Table 4.

Predictor variables for fine root mass identified by stepwise multiple regression analyses (forward and backward variable selection) grouped by study site and soil depth.

Model variables	Estimate	se	t value	p value
BEST MODEL FIT: BOTH STUDY SITES, ALL SOIL DEPTHS
(AIC = −92.75, F = 11.46, df = 3, 58, p < 0.001, R2 = 0.37)
(Intercept)	7.23	1.24 10−1	58.44	***
Na	2.49 10−5	1.23 10−5	2.02	*
mineral N	1.31 10−2	4.12 10−3	−3.18	**
Shannon-Wiener index, H	1.01	2.82 10−1	3.57	***
BEST MODEL FIT: SPIEKEROOG, ALL SOIL DEPTHS
(AIC = −79.18, F = 9.68, df = 2, 33, p < 0.001, R2 = 0.37)
(Intercept)	7.23	0.11	68.51	***
Shannon-Wiener index, H	0.80	0.20	4.03	***
aboveground biomass	0.00	0.00	1.62	n.s.
BEST MODEL FIT: WESTERHEVER, ALL SOIL DEPTHS
(AIC = −27.12, F = 6.10, df = 4, 21, p < 0.01, R2 = 0.54)
(Intercept)	6.15	3.43 10−1	17.92	***
Na	8.81 10−5	3.68 10−5	2.39	*
Shannon-Wiener index, H	1.80	7.53 10−1	2.39	*
aboveground biomass	2.93 10−4	1.39 10−4	2.11	*
total S	−2.2 10−4	1.43 10−4	−1.54	n.s.
BEST MODEL FIT: BOTH STUDY SITES, Ah Horizon (AIC = −48.39,
F = 6.90, df = 3, 30, p < 0.01, R2 = 0.41)
(Intercept)	7.36	2.40 10−1	30.60	***
mineral N	−1.38 10−2	5.88 10−3	−2.35	*
Shannon-Wiener index, H	1.05	3.80 10−1	2.76	**
aboveground biomass	−1.69 10−4	9.00 10−5	−1.79	n.s.
BEST MODEL FIT: BOTH STUDY SITES, Go/Gr Horizon (AIC = −13.29,
F = 4.05, df = 3, 24, p < 0.05, R2 = 0.34)
(Intercept)	5.05	0.28	18.343	***
Na	0.29 10−3	0.10 10−3	2.80	**
total P	−1.85 10−3	0.60 10−3	−3.06	**
aboveground biomass	0.32 10−3	0.24 10−3	1.34	n.s.

n.s. p > 0.05,

^*p < 0.05,

^**p < 0.01,

^***p < 0.001.