A unifying approach for evaluating the condition of wetland plant communities and identifying related stressors
Johnston, C.A., Zedler, J.B., Tulbure, M.G., Frieswyk, C.B., Bedford, B.L. and Vaccaro, L. (2009) A unifying approach for evaluating the condition of wetland plant communities and identifying related stressors. Ecological Applications, 19 (7). pp. 1739-1757.
|PDF - Published Version |
Download (1044kB) | Preview
*Subscription may be required
Assessment of vegetation is an important part of evaluating wetland condition, but it is complicated by the variety of plant communities that are naturally present in freshwater wetlands. We present an approach to evaluate wetland condition consisting of: (1) a stratified random sample representing the entire range of anthropogenic stress, (2) field data representing a range of water depths within the wetlands sampled, (3) nonmetric multidimensional scaling (MDS) to determine a biological condition gradient across the wetlands sampled, (4) hierarchical clustering to interpret the condition results relative to recognizable plant communities, (5) classification and regression tree (CART) analysis to relate biological condition to natural and anthropogenic environmental drivers, and (6) mapping the results to display their geographic distribution. We applied this approach to plant species data collected at 90 wetlands of the U.S. Great Lakes coast that support a variety of plant communities, reflecting the diverse physical environment and anthropogenic stressors present within the region. Hierarchical cluster analysis yielded eight plant communities at a minimum similarity of 25%. Wetlands that clustered botanically were often geographically clustered as well, even though location was not an input variable in the analysis. The eight vegetation clusters corresponded well with the MDS configuration of the data, in which the first axis was strongly related (R2 = 0.787, P < 0.001) with floristic quality index (FQI) and the second axis was related to the Great Lake of occurrence. CART models using FQI and the first MDS axis as the response variables explained 75% and 82% of the variance in the data, resulting in 6-7 terminal groups spanning the condition gradient. Initial CART splits divided the region based on growing degree-days and cumulative anthropogenic stress; only after making these broad divisions were wetlands distinguished by more local characteristics. Agricultural and urban development variables were important correlates of wetland biological condition, generating optimal or surrogate splits at every split node of the MDS CART model. Our findings provide a means of using vegetation to evaluate a range of wetland condition across a broad and diverse geographic region.
|Publication Type:||Journal Article|
|Publisher:||Ecological Society of America|
|Copyright:||© 2009 by the Ecological Society of America.|
|Item Control Page|