Wednesday, August 5, 2009 - 8:00 AM

COS 59-1: Virtual prairie dogs weigh in on the Resource Dispersion Hypothesis

Christopher X. J. Jensen, Pratt Institute, Dylan Moore, Pratt Institute, and Jennifer L. Verdolin, Stony Brook University.


The Resource Dispersion Hypothesis (RDH) predicts that as resources become more patchy, food aggregations will become more easy to defend, leading to larger group sizes and larger territories. The RDH was first stated as a verbal model and later as an optimal behavior model. Rarely engaged experimentally, the predictions of the RDH were recently extended to include increases in individual home range overlap and tested in Gunnison’s prairie dog (Cynomys gunnisoni); in response to experimentally manipulated increases in food patchiness, prairie dogs decreased their territory size, increased their individual home range overlap but did not increase their group size. To find out why prairie dog behavior is not entirely consistent with RDH predictions, we have constructed an individual-based model realized in a custom software simulation dubbed fieldTest. This program allows the user to simulate the interaction of virtual prairie dogs simultaneously following identical behavioral rules on a landscape of specified resource richness and patchiness. The focal behavior of these virtual prairie dogs is the joining and abandoning of groups, which is based on individual resource satisfaction. In contrast to optimal behavior models, organisms in our model behave in response to spatially and temporally local information rather than global knowledge.


As with the experiment conducted on actual prairie dogs, our simulation of prairie dog behavior suggests that the predictions of the RDH are only conditionally valid. On average, virtual prairie dogs do form larger groups and tolerate increased overlap of home ranges as the resource heterogeneity increases. However, there is significant variation between replicate landscapes with differing distribution patterns but identical measures of heterogeneity, suggesting that idiosyncrasies of particular landscapes will have a strong effect on the results of actual field experiments. We also find that in order for the virtual prairie dogs to enjoy the full benefits of group membership without incurring costs associated with food sharing, behavioral thresholds for joining and abandoning groups must be fined-tuned within a fairly narrow range. How (and if) such thresholds evolve in actual organisms suggests a direction for novel research. During this presentation, we will display our results, a brief overview of how the fieldTest model works, and its potential for broader application. The result of collaboration with a computer scientist, fieldTest allows for immediate visualization of group behaviors. Although we considered prairie dogs, fieldTest also has the potential to suggest new experiments testing the RDH in other group-territorial animals.