Date of Award

12-2010

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Forestry and Natural Resources

Advisor

Shelburne, Victor B.

Committee Member

Lanham , J. Drew

Committee Member

Post , Christopher J.

Committee Member

Smith , Bill R.

Committee Member

Wang , Gaofeng G.

Abstract

Multifactor ecosystem classification systems provide a three-pronged approach to identifying site units across the landscape based on repeating patterns of vegetation, soil, and geomorphology. Ecosystem classification models have been developed for a diversity of forest landscapes throughout North America, and are beneficial as an ecosystem management tool because the outcome yields data models that can be utilized by scientists and natural resource managers alike. In contrast to the enormous amount of classification studies undertaken in relatively stable, older-aged forests in eastern North America, there have been few studies that have employed multifactor classification techniques across a successional gradient, or heavily disturbed forests of the same region. The 17,500-ha Jocassee Gorges tract in upstate South Carolina represents an ideal landscape to examine both spatial and temporal variability in vegetation-environment relationships due to its myriad of landforms and long history of intense forest management over the past century. Successional vegetation patterns across this heavily disturbed, spatially heterogeneous landscape were examined using a multifactor landscape ecosystem classification (LEC) framework developed from ecosystem types described from older-aged (> 75 years) stands. Ecosystem types for three age-classes of stand development post-timber harvest (10-25, 26-50, and 51-75 years) were determined by using environmental discriminants identified in the previous older-aged (reference) stand classification, and a total of 63 plots were established in previously logged stands between April 2003 and October 2004.
Composition of ground flora and woody stem species, along with landform and soil datasets, was compared across age-classes within and among ecosystem types using non-metric multidimensional scaling, non-metric multi-response permutation procedures, and indicator species analysis. Woody stem composition remained similar between age-classes of xeric oak-blueberry and mesic hardwood-bloodroot ecosystem types, while woody stem composition was drastically different on early successional age-classes of the xeric chestnut oak-mountain laurel, submesic oak-mixed flora, and mesic hemlock-rhododendron ecosystem types. Ground flora composition differed between successional and reference age-class for each ecosystem type. Comparisons of ecosystem types across age-classes revealed the following trends: woody stem and ground flora species composition was similar between mesophytic ecosystem types, but differed between xerophytic types; by middle succession age-class (26-50 years) ground flora composition was distinct between all ecosystem types, except the submesic oak-mixed flora type which contained species diagnostic of all others; and by late succession age-class (51-75 years), both ground flora and woody stem composition differed between all ecosystem types. When ground flora and woody stems were placed into ecological species groups, canonical correlation analysis revealed similar trends in middle to late age-classes to those exhibited in reference age-classes. Overall, forest management has not had a severe effect on the disturbance regime across the Jocassee Gorges landscape to cause a significant shift in species composition within any ecosystem type. Although composition and diversity change across temporal gradients of each type, this is to be expected in a highly disturbed landscape of the southern Appalachian Mountains due to past natural and anthropogenic factors interrupting the process towards steady-state forests. Ecological classification systems are most effective in guiding ecosystem management processes when they are designed to document successional variation, as well as spatial heterogeneity, across landscapes. Adding a fourth (time-series) component to the LEC framework allows for a more accurate approach to documenting the biological diversity within a region, and serves as a more robust management tool because of its ability to predict vegetation across successional land units.

Share

COinS