20~
~ C;rid Based Modelling
8.1
Introduction
It is rarely possible in ecology to work empirically at the scales which are relevant
to the patterns and processes studied. Instead of decades and square kilometres,
field studies generally cover only square metres over a period of one to three years
(Kareiva & Anderson 1988). This is also the case for studies of macrozoobenthic
species in tidal flats. Additional problems arise in tidal flats: most organisms living
here are hard to investigate because they live in the sediment and because of their
high mobility. Moreover, frequently occurring disturbance events strongly affect
distribution and abundance, and therefore the significance of short-term studies,
which are no more than snapshots of spatial distribution, is doubtful. The problem
of how to extrapolate small-scale studies to larger scales is thus decisive for advancing the benthic ecology of soft-bottom communities (Hall et al. 1994).
A tool to this end might be ecological models which explicitly take space into
account - especially "grid-based" models, which are powerful tools (Grimm &
Jeltsch 1996). The basic idea underlying grid-based models is to divide space into
many small grid "cells" and to ignore spatial effects within the grid cells, but not
among them. Empirical ecological knowledge usually only exists at the local scale
within the grid cells: some idea exists of how the ecological state of the cell will
change under normal conditions, how it will react to disturbance events, and how it
interacts with the neighbouring cells. This local empirical knowledge can easily be
taken into account in simulation programmes by using simple "if-then" rules. The
programme then calculates how the whole region modelled will change. Chance
events on both the local and regional scale are taken into account by using random
numbers or random distributions to determine their occurrence and impact.
In terrestrial ecology, the grid-based approach has been used for different scientific and applied studies (e.g., Turner et al. 1995; Wennergren et al. 1995). In
marine ecology, however, the grid-based approach has so far only been occasionally applied. In particular, as far as questions which address the spatial and temporal dynamics of macrozoobenthos in tidal tlats are concerned, to our knowledge no
grid-based models exist.
In this chapter we shall try to apply the grid-based approach to the regional distribution of macrozoobenthos in the intertidal of the Wadden Sea. We address the
question of whether and how succession, extreme disturbance events and mutualistic interactions determine the distribution of macrozoobenthic species. This
question - that of spatial and temporal patterns and their causes (Chap. 2) - was
central to the ELA W AT project. We will also use the grid-based model to test
ecological concepts which in a spatially explicit manner refer to succession and
disturbance events, and which were considered as possible approaches to explain
the distribution of the macrozoobenthos. These are the concept of "patchdynamics" (Pickett & White 1985, Jax et al. 1993) and the theory of "mosaiccycles" (Remmert 1991; Reise 1991).
~ C;rid Based Modelling
8.1
Introduction
It is rarely possible in ecology to work empirically at the scales which are relevant
to the patterns and processes studied. Instead of decades and square kilometres,
field studies generally cover only square metres over a period of one to three years
(Kareiva & Anderson 1988). This is also the case for studies of macrozoobenthic
species in tidal flats. Additional problems arise in tidal flats: most organisms living
here are hard to investigate because they live in the sediment and because of their
high mobility. Moreover, frequently occurring disturbance events strongly affect
distribution and abundance, and therefore the significance of short-term studies,
which are no more than snapshots of spatial distribution, is doubtful. The problem
of how to extrapolate small-scale studies to larger scales is thus decisive for advancing the benthic ecology of soft-bottom communities (Hall et al. 1994).
A tool to this end might be ecological models which explicitly take space into
account - especially "grid-based" models, which are powerful tools (Grimm &
Jeltsch 1996). The basic idea underlying grid-based models is to divide space into
many small grid "cells" and to ignore spatial effects within the grid cells, but not
among them. Empirical ecological knowledge usually only exists at the local scale
within the grid cells: some idea exists of how the ecological state of the cell will
change under normal conditions, how it will react to disturbance events, and how it
interacts with the neighbouring cells. This local empirical knowledge can easily be
taken into account in simulation programmes by using simple "if-then" rules. The
programme then calculates how the whole region modelled will change. Chance
events on both the local and regional scale are taken into account by using random
numbers or random distributions to determine their occurrence and impact.
In terrestrial ecology, the grid-based approach has been used for different scientific and applied studies (e.g., Turner et al. 1995; Wennergren et al. 1995). In
marine ecology, however, the grid-based approach has so far only been occasionally applied. In particular, as far as questions which address the spatial and temporal dynamics of macrozoobenthos in tidal tlats are concerned, to our knowledge no
grid-based models exist.
In this chapter we shall try to apply the grid-based approach to the regional distribution of macrozoobenthos in the intertidal of the Wadden Sea. We address the
question of whether and how succession, extreme disturbance events and mutualistic interactions determine the distribution of macrozoobenthic species. This
question - that of spatial and temporal patterns and their causes (Chap. 2) - was
central to the ELA W AT project. We will also use the grid-based model to test
ecological concepts which in a spatially explicit manner refer to succession and
disturbance events, and which were considered as possible approaches to explain
the distribution of the macrozoobenthos. These are the concept of "patchdynamics" (Pickett & White 1985, Jax et al. 1993) and the theory of "mosaiccycles" (Remmert 1991; Reise 1991).