| int | x1 | x2 | x3 | x4 | k | log() | AIC | Δi | wi | R2 |
---|
| −0.61 | 0.02 | −0.04 | 0.05 | | 5 | 141.68 | −272.17 | 0.00 | 0.47 | 0.35 |
−0.61 | | −0.04 | 0.05 | | 4 | 140.01 | −271.23 | 0.94 | 0.29 | 0.31 |
−0.61 | 0.02 | −0.04 | 0.05 | 0.00 | 6 | 141.69 | − 269.67 | 2.50 | 0.13 | 0.35 |
−0.60 | | − 0.04 | 0.06 | 0.01 | 5 | 140.06 | −268.92 | 3.25 | 0.09 | 0.31 |
SW | 1.00 | 0.61 | 1.00 | 1.00 | 0.23 | |
β | −0.61 | 0.02 | −0.04 | 0.05 | 0.00 |
SE | 0.01 | 0.01 | 0.01 | 0.01 | 0.03 |
- Model averaging was carried out with a 95% confidence subset of models. For each model of the subset, we reported parameter estimates, total number of estimable parameters (k), the log-likelihood log(), AIC criterion, Δi = AICi – minAIC, Akaike weight (wi), and adjusted R2. Models are ordered in terms of Δi for AIC. At the bottom of the table, we reported model-averaged estimates β with their standard errors (SE) and their sum of weights (SW), for the four variables (quantitative variables: x1 - DW in tree, x2 - DW on ground, x3 - DW Removed; categorical variable: x4 - land management type (protected area/grazed woodlands-intercept-))