Understanding causes and effects of ecosystem dynamics is much like finding needles in a haystack. Most often we try to simplify the processes, looking for predictors that affect mean values of a population, e.g. mean mortality risk as a function of the distance to a road. Such approaches, however, do not consider individual behavior and interactions with the environment from which the dynamics arise. Agent Based Models (ABMs) allow individuals to interact with each other and with the environment, giving new insight into population dynamics and ecosystem functioning. In a previously developed SDM we have predicted the distribution of the Tawny owl (Strix aluco) throughout Denmark at the scale of individual breeding territories. We speculated that some of the areas predicted to be occuued by the SDM were out of reach for the species due to behavioral traits and population dynamics and that our SDM predictions could be fine tuned by incorporating it into an ABM. We therefore built an ABM to simulate the distribution of the Tawny owl in Denmark, using the model predictions from the SDM to form a spatially explicit map of more or less suitable areas. The landscape consisted of patches (square grid cells) of 600 by 600 meters covering all parts of Denmark. Each patch was assigned a quality defined as the likelihood of a patch being a breeding territory for the tawny owl as predicted by the SDM. We allowed the individuals to perceive the landscape as a mosaic of more or less suitable territories, and disperse, settle and die according to their immediate surroundings (environment and presence of other individuals). Combining predictions from a SDM may simplify complex behavioral rules that would otherwise have to be incorporated in the dispersal functions in the ABM . We simply created a landscpe of more or less suitable patches and let the individuals be most likely to move to better patches. Our most conservative AGM predicted local extinctions of the Tawny owl in parts of Denmark where it is supposed to be present according to our SDM. Less conservative estimates of adult mortality allowed the population to persist at higher numbers and throughout a larger range, and only unrealistically high mortality and reproductive rates allowed the population to persist in all the areas the SDM had predicted to be occupied. Our results suggest that ABMs can be used to fine tune predictions of SDMs.

Username
Password