Remote Sensing of Sinkholes in Barren County, Kentucky

by

Chris Jung

ES 775 Advanced Image Processing, Emporia State University

This photo of a collapsing sinkhole was downloaded from the USGS website.


Abstract
Karst
Study Area
Natural Color Composite
False Color Composite
Ratio and Reclass
Conclusion
References


Abstract

Remote sensing is the science of acquiring information about an object without being in direct contact with the object (Jensen, 2000). In this project remotely sensed Landsat Thematic Mapper (TM) data was manipulated using IDRISI software to differentiate between the land surface and small circular water filled sinkholes. The study area is a heavily karst area located in rural Barren County, Kentucky.

Return to the top


Karst

Karst features are prevalent in many of Kentucky’s limestone formations. As water flows through limestone, the rock is dissolved and small fractures become larger conduits, which in turn transport soils and surface water through the karst system. Sinkholes that form in karst conduit flow conditions are either characterized as collapsing or dissolution sinkholes. Collapsing sinkholes form when the dissolution of underlying bedrock forms a cavity and the roof of this cavity collapses. The image that appears at the beginning of this website is an example of a collapsing sinkhole. A dissolution sinkhole, which is more common than a collapsing sinkhole, forms when rock is dissolved and soils are slowly slumped away forming a depression in the land surface.

Topographic/Hillshade map (Right) and Percent slope map (Left)

Land management in karst sites has become increasingly more important as the need for fresh drinking water sources has increased with an increase in world population. In the United States over 40 percent of the drinking water comes from karst aquifer systems. Controlled management strategies around sinkholes are important to prevent the direct contamination of water supplies, destruction of cave ecosystems, and damage to developments due to land subsidence.

Return to the top


Study Area

The study area for this project is located in Barren County Kentucky, just south of the Mammoth Cave National Park. Public water systems supply 88 percent of the drinking water to the 37,163 people living in the county. The area is used mostly for agricultural purposes. The karst features in the area consists of sinkholes, sinking streams, and caves (Carey, D.I., Stickney, J.F. 2002).

Aerial map

Return to the top


Natural Color Composite

The selected site was first determined to be in Path 20 and Row 30 of the Landsat TM data (1990 +/- 3 years). The data were retrieved from http://www.landsat.org. The images were converted from TIFF image format to an IDRISI image using the import conversion option of IDRISI. The image was further cropped using the image windowing function to reduce the image to include only the area of study.

123 (BGR) Natural Color Image

Natural color composite images most closely resemble the range of vision of the human eye. The composite image was generated using bands 1, 2, and 3 (BGR) and the IDRISI Composite function. The image has a blue hue because blue light is easily scattered by atmospheric conditions. The forested areas are the darkest, while agricultural fields are lighter and appear in block patterns. The small town of Park City is located in the northwest quadrant of the image. This composite image is not well suited for identifying water filled sinkholes.


False Color Composite

Several false color composite images were produced for the area of study. Three composites were selected based on their clarity and ability to discern ground features (234, 354, & 742 (BGR)).

False Color Composites: 234 (Left), 354 (Center), 742 (Right)

False color composite 234 is the standard false color composite image. Dense areas of vegetation are represented by dark shades of red while agricultural land is represented by light red. Water in the image is identified as dark green to blue. In the 354 composite there is a large contrast between the orange color of the trees and the light green color of the agricultural areas. The small ponds and water filled sinkholes are also a contrasting dark blue to the surrounding green. In the final composite 742 the forested areas are light green and varying crop uses can be determined by varying shades of purple and green. The water in the final composite is dark red.

Return to the top


Ratio and Reclass

Several ratio images were created for the project using the image overlay function of IDRISI. Initially, a 4/3 ratio was selected for the final analysis of sinkholes in the study area, but after much trial and error the 4/1 ratio was selected as the best for discriminating between the land and water.

4/1 Ratio (Left) and Ratio Reclass (Right)

The 4/1 ratio image was then reclassed to give all the data that were not water a value of zero and the water a value of one. The background of the image was then changed to create greater contrast. As seen in the reclassed image the study area has many sinkholes.

Final Image of Sinkholes

Return to the top


Conclusion

The initial purpose of this project was to discern cave entrances in the Mammoth Cave National Park. The constant temperatures found in the cave should be detectable through remote sensing with a large enough contrast in temperature between the entrance of the cave and the surrounding ambient temperature. However, several attempts were made, and it was concluded that Landsat TM data does not have the resolution capabilities to identify temperature changes across an area as small as a cave entrance. The project was refocused on locating water filled sinkholes of Barren County, Kentucky. The sinkholes were adequately identifiable from the surrounding land cover using a Landsat TM band 4/1 ratio. Further reclassing of the image eliminated unwanted data and improved the visibility of the sinkholes.

Return to the top


References

Carey, D.I., Stickney, J.F. 2002. Groundwater Resources of Barren County, Kentucky, Kentucky Geological Survey, http://www.uky.edu/KGS Retrieved on 11/27/09.

Clark Labs. Geographic Analysis and Image Processing Software. http://www.clarklabs.org/

Jensen, J.R. 2000. Remote Sensing of the Environment. Prentice Hall, pp. xiii, 192-197

Landsat.org Web Site. http://landcover.org Retrieved on 11/27/09.

USGS Website Website. http://ga.water.usgs.gov/edu/earthgwsinkholes.html Retrieved on 11/27/09.

Return to the top


This web page was created to fulfill the requirements for ES 775 Advanced Image Processing at Emporia State University.
Return to student webpages.
This page was created on 12/09/09.