106
Miguel A. Zavala
8.2 Background
8.2.1 Models of Forest Dynamics
As in other sciences, mathematical modelling has played a major role in the
development of ecology (e.g. MacArthur 1972; Tilman 1982). In plant ecology, models of coexistence have been used primarily to explain or to predict
community pattern and dynamics in relation to observations carried out at
the population level (for a review see Czaran and Bartha 1992; Pacala 1997).
These models have typically been classified in two broad categories depending on whether competition among individuals is evaluated by a measure of
local population density (Horn 1975; Pacala and Silander 1985) or through
the effect of resources on plant performance (Tilman 1982; Shugart 1984).
For the purpose of this contribution I will refer to the latter type of forest
dynamics models - models that consider explicitly species' responses to
changes in resource availability (Shugart 1984; Pacala et al. 1996).
Mechanistic or resource-based models of forest dynamics can be broadly
defined as stochastic formulations that compute the performance of each
tree as a function of resource availability and keep track of the variations in
resource levels induced by supply rates and tree consumption (Shugart 1984;
Botkin 1993). The original JABOWA (Botkin et al. 1972) and FORET (Shugart
and West 1977) models have been modified and applied to a wide range of
forest types, such as subtropical forests (Shugart 1984), temperate forests
(Prentice and Leemans 1990) and chaparral (Malanson and O'Leary 1995).
Thise type of models considers individual trees that compete for resources
within a cell scaled to the size of a dominant tree's canopy. Within this cell,
resource availability is spatially constant and varies in time according to resource supply and tree interception (light) or tree uptake (water and nitrogen). Different species respond differentially to the resource levels based on
their growth functions (Shugart 1984) or both growth and mortality functions (Prentice and Leemans 1990). A recently developed model of forest dynamics has introduced important structural differences in regard to the
widely used JABOWA-FORET simulators. SORTIE (Pacala et al. 1996) is a
spatially explicit, calibrated model in which light availability is computed
separately for each individual and considers the shade cast by each neighboring tree and daily solar trajectory. Recruitment is also truly spatial and
depends on the spatial distribution of the parental trees and their reproductive potential. Analyses of this model have shown that spatial detail is critical
to adequately describe some aspects of forest dynamics such as standing
crop or species diversity (Pacala and Deutschman 1995). Likewise, SORTIE
has exemplified the need for developing models intimately calibrated to data
and simple enough to be predictive and biologically interpretable.
Individual-based simulators have the advantage of tracing patterns at the
stand level to individual level processes, hence allowing us to interpret
Miguel A. Zavala
8.2 Background
8.2.1 Models of Forest Dynamics
As in other sciences, mathematical modelling has played a major role in the
development of ecology (e.g. MacArthur 1972; Tilman 1982). In plant ecology, models of coexistence have been used primarily to explain or to predict
community pattern and dynamics in relation to observations carried out at
the population level (for a review see Czaran and Bartha 1992; Pacala 1997).
These models have typically been classified in two broad categories depending on whether competition among individuals is evaluated by a measure of
local population density (Horn 1975; Pacala and Silander 1985) or through
the effect of resources on plant performance (Tilman 1982; Shugart 1984).
For the purpose of this contribution I will refer to the latter type of forest
dynamics models - models that consider explicitly species' responses to
changes in resource availability (Shugart 1984; Pacala et al. 1996).
Mechanistic or resource-based models of forest dynamics can be broadly
defined as stochastic formulations that compute the performance of each
tree as a function of resource availability and keep track of the variations in
resource levels induced by supply rates and tree consumption (Shugart 1984;
Botkin 1993). The original JABOWA (Botkin et al. 1972) and FORET (Shugart
and West 1977) models have been modified and applied to a wide range of
forest types, such as subtropical forests (Shugart 1984), temperate forests
(Prentice and Leemans 1990) and chaparral (Malanson and O'Leary 1995).
Thise type of models considers individual trees that compete for resources
within a cell scaled to the size of a dominant tree's canopy. Within this cell,
resource availability is spatially constant and varies in time according to resource supply and tree interception (light) or tree uptake (water and nitrogen). Different species respond differentially to the resource levels based on
their growth functions (Shugart 1984) or both growth and mortality functions (Prentice and Leemans 1990). A recently developed model of forest dynamics has introduced important structural differences in regard to the
widely used JABOWA-FORET simulators. SORTIE (Pacala et al. 1996) is a
spatially explicit, calibrated model in which light availability is computed
separately for each individual and considers the shade cast by each neighboring tree and daily solar trajectory. Recruitment is also truly spatial and
depends on the spatial distribution of the parental trees and their reproductive potential. Analyses of this model have shown that spatial detail is critical
to adequately describe some aspects of forest dynamics such as standing
crop or species diversity (Pacala and Deutschman 1995). Likewise, SORTIE
has exemplified the need for developing models intimately calibrated to data
and simple enough to be predictive and biologically interpretable.
Individual-based simulators have the advantage of tracing patterns at the
stand level to individual level processes, hence allowing us to interpret