干旱区生态环境知识资源中心

Question: Can population-level patterns be used to model the species-time relationship? Which non-random patterns in population time-series are necessary for modelling the species-time relationship?

Statistical modelling methods: The presence of aggregation, autocorrelation, and interspecific association was determined using Morisita’s I,, Moran’s I, and Ive’s C respectively. Models for the species-time relationship were constructed from these sub-patterns using a combination of analytical models and randomization methods.

Data studied: Observational time-series of rodents and annual plants in the Chihuahuan Desert.

Conclusions: Aggregation was observed in the majority of population time-series. Most rodent species, but fewer than 10% of plant species, exhibited significant temporal autocorrelation in abundance. Models that included temporal autocorrelation as well as aggregation provided the best fit to the species-time relationship. The species-time relationship is intimately connected to the population dynamics of individual species. Models that attempt to connect the apparently general behaviour of the species-time relationship to the complex dynamics of populations are important for understanding the dynamics of ecological communities.