Ken Spitze
Win Chromec, PhD
Greg Wetherbee
Jeff Pietsch

Using GIS in Soil Erosion Modeling at Rocky Flats

ABSTRACT

To accomplish closure of the U.S. Department of Energy, Rocky Flats Environmental Technology Site (RFETS) by the year 2006, it is necessary to demonstrate that there will be no long term environmental or health effects to surrounding communities from residual contamination. The Actinide Migration Evaluation Project has determined that erosion of actinide contaminated soil and transport of sediment offsite over an extended period of time is a potential concern. This paper will show how GIS was used as an effective tool for determining key parameters for input into the erosion modeling software and to effectively display the final results produced by that software. GIS was used to simplify the complexities of the real world environment to satisfy the simplifying assumptions and requirements of the modeling software, while still retaining the ability to display the rich detail of the results in a meaningful way within the geographic setting.

INTRODUCTION

RFETS is a U.S. Department of Energy, contractor run site, that was part of the nationwide complex for nuclear-weapons production. It consists of 6,550 acres of federally-owned land, and is located 16 miles northwest of Denver, Colorado at the base of the foothills of the Rocky Mountains.

The mission of the plant was to fabricate components from plutonium (Pu), uranium (U), beryllium (Be), and stainless steel. As a result of those activities, radioactive and nonradioactive wastes were generated at RFETS. Historically, those wastes were either disposed of on-Site, stored in containers on-Site, or transported off-Site for disposal. Early on-Site waste storage practices resulted in the release of both radioactive and nonradioactive contaminants to the environment.

In October 1989, the facility was added to the Superfund National Priorities List. Since 1992 the mission of RFETS has been environmental restoration, waste management, and decontamination and decommissioning of buildings. Current schedules call for closure of RFETS by December 2006.

THE PROBLEM

Potential contamination of surface water is a significant issue in the environmental restoration of RFETS. The great majority of surface soil contamination at RFETS is due to three incidences. Building fires in 1957 and 1969 released unknown quantities of Pu to the environment that was spread to the east and southeast by prevailing winds. The most significant release resulted from leakage of Pu- and U-contaminated industrial oils and volatile organic solvents stored in over 5,000 steel drums placed directly on the soil surface at the eastern edge of the Industrial Area from 1958 to 1967. Approximately 420 of the drums leaked, releasing an estimated 5,000 gallons of liquid containing between 86 and 126 grams of Pu. High winds and heavy rains redistributed a significant amount of the contaminated soil during the subsequent removal of the drums, attempted remediation of surface soil, and site preparation for an asphalt pad.

Dose-based surface soil action levels (AL) for remediation of radionuclides have been developed, however, the AL calculations do not consider erosion of the contaminated soil or transport of the sediments in surface water. The Rocky Flats Cleanup Agreement (RFCA) contains ALs and standards for both surface soil and surface water that must be met at closure. Therefore, a study was undertaken to determine the long-term effects of diffuse radionuclide contamination in surface soils on surface water quality and the migration of actinides to down stream locations.

THE MODEL

The Water Erosion Prediction Project (WEPP) Model was developed by the United States Department of Agriculture (USDA), the United States Forest Service (USFS), the United States Department of the Interior (USDI), and other cooperators. WEPP model was chosen for its accurate representation of the complex climate, topographic, plant, and soil relationships that determine erosion rates, and its ability to identify areas of erosion and deposition on the topography. It is a new generation of process-oriented, computer implemented, erosion prediction technology, developed from the principles of modern hydrologic and erosion science. The WEPP model is a distributed parameter, continuous simulation computer program which predicts:

1. Soil loss and sediment deposition from sheet and rill flow on hillslopes;

2. Sediment loss and deposition from concentrated flow in small channels;

3. Sediment deposition in impoundments; and

4. Sediment leaving the watershed by channel flow.

 The model includes a stochastic weather generator that uses meteorological data from weather stations located in all 50 states and internationally to produce climate input files that can be up to 999 years in length. WEPP estimates infiltration and runoff using detailed, site-specific soil and vegetation growth, cover, and decomposition information. The model computes both spatial and temporal distributions of soil loss and deposition on a hillslope or in a watershed, and estimates flow and sediment transport for ephemeral and intermittent flow channels with watershed areas up to 60 km².

The site-specific WEPP model is built from a series of watershed sub-basins and hillslopes, which become the defining boundaries for the modeling effort. Hillslopes can be subdivided into elements, called Overland Flow Elements (OFE). OFEs are regions of homogeneous soils, cropping, vegetation or management on a hillslope. Each hillslope in the model may have as many as ten OFEs. Hillslopes and OFEs are modeled as rectangular representations, having uniform width and length. The model has not yet been adapted to a grid structure. This presents problems in reducing real world data to meet the model data conventions while retaining a measure of the variability of the natural setting, and in transforming the output of the model back to the geographic setting as a useful and meaningful representation. Arc/Info was found to be an effect mechanism in resolving both issues.

GIS DATA INPUT

In order for WEPP to work effectively, it is important that the real-world observations and data used as input to WEPP be as representative as possible. This was challenging, especially with the degree of observed geographical variability of the landscape at RFETS, and the model requirement that the parameters for each hillslope be constant laterally (perpendicular to the fall line). Model input data requirements, availability, and sources are summarized in Table 1.

Watershed sub-basins and hillslopes boundaries were developed using a combination of hydrologic modeling, performed by ArcView Spatial Analyst, and ground-truthing. Sub-basins were calculated from a Digital Elevation Model (DEM). Hillslope boundaries were determined from these sub-basins. The hillslope boundaries were entered into Arc/Info as polygons, and lengths and hillslope areas were calculated for each hillslope.

The second step was to determine the number and size of OFEs within each hillslope. The OFEs were derived from intersecting soil and vegetation coverages already existing over RFETS, and then working with the team experts to simplify the resulting coverage. These coverage boundaries were extracted from the intersected coverages, areas calculated and OFE lengths computed based on the width of the hillslopes. The resulting OFE polygons were then overlain on the hillslope coverage.

Next, input data had to be calculated and data files generated. Input data files necessary to run the WEPP Model are divided into four categories (Table1): 1) The climate file that provides meteorological data used by the plant growth and the erosion modules; 2) The slope and soil input files required for running the plant, hydrologic and erosion modules; and 3) The channel files required to run the watershed simulation component. Extensive site-specific data was used for input parameters and for calibration.

Developing representative slope value parameters were particularly challenging for input to the model. It was decided to derive the values from the RFETS DEMs, in which slopes are constantly varying in order to produce slope values that were as realistic as possible. It was determined from a review of hillslope and OFE boundaries that the best representation of the slope values could be made by placing multiple traverse lines perpendicular to the elevation isopleths on each OFE, and then sampling each traverse line at regular intervals for the instantaneous slope. These values were then averaged laterally across the slope, and slope files automatically generated in a format that could be input directly to WEPP.

OUTPUT

The results of the WEPP model include estimates of net erosion or deposition for each percent interval of the OFE length. Reports are also generated estimating total runoff and sediment delivery by each hillslope to drainages. The reports include total volume of runoff, mass of eroded soil leaving the hillslope, the particle size distribution, organic matter content, and surface area enrichment, and the volume of channel flow and sediment movement out of the watershed. Like the input, the output results varied only vertically down each hillslope. The challenge remained how to effectively show these results in a meaningful way geographically.

GIS VISUALIZATION

Even though the tabular reports were very useful in analysis of total sediment erosion and deposition, it was recognized that a visual representation of the output in a geographical context was necessary to easily identify areas with the greatest susceptibility to erosion. WEPP does not have this capability. It was determined that GIS would be a useful tool for displaying the results of the model. The output of the model was read into an erosion/sediment database and linked to locations in each OFE by creating a series of evenly spaced transects duplicated across each OFE. Arc/Info Network was then used to generate 100 evenly spaced points per transect.

Duplication of the transects and points allowed the data results to be repeated laterally across each OFE. TIN was then used to create an erosion surface from this data which followed the geographic constrains, yet was laterally homogeneous and vertically varying according to the model output. The TIN data was then converted into a grid and the resulting values analyzed. A continuously varying shadeset was created ranging from dark blue to white to dark red to best depict the TIN data. Areas of greatest erosion are shown as dark red, areas of no erosion as white, and areas of greatest deposition as dark blue.

A Pu mobility map was also created to show areas of greatest potential contaminant movement. The map was generated by multiplying the erosion grid by the surface grid of Pu activity, which was developed by Kriging observational data and mapping the resulting surface. The Pu mobility map is used to depict areas of potential concern, where the combination of predicted erosion and Pu levels indicate a high probability of surface water degradation.

CONCLUSION

GIS was found to be a significant tool in reducing complex amounts of information to fit within the constraints of the erosion modeling software. It was also found to be a valuable asset in demonstrating the results of the modeling in ways that are more meaningful to scientists, managers, and stakeholders. These results will be used to aid in remediation decisions, including the selection of surface soil cleanup levels, areas of surface soils to be remediated, and institutional and engineering controls on the potential movement of soils and radioactive contaminants on- and off-Site.

Ken Spitze
President
Destiny Resources, Inc.
1658 Cole Blvd. Suite 205
Golden, Colorado 80211
Ph: (303) 232-6515
Fax: (303) 232-6618
Email: kspitze@destinyresources.com
Web: http://destinyresources.com

Win Chromec, PhD
Rocky Mountain Remediation Services

Greg Wetherbee
Wright Water Engineers

Jeff Pietsch
Destiny Resources, Inc.