The patch mosaic and ecological decomposition across spatial scales in a managed landscape of northern Wisconsin, USA

Document Type

Article

Publication Date

1-1-2002

Abstract

Understanding landscape organization across scales is vital for determining the impacts of management and retaining structurally and functionally diverse ecosystems. We studied the relationships of a functional variable, decomposition, to microclimatic, vegetative and structural features at multiple scales in two distinct landscapes of northern Wisconsin, USA. We hoped to elucidate any characteristic resolutions of structure-process relationships on these landscapes, and to determine the validity of extrapolation of structure-process associations across scales and management regimes. We used a combination of ANOVA, wavelet, canonical discriminant, and correlation analyses and asked specifically whether: 1) specific combinations of microclimatic, structural, and vegetative features were consistently associated with differences in decomposition among management zones along transects (i.e., within landscapes); and 2) factors influencing decomposition were consistent among resolutions of analysis and depths within and between landscapes. Decomposition was greater on the pine barrens than the small block transect and greater at 4 cm depth than at the surface for both landscapes. Significant differences in decomposition occurred among management patches for both transects and depths, except 4 cm along the pine barrens. In general decomposition was faster for patch types with greater overstory cover at 1 cm (both transects) and lower at 4 cm (small block). Canonical discriminant analysis also separated management patches by overstory cover for both transects. Secondary vectors also separated patches along both transects by microclimate, independent from overstory effects on those variables. Dominant resolutions in the patterns of decomposition differed between transects and depths: 80 m (pine barrens, 1 cm); 500 m (pine barrens 4 cm); 160 m (small block, 1 cm); and 750 m (small block, 4 cm). At the 1 cm depth, the strongest correlations of wavelet transforms of decomposition with structural, microclimatic, and vegetation variables often differed in resolution from those dominant in the decomposition patterns. At the 4 cm depths, many of the strongest correlations occurred at the maximum resolutions examined. Although many important correlates differed between transects and depths within a transect, there were some consistencies. On both transects, surface and soil temperatures were strongly correlated (|r| > 0.40) with decomposition; soil temperatures were stronger correlates along the small block. The direction of association between decomposition and temperatures changed with depth, being negative at 1 cm and positive at 4 cm for both transects. Overstory was an important correlate (|r| > 0.50) for 3 of 4 transect-depth combinations. On both transects, correlations between decomposition and overstory peaked at different resolutions and were different signs (positive at 1 cm and negative at 4 cm) for the two decomposition depths. Along the pine barrens but not the small block, there were two peaks of resolutions of correlation that appeared consistently across variables. Thus, correlates of decomposition changed with scale as well as depth and management regime. This suggests that factors other than management may still be maintaining decomposition patterns on the landscapes. Further, patterns in and relationships to process variables should be examined at multiple scales to develop a comprehensive understanding of the mechanisms driving functional heterogeneity.

Publication Title

Basic and Applied Ecology

Link to Full Text

Share

COinS