Each facility submitted either a paper or digital facility map with well locations. The maps were field-checked to determine suitable locations for GPS control points. They were then rectified using the surveyed control points and Esri's ArcInfo software by transforming them into UTM coordinates. This project took three years to complete and included 58 regulated facilities, 1600 wells, and 162,000 water quality records. Transforming the facility maps into a real world coordinate system allowed staff to integrate these coverages with those developed from other state and federal agencies in assessing the impact of these facilities on the environment of Northwest Indiana.
After receiving funding for this project, work began in earnest in 1993 to develop the GIS. A major effort began in the Hazardous Waste Geology Section of the Office of Solid and Hazardous Waste Management (OSHWM). Staff began creating a series of base map coverages for regional groundwater analysis in Northwest Indiana. These base maps included data for both possible contaminant sources and receptors. A large amount of groundwater data existed for this region in the EPA Region V groundwater monitoring data set, known as GRITS (Groundwater Information Tracking System). This data represented monitoring over a two year period for 1500 wells located on 58 facility sites, and represented a 162,000 record table. However, the monitoring wells did not have any locational attributes assigned to them. Secondly, the facility maps from which the digitized locational data could be obtained were all in unique coordinate systems (plane coordinates), created by the staff that surveyed the facility. Therefore, the top priority for staff in the Geology Section was to define accurate locations for all of the monitoring wells in the GRITS data set within a common coordinate system.
Although the well coordinates would be derived from pre-existing information, if this could be accomplished with a permissible level of error, then this would be acceptable. It was decided to use this method and define at least two control points at each facility with GPS techniques. Usually, at least three sets of coordinates are used as control to transform from one coordinate system to another. Two sets of coordinates will define the plane, but a third set of coordinates statistically defines the distortion and displacement, known as the Root Mean Square error (RMS error), created by the process of transformation. This defines the accuracy of the control coordinates in relation to the source map.
Unfortunately, in this project, a number of the facility maps to be used were of poor quality and impeded the use of RMS error. Two of the largest facilities' drawings, submitted digitally, were too large and detailed to create as one file. Therefore these facilities digitized and created at least two drawing files per facility map. When these files were appended, significant distortion and displacement occurred. Many facilities submitted photocopies of the original hardcopy maps, which distorted scale and dimension. A few of the facility maps submitted were not maps but drawings, with no scale or spatial relevance. And a final set were outdated maps, having been created by the last facility to own the property, since abandoned and under the jurisdiction of the U.S. EPA as a Superfund Site, and managed by IDEM. With these considerable source errors inherent in the facility maps, RMS error proved meaningless in many instances. Therefore, it was decided to only collect two points per facility. In instances where facilities were large and digital data was available, at least three control points were used.
Considering the error already inherent in the transformation process through digitization, scale, and possible source map error, it was decided to purchase survey-grade GPS equipment to minimize the amount of error introduced through data collection. Requisitions for GPS equipment were initiated in 1993. Requests were made for 3 Trimble 4000SE Land Surveyors and peripherals. The 4000SE is single-frequency survey-grade receiver with technical specifications of (1cm + 2ppm x baseline length) horizontal accuracies and (2cm + 2ppm x baseline length) vertical accuracies. Equipment was purchased in 1994. From April of 1994 through August of 1996, eight surveys were conducted at 58 facilities in the region of Lake, Porter, and LaPorte counties.
GPS is a relatively new technology, having been developed by the Department of Defense in the early 1970's. In the early 1980's, it began to have practical civilian applications and equipment was developed and sold commercially for civilian use. However, it is not a failsafe technology. It does have its limitations and drawbacks. First and foremost is where you can use GPS. A GPS receiver acquires radio signals transmitted from satellites, which it uses to determine distances to those satellites, and from that information its location. However, these signals can be blocked by obstructions such as buildings or trees, and they can be corrupted by reflective objects located near the GPS units. Secondly, it is only as accurate as the control you are utilizing. One may be able to conduct an inherently precise survey. If this survey is adjusted to poor control, however, the precision is meaningless. Therefore, the most important aspect of GPS surveying is the initial pre-planning. Staff would visit the area approximately one week prior to an event to verify existing control and to visit the facilities to be surveyed to ground truth the facility map and establish survey locations. All control points used in this project were National Geodetic Survey monuments, 1st and 2nd order, constrained to the NAD83 datum. Facility survey locations were generally established at monitoring wells located on site. However, if this was not possible due to possible satellite signal interference, a property corner or other easily rectifiable feature was used for the survey location.
The surveys were usually 2 to 3 day events involving approximately 10 to 12 people. The project coordinators were able to utilize six receivers in the field by borrowing 3 Trimble 4000SE Land Surveyors from the Office of Environmental Response. Once the survey was completed the data was processed back in Indianapolis, along with running it through a least squares adjustment. However, due to the fact that the three counties surveyed are approximately 160 miles from Indianapolis, a laptop computer was eventually taken into the field to process the data after each field day to determine if any points needed to be reoccupied.
Digital map data, usually submitted as either an AUTOCAD .DWG file or a DXF file, was converted into coverages using the same process described previously. The coverages created were rectified to the UTM coordinate system by a transformation. This is performed by using the TRANSFORM command in ArcInfo. A transformation rotates, scales and translates each feature from one coordinate system into another. The original coverage coordinates are moved by using a second coverage, created with a set of control points, as a template. When staff first began conversion of digital data into UTM coordinates, they were using CAD software to scale, move, and rotate each drawing. This was time-consuming and inefficient. Scaling calculations had to be performed by hand, and each command had to be invoked separately. Additionally, a compounded amount of displacement error was created each time one of these commands was used. The TRANSFORM command in ArcInfo invoked these processes intrinsically, and if three or more control points were used, an RMS error was generated to alert staff to the amount of distortion which occurred during this process (Esri, 1992).
The final procedure in this process was to define the projection of the coverages. All coverages were rectified to the Universal Transverse Mercator coordinate system, zone 16. The datum used was NAD83. The data was then linked to the GRITS data set by creating key fields in the point-attribute-tables of the well coverages. These key fields included a facility ID number, and the monitoring well ID.
A baseline confirmation survey was conducted with survey-grade GPS receivers in the spring of 1997 to determine the accuracy of the derived locational data method used by IDEM (Koelpin & Others, 1997). Part of this study was to compare the derived locational data method used by IDEM to direct GPS mapping-grade methods used by the EPA. A number of the monitoring wells with derived locations in the IDEM data set had been surveyed with mapping-grade GPS units by the U.S. EPA in another project in the area. Due to time and budget constraints, a subset of 10% of these were selected for the baseline survey. IDEM averaged a difference of 17.8 meters from the confirmed locations. EPA averaged a difference of 9.17 meters from the confirmed locations.
While IDEM's derived locational method was not as accurate as EPA's direct mapping, the number of wells assigned locations was almost five-fold. IDEM took approximately three years to complete their project. EPA conducted their survey over a two year period. Secondly, the 17.80 meters of difference in the IDEM project was well within EPA's stated goal of 25 meters or less for locational data accuracy standards, which was the benchmark used in this project (US EPA, 1992).
Survey-grade GPS equipment may be unnecessary for locational data control in this type of project. Sub-meter mapping grade GPS equipment would have met the needs of the staff and would have provided acceptable data cheaper and in a shorter amount of time. Hazardous Waste Geology is presently purchasing new sub-meter mapping-grade equipment. The survey-grade GPS equipment is now being utilized for engineering and compliance efforts in topographic surveying. So, it is now serving a multi-purpose role.
Even with the cost of survey-grade equipment, the cost of completing this project was one-quarter to three-fourths that of similar work being contracted in another region of the state, depending upon how staff time and equipment costs are handled.
A second step IDEM is taking to improve locational data development is in the creation of agency-wide GIS and GPS data standards. As data development continues in the agency and more areas become involved in data-sharing, it is imperative that there is consistency in the work conducted within the agency. Therefore, a committee has been created to develop a set of standards for documentation purposes that will be included with every source GIS coverage created. These will then be archived as source data in a central repository so that as personnel pull this information into their various projects, there will be reliable information on source and scale of this data.
A third step IDEM is taking to improve locational data development is in the standardization requirements for data submitted by the regulated community. Although not mandated by regulation, it is hoped that these standards will be received in the spirit with which they were created. As GIS development moves across the state for IDEM, the agency will need to rely on data submitted by our customers, as the agency does not have the resources to capture all necessary data . With a consistency in locational data submitted to our agency, there will be an improvement in the efficiency of our analyses and reviews. These standards will include all data submitted in a digital format, consistency in scale of facility maps, and standards and specifications for all locational data collected by outside contractors.
A final step which will improve locational data is the development of a HARN (High Accuracy Reference Network) in Indiana. As of now, the best control being used for most GPS work is at a precision of 1:100000, whereas the HARN points will have a precision of approximately 1:1000000. By creating a network of high precision control points, and adjusting the existing NAD83 datum to this network, both precision GPS work and traditional survey methods will be improved across the state. This survey is scheduled for the summer of 1997.
Koelpin, R.U., Goldblatt, I.A., 1996, GIS Development in the Hazardous Waste Geology Section, Office of Solid and Hazardous Waste, Indiana Department of Environmental Management, in Proceedings 1996 Annual Users Conference, May 1996, Environmental Systems Research Institute, Redlands, California
Koelpin, R.U., Murray, B.R., Miller, K.H., 1997, A comparison of Three Vintages of Groundwater Monitoring Well Locations in the Indiana Department of Environmental Management, Hazardous Waste Geology Section's GIS, in Press, 1997 Annual Users Conference, July 1997, Environmental Systems Research Institute, Redlands, California
United States Environmental Protection Agency , Administration and Resources Management, 1992, Locational Data Policy Implementation Guidance, Guide to the Policy, March 1992, 220 B-92-008, Washington, D.C., United States Environmental Protection Agency, Administration and Resources Management
Roger Koelpin, Geologist
Hazardous Waste Geology Section, Office of Solid and Hazardous Waste
Indiana Department of Environmental Management
100 North Senate Avenue, P.O. Box 6015
Indianapolis, Indiana 46206-6015
Telephone:(317)-232-8726
Fax: (317)-232-3403
Kevin Miller, Geologist
Hazardous Waste Geology Section, Office of Solid and Hazardous Waste
Indiana Department of Environmental Management
100 North Senate Avenue, P.O. Box 6015
Indianapolis, Indiana 46206-6015
Telephone:(317)-233-5298
Fax: (317)-232-3403