Measuring Forest Integrity At Multiple Spatial Scales: Evaluating Indicator Accuracy, Sensitivity, And Specificity In Regional Assessments

J. Christopher Haney (Presenter), Ecology and Economics Research Department, The Wilderness Society

Carol R. Foss, Audubon Society of New Hampshire

Geetha M. Jayabose, Department of Natural Resources, Cornell University

Cindy DeGrood, The Wilderness Society

Changes in biotic integrity of forests can be referenced against benchmarks consisting of “a balanced, integrated, adaptive community of organisms having a species composition, diversity, and functional organization comparable to natural habitat of the region” (Karr and Dudley, 1981).  Biotic integrity of forest ecosystems can be measured along five axes, any of which may respond to natural and anthropogenic stressors: 1) total cover, 2) degree of fragmentation or patch continuity, 3) age, 4) composition, and 5) disturbance regime.  Ideally, indicators for assessing these components of forest integrity display statistical reliability, including accuracy, sensitivity, and specificity (Murtaugh, 1996).

We evaluated whether avian incidence could assess forest integrity at various scales.  A screening procedure identified 14 species that gave accurate (explained >= 50% variance), sensitive (true positives), and specific (true negatives) predictions of forest cover across large portions of the eastern U.S. (New England to Tennessee).  Most indicators continued to perform well when tested in “new” regions adjacent to the original model’s study domain.  To verify reliability of indicators at smaller spatial scales, we tested predictive accuracy at the watershed level (24 separate USGS HUC units).  A majority of species indicators continued to perform well, in some instances explaining even more variance (>= 80%) in forest cover at this scale.  Avian indicators predicted forest cover and forest core area with somewhat greater accuracy than extent of forest fragmentation and edge.  Our analyses suggest that some ecological indicators are sufficiently reliable to assess regional forest condition over multiple spatial scales.

Workshop II: Monitoring Forest Changes

Online presentation

oak leaf cluster, logo of Sustaining Southern Forests

    John M. Pye
a conference sponsored by the Southern Forest Resource Assessment