On the Ecology of Mountainous Forests in a Changing Climate: A ...

202 Appendix
80% of the foliage weight obtained for group 2 (Tab. A-3). The relationships between
foliage weight and diameter as well as a comparison of the new parametrization with the
one used by Kienast (1987) are plotted in Fig. A-1.
Tab. A-3: Regressions of foliage wet weight (L) vs. diameter at breast height (D) for several species
groups as derived from the data in Burger (1945-53), where L = kA 1·D kA 2. Linear regressions were calculated
based on the transformation Ln(L) = Ln(kA 1 ) + kA 2·Ln(D). n denotes sample size.
Species kA 1 kA 2 r 2 group no. n
Abies alba, Picea excelsa, Pinus cembra, P. montana 0.23 1.56 0.93 5 130
P. silvestris 0.17 1.40 0.76 4 210
Fagus silvatica, Quercus spp. 0.06 1.70 0.93 3 144
Larix decidua 0.10 1.43 0.87 2 99
Betula sp., Salix sp. (no data) 0.08 1.43 – 1 –
L (kg)
Picea: foliage data
gFolW (kg)
40.00
300
35.00
30.00
25.00
1
2
200
20.00
3
15.00
10.00
5.00
0.00
0 20 40 60 80
D (cm)
4
5
100
0
0
20
40
60
80
Kienast’s formula
regression, group 5
measurement
100 120 D (cm)
Fig. A-1: Left: Relationship between foliage dry weight (gFolW) and diameter at breast
height (D) of the species groups defined in Tab. A-3. Right: Measured foliage fresh weight
of Picea excelsa (dots), the FORECE parametrization (Kienast 1987) and the parametrization
for group 5 derived from the original Burger data (Tab. A-3).
Grouping of the species
To assign every species to one of the five groups defined above (Tab. A-3), the ranking
of the species with respect to their capability of casting shade from Ellenberg (1986) and
the values of the parameters A1 and A2 used by Kienast (1987) were consulted (Tab. A-
4). These data sources allowed to rank the 30 tree species with respect to the sType parameter.
kDm, kHm & kAm parameters
Probably the best way to determine these three parameters would be based on diameter,
height, and age distributions of trees in old-growth forests that are not subject to strong
environmental stress. A theoretical distribution could be fitted to such data, e.g. an exponential
distribution in the case of maximum age (cf. Eq. 3.29f. in section 3.3.1), and the
parameters could be estimated from these distributions. However, data from old-growth
European forests is scarce (e.g. Leibundgut 1993). Moreover, it would be difficult to obtain
these data for many tree species, let alone for all the 30 species used under European
conditions in FORCLIM. Therefore, a simpler and necessarily less accurate approach had
to be adopted to derive an estimate of these parameters.