Model code | Decision criteria | Correct (%) | AIC | Mean probability of mortality (SE)z | t(2),146 | P-value | |
---|---|---|---|---|---|---|---|
Live | Dead | ||||||
ModelA | Percent crown diebacky, DBHx, bark splitsw | 83.8 | 3.15 | 0.23 (0.03) | 0.73 (0.24) | −11.04 | < 0.001 |
ModelB | Vigor ratingv, DBH, bark splits | 86.5 | 1.30 | 0.22 (0.03) | 0.76 (0.20) | −13.22 | < 0.001 |
ModelC | Bark splits, DBH, woodpecker activityw | 75.7 | 2.85 | 0.29 (0.03) | 0.63 (0.23) | −7.26 | < 0.001 |
ModelDu | DBH | 67.6 | −1.32 | 0.44 (0.02) | 0.58 (0.17) | −4.62 | < 0.001 |
ModelE | Ten-year growth ratet | 82.9 | 0.53 | 0.04 (0.02) | 0.10 (0.00) | −1.14s | 0.261 |
ModelF | Five-year standard errort, five-year growth rater | 82.9 | 0.53 | 0.05 (0.02) | 0.00 (0.00) | −1.14s | 0.261 |
ModelGu | DBH | 67.6 | −1.32 | 0.44 (0.02) | 0.58 (0.17) | −4.62 | < 0.001 |
ModelHu | Bark splits, DBH, woodpecker activityw | 75.7 | 2.85 | 0.29 (0.03) | 0.63 (0.23) | −7.26 | < 0.001 |
↵z Calculated from the probability of mortality for each tree at the node terminus.
↵y Range = 5%–100%.
↵x Diameter at breast height (cm).
↵w Presence, absence (1,0).
↵v Categorical 1–5 (1 healthiest, 5 poorest health).
↵u ModelD, ModelG, and ModelH included other variables as inputs but only DBH met the pruning protocol.
↵t Mean ring basal area growth (mm2).
↵s df = 33.
↵r Mean ring basal area standard error.