Edge effects in computer models of spatial competition. J.W. Haefner, G.C. Poole, P.V. Dunn, and R.T. Decker. Ecological Modelling 56: 221-244. 1991.
Spatially explicit models of biological growth processes make
assumptions about the interactions near the edges of the finite plots
modeled. The most realistic approach is to embed the plot of interest
in a larger space, but this is computationally costly and only moves
the edge further away. To determine the bias introduced by using
computationally efficient but less realistic methods, we used Monte
Carlo simulation to compare four approaches to modeling the effects of
boundaries on sessile organisms growing into asymmetric shapes under
crowed conditions in finite regions. The four algorithms compared were:
(a) embedding the modeled area in a space larger than the plot, (b)
reflecting the area over the boundaries, (c) mapping the area onto a
torus (translation), and (d) linearly expanding (extrapolating) the
proportion of occupied area outside from that occupied inside. We
studied the behavior of each algorithm in different ecological
situations that varied the intensity of interactions and maximum sizes
of individuals.
|
Publications 
Download Haefner et al. 1991