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

6 Chapter 1
ecological factors and emphasizing those aspects of forest ecosystems that are relevant for
managers, such as stand structure and wood volume. Most of the models neglect climatic
effects completely or treat them only marginally. Hence their application to study climatic
change appears to be questionable.
Individual-based models: Yield tables commonly used in forest management are a
prominent type of static single tree models for monospecific stands (e.g. Schober 1987).
Bossel et al. (1985) and Bossel (1987) developed the dynamic model SPRUCE to simulate
the effects of air pollution on tree growth; a disadvantage is that SPRUCE was restricted to
single species stands. Bossel et al. (1991) developed a similar model for tropical forests
that explicitly simulates every tree in five distinct canopy layers; yet it still does not allow
for changes of species composition. Single tree models that were built to simulate mixed
species stands include the classic matrix model by Horn (1975a,b), which was used to
project the species composition of the Hubbard Brook Experimental forest in New
Hampshire from simple field measurements. The development of mixed-species, mixedage
stands as a function of their environment was simulated with a very detailed spatial
model called FOREST (Ek & Monserud 1974). The size and location of each tree were
kept track of; thus shading and competition could be modelled realistically. This detail
made simulation studies extremely tedious, but it did not offer clear advantages over nonspatial
models (cf. Shugart 1984). Simpler approaches that also consider tree position explicitly
include the geometric models of Galitsky (1990) and Faber (1991). Their main
emphasis was to investigate the mechanisms underlying competition for space and not to
simulate realistic forest dynamics.
Another type of individual-based forest models was introduced by Siccama et al. (1969).
Based on the theory of gap phase replacement described by Watt (1925, 1947), they developed
a stochastic succession model of the Hubbard Brook forest. The model simulates
the establishment, growth, and mortality of trees on small patches, a patch being the area
that can be dominated by a large canopy tree. Within a patch, the location of a tree thus
could be neglected, which avoided the need to use a distance-dependent approach. Botkin
et al. (1970, 1972a,b) presented JABOWA, the prototype of these “forest gap models”.
The models include many biotic and abiotic influences on establishment, growth, and
mortality of trees. These three processes operate on different spatial and temporal scales;
forest gap models couple them explicitly and allow to study their effects on long-term
forest dynamics (Shugart & Urban 1989). Moreover, the models integrate processes on
different organizational levels, such as the growth of individual trees, competition of tree
populations at the patch level, and ecosystem characteristics at the scale of many patches.