MCA: New approaches to the detection of niche differentiation patterns in ecological communities

Biodiversity in nature can be puzzlingly high in the light of competition between species, which arguably should eventually result in a single winner. Coexistence mechanisms shape the dynamics of communities and ecosystems, and the services ecosystems provide. Uncovering what these mechanisms are has been a long-standing challenge in ecology, as the experiments that would reveal them are logistically challenging to carry out in real-world ecological communities.

For example, trees can live longer than researchers, so carrying out experiments long enough to assess competition among tree species is quite a challenge. Hence, one approach ecologists use to reveal what “niche differences” are allowing competing species to coexist is to look at patterns of which types of species are present together. This project will significantly advance this approach in ecology, by tailoring new modern machine learning tools from data science to detect the patterns in ecological communities.

Additionally, this study would result in the training of graduate students, including individuals from under-represented groups, and outreach to female middle school students.

Specifically, this project will develop tools that account for recent advances in the theoretical understanding of what patterns niche differentiation will create in communities. Historically, ecologists expected coexisting species to be distinctive in traits indicative of their strategies. Recently ecologists have realized that when more species are present than can coexist stably through their niche differences, additional coexistence is enhanced by similarity.

Hence, they have come to expect clusters of species in trait space, especially in highly diverse communities. Modern data science offers a host of potential approaches for detecting species clusters. However, it does not provide a one-size-fits-all approach.

This project will tailor cluster detection tools to the detection of clustering indicative of niche differences, and ground-truth them on simulated communities. It will involve training in data science of an ecologist who has contributed to recent theoretical developments in our understanding of competitive communities, and support initial collaborative work by the ecologist with a data scientist to build new pattern detection tools for ecologists.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.