HOME     SCHEDULE     AUTHOR INDEX     SUBJECT INDEX         

PARENT SESSION
Oral Session #100: Landscape Ecology.
Presiding: T. Crist
Friday, August 9. 9:00 AM to 12:00 PM. Gila Meeting Room, TCC.


Landscape pattern and stream habitat quality in Southeastern Michigan.

Allan, David*,1, Cifaldi, Rebecca1, Infante, Dana1, 1 University of Michigan, Ann Arbor, MI

ABSTRACT- The ecological integrity of streams is affected by diverse human actions throughout the watershed. We explore the effectiveness of different landscape measures in predicting habitat condition for headwater streams in two watersheds of Southeastern Michigan. Using 109 sub-catchments (avg. area 43 km2) as landscape units and 26 metrics identified in prior studies as effective indicators of landscape pattern, we identified five significant principal components axes that explained ~70% of the variation in land cover spatial patterns across the study landscape. The first two axes identified a fragmentation gradient and a gradient in patch size variation. These axes tended to separate the Raisin River, a mostly agricultural basin, from the Huron River, with its more even mix of agricultural, urban, and natural land covers. The third-fifth axes included measures of patch interspersion and metrics specific to a given land cover type. Overall, the landscape analysis successfully explained much of the variation across the sub-catchments, identifying axes and metrics often consistent with prior studies investigating much larger landscape units. Differences between this and prior work could be due to differences in spatial scale, number of land cover types and the original variable pool. We next performed a similar factor analysis of in-stream habitat measures that best depicted variation from some 50 streams of the 109 sub-catchments, and then used multiple linear regression to predict habitat condition from landscape condition. Channel shape and substrate were best predicted by subcatchment-scale geology and land cover/use. These landscape measures also successfully predicted the visually-assessed habitat metrics, as did the land cover/use within the riparian corridor and pattern metrics at the subcatchment scale. While landscape pattern analysis provided useful insight into the main gradients that characterize these two watersheds, it was of limited additional value in predicting stream habitat condition.

KEY WORDS: landscape pattern, habitat quality, stream, river