Environmental Engineering Reference
In-Depth Information
long-eared owl (
Mustela nivalis, Asio otus
), with respect to their different ecologi-
cally relevant properties (e.g. territoriality, food specialization, migration beha-
viour). Simulations take place on a grid map, with an extent of 150 ha and a spatial
resolution of 30
30 m. The resources for the rodents are calculated for each grid
cell and exhibit seasonal dynamics and allow for a feedback process with exploita-
tion by rodents.
The model represented dynamics and interactions on different integration levels
including individual life history traits, population development and community
interactions. The individual level included e.g. ontology, reproduction sequences,
weight development, habitat use and interaction with other rodents and predators.
As a result of the individual interactions cyclic population dynamics emerged, as
they are typical for Scandinavian rodent communities with an average cycle length
slightly below 4 years. These cyclic population level dynamics were not imple-
mented in the programme specification but are emergent properties produced by
the model components during execution (Breckling et al. 2005; Reuter et al.
2005, 2008).
The model also allows analyzing the population structure with respect to
age structure, reproduction rates and mortalities for different phases of the cycle.
Specialized predators shared the cycles frequency with the rodents but the phase
lags behind. The factors that cause this sudden decline of rodents are believed to be
of crucial importance for the whole system dynamics (Batzli 1996).
By analyzing mortalities in the phase between maximum abundance and the
following minimum, the model gave new insights into the driving forces: The
overall model results showed that mortalities due to intrinsic factors like senes-
cence did not increase distinctly. In contrast, trophic interactions that are based on
the lack of food and predation pressure contributed to about 90% of mortality of
rodent individuals in this phase. The results clearly emphasize that food web
interactions constitute the essential driving force of the cyclic dynamics. Further
investigations however yielded another surprise: The trophic control and its
relative strength, which was calculated as the difference between normalized
mortality rates in relation to either of the two factors, varies unpredictably in
the model (Fig.
12.5
) and cannot be correlated to specific properties of the
respective cycle.
With respect to the identification of these two factors that have a varying impact,
the model results give important new conclusions for the discussion of the driving
forces in Scandinavian rodent cycles and may thus help to explain the differing
results of numerous empirical investigations.
12.5 Conclusions
The relation of modelling on the one side and empirical information and biological
knowledge on the other is different for IBM than it is for other modelling
approaches. While the modelling approaches usually employ a specific abstraction
