| I investigated the relationships between spatial structure and understory vegetation at various scales in a managed landscape in northern Wisconsin. My objectives were (1) to investigate relationships among patch type, understory vegetation, and site factors, (2) to identify relationships between understory vegetation and landscape structural features across multiple resolutions within a single landscape, and (3) to compare multi-scale relationships between vegetation and landscape structure among four different landscapes.; To address the first objective, I sampled 77 patches representing seven patch types. Within each patch interior, I sampled overstory, soils (pH, moisture, organic matter, carbon, nitrogen), understory vegetation, duff depth, canopy cover, litter, and woody debris. ANOVA showed no differences in plant species richness and H′ among the patch types, except between clearcuts and hardwoods. MANOVA demonstrated that canopy cover differed more than other variables among patch types. Quantitative relationships (determined by regression and ordination) were weak, but improved when patch types were grouped by overstory characteristics. Patches occurred along a gradient of canopy cover, soil moisture, and duff depth.; A 3575-m transect was placed within the pine barrens landscape for Objective 2. Understory vegetation and site variables were recorded in 1 x 1m quadrats located every 5m along the transect. Many species' distributions corresponded with landscape structure. TWINSPAN and DCA identified three overlapping species groups. Wavelet transforms showed varying relationships between landscape structure and plant diversity at different resolutions. Effects of roads were apparent at a wide range of resolutions. At broad scales, cumulative effects were apparent, emphasizing the importance of a broadscale perspective.; For Objective 3, I sampled 3000m transects in four landscapes: POA (pine-oak-aspen forest), LB (large-block pine-oak), SB (small-block pine), and PB (pine barrens). Sampling methods were identical to those for Objective 2. PB was more diverse than the other transects. Landscape-level species diversity was more correlated with number of roads/edges along the transects than to number of patches. The landscape matrix and mosaic also influenced plant diversity. Wavelet analysis demonstrated parallels between diversity pattern and patch structure across multiple scales. Wavelet variances indicated diversity patterns were dominated by >1 scale, suggesting that single-scale analysis may give incomplete information. |
