Forest Patch Size , IAnd Use , and Me & Forest Herbs inn the French Broad River Basin , North Carolina

The effect of forest fragmentation on cove-forest herbs was studied in the Southern Blue Ridge Provjnce. Patches of mesic forests were sampled with 4 ha study plots. The coverage and density of herb species were greater in large patches (>200 ha) than in small patches (cl0 ha). Several ant-dispersed species, such as Disporum mud&urn and Uvularia gmndijlorw, were more likely to be absent from small patches than from large patches. Wind-dispersed species, such as ferns and composites, were not a&ct,ed by patch size and isolation. Small patches had reduced amounts of organic matter in the soil, suggesting that small patches have experienced more disturbance than large patches. Ctherwise, there were no other differences in soil characteristics between patch sizes. Mechanisms hypothesized to have a&cted these populations include (a) disruption of population dynamics due to habitat f?agmentation, (b) habitat degradation, and (3 anthropogenic disturbance via land use. Disturbances may have affected herb populations directly by increasing mortality rates and by degrading habitat. These habitat changes were confounded by the small size and isolation of small forest patches. I N T R O D U C T I O N Land use is the leading cause of landscape changes Landscape-level changes often-result in habitat loss and fragmentation (Skole et al. 1994, Turner et al. 1994, Sinclair et al. 1995). Both worldwide and in the United States, land cover is altered principally by direct human use-through agriculture, pasture, forestry, and development (Meyer and Turner 1992). Land use patterns affect both terrestrial and aquatic systems (Beiners et al. 1994, Cooper 1995) and influence biodiversity for several reasons (Turner et al. 1997). First, land-use activities may alter the relative abundances of natural habitats and result in the establishment of new landcover types. Species richness may be enhanced by the addition of new cover types, but natural habitats are often reduced, leaving less area available for native species (Walker 1992). Exotic species may become established and outcompete the native biota. Second, the spatial pattern of habitats may be altered, often resulting in fragmentation of once-continuous habitat. Species using a particular habitat may respond individualistically to these changes. Species that differ in their patterns of survival, fecundity, and dispersal, as well as in habitat needs, wih likely differ in their response to habitat loss and fragmentation. Species having specialized habitat needs, requiring a large area for home range, or having limited vagility, will be impacted more than generalist species that disperse well and can live in small isolated patches (Terborgh 1992, Dale et al. 1994). Clearly, the conservation of native species and their habitats requires a landscape-level solution (Franklin 1993,1994; Tracy and Brussard 1994). Ecologists are beginning to develop techniques for assessing the complex impacts of landscape changes on biodiversity. Generalizations are difficult to obtain because habitat fragmentation affects species in different ways (Pearson et al. 1996). Assessing spatial and temporal changes in biodiversity by monitoring a suite of species having different habitat requirements has emerged as a useful technique. For example, patterns of deforestation in Bondonia, Brazil, were interpreted from the perspective of species having different area requirements and dis382 CASTANEA VOLUME63 Figure 1. Forest cover in a portion of the French Broad River basin The region shown ia north of Asheville in northern Buncombe and southern Madison Counties, North Carolina. This map shows nonforest (white) and forest @ray) land covers for a 35.9 X 25.6 km region. Streams ars shown as black lines. Numbers show locations of study plots. Plots l-8 were located in small patches of forest. Plots 9-14 were located in large patches. persal capabilities (Dale et al. 1994). In lowto mid-elevation forests in the Pacific Northwest, risk to bird species under four different management scenarios was evaluated by quantifying suitable habitat for each species by using habitat maps, species-habitat associations, and other life-history requirements (Hansen et al. 1995). Avian diversity has been used as an indicator of the vertical and horizontal structural complexity of vegetation over a broad region (Flather et al. 1992). Butterflies (Kremen 1992,1994) and birds and butterflies (Debinski and Brussard 1994) have been used as indicators of biodiversity. While species’ responses to landscape patterns are complex, the collective response of many species produces the patterns of biological diversity observed on landscapes. Moreover, landscapes have patterns that change in both space and time. Disturbance influences species diversity in many landscapes, and a better understanding of the interaction between spatial pattern and disturbance is needed (Roberts and Gilliam 1995). While the southern Blue Bidge Province currently has a high percentage of forest cover (approximately 70% forested, SAMAB 1996), forest cover in some regions, particularly in large river valleys, is fragmented by agricultural and suburban land uses. The central portion of the French Broad River basin, North Carolina is one such region (Figure 1). The present-day landscape consists of a mosaic of forest patches interspersed with agricultural and suburban land uses. This region has experienced sustained agricultural use for more than 100 years (Eller 1982), and more recently has experienced a growth of urban and suburban land uses. This region has demonstrated the unlikely combination of increasing human population size and increasing forest cover since 1950. Agricultural land uses have declined since the turn of the century and land formerly in agriculture has reverted to forest, especially since 1950 (Wear