A quantitative analysis of forest island pattern in selected Ohio landscapes

The purpose of this study was to quantitatively describe the various aspects of regional distribution patterns of forest islands and relate those patterns to other landscape features. Several maps showing the forest cover of various counties in Ohio were selected as representative examples of forest patterns to be quantified. Ten thousand hectare study areas (landscapes) were delineated on each map. A total of 15 landscapes representing a wide variety of forest island patterns was chosen. Landscape placement was dependent upon the grain and intensity inherent in the forest island patterns.

Raw data taken from each landscape included measurements of the perimeter and area of each woodlot and measurements of the distances between woodlots. The raw data were converted into a series of continuous variables which contained information pertinent to the sizes, shapes, numbers, and spacing of woodlots within a landscape. Other features of landscapes (topography, soils, glacial history, original vegetation, and potential natural vegetation) were recorded as non-continuous nominal variables.

The continuous variables were used in a factor analysis to describe the variation among landscapes in terms of forest island pattern. Most of the variation among landscapes was accounted for by two factors. One factor was strongly related to the portion of the study area covered by forest, as measured by the variable percent forest cover. The other important factor was the degree of forest fragmentation, as measured by either the total number of woodlots in a landscape (island density) or landscape DI (a function of forest perimeter/area ratio).

The factors pertaining to forest cover and fragmentation were used as the basis of a two-dimensional ordination of landscapes. The ordination proved to be a useful graphic summary of landscape variation on which to test hypotheses concerning forest island patterns.

The environmental features of landscapes (as measured by nominal variables) were related to forest island patterns by: (1) overlaying environmental information on to ordination plots; and (2) using discriminant analysis with the environmental features as the criteria of discrimination. The results showed that forest island patterns are related to topography and other environmental features correlated with topography. Landscapes with smooth topography and arable land had but a few percent forest coverage divided among a relatively low number of small woodlots. With increasing topographical roughness, the size and/or density of forest islands increased. The greatest degree of forest fragmentation was associated with 20-30 percent forest cover in hilly topography. Since forest pattern is a result of the cumulative effects of land use over time, these results were interpreted to mean that the patterns of land use in Ohio were somewhat dependent on topography and related factors.

The ability to quantify landscape pattern on a regional basis has applications to regional ecology. Mathematical models can be developed to predict changes in landscape pattern for given land use scenarios, and parameters of landscape pattern can be analyzed as indicators of habitat and resource distribution and long-term ecosystem stability. The ability to measure forest island pattern and other aspects of landscape pattern can be one part of quantitative approaches to solving regional ecological and land use problems.