ArcUser Online

Exploring the Karst Landscape of Cockpit Country
By Parris Lyew-Ayee, School of Geography and the Environment, University of Oxford

Editor's note: Parris Lyew-Ayee, a doctoral student in the School of Geography and the Environment, University of Oxford, United Kingdom, is using GIS in his study of karst geomorphology not only as a research tool but also to manage and optimize his fieldwork. He is working in a unique region in the interior of western Jamaica called Cockpit Country. The area's thousands of depressions, caused by the erosion of limestone, were thought to resemble cockfighting arenas of 17th century England. Its rugged terrain has limited human incursions and helped preserve the area's biodiversity.

With its deep karst depressions surrounded by residual hills, the Jamaican Cockpit Country is the type area for the cockpit variety of karst landscape. GIS was used to conduct a comprehensive morphometric [shape measurement] survey of the landscape and its features to unambiguously characterize the nature of cockpit karst in its type area and compare these landscapes with noncockpit karst areas as well as distinctly nonkarst areas from other parts of the planet. Such detailed morphometric investigations would provide information about how such landscapes evolved and the factors that control their development.

IKONOS satellite imagery of part of Cockpit Country was draped on a triangulated irregular network (TIN) representation of the terrain to provide photorealistic depictions of the study areas that were useful both before and after fieldwork.

Digital terrain analyses, ranging from geostatistical investigation of the variography of the terrain to analysis of the scaling properties of karsts, as well as digital landform analysis involving the automatic delineation of karst features, were all conducted using GIS. [Variography is an aspect of geostatistics concerned with interpreting variograms, which plot variance as a function of distance of separation.] Additionally, virtual reality scenes were also created to represent the landscape in a photorealistic manner that furnished perspectives of the landscape that would be unavailable from the ground.

Morphometric analysis of karst terrain has been going on since the 1970s, although the use of sophisticated computer-based techniques has not been rigorously applied to karst areas. This research involves many original methods used to describe the landscape and landforms of the Cockpit Country of Jamaica as well as adapting techniques used in other branches of geomorphology, landscape ecology, and mineral exploration. The unique characteristics of the cockpit karst landscape can then be quantified, and methods aimed at conservation and environmental management of such areas can be devised.

This work required topographic data that was supplied through digitizing 1:12,500-scale contour maps. Creating digital elevation models (DEMs) from sources such as imagery or radar was prohibitively expensive. Instead, GPS spot heights collected in the field were integrated with the contour data and converted to vector points so that DEMs could be created. The variography of the topographic data could also be examined using geostatistics. Semivariograms of the different study areas could be produced for comparative analysis and as inputs for kriging interpolation of the data.

Once DEMs were created, the primary topographic attributes, such as mean elevation, slope, aspect, and curvature, could all be computed. Neighborhood analyses of DEMs also allowed extraction of additional information from the DEMs. The range of elevation at different neighborhood sizes revealed key differences between cockpit karst and noncockpit karst areas. As a result, cockpit karst areas could be readily identified—especially in areas of mixed topography—on the basis of their unique characteristics, and a semiautomated method of mapping cockpit karst could be developed.

Digital landform analysis involved the added complication of delimiting karst features from the overall landscape. Sinks and summits in the terrain were identified from the DEMs using hydrological modeling techniques that allowed the patterns and distributions of these features to be described. Delimiting the hills and depressions was a greater challenge. Isolating these features, using manual digitizing and supervised image classification proved unsatisfactory, so the Compound Topographic Index was used. This method involves determination of the slope and catchment areas from the original DEMs. Areas of steep slope and small catchment areas were positive relief features (hills), and areas of gentle slope and large catchment areas were negative relief features (depressions). The digital identification of sinks, summits, hills, and depressions was were all verified by ground-truthing.

The morphometric properties of the delimited landforms could then be determined. Attributes, such as area, height of hill/depth of depression, elongation, degree of circularity, surface area, and slope, were all ascertained. Additionally, the degree of isolation of each set of features could also be calculated. Again, unique properties of cockpit karst areas could be described using the terrain's constituent landforms.

The quantification of such landscapes using GIS enabled further research in determining the origins of such places and can be used in other fields such as investigating the connection between the high biodiversity of this area, its geomorphology, and the application of this information to karst terrain management. Many key assumptions about the nature of such places can be reevaluated on the basis of thorough GIS-based morphometric analysis. Other spectacular karst landscapes exist in southeast China and in other parts of the Caribbean from Belize to Barbados. The extent to which these landscapes are geomorphologically similar to each other can now be assessed.

GIS was an invaluable tool in conducting this research. With GIS, a wide-ranging database could be assembled that included topographic data and its associated derivatives, vector geology maps, and Space Imaging's IKONOS imagery. While many specialized software packages exist, GIS allowed software customization for performing detailed analyses while retaining other aspects of functionality. The GIS methodology can be easily exported elsewhere without the need to develop entirely new software and would enable comparative analyses of different areas based on the same analytical framework.

The use of GIS in this research went beyond analysis of topographic data and DEMs. GIS was also constructively used to plan the logistics of fieldwork and the sampling design for the geological portion of the research in Cockpit Country, taking into account issues of accessibility to the study areas by using satellite imagery, georeferenced topographic sheets, digitized maps, and terrain roughness. Fieldwork information could be recorded on a GPS instrument and readily transferred to the GIS for subsequent analysis. The versatility of GIS for prefieldwork planning and postfieldwork analysis, as well as detailed morphometric investigation of the area, enabled a comprehensive study of Cockpit Country to be carried out efficiently.

The entire GIS portion of the research was conducted using ArcView 3.2 and ArcGIS 8.2 software packages, which were able to integrate GPS data collected in the field as well as IKONOS satellite imagery with the rest of the dataset. Key software extensions used were ArcView Spatial Analyst, ArcView 3D Analyst, and ArcGIS Geostatistical Analyst.