IBMs contain at least two types of entities: individuals and their environment. In the simplest case, there is only one type of individual, for example, one plant species, and a homogeneous environment in which the individuals are located. More complex IBMs consider more than one type of individual, for example, a plant species, a herbivore, and a predator, and a more complex environment, for example, a heterogeneous landscape consisting of different types of habitat.
The entities of the IBM are characterized by a set of state variables. Individuals might have one or more state variables, for example, numbers representing position, age, sex, size, stage, rank, condition, energy reserves, memory of suitable habitat, etc. It depends, as with any model, on the question addressed with the model which state variables are included, but the general guideline is to keep the representation of the individual as simple as possible. The environment often is represented as a grid of quadratic cells, which represent spatial units of the landscape and may be characterized by the state variables habitat/nonhabitat, cover, biomass, moisture, temperature, soil type, exposure to wind or predators, etc.
A further important design decision of an IBM is its spatial and temporal resolution and extent. The spatial resolution is determined by the size of the grid cells, which in ecology can range from square centimeters to square kilometers, and spatial extent is set by the number of grid cells, for example, 100 x 100 cells. The temporal resolution is given by the length of a time step, which often is a year, a week, a day, or even less than an hour. Different parts of a year may also be represented with different resolutions. The temporal extent is set by the time horizon considered, for example, 100 years.
Was this article helpful?