Appendix B. Tables showing (1) final logistic regression models used to predict occurrence of river otter latrine sites o Yellowstone Lake and tributary streams; and (2) intercept coefficients for regressions of δ15N (‰) vs. fecal deposition rate (estimate ± SE) for plants growing on otter latrine sites compared with δ15N (‰; mean ± SE) of plants growing on non-latrine sites.
TABLE B1. Logistic regression models used to predict occurrence of river otter latrine sites on Yellowstone Lake (A) and tributary streams (B), summer 2002 and 2003. Lake shoreline model correctly classified 92.3% of non-latrine (n = 53), 77.4% latrine (n = 31), and 86.7% overall. Stream model correctly classified 83.7% non-latrine (n = 49), 68.6% latrine (n = 35), and 77.4% overall.
Variable |
Coefficient |
SE |
Significance |
A) | |||
Shading |
0.057 |
0.016 |
0.000 |
Alder |
0.662 |
0.345 |
0.055 |
Grasses |
0.830 |
0.294 |
0.005 |
Rock Formation |
5.100 |
1.477 |
0.001 |
Lagoon |
6.587 |
1.715 |
0.000 |
Tributary Stream |
3.320 |
1.837 |
0.071 |
B) | |||
Shading |
0.017 |
0.008 |
0.027 |
Other Understory |
-0.508 |
0.209 |
0.015 |
Forbs |
-0.641 |
0.219 |
0.003 |
Brush |
-0.430 |
0.214 |
0.045 |
Sand Substrate |
0.466 |
0.151 |
0.002 |
Large Rock Substrate |
0.391 |
0.249 |
0.117 |
TABLE B2. Intercept coefficients for regressions of δ15N (‰) vs. fecal deposition rate (estimate ± SE) for plants growing on otter latrine sites compared with δ15N (‰; mean ± SE) of plants growing on non-latrine sites.
Latrine Y intercept |
Non-latrine |
|||||||||
Sample |
n |
δ15N |
Lower 95% CI |
Upper 95% CI |
n |
δ15N |
Lower 95% CI |
Upper 95% CI |
P value |
|
Spruce |
23 |
-1.6 ± 0.9 |
-3.4 |
0.2 |
20 |
-2.2 ± 0.5 |
-3.2 |
-1.2 |
0.53 |
|
Whortleberry |
11 |
0.1 ± 0.8 |
-1.8 |
1.9 |
7 |
-0.0 ± 1.0 |
-2.4 |
2.3 |
0.93 |