Tue, Aug 16, 2022: 8:00 AM-8:15 AM
513C
Background/Question/MethodsThe controversy about whether ecological communities are random assemblages of species or rather well-structured species associations goes back to Gleason and Clements in the early years of ecology. Species tolerating the same environmental conditions can potentially colonize and thrive in the same habitat. This is central to species distribution models and, therefore, to future projections of biodiversity loss and range shifts under climate change. Recent research efforts incorporate species inherent interactions and dependencies in the framework of species distribution models, which, initially, were based on the species-independent assumption. However, can we reliably quantify the propensity of two species to co-occur together in the first place? In this work, we address the issue by introducing a sociability index that aims to characterize the affinity of any given species to co-occur in association with others. This is relevant because it leads to forming a quantitative criterion to classify species in a community into those that only respond to the environment and behave effectively independently from each other, from those that, additionally, are constrained or favored by the presence of others.
Results/ConclusionsOur main result is that the question about quantifying the propensity of any pair of species to co-occur together strongly depends on the spatial scale we consider. The index characterizing species sociability in a community is scale- and context-dependent, and not an intrinsic property of that species. This is not surprising, as most biodiversity descriptors are also scale-dependent. We study these scale relations through co-occurrence-frequency curves, a novel approach to represent the probability of a species to co-occur versus its overall frequency. These informative patterns naturally lead to the definition of our sociability index for every species in the data set. This index applies to presence/absence data, this is, species occurrences across a number of local sites. We develop these ideas by using both model-generated metacommunities and empirical data from two very different sources: littoral benthic organisms along the North-Mediterranean coast and tree communities from Barro Colorado Island. We calculate the expected co-occurrence-frequency curve under different model hypotheses. By comparing empirical and model generated curves, we may better interpret why particular species score significantly higher in their sociability indexes, or, in other words, what drives them to co-occur more often than what would be expected under the null model hypothesis.
Results/ConclusionsOur main result is that the question about quantifying the propensity of any pair of species to co-occur together strongly depends on the spatial scale we consider. The index characterizing species sociability in a community is scale- and context-dependent, and not an intrinsic property of that species. This is not surprising, as most biodiversity descriptors are also scale-dependent. We study these scale relations through co-occurrence-frequency curves, a novel approach to represent the probability of a species to co-occur versus its overall frequency. These informative patterns naturally lead to the definition of our sociability index for every species in the data set. This index applies to presence/absence data, this is, species occurrences across a number of local sites. We develop these ideas by using both model-generated metacommunities and empirical data from two very different sources: littoral benthic organisms along the North-Mediterranean coast and tree communities from Barro Colorado Island. We calculate the expected co-occurrence-frequency curve under different model hypotheses. By comparing empirical and model generated curves, we may better interpret why particular species score significantly higher in their sociability indexes, or, in other words, what drives them to co-occur more often than what would be expected under the null model hypothesis.