Table 3.

Outputs from linear mixed models testing predictors of bud development rate for individual trees

	Coefficient	SE	t value
(A) Oak
Intercept	0.045	0.003	12.978
Altitude	0.001	0.001	0.948
dbh	0.001	0.001	1.388
Soil: sandd	0.002	0.002	0.863
Soil clayd	0.005	0.003	1.606
Habitat: 19th C plantationa	0.002	0.003	0.610
Habitat: regenerationa	0.002	0.003	0.622
Habitat: ancient seminata	0.004	0.002	1.611
Observer BPb	0.000	0.002	0.008
Observer SJCb	0.001	0.002	0.228
Observer SLb	0.010	0.002	4.903
Observer ZDb	0.004	0.003	1.315
Yearc	0.015	0.002	8.710
(C) Beech
Intercept	0.045	0.005	9.075
Altitude	0.003	0.003	1.013
dbh	−0.006	0.002	−3.928
Soil: sandd	−0.005	0.004	−1.386
Soil clayd	0.010	0.007	1.559
Habitat: 19th C plantationa	0.011	0.004	2.412
Habitat: regenerationa	0.014	0.005	2.743
Habitat: ancient seminata	0.006	0.004	1.615
Observer BPb	0.006	0.005	1.233
Observer SJCb	0.010	0.004	2.329
Observer SLb	0.009	0.004	2.036
Observer ZDb	0.023	0.006	3.643
Yearc	−0.021	0.003	−6.018
(E) Sycamore
Intercept	0.053	0.005	11.423
Temperature	0.003	0.002	1.305
dbh	0.002	0.001	1.914
Soil: sandd	0.001	0.004	0.288
Soil clayd	0.007	0.005	1.349
Habitat: 19th C plantationb	0.004	0.004	1.113
Habitat: regenerationa	0.006	0.004	1.479
Habitat: ancient seminata	0.000	0.004	0.025
Observer BPa	0.011	0.004	2.554
Observer SJCb	0.006	0.003	1.742
Observer SLb	0.008	0.004	2.025
Observer ZDb	0.001	0.005	0.246
Yearc	0.012	0.002	5.271

All models contained the random effect “tree ID” and fixed effects: tree size (dbh), soil type (corallian limestone, sand, clay), habitat type (twentieth‐century plantation, nineteenth‐century broadleaf plantation, secondary regeneration, and ancient seminatural woodland), observer, year, and either altitude or ambient spring temperature (see Statistical Methods). All models correct for spatial autocorrelation by accounting for the spatial location of each tree (see Statistical Methods for further details).

^aTwentieth‐century plantation set to zero.

^bObserver AH set to zero.

^c2013 set to zero.

^dCorallian limestone set to zero. Bold indicates the effects that are more than twice the standard error around the estimate.