Fayetteville Shale Water Resources Decision Support System


Domestic natural gas operations face constraints to development of domestic natural gas shale resources, and  the water issues that impact access to natural gas resources as well as their production.  While our development efforts will specifically address operations in the Fayetteville Shale, we expect that the results will be applicable and exportable to a large number of oil and gas plays.

It is generally recognized that activities related to oil and natural gas exploration, drilling, completion and production affect surface and ground water resources to some degree. Water availability, due to significant use of water for the fracturing process and the consequent disposal of wastes, are a primary concern, along with potential impairment of water quality.  While natural gas operations use a relatively small amount of available water in the Fayetteville Shale play, the combined effect of public, agriculture and natural gas production could be significant.

Natural gas-shale activities are expected to increase in the near future. A substantial body of literature has provided evidence that water supply related issues have at least the potential to be a major constraining factor in the growth of shale gas development.  Water resources are affected by numerous competing factors, such as water demand for agriculture and human consumption, along with potential changes in supply levels brought about by processes such as drought and changing climate.

While specific issues are framed by the existing knowledge base, there is currently a lack of combined drilling and hydrology science-based informational and operational support systems and strategies. These structures are needed to help the gas industry and the regulatory agencies face these water resource and water management issues.   The lack of such strategies impairs the implementation of reliable regional and basin-oriented shale gas development plans to support regulatory streamlining and permitting associated with shale gas development.

Current State-of-the-Art

While there are currently no known, public resources to model the effects of water usage in natural gas shale drilling operations on water availability and water quality, there are many resources and regulatory activities in the Fayetteville Shale Play which can be leveraged to meet the objectives of this proposal.  A summary of these resources follows.

Oil and gas exploration and production is regulated by two agencies in Arkansas.  The Arkansas Department of Environmental Quality (ADEQ) has responsibility for the permitting of the reserve pits for storage of drilling fluids and for the ultimate disposal of these fluids. Part of the application package requires USGS topographic maps of the reserve pit location and the land application site for the disposal of the drilling fluids.  In addition a county map showing land use (buildings, roads, etc.), soil type, location of springs and surface water is required. 

The Arkansas Oil and Gas Commission (AOGC) has the responsibility for issuance of permits pertaining to seismic exploration, drilling, and jointly with the Arkansas Department of Environmental Quality, disposal of salt water or fluids used for enhanced recovery (which would include fracture fluids used in the Fayetteville Shale play).  With regard to seismic exploration, the AOGC requires a topographic map of the region of interest with lines showing where the explosives will be placed.  The inclusion in this project of the locations of streams, rivers, ponds as well as known features relating to underground water sources and supplies, including the location of known and previously mapped inlets and outlets of subsurface streams will lead to the ability to better protect those resources during the seismic exploration phase of lease development.

The University of Arkansas and Argonne National Laboratory, sponsored by the US Department of Energy through the Low Impact Natural Gas and Oil (LINGO) Program (Probabilistic Risk Based Decision Support for Oil and Gas Exploration and Production Facilities in Sensitive Ecosystems), developed the Fayetteville Shale Information Website and the Fayetteville Shale Infrastructure Placement Decision Support System. 

The Fayetteville Shale Information Website is a public information site jointly developed by Argonne National Laboratory and the University of Arkansas and hosted at CAST. It enables readers to learn about the natural gas resources available in the Fayetteville Shale formation in Arkansas and explains the steps followed by natural gas development companies, from gaining access to the land through sending the gas to the marketplace. For each step in the process, the site provides information about the state and federal regulatory requirements that developers must follow. The site also describes some of the technologies that can be used to minimize the environmental impacts of natural gas development and provides current interactive maps showing the locations of active drill sites and permitted sites. 

Because of the large volume of "frac water" required at drill sites, water availability and the state and federal regulations regarding water use are major considerations in drill site selection. However, while the website does address several water issues related to the development, a thorough treatment of the availability of water resources and the regulations associated with its use was beyond the scope of the LINGO project.

The Fayetteville Shale Infrastructure Placement Analysis System was designed with input from Chesapeake Energy, Southwestern Energy Company, Arkansas Oil and Gas Commission, Arkansas Department of Environmental Quality, US Fish and Wildlife Service, and many other agencies and producers. A key part of the LINGO project was a series of public meetings and private meetings with the producers and agencies that were held to determine what elements of the process could be improved through decision support tools.  At these meetings, discussions were positive and indicated a strong willingness of the industrial and regulatory parties to collaborate with each other and the project team to help create a tool that will be beneficial in the development of the Fayetteville Shale Play.  The major themes that emerged as areas where the greatest benefit to the stakeholders would be felt were education and integration. There was also general agreement that more efficient communication between the regulators and industry would be a significant benefit to all.

The result was a limited-access online map-based siting decision support system, targeted at operators, regulators, and other primary stakeholders.  The decision support system (essentially an online Geographic Information System based on ESRI's ArcGIS Server 9.3) is hosted at CAST and leverages public data available through federal and state agencies (Table 1).   In addition to aggregating these public data sources, the system hosts several datasets developed specifically for the LINGO project.  One prominent dataset is a modern terrestrial habitat model developed by CAST and the Arkansas Natural Heritage Commission which can be used to predict possible conflicts with threatened and endangered (T&E) species - without revealing sensitive information about known T&E spec ies sitings.  An operator seeking a drilling permit can securely log into IPAS, use graphic tools to design and place infrastructure (drill pads, gathering lines, reserve pits and access roads) then run a series of models against the included data layers to test for potential impacts (e.g. upslope from sensitive waterways, proximity to possible T&E species habitats, proximity to highly erodible soils, USACE 404 water crossings and more). If potential negative impacts are returned the operator may move the placements to reduce the potential of negative impact.  After a suitable placement plan or plans are sited, the operator may save the results to the secure database and electronically notify the appropriate regulatory agency of the scenario.  The regulator may, then in turn, log into the system, load the operator's proposed placement and look at the model results.  IPAS is intended as a planning tool only and explicitly takes into account the spatial uncertainly of both the infrastructure placement and the underlying data layers.

Table 1.  Partial List of Public Data Layers available in FSDSS.



US Geological Survey 

National Elevation Dataset

US Geological Survey 

Digital Raster Graphics (digitized 1:24K topographic maps)

US Geological Survey and US Environmental Protection Agency 

National Hydrology Dataset (1:24,000 scale)

Arkansas Geographic Information Office 

Arkansas Road Centerlines

Arkansas Geographic Information Office 

Public Land Survey System (Township, Range and Section corners)

Arkansas Geographic Information Office 

2006 Orthophoto Image Base (0.33 - 1.0 meter GSD)

Arkansas Geographic Information Office 

Arkansas political boundaries (county, city)

Arkansas Natural Resources Commission 

Watershed boundaries

US Forest Service 

Public forest boundaries

Bureau of Land Management 

Public land boundaries

Arkansas Oil and Natural Gas Commission 

Existing drill pad and well locations (permit status and production history) and production history

Arkansas Oil and Natural Gas Commission 

Locations of major gas transmission lines

Arkansas Department of Environmental Quality

Locations of reserve pit locations and permit status

Arkansas Natural Resources Commission

1999, 2004 and 2006 Land Cover and Land Use

Natural Resources Conservation Service (US Department of Agriculture)

Soil Survey Geographic Data (SSURGO)

Another resource relevant to water availability and disposal in the Fayetteville Shale is the Arkansas Watershed Information System, a comprehensive statewide electronic watershed atlas consisting of a series of practical maps and reports for the 308 watershed units across the state. The system was funded by the Arkansas 85th General Assembly through the Arkansas Natural Resources Commission and helps to provide citizens and decision makers with easily accessible information via the internet.

The Arkansas Watershed Information System provides timely and accurate watershed information to federal, state and local agencies (including natural gas shale regulators) in the form of online map services, high quality maps and tabulated reports. This information helps local communities to more effectively respond to federal requirements and improve their decision making.  The 308 ten-digit watershed units and the 1,556 twelve-digit sub-watersheds have recently been delineated for the entire state by the Arkansas Office of the Natural Resources Conservation Service. The Arkansas Watershed Information System provides more than 8,500 thematic maps and 22,300 summary reports in an accessible and easy to use format. Available maps and reports include:  population density, population change from 1990 to 2000, land use in 2004 and 2006, land use change from 1999 to 2004, septic suitability, soil productivity, hydric soils, elevation, slope, business inventory, sub-watersheds, aerial imagery from 2006 and roads by federal classification - this kind of information is important regarding permitting for land disposal of water based drilling fluids, as currently allowed by the ADEQ.

Finally, a variety of simulation models which compute water flows, levels, quality, and other parameters based upon physical laws and conceptual relationships are available. The models are calibrated by comparing computed results to observed values with subsequent adjustments to mathematical coefficients to attain closer agreement. The models are verified by comparing historical events (which are not used in the calibration) to computed results. SWAT (Soil and Water Assessment Tool) (Arnold et al., 1998) is one such physical based model developed to simulate at a river basin or watershed scale.  SWAT was developed to predict the impacts of land management practices on water, sediment and agriculture chemical yields.  To meet the objectives of the tool, the model requires specific information regarding weather, soils, topography, vegetation and land management practices that occur in the watershed.  Utilizing these sources of information and the SWAT approach to simulation allows the user to model watersheds where empirical data availability is minimal or lacking; and also allows the user to manipulate variables (land use, climate, vegetation, etc.) and assess the changes impacts on water quality or other variables.   The SWAT model is a continuous time-step model allowing the user to simulate long-term impacts of management decisions. 

The SWAT model allows for several different physical processes to be simulated within a watershed.  Within each watershed, the tool allows the user to segment multiple sub-watersheds or sub-basins based upon a multitude of factors that can include, but are not limited to: 1) multiple tributaries within the large watershed, 2) different areas of activity, and 3) other hydrologic parameters of interest.  All activities within the watershed are organized into categories: 1) climate, 2) hydrologic response units (HRU's), 3) ponds/wetlands, 4) groundwater, and the 5) main channel that drains the sub-basin (Neitsch, et al.,   2005).  These categories can then be lumped based upon land cover, soil and management activities.

The water-balance approach of the model allows users to accurately predict movement of chemicals, sediments or other variables within the landscape.  The model functions based upon basic hydrologic principles including interactions between precipitation, evapotranspiration and terrestrial influences (soils, land use, etc.).  Parameters include climate, hydrology, land cover and plant growth, erosion, nutrients, pesticides and management.  These parameters can be adapted based upon the industries specifications and the landscapes being simulated.

This model supports a number of federal regulatory and assessment objectives while setting the standard application methodology to be adopted across federal, state, and local agencies. For example, the US EPA has adopted SWAT as one of the main supporting tools to TMDL projects. TMDLs are "budgets" setting the amount of pollutants that waters can receive and be considered by EPA as safe for various designated uses such as fishing, swimming, drinking, etc. Our team developed the BASINS software framework for EPA (Di Luzio, et al.,   2002), which includes SWAT model and a variety of tools to help developing TMDL projects.  SWAT is also at the core methodology for the USDA NRCS CEAP (Conservation Assessment Project) (Mausbach and Dedrick, 2004), designed to estimate the environmental benefits for conservation practices applied to cropland.  An extended number of studies and applications of these models across the world are documented (BREC and USDA ARS, 2009; Gassman, et al.,   2007).

Any user of SWAT knows that it is a complex tool with literally thousands of user defined parameters.  There are dozens of extensions which add to the tool's utility but also to its complexity (Gassman, et. al., 2007).  The team we have assembled for this project is perfectly positioned, with expertise in both SWAT and web-service GIS, to reduce this complexity and make this simulation tool available to stakeholders in the Fayetteville Shale Play.

Current R&D and Proposed Solution

In the past 10 years, access to geospatial data - which is critical to virtually all logistic intensive activities such as natural gas production - has increased dramatically.  Enterprise GIS systems based on geospatial databases such as Oracle Spatial and ESRI ArcSDE are common.  Companies like Google, Microsoft and Navtec provide - often free - access to terabytes of useful geospatial data and have even provided application specific tools (routing and location based service listings are among the most common).  However, most sophisticated GIS analysis, such as that undertaken by oil and gas companies during the exploration and development of a play, is still relegated to the desktop workstation with data that is downloaded and stored locally.  While there are good reasons for this desktop-oriented analysis, it does not promote data sharing and collaboration.   For example, a producer might perform a very sophisticated GIS analysis using land cover data, soils data, parcel maps and high resolution topography (none of which, incidentally, is available in consumer mapping applications) to support a proposed siting.  Along with the many data layers, the analysis likely involves a variety of complex geoprocessing tasks.  When complete, the output of the GIS is captured in a static map, tabulated for a report and sent to the regulator for review.  If the regulator has questions concerning the output of the model, she is unable to immediately see the effects of changed assumptions or placement and indeed is unable to even to reproduce the results.  The iterations that follow slow the permitting process and obscure potentially important information - especially when more than one regulatory agency is involved. The Fayetteville Shale Decision Support System (FSDSS) is an example of an attempt to avoid this wasteful and unnecessary iteration by moving standard GIS analysis (not just data sharing) to a web service. In this configuration, both producers and regulators are able to perform a complex GIS analysis through a light-weight but fully functional browser interface.  They access the same, presumably authoritative, geospatial data and execute the same sophisticated geoprocessing models on the robust server.  This promotes trust in the results and requires less design iterations. However, even in modern enterprise and web-based systems like FSDSS, complex models like SWAT remain inaccessible due to the number and complexity of required inputs and the difficulty of effectively displaying the results.

Our solution will be a large step towards bridging that gap for gas producers and regulators working in the Fayetteville Shale.  Details on how we plan to accomplish this integration are provided in Section 3.  In general, however, the idea is to develop the software necessary to let the FSDSS provide most of the complex inputs to a SWAT simulation.  For example, HRU determination and characterization is difficult because it relies on several data layers with many, complex attributes.  FSDSS understands the data layers, and with the proper coding, could provide them aggregated to the sub-watershed or sub-basin of interest.  Likewise, daily temperature and estimated precipitation, which are difficult to collect and package into a format needed for the simulation, could be automatically provided in the right format by FSDSS.  Again, we emphasize the value of the enterprise web-service architecture: once the necessary integration code is developed it can be distributed to all producers and regulators in a single, server installation.  Not all inputs can be provided by the FSDSS, of course, so a careful and scientific investigation to determine, watershed by watershed in the Fayetteville Shale Play, the combination of parameter defaults and input ranges that provide meaningful and robust (i.e. insensitive to parameter changes) results.  


This project will build on two years of active development of the Fayetteville Shale software tools at the University of Arkansas; two years of close collaboration between the Argonne National Laboratory, the University of Arkansas, and state agency, federal agency and private stake-holders in the Fayetteville Shale; and many years of hydrology modeling by the Texas AgriLife Blackland Research and Extension Center (BREC).  The current decision support tools already include many of the data layers required to support the proposed enhancements.  The decision support aspects of the proposed project deliverables will leverage significant software development which already provides a stable and extendable, interactive web-based GIS and allows secure transactions within the system (allowing, for example, private operations to maintain proprietary information).   The effort needed to enhance the LINGO project deliverables to 1) include more in-depth water-related technical and regulatory information and 2) to integrate a proven water modeling tool (SWAT) - though technically and scientifically challenging - is well understood by the team members.  Because SWAT has been accepted by EPA, USDA/NRCS and other federal agencies as a valid simulation tool, its inclusion in this DOE project will increase the credibility of the predictions and management alternatives. 

Small Water Body Extraction

While the Arkansas Watershed Information System currently incorporates the USGS National Hydrology Dataset, potentially large volumes of rainwater is captured and stored in small retention ponds (less than 1200m2).  One task involved in this project will use object based remote sensing methods developed at the Center for Advanced Spatial Technology to identify small retention ponds (as small as 300 m2)and their drainage areas using current, high-resolution, color-infrared aerial imagery (from the Arkansas Digital Orthophotography Program from 2006 and the USDA National Agriculture Imagery Program from 2006 and 2009) of the entire Fayetteville Shale area.  These small ponds are designed to retain water for agriculture uses and, based on studies at the University of Arkansas, can cover as much as 2% a watershed and drain up to 8% of the area.  This significant amount of stored water is not represented in national datasets such as the National Hydrography Dataset (NHD) but could have significant impact on SWAT model outcomes.  Drainage areas for these retention ponds will computed using the 5m resolution DEM developed as part of the LINGO project.  This information will be used as input to the new SWAT models to better estimate and predict water availability at the basin level.  In this subtask we also propose to measure the location and size of retention ponds at this larger scale using recently developed techniques which utilize image segmentation coupled with object-based classification.  These methods make it possible to extract land-cover information from these low spectral-resolution but high spatial-resolution aerial photographs. These techniques have been successfully employed in high resolution imagery to extract impervious surfaces (Cothren and Gorham, 2005). The methodology proposed here for the extraction of retention ponds and otherall water bodies will employ these recently developed techniques (Figure 4).


(left) Retention ponds in this North Central area of Arkansas in the White River Basin as seen in 1-meter resolution color infrared imagery.  None of these water bodies appear in the NHD. (right) Object oriented classification of the color-infrared imagery.  Surface water objects appear in shades of blue and range in size from 250m2 to 12,000m2.



Arnold, J.G., Srinivasan, R., Muttiah R.S., Williams, J.R., 1998. Large area hydrologic modeling and assessment part I: model development. J. American Water Resources Association 34(1), 73-89.


Cothren J. and Gorham, B., 2005. Automated Feature-Extraction: Software Advances Extract Impervious Surfaces from Satellite Imagery.  Earth Imaging Journal, 2, 32-34.

Di Luzio M, R. Srinivasan, and J. G. Arnold, 2002. Integration of Watershed Tools and SWAT Model into BASINS. Journal of American Water Resource Association. (38) 4: 1127-1141.

Gassman, P.W., M.R. Reyes, C.H. Green & J.G. Arnold.  2007.  The Soil and Water Assessment Tool: Historical development, applications, and future research directions.  Trans. ASABE.  50(4):1211-1250.

Mausbach, M. J. and A. R. Dedrick, 2004. "The Length We Go-Measuring Environmental Benefits of Conservation Practices". J. of Soil and Water Conservation, Volume 59 (5):96-103

Tarboton, D. G., 1997. "A New Method for the Determination of Flow Directions and Contributing Areas in Grid Digital Elevation Models," Water Resources Research, 33(2): 309-319.  





Greg Thoma
Deparemtn of Chemical Engineering
479.575.7374, Turn on JavaScript!

Jackson Cothren
Department of Geosciences
Center for Advanced Spatial Technologies
479.263.3911, Turn on JavaScript!


 US Department of Energy