Date of Award:
Master of Science (MS)
To meet the requirements of the Clean Water Act (1972), natural resource managers need to be able to detect biological degradation in wetland ecosystems. Biological indices are commonly used by managers to assess wetland biological condition. The accuracy and precision of wetland condition assessments are directly related to the performance of these indices, and biological index performance is thought to be related to how well an index controls for the effects of environmental attributes on biological assemblages. Many plant-based biological indices control for environmental and biological variation through the use of classification schemes that are based on geographic location and dominant vegetation type. However, the use of classification schemes tends to produce indices with limited applicability and may not adequately control for natural variation. The goal of my research was to use modeling techniques, as an alternative to classification, to account for biological variation associated with natural environmental gradients and to improve the performance of previously developed indices.
I developed two types of model-based biological indices to quantify the biological condition of Ohio wetlands: a vegetation-based index of biological integrity (MVIBI) based on several attributes of wetland plant assemblages, and several indices of plant assemblage taxonomic completeness. I evaluated the accuracy and precision of the MVIBI relative to previously developed indices, and determined that the use of modeling techniques can significantly improve the performance of plant-based indices of biological integrity. Due to increases in accuracy and precision, use of the MVIBI should improve manager’s confidence in wetland biological condition assessments. The indices of taxonomic completeness exhibited poor performance relative to similar indices developed for other types of biological assemblages (i.e. aquatic insects, fish). I attribute poor index performance to my inability to accurately predict individual species occurrence, which is likely a result of plant communities being heavily structured by random disturbance events and biotic interactions that are difficult to account for. My results should help inform index developers of ways to potentially improve wetland condition assessment indices.
Jones, Robin C., "Modeling to Improve Vegetation-Based Wetland Biological Assessment" (2014). All Graduate Theses and Dissertations. 2082.
Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .