SPATIAL-ANALYSIS OF SANDHILL CRANE NESTING HABITAT

Ecologists need to understand how species' habitat requirements change across spatial scales, and how scale influences spatial analyses across heterogeneous landscapes. We used a geographic information system (GIS) to analyze nesting greater sandhill crane (Grus canadensis tabida) use of habitat at multiple spatial scales. We collected data on a 11,487-ha portion of Seney National Wildlife Refuge (NWR), Michigan, during 1984-87. Percent composition of 17 habitat variables (4 upland classes, 6 wetland classes, 6 water regimes, and total wetland) was compared around nest sites and random points for 5 circular buffers with radii of 50, 100, 200, 419, and 709 m. Cranes selected (P < 0.01) nest sites in or near seasonally flooded emergent (nonwoody) wetlands and avoided (P < 0.01) forested uplands. There was no (P > 0.01) habitat selection beyond 200 m from a nest. Beyond this distance our analysis was inconclusive, in part, because larger buffer scales increased heterogeneity and overlap among nest and random buffers. Observers should consider scale of analysis when investigating spatial patterns of use of habitat.