Table 1. Comparison of linear mixed-effects models for predicting lifespan of mountain pines.
| Model | Variables | AICc | ΔAICc | Akaike weight (%) |
|---|---|---|---|---|
| Model 32 | Early growth + EW + NS + elevation + slope | 16.8 | 0.00 | 59.9 |
| Model 16 | Early growth + EW + NS + slope | 19.1 | 2.31 | 18.9 |
| Model 24 | Early growth + EW + NS + elevation | 20.2 | 3.41 | 10.9 |
| Model 8 | Early growth + EW + NS | 22.2 | 5.44 | 3.9 |
| Model 28 | Early growth + EW + elevation + slope | 22.7 | 5.89 | 3.1 |
| Model 20 | Early growth + EW + elevation | 23.0 | 6.25 | 2.6 |
| Model 4 | Early growth + EW | 28.0 | 11.23 | 0.2 |
| Model 12 | Early growth + EW + slope | 28.4 | 11.57 | 0.2 |
| Model 30 | EW + NS + elevation + slope | 30.5 | 13.72 | 0.1 |
Shown are the predictor variables of the mixed-effects models (Eq 1), the AICc (Eq 2), the ΔAICc (AICc difference to model with lowest AICc) and the Akaike weight (Eq 3). The models were fitted to data from 160 trees in 20 plots using maximum likelihood (ML). Only models with an Akaike weight ≥ 0.1% are shown (9 of 32 models). Abbreviations of variables: Early growth = mean ring width over the first 50 years (mm yr-1); EW = east-west gradient; NS = north-south gradient; elevation = elevation (m a.s.l.); slope = slope steepness (°).