Table 5.

ANOVA for full and simplified regression models regarding efflux estimation at collar level (LOG(Efflux), μmol m-² s⁻¹) depending on predictive variables: DBH (cm), distance of the collar to the nearest tree (Dist, m) and air temperature (Tair, °C). n = 169 measurements. In bold significant coefficients.

FULL MODEL	Sum of Squares	Df	Mean Square	F-Snedecor	Sig. Level	R2	SEEa
Model	26.6547	6	4.44246	8.2	0.0000	20.4	0.74
Residual	87.8169	162	0.54208
Total (Corr.)	114.472	168
PREDICTIVE VARIABLES	Estimate	Standard Error	t-Student	Sig. Level	CFb
CONSTANT	8.4885	1.9533	4.3458	0.0000	1.3150
DBH	0.0008	0.0064	0.1299	0.8968
Dist	−0.0089	0.0090	−0.9984	0.3196
Tair	−0.2060	0.0594	−3.4678	0.0007
DBH*Dist	0.0000	0.0000	−0.8301	0.4077
Dist^2	0.0000	0.0001	0.6812	0.4967
Dist^3	0.0000	0.0000	−0.5808	0.5622
SIMPLIFIED MODEL	Sum of Squares	Df	Mean Square	F-Snedecor	Sig. Level	R2	SEEa
Model	24.8898	2	12.4449	23.06	0	20.8	0.73
Residual	89.5818	166	0.53965
Total (Corr.)	114.472	168
PREDICTIVE VARIABLES	Estimate	Standard Error	t-Student	Sig. Level	CFb
CONSTANT	8.4835	1.8490	4.5882	0.0000	1.3053
Dist	−0.0044	0.0008	−5.2780	0.0000
Tair	−0.2105	0.0587	−3.5831	0.0004

^aStandard error of the estimation (residual deviation, log-units).

^bSprugel (1983) correction factor (already included in the final model, Eq. (3)): CF = exp (SEE²/2).