Ecological Archives A016-039-A3

Mary Ann Cunningham and Douglas H. Johnson. 2006. Proximate and landscape factors influence grassland bird distributions. Ecological Applications 16:1062–1075.

Appendix C. Regression coefficient estimates demonstrated the strength of tree cover effects across scales.

Regression coefficient estimates, with 95% confidence intervals, for tree cover at proximate scales (50 to 100 m) and landscape scales (200 to 1600 m). Coefficients are from the landscape-only models (e.g., year + tree200).

Regression coefficient estimates demonstrate the strength of tree cover effects across scales. Unlike AIC, the strengths of effects can be compared by the magnitude of parameter estimates, so the scale of vertical axes of graphs below is constant, except for a few species with large coefficients. Seven of the 19 species (Blue-winged Teal, Upland Sandpiper, Marbled Godwit, Marsh Wren, Savannah Sparrow, Grasshopper Sparrow, and Bobolink) had strongest responses at smallest scales (≤ 400 m). Five species (Wilson’s Phalarope, Clay-colored Sparrow, Field Sparrow, Vesper Sparrow, and Brown-headed Cowbird) had stronger responses at the largest scales (1200-1600 m). Five species (Killdeer, Wilson’s Snipe, Sedge Wren, Red-winged Blackbird, and Western Meadowlark) had relatively strong responses at both small and large scales, and two (Mallard and Common Yellowthroat) had negligible responses to tree cover at nearly all landscape scales. In general, scales that produced small ΔAIC values in landscape models (Appendix A) corresponded to the scales with large parameter estimates (below).



[Back to A016-039]