A Big Database and Big Models Pave the Way for Big Questions in Ecology

This post was provided by Sean McMahon.

Sean is an Associate Editor for Methods in Ecology and Evolution and is a staff scientist at the Smithsonian Institution based at the Smithsonian Environmental Research Center.  His research focuses on the ecological mechanisms that structure forest communities, with interests spanning the fields of demography, physiology, and remote sensing.

The 100th anniversary of the Ecological Society of America was celebrated in Baltimore, Maryland at their Annual Conference in August. This year a record 10,000 ecologists attended the six day event. ESA conferences now boast a staggering number of scientific presentations, ranging from numerous plenary talks, organized oral sessions and regular oral presentation sessions to lightening talks, posters, workshops and mixers. It was both exhilarating and overwhelming, but featured a truly impressive amount of science.

As the sheer magnitude of the event made attending even a fraction of the talks impossible, it feels odd to highlight any particular presentations. Two talks, however – both on the final morning of the conference – did strike me as worth mentioning; not because they featured groundbreaking science, or novel insights, but because they reflect potentially powerful new platforms from which groundbreaking science might develop.

Organizing the herbarium of the new world

Brian Enquist of the University of Arizona, Tuscon opened the Friday organized oral session on “The macroecology of botanical diversity: History, new insights and the central informatics barriers” with an introduction to the Botanical Information and Ecology Network (BIEN), a workflow to standardize and integrate global plant collections and databases.

After thanking the organizers for the coveted Friday 8am slot, Brian demonstrated how the workflow, built by over 50 researchers, collects records of specimens and locations, traits, and genetic code, to build species range maps for approximately 100,000 species across the Americas. Members of BIEN have also created a standardized species list and multi-gene phylogeny for all New World species as well as continental-scale taxonomic, phylogenetic, diversity, and trait maps.

New products from BIEN are detailed diversity maps created using high performance computing through iPlant. These maps have been combined with phylogenetic information and reveal novel patterns of new world diversity. Further, these maps form the basis of a new app, called Plant-O-Matic (available in the iTunes store and elsewhere). This app, recently highlighted in Nature Plants, delivers to the user a list of all plants that are likely to be found in a 100km radius. This should not only offer the public a way to learn about plant distributions and relationships, but also encourage citizen scientists to contribute to the growing BIEN database.

Earth systems models for ecologist who refuse to learn Fortran

The final talk of the organized oral session “Creative Approaches for Addressing Ecological Uncertainty in Earth System Models” was given by Mike Dietze of Boston University (who appropriately finished his talk by declaring the 100th anniversary annual conference officially closed!). Mike presented The Predictive Ecosystem Analyzer (PEcAn), a toolbox for model-data ecoinformatics. He explained how PEcAn’s simple web-based interfaces are designed to bridge the gap between earth systems model developers and the potential users who are not necessarily playfully coding away in Fortran.

A workflow of PEcAn, where experiments and observational studies can improve the uncertainty in predictive models of vegetation dynamics. © Michael Dietze

A workflow of PEcAn, where experiments and observational studies can improve the uncertainty in predictive models of vegetation dynamics. © Michael Dietze

In fact, Mike sees PEcAn as a way to expand the range of scientists who use earth systems models. The user-friendly toolbox can help anyone who is collecting data that they think could, or should, inform earth system and ecosystem models to play an active role in the comparison between models and data. Experts in sub-disciplines and sub-regions can check how the community of models are doing for the processes and the places that they study. Mike envisions a future where scientists from many disciplines who might conclude a paper or proposal with some variant of “these results should inform models” are able to take the lead in actively implementing this promise.

PEcAn has an expanding menu of Earth Systems Models that it supports, each of which uses distinct approaches to answering similar questions about how climate, the physical world, and the world’s vegetation interact. Users can efficiently design model runs that project vegetation forward under different scenarios or evaluate alternative hypotheses about how ecosystems function. Mike also envisions analyses designed to characterize model uncertainties as playing a role in helping to focus data syntheses and new field observations and experiments on processes with the greatest uncertainty.

These two platforms, and others like them, are laying the groundwork for future breakthroughs in ecological studies. It is exciting to begin to learn about them and about their capabilities. Hopefully, both of these will make significant contributions in the coming years and I wish Mike, Brian and everyone connected with Plant-o-Matic, iPlant and PEcAn the best of luck.

Advertisements

One thought on “A Big Database and Big Models Pave the Way for Big Questions in Ecology

  1. Pingback: A Big Database and Big Models Pave the Way for Big Questions in Ecology | Enquist Lab

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s