Evaluating and revising a long-term monitoring program for vascular plants: Lessons from shenandoah national park

Monitoring programs have been established by agencies to assist in the conservation of biological diversity and to assess the impacts of a variety of threats to species and natural systems. However, poorly designed programs may fail to detect changes in the system they were intended to monitor. We evaluated and revised the long-term monitoring program for vascular plants at Shenandoah National Park (SHEN) and used our evaluation and revision to illustrate common problems associated with monitoring programs. Our evaluation of data collected through SHEN's 14-year monitoring program indicated that the design and sampling effort lacked the statistical power to detect changes in stem density of shrubs or saplings, although there was sufficient power to detect changes in tree basal area for some species. The inability of the program to detect changes was due, in part, to the lack of specific management and sampling objectives, stratifying the park according to forest cover types that changed over time, errors associated with misidentification of species, and an inappropriate temporal schedule for visiting sampling plots. Our revised sampling design attempted to correct these deficiencies by using strata that are unlikely to change over time (elevation, bedrock geology, and aspect), stream-lining data collection to allow a greater number of sampling plots, and visiting all plots over a 2-year period (interpenetrating sampling) to increase sample size. Long-term monitoring programs should: (1) establish specific, quantifiable, and attainable management and statistical sampling objectives; (2) use permanent characteristics to define strata (e.g., elevation or bedrock geology), if stratification is warranted; and (3) perform a statistical evaluation of the monitoring program once sufficient data have been collected. Prompt evaluation of an implemented monitoring program will allow timely modifications of either sampling objectives or the sample design with minimal delay in collecting meaningful long-term trend information.