Table 2.

Model details for the site average shift rate

Parameter	Estimate	Std. error	t value	Pr (>|t|)
Model 1
Intercept	11.24	2.46	4.58	<0.001
scale (Loss)	−5.80	2.93	−1.98	0.06
scale (Cover)	6.85	2.69	2.54	0.02
scale (T)	12.94	3.24	3.99	<0.001
scale (Loss) × scale (T)	15.98	4.52	3.53	0.001
scale (Loss) × scale (Cover)	5.84	3.61	1.62	0.11
Model 2*
Intercept	11.09	2.51	4.43	<0.001
scale (Loss)	−7.78	2.71	−2.87	0.007
scale (Cover)	5.87	2.68	2.19	0.03
scale (T)	11.61	3.20	3.63	<0.001
scale (Loss) × scale (T)	15.79	4.62	3.42	0.002
Model 3
Intercept	8.68	2.97	2.92	0.006
scale (Loss)	−3.36	3.25	−1.04	0.31
scale (Cover)	6.99	2.64	2.65	0.01
scale (T)	13.60	3.21	4.24	<0.001
Type_plant	2.88	4.62	0.62	0.54
scale (Loss) × scale (T)	18.16	4.76	3.82	<0.001
scale (Loss) × Type_plant	−9.75	4.44	−2.20	0.03

Details of the three best fitting models (Table 1) for the site average shift rate (n = 43). Predictor variables are scaled (cf. the scale() function in R) for comparison purposes

*See Supplementary Tables 1 and 2 for unscaled estimates and weighted coefficients of model 2, the most parsimonious model

T: baseline temperature, Loss: forest loss percentage, Cover: forest cover percentage, Type: taxa type (animal or plant)