleum, minerals, and sediment sample types. These added analyte types are present as lookup information for the field observations. This added flexibility creates a more robust database, with reduced possibility for data entry error.
EISPTY Table: This EISPTY table is used as a lookup table in the ST and STN tables to identify point types. The EISPTY table contains lookup parameters for various point types provided in the UTDMS. Example parameters include sphere, gear type, channel bank, basis for analysis, chemical type, sample matrix, drilling method, data qualifier, data use, data validation codes, floodplain or land, data import status, United States Geological Survey map type, perforation descriptors, quality assurance/quality control (QA/QC) level, sample type, well use, horizontal reference datum, horizontal coordinate method, coordinate position, elevation reference datum, elevation method, and coordinate zones.
EISPAR Table: The EISPAR table provides data flag information used in quality assurance procedures. When combined with the EISPTY tables, various lookups are created that are used throughout the RDBMS.
 
 
The goal of this attribute database structure is to promote data integrity and completeness. No data item, such as a sample identifier or a well name, is stored redundantly in the database. Instead, a unique numeric key is used in the PT_KEY table to prevent update anomalies. A unique numeric key is automatically created in all tables as data are entered. (An update anomaly occurs when a meaningful data item, such as a well name, is stored in more than one table but the well name is changed in only some of the tables.) All data entered into the RDBMS has an identifiable source document, and the data is stored exactly as shown on that source document. If data conversion was necessary, the converted data are stored in additional columns in the database. It was of utmost importance that direct, real?time, and end?user access to the data was developed through RDBMS and client/server solutions that mask data complexity and promote data exploration and interpretation.
 
GIS Design
 

NRIS is a clearinghouse for all GIS data created in the state. Much of the data that existed up to that point could be acquired through their map object oriented web site. Stream flow information, census information, the MWCB program files, TIGER files, small scale ownership coverages, and USGS quads ranging in scales from 1:24000 to 1:2,000,000 are just some of the data that is available to the public. Access to these types of data clearinghouses greatly reduced our data acquisition costs. Although the state is fortunate to have the NRIS clearinghouse, much of Montana is rural which results in a lot of data that has not yet been created. For example, the USGS is just starting to provide digital ortho quarter quads for portions of the state.

Due to the readily available data from the NRIS site, much of the data had already been registered to the state mandated coordinate system which is State plane NAD83 Feet. Other data came from Federal agencies, many of which have a custom coordinate system. For example, the Region VIII EPA, which would be final clearinghouse of all the data, asked that all data be delivered in the Albers projection using a custom Northing and Easting, as well as a custom point of origin. On the other hand, the USGS provides imagery such as digital raster graphics (DRGs) in Universal Transverse Mercator (UTM) NAD83 meters. The MWCB provided mine locations in both latitude/longitude (with an unknown datum) as well as Township, Range, Section, Quarter Section format. Other Superfund data in the state of Montana used the UTM NAD83 meters projection. Due to an expedited schedule, TtEMI was tasked with finding the most efficient route to convert data so that it could be easily integrated and quickly available. Reviewing the needs of the scientists showed that calculations such as length, area, distance to stream, and distance of hauling routes would be the basis of their analysis. In the selection of a projection that maintained both area and length for a site that was only approximately 51 square miles running north to south, we found that the difference in the various projections were approximately 0.1%. Since there appeared to be no large difference in calculations for this specific area, the most efficient route to integrating the data was to leave the raster data in its original projection and convert all of the vector data. As a result, the selected projection was UTM Zone 12, NAD83 meters.

 
Step 4: Gathering GIS data needs

With an established base map of the site, data gaps were more easily identified. The most glaring GIS data gap was the actual location of abandoned mines. Initially, the EPA utilized all historical records, interviews with MDEQ employees and private consultants, and historical aerial photography to identify as many mine sites as possible. In addition, a majority of the USGS and USFS mine locations were provided in latitude/longitude coordinates. Many of these coordinates were calculated manually from USGS 7.5 quad maps. Although many of the lat/long locations turned out to inaccurate, we did get the names and an original count of 60 abandoned mines on record. Because the mine site locations would be the basis of analysis and assessment of risk, it was a priority to have an accurate location of the abandoned mines and adits. Since the watershed was situated on mostly USFS land, the USFS provided a source for historical aerial photography. The intention was to save time and resources by using the aerial photography to locate additional waste rock piles at known abandoned mine locations instead of having field teams hike to each site in search of these locations. Unfortunately, the aerial photography was flown at such a high elevation it provided little ground resolution and the black and white nature of the photos left room for an unacceptable margin of error. Other issues for aerial photography included using recent imagery since several sites had recently been reclaimed. The more difficult aerial photography issue was the thick canopy of evergreen trees that shrouded the watershed. Considering the size of the base map area and the fact that accurate extents of mine locations can only be found when there isn't snow on the ground, the field team had approximately three months to locate the abandoned mines and adits. It would have been very difficult to accomplish this by simply canvassing the area on foot for these types of sites.

One solution to this challenge was to first focus on the 'priority mine sites'. Acquiring color aerial photography that was flown at a low elevation (7,000 feet) and scanned at a high resolution (2500 microns) enabled the field teams to pinpoint their canvass areas more effectively. Field teams then used a combination of the written historical information for the mines, the projected lat/long locations, and a global positioning system (GPS) unit to accurately locate mines in the area. Again, the thick canopy and steep terrain of the watershed provided challenges to collecting GPS data.

Gathering GIS data in the field also meant having accurate property ownership coverage. Prior to any fieldwork, the EPA had contacted private property owners requesting access to the abandoned mines on their properties. Homes outside of the Rimini community were built on land located in USFS managed land, or private land in Lewis and Clark or Jefferson Counties. Creation of the ownership coverage required research at the county assessors office where many properties were still maintained in metes and bounds descriptions.

 
Step 5: Providing mapping products for preliminary assessment

Establishing a base map with mine locations sped up the field effort and increased the accuracy of the map products and related assessments. Field maps were created for daily use, modified and edited, and re-created for quality control. The primary focus of the field effort was to compile a comprehensive physicochemical inventory of abandoned mines in the watershed and identify mining waste features that could have a potentially negative impact on surface water, groundwater, ecology (terrestrial and aquatic), and human health. During site reconnaissance, EPA identified a total of 150 abandoned mine sites in the watershed. The purpose of the site reconnaissance was to visit the abandoned mine sites that could be accessed, and measure and document physical characteristics of the site features. Field personnel also conducted an inspection of every tributary of upper Tenmile Creek to identify mine sites that were in direct contact with these tributaries, or that could come into direct contact during runoff events.

The following summarizes the information that was gathered during the survey of each mine site and added to the RDBMS and GIS.

· General mine site survey information - The mine site name, subarea name, the AMRB preliminary assessment number (PA#), and numerical listing by the Montana Bureau of Mines and Geology (MBMG#). The date and time the survey was conducted, a listing of survey team personnel, and weather conditions were recorded.
· Mine site location - A brief description including directions from the community of Rimini, accessibility, mine site elevation, latitude, and longitude. Information regarding hazards associated with structures and historical uses of the structures also were listed.
· Mine site layout and size - A detailed sketch of the mine site layout was created and included mining features, such as potential waste sources, adits, shafts, and staining, and environmental features, such as stressed vegetation, surface water sources, groundwater seeps, ponds, and wetlands. Locations of potential borrow materials were also noted.
· Mine site features - Information was noted regarding features identified at the site. The recorded information included:

1. Distance of the feature from creek or drainage

2. Each feature's interaction with surface water
3. Estimation of each feature's area, volume, and/or discharge
4. Feature geometry - dimensions, shape, aspect, and slope
5. Feature stability (slope conditions, erosion, etc.)
6. Rock mineral identification (visible determination of mineralogy, observation of pyrite, etc.)
7. Surface water field parameters
8. Types of samples collected and their location
9. General comments related to the feature including the presence and condition of vegetation, color, debris, and degree of staining
The collection and retrieval of this data set the groundwork for analysis by each discipline.
 
Step 6: Using the GIS as a tool
The capacity to use GIS was different for each data user group. The EPA initially used the information as a means of educating the public at public meetings. The data was used to display the National Priority List (NPL) designations, point source pollution locations, and to show the threats to human health in the form of hot linked images of raw pollution from high priority abandoned mine sites. TtEM Inc. used the GIS as a method of tracking down waste rock piles from low level, high-resolution color aerial photography. With a ground resolution of 1:600, any exposed waste rock piles larger than 50 feet in diameter could be pinpointed for field reconnaissance teams (Figure 4). The GIS team created three-dimensional, color-coded graphical displays of historic arsenic, cadmium and lead concentrations for each tributary to Tenmile Creek. Scientists could easily see that metal concentrations in surface water were decreasing in the downstream reaches (Figure 5).
 
 
 
One important product provided by a GIS system was the establishment of hydrologic units within the watershed. Using 20-foot contours and Esri's Spatial Analyst Tool, the 51 square mile Tenmile Creek watershed was broken down into 7 sub-watersheds or 'Subareas' (Figure 6). This provided the engineers and scientists the opportunity to look at each individual component of the watershed to assess its impact on the whole. In depth studies were performed at each subarea, including an assessment of groundwater, surface water, stream sediments, mine waste material, and residential soils, proximity to the creek and sample comparisons to established regulatory standards for drinking water and aquatic life.
 
 
Based on priorities outlined in the beginning of the project, TtEM Inc. developed a process of site prioritization and categorization to guide and streamline the evaluation and selection of remedial alternatives that would address the worst contaminant sources first. With the information gathered at each site, the EPA could conduct a precursory analysis of each site rather than a full investigation and alternative analysis for each individual mine site. This process was referred to as Mine Prioritization and served as criteria for ranking the risk to human health and the associated methods of reclamation. The process for establishing mine prioritization is a two-step process.
 
· First, a numeric score was assigned to each site based on impact of the solid media (waste rock and tailings) at the site on actual and potential direct contact risks, surface water, and groundwater. Impacts associated with adit discharges were addressed separately. Each mine site was ranked relative to other mine sites based on available site data and field observations.
 
· In the second step, with the help of team chemists, hydrologists and GIS personnel, the sites were grouped into five categories based on overall site score. The categories range from Category A (sites exhibiting little potential for impact on human health and the environment) to Category E (sites having the greatest potential for impact). In the detailed analysis of alternatives for the waste rock and tailings media, all potential remedial alternatives were evaluated for each category of sites. Since sites with similar features and impacts generally were scored similarly and therefore grouped together, the evaluation of alternatives was streamlined and made more efficient by evaluating those sites as a single category. Preferred remedies for each category of sites were presented in a proposed remediation plan. Remediation cost estimates were developed by subarea and site category by summing all cost elements for those mine sites in each site category for each subarea.
 
As a result, the following potential remedial alternatives for waste rock and tailings were developed and retained for detailed evaluation:
Alternative WR1: No action
Alternative WR2: Surface Controls
Alternative WR3: Containment (with an earthen cap)
Alternative WR4: Excavation and On-Site Disposal at the Luttrell Repository
 
For Categories A and B waste rock and tailings, all alternatives (including no action) provide overall protection of human health and the environment and meet all evaluation criteria to approximately the same extent. While cleanup actions at these sites would likely reduce some potential risks, the overall risks to human health and the environment at these sites are small and currently considered acceptable. The environmental disturbance required to reach these sites under Alternatives WR2,WR3, and WR4 likely offsets the benefits gained by implementing these alternatives.
 
For Category C sites, the no action alternative is not expected to provide overall protection of human health and the environment. These sites currently pose low to moderate risks to human health and the environment. While Alternatives WR2, WR3, and WR4 all satisfy the evaluation criteria, Alternatives 3 and 4 are more costly than Alternative WR2. Although Alternative WR2 provides only slight to moderate reduction in risks, the risks associated with Category C sites require only a moderate reduction in order to be at acceptable levels. The environmental disturbance required to reach these sites with the larger construction equipment associated with Alternatives WR3 and WR4 likely offsets the benefits gained by these alternatives.
 
For Category D and E sites, the no action alternative and Alternative WR2 are not expected to significantly reduce risk to human health and the environment. These sites currently pose moderate to high risks to human health and the environment. Both Alternatives WR3 and WR4 satisfy the evaluation criteria, however Alternative WR4 is the more effective than Alternatives WR3. In addition, for most of these sites Alternative WR4 is expected to be less costly than Alternative WR3.
 
V. Conclusions
 
The GIS that was designed for the Tenmile Creek watershed was created as a collaborative effort from the grassroots level to the Federal government. The purpose of the GIS was to ensure that all key contaminant release factors were considered and that the cleanup reduced as much risk as possible in the most cost effective manner. Using GIS helped the EPA to focus on a watershed approach to cleanup and to consider other needs within the watershed. For example, the state of Montana (through DEQ) is obligated to develop a total maximum daily load (TMDL) allocation plan for the upper Tenmile Creek watershed by 2003. The TMDL must address metals, habitat alteration, and turbidity in the stream reach above Rimini and metals and flow alteration in the reach from the Helena water treatment plant to Rimini. EPA's cleanup action will address metal load reduction. In addition, EPA has collected site-monitoring data that will be useful to DEQ in completing the remainder of the TMDL analysis and preparing the TMDL plan.
 
Information and data used in the GIS will be invaluable for presenting information that is understandable to engineers, scientists and the general public in documents such as the Remedial Investigations report and Feasibility Study. The Upper Tenmile Creek GIS will continue to be an essential tool to evaluate potential remedial action alternatives.