Variation of cork porosity along the stem in harvested cork oak (Quercus suber L.) trees

Table 3 Statistics of the mixed-effects linear models

Model	Model equation	α₀	α₁	α₂	σ_Treej(μ_j)	σ_Treej(μ'_j)	σβ_Treej	σβ'_Treej	σ_εij	AIC	BIC	LogLik
M0	CP_ij = α₀ + α₁ h_Hij + μ_Treej + ε_ij	12.99	−1.44**		3.06				3.05	904.8	917.1	−448.4
M1	CP_ij = α₀ + α₁ h_Hij + μ_Treej + β_Treej · h_Hij + ε_ij	12.98	−1.42**		3.36		0.29		3.04	908.2	926.7	−448.1
M2	CP_ij = α₀ + α₁ h_Hij + α₂ A_maxij + μ_Treej + ε_ij	9.01	−0.62*	0.09**	2.30				2.74	857.9	873.3	−423.9
M3	CP_ij = α₀ + α₁ h_Hij + μ’_Treej + ε_ij	12.98	−1.42**			3.03			3.04	904.6	917.0	−448.3
M4	CP_ij = α₀ + α₁ h_Hij + μ’_Treej + β’_Treej · h_Hij + ε_ij	12.97	−1.40**			3.32		0.29	3.02	908.0	926.5	−448.0
M5	CP_ij = α₀ + α₁ h_Hij + α₂ D_hij + μ’_Treej + ε_ij	13.60	−1.50**	−0.01*		3.03			3.04	906.5	921.9	−448.2

**Significant at the 1% level
*Significant at the 5% level
The response variable was the porosity coefficient (CP). Models M0, M1 and M2 considered the random effects of tree size (i.e. stem diameter at breast height, D_bh). In M2, the fixed part of the model included the height in the stem and the size of lenticular channels (i.e. A_max). Models M3, M4 and M5 considered the random effects of trees’ intrinsic porosity (i.e. the porosity coefficient at breast height, CP_bh). In M5, the fixed part of the model included the height in the stem and the stem diameter over cork, D_h

ISSN: 1297-966X