Recent growth changes in Western European forests are driven by climate warming and structured across tree species climatic habitats

Table 5 Mathematical formulation, significance and magnitude of the temporal changes in BAI estimated for the 8 species under study over 1980–2007

Environmental context	Species	Trend complexity	Significant F tests (%)	Relative BAI change 1980–2007 (%)	95% bootstrap bilateral confidence interval (%)	Average BAI 1980–1985 (m²/ha/year)	Absolute change in BAI 1980–2007 (m²/ha/year)
Mountain	P. abies	2	97	+42	[29; 54]	0.78	+0.33
Mountain	A. alba	1	99	+19	[10; 29]	0.68	+0.13
Generalist	P. sylvestris	1	7	−2	[−12; 9.4]	0.45	−0.01
Generalist	F. sylvatica	2	70	+5 (+13/−8)	[−6; 17]	0.51	+0.03 (+0.29/−0.18)
Temperate lowland	Q. petraea	3	86	−3 (+12/−15)	[−14; 9]	0.44	−0.01 (+0.25/−0.31)
Temperate lowland	Q. robur	1	14	+4	[−6; 13]	0.42	+0.02
Mediterranean	Q. pubescens	1	97	−17	[−25; −9]	0.26	−0,05
Mediterranean	P. halepensis	3	92	−24 (+19/−45)	[−34; −12]	0.37	−0.09 (+0.42/−1)

Trend complexity from 1 to 3 refers to the polynomial degree retained for calendar year effect (Eqs. 5 to 7). The proportions of significant F tests of nested models on the bootstrap fits are reported (section 4.1), as well as retained trend complexity (complexity 0 being in absence of calendar year effect in the nested model). Q robur and P sylvestris are the only species for which no trend is significant. The relative and absolute (applying the relative BAI change to the average initial species BAI, computed from the first 2/3 years depending on data availability) changes in BAI between 1980 and 2007 are indicated. Numbers in brackets refer to trends showing a maximum, for which we also indicated the relative increase and the following relative decrease (trend complexity 2 and 3)

ISSN: 1297-966X