USGS - science for a changing world

Kentucky Water Science Center

Ground-Water Resources Program Karst Hydrology Initiative

Karst is a complex geological environment where surface water and ground water are highly interconnected and distinctive physiographic and hydrologic features develop as a result of dissolution of soluble bedrock such as limestone. Karst topography occurs over nearly 15 percent—or about 1.2 million km2—of the continental United States, and in about 40 percent of the land area located east of the Mississippi River. Typical karst physiographic features include sinkholes, sinking or losing streams, caves, underground streams, and springs. Related karst hydrologic features include (1) disrupted surface drainage patterns or the absence of surface streams, (2) internal drainage of storm-water runoff by sinkholes and sinking, (3) aquifers characterized by rapid and turbulent subsurface flow through pipe-like or channel-like solutional openings called conduits, and (4) resurgence of subsurface waters at large karst springs.

Big Spring Huntsville, Alabama, which served as the original source of the City of Huntsville’s public water supply.

Button for reversing picture slide show              Forward button for advancing slide show

Because of the presence of conduits, karst aquifers possess a number of unique and problematic hydrogeologic properties compared to typical granular and fractured-rock aquifers. For example, the flow of water through conduits is rapid, turbulent, and highly anisotropic. Flow velocities in well-developed and well-integrated conduits ranging on the order of 100s-1000s of feet per day are not uncommon. Under these conditions, Darcy’s Law, the fundamental equation used to characterize ground-water flow in most aquifers, does not apply. As a consequence, conventional methods of hydrogeologic characterization based on the application of Darcy’s Law—such as hydraulic tests of wells, water-table or potentiometric-surface mapping, and numerical ground-water modeling—often provide erroneous information about karst aquifers.

Better methods of regional characterization of complex karst aquifers and improved karst analytical and data-management tools are among the goals of the Ground-Water Resources Program (GWRP) of the U.S. Geological Survey (USGS). The GWRP encompasses regional studies of ground-water systems, multidisciplinary studies of critical ground-water issues, improved access to ground-water data, and hydrogeologic research and methods development. The program provides unbiased scientific information and many of the analytical and data-management tools that are used by Federal, State, and local management and regulatory agencies to make important decisions about the Nation's ground-water resources.

The Karst Hydrology Initiative (KHI) project is a study effort being conducted by the GWRP to develop improved methods of using Geographic Information System (GIS) technology to collect and synthesize regional karst hydrogeologic data and improve hydrologic analysis in well developed karst terrains, especially with regard to water budgets. Work being done as part of the KHI project is also providing information to address three research areas of high priority to the GWRP including: (1) effective characterization of ground-water flow in shallow aquifer systems; (2) better understanding of ground water and surface water interaction; and (3)improved regional hydrogeologic frameworks.

All research and study activities conducted as part of the KHI project are directed by the Kentucky Water Science Center (KWSC) and are focused on the major karst aquifers within the Interior Low Plateaus region of the central United States (figure 1). The regional study area includes large parts of Kentucky, Alabama, Indiana, and Tennessee underlain by karstic Mississippian, and to a lesser extent, Ordovician limestone rocks. Major environmental or water-resource management issues in the regional study area include: 1) characterization of the availability and quality of ground water; (2) delineation of spring recharge areas or source-water protection areas; (3) determination of water budgets for spring basins and karstic watersheds; (4) availability of hydrologic and water quality data needed to support surface and ground-water regulatory programs and stream-use designations; (5) reduction or elimination of non-point source pollutants in runoff drained by sinkholes or sinking streams; and (6) protection of cave habitats and threatened or endangered cave species, and (7) identification and mitigation of karst geohazards, such as flooding and the formation of cover-collapse sinkholes. The ability to evaluate and manage these resource and geohazards issues is dependent on the availability of suitable karst hydrogeologic data, on the interpretation of these data, and the effectiveness of analytical, resource-management, or decision-support tools that utilize these data.

Additional information about karst hydrogeology, karst in Kentucky, and about the KHI project, can be obtained by contacting Chuck Taylor, KWSC, (phone) 502-493-1931, (email)

A Compilation of Provisional Karst Geospatial Data for the Interior Low Plateaus Physiographic Region, Central United States

U.S. Geological Survey Data Series 339 Report is the first of several planned reports to be released as products of the GWRP KHI Project. The report describes GIS methods that were used to compile a regional karst geospatial dataset for the Interior Low Plateaus study area (figure 1). Specific karst features that have been digitally mapped include sinkholes, sinking or disappearing streams, and their catchments; karst springs inventoried in the USGS National Water Information System (NWIS) database; relic stream valleys; and plotted karst flow paths inferred from the results of previously-reported water-tracer tests conducted in various parts of the regional study area.

Because knowledge of the distribution and interrelation of regionally-mapped karst features is potentially valuable to water-resources managers, environmental regulators, geotechnical or environmental consultants, and others, the geospatial data files that were compiled during the KHI project during 2005-2007 are being released as a provisional dataset. Visualization and manipulation of the geospatial data is accomplished using ESRI’s® ArcReader ArcGIS Desktop Viewer® (available for download at

To facilitate use in the context of a watershed management framework the karst geospatial data files are compiled by eight-digit Hydrologic Unit Codes (HUC-8s). Figure 2 shows the location and boundaries of HUC-8s within the regional study area, and Table 1 provides a list of the same. An example of the karst data that may be available for each HUC-8, and that can be displayed and visually evaluated using the ArcReader user interface, is shown in figure 3A-B. All other data content, and ArcReader display and manipulation capabilities, are discussed in detail in Data Series 339 Report.

At present, the karst geospatial data files are available by request only on DVD storage media. Please send written inquires by email to, or, or call Charles J. Taylor at the Kentucky Water Science Center at 502-493-1931. At some future date, we anticipate that the geospatial data files may also be made available by direct download through the Internet. Efforts are also being made to enlarge and update the content of the dataset as new karst geospatial data become available and as funding and available human resources permit. Please check this webpage for periodic updates.

Karst researchers, or others, who have karst hydrogeologic mapping or water-tracer data that they may be interested in contributing to this regional database are encouraged to contact the principal investigator (Charles J. Taylor).

Map of the study region

Figure 1.

Map showing location and boundaires of HUC-8s

Figure 2.

Screen capture showing the ArcReader user interface

Figure 3. (A)

Screen capture showing digital map image

Figure 3. (B)

Table with the list of HUC-8s included in the regional study area

Table 1.

Application of GIS Hydrologic Data Models to Karst Terrains

The Graham Springs karst basin as represented by a karst HDM created using ESRI Arc Hydro®. Shaded-relief topography is derived from 10-meter DEM data. Karst basin boundaries are indicated by bold red line; tracer-inferred flow paths by dashed purple lines, and sinking streams by bold blue lines. Note the large area of sinkhole-dominated topography in the center of basin. The karst HDM is a dynamic GIS-based digital representation of surface and subsurface drainage features and the basin's physical hydrologic framework. Karst HDMs have great potential as a useful new water-resources data analysis and management tool.

Hydrologic Data Models (HDMs) are digital representations of surface drainage networks created from DEM datasets using GIS terrain-processing methods and specialized software such as ESRI Arc Hydro (Maidment, 2002). HDMs are being increasingly applied as hydrologic tools to address complex water-resources management issues. However, the creation of HDMs is highly problematic in karst terrains where surface drainage is disrupted or captured by sinkholes, sinking streams, and subsurface conduits.

To investigate whether HDM processing methods and software could be suitably applied in karst, a demonstration project was conducted during 2007-08 as part of the USGS Ground-Water Resources Program Karst Hydrology Initiative (Taylor and others, 2008). The Graham Springs karst basin in south-central Kentucky (fig. 1) was selected as a test case. This large (360 km2) well developed karst basin is characterized by extensive sinkhole topography, sinking stream recharge, and complex subsurface flow routes that discharge to multiple karst underflow and overflow springs. These characteristics are typical of the karst hydrology exhibited in many parts of the eastern and central United States.

Map of the Graham Springs karst basin.

Figure 1. Map of the Graham Springs karst basin as delineated by Ray and Currens (1998, 2000). Note basin boundary lines (dashed), tracer- inferred subsurface flow paths (red), sinking streams (light blue). Water drained by the basin is discharge from multiple karst springs (open and closed blue dots, to left).

Conventional HDM processing methods generally require the use of an artificially-smoothed input DEM which unfortunately results in the elimination of closed natural depressions such as sinkholes. For this demonstration project, new GIS-data processing methods developed and described by Taylor and Nelson (2008) were applied to create a topography input dataset that preserves sinkhole and sinking stream catchments identified in the Grahams Spring karst basin (fig. 2). In the screen-capture image below, the larger sinkhole catchments in the basin (bold green boundary) are indicated by black-bounded and orange-shaded polygons. Along with the sinking streams, these sinkhole catchment polygons define internally-drained contributing areas (sub-basins) in the karst HDM.

Screen capture image showing aggregrated karst topography dataset.

Figure 2. Screen capture image showing aggregrated karst topography dataset used in the karst HDM created for the Graham Springs karst basin (bold green line). Black polygons and orange-shaded areas represent sinkhole catchments identified in the area.

The Digital Drainage Network for the karst HDM as created using hydrologic feature classes defined in the Arc Hydro schema (fig. 3) to represent major surface and subsurface karst drainage features. Key hydrologic feature classes defined in Arc Hydro include HydroPoints (used to represent karst springs), Watersheds (sinkhole and sinking stream catchments), and HydroEdges (surface stream reaches and subsurface tracer-inferred flow paths). The HydroID is a unique identifier for each hydrologic feature entered in the geodatabase, and is used along with Arc Hydro flow attributes (NextDownID, FlowDir, et cetera) to establish the proper topological connections between all individual surface and subsurface features that comprise the digital drainage network.

Diagrammatic sketch of the hydrologic feature classes and attributeds

Figure 3. Diagrammatic sketch of the hydrologic feature classes and attributes defined in the Arc Hydro schema (modified from Maidment, 2002).

Using Arc Hydro in combination with manual and automated terrain-processing methods, an aggregated polygon-and-vector drainage network is created for the HDM that provides an effective digital representation of the physical hydrologic framework of the Graham Springs karst basin. The drainage network links internally drained sinkhole catchments, dye-tracer-input sites, and sinking streams (inputs) to karst springs or surface streams (outputs) by way of tracer-inferred flow paths (throughputs). Specific hydrologic connections between individual sinkhole or sinking stream catchments and tracer-inferred flow paths were created in Arc Hydro using schematic links and nodes (fig. 4).

Figure 4. Screen-capture image of part of the Graham Springs karst basin HDM, illustrating use of schematic links (black lines) and nodes (dots) to link individual sinkhole catchments to tracer-inferred flow paths (dashed purple lines) by way of specific HydroJunctions (blue diamonds) created using Arc Hydro functions.

Temporally/Spatially Variable Flow is also digitally represented in the karst HDM using available ArcGIS flow network tools. For example, under low-to-base flow conditions (fig. 5a) the perennial or underflow discharge from the basin is illustrated by the yellow flow-path line, and a flow blocking flag (“x”) is used to deactivate one of several known overflow spring routes (purple line). Under high flow conditions (fig. 5b), the overflow spring route is reactivated (yellow line) by removing the flow blocking flag.

Screen-capture images illustrating digital representation of temporally/spatially Screen-capture images illustrating digital representation of temporally/spatially

Figure 5a. (left), Figure 5b. (right). Screen-capture images illustrating digital representation of temporally/spatially variable karst flow. See accompanying text for details.

While conventional HDMs are often used to predict, or automatically calculate, discharge through surface stream networks using simple relations between precipitation, surface runoff, slope, and area, discharges within the subsurface conduit-drainage networks of karst basins cannot be automatically calculated or predicted at present given available model algorithms and software functionalities. However, time-series data, such as discharge measurements, can be stored and linked to any geodatabase feature. This enables the HDM to serve as a dynamic, GIS-based tool for storing, retrieving, and visualizing data; identifying flow routes, upstream water sources, and downstream water receptors under variable flow conditions; and for tracking and assessing temporal or spatial changes in field-monitored hydrologic and water-quality parameters. Thus, although the geospatial data input and processing requirements are intensive, and some practical and technical limitations exist, karst HDMs have good potential for use as a valuable new karst water-resources assessment and management tool.

Additional Information about this demonstration project is provided in the paper by Taylor and others (2008) included in “Sinkholes and the Engineering and Environmental Impacts of Karst”, Proceedings of the 11th Multidisciplinary Conference, American Society of Civil Engineers, Geotechnical Special Publication No. 183, p. 146-155. Copies of the proceedings, or the paper itself, are available for purchase at


Accessibility FOIA Privacy Policies and Notices

Take Pride in America logoU.S. Department of the Interior | U.S. Geological Survey
Page Contact Information:
Page Last Modified: Monday, June 08, 2009 08:08:35 AM