Geoscience Reference
In-Depth Information
Simulating the expansion of agriculture in Europe during the Neolithic, Parisi
et al.
[PAR 08] developed a more sophisticated model incorporating a process of
demographic growth and a diffusion process in which the decision to migrate
depends on the state of neighboring cells. The cells are characterized by a set of
parameters describing their agricultural potential, which determines their carrying
capacity. The initial situation is assumed to be 9,000 BP and only one cell (assumed
to correspond to South-Western Anatolia) is occupied by a population of farmers.
When the number of farmers exceeds a certain threshold, part of the population
migrates to a neighboring cell, chosen at random but on the condition that it is empty
and that it has an agricultural potential. The rate of demographic growth in a cell is
proportional to the number of neighboring cells possessing an agricultural potential.
This model is able to reproduce the successive phases of the spreading of agriculture
in Europe, broadly speaking. Unlike the previous model in which the space was
homogeneous, the environmental constraints of Europe in this case are introduced
through the agricultural potential. The authors have then traced the genealogy of the
population groups and have been able to put it in parallel with the different language
groups in Europe. In this example, the geography of Europe intervenes at the input
of the model (through the parameters of agricultural potential) while the linguistic
areas emerge, identified from the genealogy of the groups, the oldest separations
being interpreted as corresponding to the strongest linguistic differentiations. The
interpretation is done from the time elapsed since the different “clusters” of
separation.
To make a cell change state according to the states of the neighboring cells is the
basic rule of CA. Parisi
et al's
model results in fact of the coupling of a logistic
growth model (see Chapter 2, section 2.1) and a cellular automaton where the cells
are characterized, on the one hand, by their agricultural potential and, on the other
hand, by their human occupation. This type of model has experienced a large
development during the past 20 years in geography, notably in the modeling of the
transformations of urban space.
4.3.2.
Cellular automata approaches: the case of land use changes
CA is actually considered by many geographers as a geographical tool
by
essence
, due to the analogy between the geographical space and the automaton
representation in two dimensions with a grid. Hence, Tobler [TOB 79] mentioned a
“cellular geography” and Couclelis [COU 85] mentioned a “cellular world” which
allows us to reflect on and model the links between the micro- and
macrogeographical dynamics. Thus, space is conceptualized in the form of a grid
and each grid cell is represented by a state (living or dead in the “game of life”,
agricultural potential in the example mentioned above, and type of land used in
