A comparison of ground flora sampling methods to assess recovery of bottomland forest restorations.

Abstract

We compared ground flora sampling methods for assessing the state of multiple bottomland forest restorations 6-21 years after restoration implementation in northeast Indiana, United States. Data from standard fixed-area plots of appropriate size and complexity for forest assessments were compared with data from smaller fixed-area plots of different shapes and plotless sampling methods. The methods were compared for their ability to (1) assess species richness and abundance, (2) detect and monitor invasive and dominant species, and (3) understand community composition. We assessed the biases of the compared methods, identified the training and skills needed to conduct sampling, and examined sampling costs in relation to total monitoring costs. Results show that smaller plots are able to detect more species per unit area sampled without significant differences in floristic quality measures. Data from smaller plots were sufficient for comprehensively describing site conditions even when less of the total site area was sampled. Although multivariate analyses of data from smaller plots yielded greater within-group dissimilarity than data from larger plots, multiple response permutation procedure analyses indicated no significant differences between nonmetric multidimensional scaling solutions based on data from the different sampling methods. Regardless of the sampling methods used to collect data, use of multivariate analyses identified a gradient of time since restoration was implemented as the dominant factor relating to differences between community composition. Sampling costs spanned a range of one order of magnitude but generally represented less than 6% of total assessment costs. Results suggest that when selecting sampling methods, matching monitoring effort to specific measurable management endpoints is more important than sampling cost.