Broadening Our View of Earth’s Biodiversity Will Help Us Maintain Biosphere Integrity
Evan P. Economo
Okinawa Institute of Science and Technology
“This global biodiversity map will help us more effectively conserve the organisms that keep the world’s ecosystems running”
The biodiversity crisis is increasingly recognized as a major challenge for the planet as serious as climate change. In a time of dramatic global change, protecting the diversity and abundance of plants, animals, and other organisms is critical for biosphere integrity and resilience of the Earth system. In recent years, this rising concern has been reflected in high-profile meetings such as United Nations Biodiversity Conference (COP 15), new regulatory frameworks for corporations, and increasing engagement of the public.
As a biodiversity scientist, this newfound attention is encouraging, but also a little unsettling as we might not be fully ready to provide the answers society needs. In particular, we lack critical baseline information about what species exist, where they are found in the world, and what ecological roles they play. We are missing a comprehensive “map of life” akin to our global maps of Earth’s geophysical features. This makes it challenging to target our conservation resources in the most effective way because we can’t save species if we don’t know where they are. In the past few decades, we have made great progress toward resolving large-scale biodiversity patterns, but this progress has been highly skewed toward the species that get the most attention from both scientists and the public: vertebrates. We can now identify which geographic regions harbor the most threatened mammals, birds, and reptiles, and as consequence, vertebrate data guide the billions of dollars spent on conservation each year.
However, vertebrates represent only a small (albeit charismatic) branch of the tree of life, less than half a percent of the millions of species on Earth. And it is the understudied majority—worms, beetles, snails, mites, and other invertebrates—that keep Earth’s ecosystems running, and we don’t have a clear picture of how they are distributed around the globe. The scale of the challenge is immense; for example, we have millions of insects to document compared with tens of thousands of vertebrates, with far fewer people and resources devoted to their study.
Over a decade ago, my colleagues and I set out to address this gap in knowledge to resolve large-scale patterns of invertebrate diversity. As a test case, we chose an exemplar invertebrate group—ants (Photo 1). Ants might be unwelcome guests at your picnic; but they are dominant players in most terrestrial ecosystems, important for everything from nutrient flows to plant reproduction, and they impact the ecosystem services relevant for humans. More generally, ants are a surrogate for other less-studied insects and other invertebrates, and if we can figure out how to resolve a map of life for ants, we should be able do the same for those groups as well.
Photo 1. One of the lead authors of our nominated article, Prof. Benoit Guénard, collects ants with some young assistants. The nominated study synthesized data from 250 years of field collections such as this one, and the general public is increasingly contributing to the generation of biodiversity data.
Carl Linneus, the father of our modern scientific system of classifying life, described the first ant species in 1758. Since then, we have accumulated a great deal of information about ants and other organisms, but this knowledge is fragmented in obscure academic papers, museum collections, and databases, limiting our ability to see the big picture. Our first step was to consolidate these data on ant diversity into a single database; including reading more than 10,000 papers—an effort that took many years. Second, this fragmented, and error-prone raw material needed to be synthesized into a unified picture. For this, we turned to modern data science, informatics, and machine learning. We built computational pipelines to do everything from interpret the scribbles of a 19th century naturalist, to identifying mistakes, to modeling the regions in which each species can live based on their known climatic tolerances. A third problem is that even if we consolidate our current knowledge, our exploration of the world has been highly uneven; while many areas of the world are sampled intensively, others have barely been explored. To account for this, we used models that integrate the effects of a variety of geophysical and biotic variables to predict what we may be missing in underexplored areas, essentially “fixing” the holes in our map.
These machine learning models also allow us to identify hotspots of future discovery; areas that we predict will yield geographically restricted, vulnerable species that have not been documented yet. I like to think of this as a “treasure map” that can guide the next round of field exploration and discovery, providing a new link between the early explorers and modern machine learning-driven data science.
Our nominated paper presents the culmination of this decade-long effort. We present a global “map of life” for ants constituting the known distributions along with predictions for our knowledge gaps. This can be directly integrated into global conservation planning alongside vertebrates and inform efforts such as 30-by-30 (30% of land protected by 2030). Our analysis shows that for all groups of organisms, including ants, only 15-20% of critical biodiversity areas have any sort of protected status.
The cover image of the journal that accompanied our article depicted a copper engraving of the world from 1689 (Figure 1). At that time, our view of the continents was biased, skewed, and incomplete, pieced together from fragments of information and some educated guesses. But such maps provided a framework to keep exploring and refining our understanding of the world.
Figure 1. The cover of Science Advances depicts a copper engraving by Gerhard von Schagen from 1689. Such maps were imperfect, but facilitated further exploration and learning. This is akin where we are now for the biodiversity of less-studied groups such as ants.
This is where I think we are now with biodiversity. We are starting to resolve a broad view of life on Earth, and while this picture is skewed and imperfect, we can use it to keep learning and iterating. But we need to work faster than we have before, because using traditional approaches alone it will take centuries to resolve a map of life encompassing all animal groups, far too long given the pace of the biodiversity crisis. Our work and the work of others in our field show that we can shorten this timescale to decades by harnessing modern computational science to accelerate research. If this can be combined with a much-needed resource investment in fundamental discovery, we have a good chance of revealing Earth’s biodiversity in time to protect it.
