Dale et al. (1997) used a GIS to deductively predict potential cedar barrens within the Department of Energy's Oak Ridge Reservation by intersecting soil, slope, geology, and land use characteristics. They defined potential cedar barren habitat as the intersection of specified Mollisols (lithic Rendolls) or lithic and typic Hapludalf soil series, along with thin- and medium-bedded Chickamauga limestone residuum and colluvium geology, slopes less than 25%, and land use/land cover types of urban, pasture, barren land, mixed forest, or transitional areas. The spectral signature for exposed limestone within barrens resembles that of urban gravel or concrete (Dale et al. 1997). Dale et al. (1997) found that Mollisols were always associated with limestone cedar barrens, indicating that these sites have long been dominated by herbaceous vegetation. However, some cedar barren communities were located not on Mollisols, but on Alfisols, indicating a mixed history of vegetation cover.

We hypothesized that limestone barrens at Fort Knox would most likely occur on Mollisols or Alfisols, but would not be found on Ultisols. From these edaphic principles, we hypothesized that the Corydon (RoE) and Fairmount (FkF, FnF) lithic Mollisols, the Cynthiana (FlE) lithic Hapludalf, the Caneyville (CbD, CnD, and CnE), Fredonia (FdC), and Hagerstown (HnB, HnC, and HnD) typic Hapludalfs would be associated with cedar barrens. Thus, potential cedar barren communities at Fort Knox were deductively defined as one of the above soil types and either transitional, barren, urban, maintained grass, cropland, or lawn grass land use categories, with slope less than or equal to 30%. Geologic categories required for potential barrens were predominantly thin- to medium-bedded limestone types, including Harrodsburg, Salem, Ste. Genevieve, or St. Louis formations, or argillaceous or crinoidal limestone.

Fort Knox has established a preserve for cedar barrens in the southernmost tip of the reservation, and the management plan includes a coarse map of cedar barrens within the Cedar Creek Glades Preserve (White 1993). We made iterative edaphic and physiographic refinements in the cedar barrens habitat specification based on comparison of predicted prospective barrens with mapped barrens locations within the barren preserve.

The addition of Garmon soils underlain by Harrodsburg or Salem limestone resulted in a site-specific accuracy (Mead and Szajgin 1982) of predicted barrens locations of nearly 100% for all known cedar barrens inside the Cedar Creek Glades Preserve. The Garmon soil series, a dystric Eutrochrept Inseptisol, is reported by Baskin et al. (1994) as the primary series on which xeric limestone prairies occur in Hardin and LaRue counties, KY. The soils mapped as Garmon in the Hardin and LaRue county soil surveys contain thin bands of Caneyville-Rock outcrop complex (Arns et al. 1979). These Caneyville-Rock outcrop inclusions only occur where Salem and Harrodsburg limestone were present. Thus, the Garmon (GmE, GmF) dystric Eutrochrept was included in the cedar barren specification, but only where underlain by Harrodsburg or Salem limestone.

For further information contact Linda Mann, Bill Hargrove, or Tom Ashwood


Arns, F. S., M. J. Mitchell, F. C. Watts, and B. L. Wilson. 1979. Soil Survey of Hardin and Larue Counties, Kentucky. Soil Conservation Service, USDA, Washington, D.C.

Baskin, J. M., C. C. Baskin, and E. W. Chester. 1994. The Big Barrens Region of Kentucky and Tennessee: Further Observations and Considerations. Castanea 59:226 –254.

Dale, V. H., A. W. King, L. K. Mann, R. A. Washington-Allen, and R. A. McCord. 1997. Assessing land-use impacts on natural resources. Environ. Manage. (in press)

Meade, R. A. and J. Szajgin. 1982. Landsat classification accuracy assessment procedures. Photagrammetric Engineering and Remote Sensing 48(1):139 –141.

White D. L. 1993. Management and Monitoring Plan for a Limestone Hillside Glade System, Fort Knox Military Reservation, Kentucky (January 1993). Kentucky State Nature Preserves Commission, Frankfurt, Kentucky.

Return to the top of this page

Return to the Specific Models page

Return to the Web Site Map

Return to the Ecological Modeling home page