Biologists Meet Development: Leveraging Multimodal Data for Ecological Conservation and Community Benefit

Gabriel Englander (Development Economist at the World Bank Research Group) and Sean Luna McAdams (Sr. Program Manager at CEGA) summarize insights from CEGA’s recent Measuring Development (MeasureDev) conference, co-hosted with the World Bank’s Development Impact department (DECDI), Data Analytics and Tools Unit (DECAT), the Development Research Group (DECRG), and the Institute for Economic Development in Washington, D.C. MeasureDev 2025 showcased efforts that combine cutting-edge computational approaches, biology, and social science to better understand the vital relationship between our natural environment and human prosperity.

Co-hosted by CEGA and the World Bank, the eleventh annual Measuring Development Conference: Biodiversity on Land and at Sea focused on the use of biodiversity panel data and causal inference to improve the effectiveness of wildlife conservation efforts and policy. Three insights emerged over the course of the day that together illustrate how promising research collaborations with governments, the private sector, and NGOs can have significant positive impacts on natural ecosystems and the communities that depend on them.

Data Disparities at Land and Sea

It’s easier to observe animals and plants on land than in the ocean, so the conference had twice as much terrestrial research as marine. There’s no marine equivalent of the North American Breeding Bird Survey, which contains decades of consistently measured annual counts of bird species at a 0.5 km² resolution along survey routes. Harnessing new data sources to produce publicly-available, research-grade panel data — like tracking and predicting bat movements using weather radar — is farther behind at sea.

Instead, Global Fishing Watch (GFW) is “measuring the hunters” — counting fishing vessels and their activities — with remotely-sensed data from vessel transponders, satellite radar, and optical imagery. David Kroodsma, Chief Scientist at GFW, shared their ongoing work on the Open Ocean Project. “At its heart, it’s really about combining the GPS data with imagery to map as much and say as much as we can about what people are doing [at sea],” said Kroodsma. “And this includes not only fishing… but also wind, oil, sandmining, aquaculture, and shipping.”

In the coming years, we expect to see a much richer information environment on marine ecosystems. Marine biologists and conservation professionals can exploit recent advances in autonomous, decentralized sensors to begin filling the tremendous data gap in direct observation of animal populations and activities, supplementing GFW’s efforts on human economic activities at sea.

New Data, Faster, with AI Tools

The analog method for measuring biodiversity is “transect surveys.” On land, a person walks along a line and records the species within a fixed distance from the line. At sea, a vessel drags a net in a straight line and the crew records the number or weight of species caught. In both, scientists extrapolate the sampled data over the survey region to produce population estimates.

Transect surveys are the traditional gold standard measurement tool because they apply a known, constant methodology. “Citizen science” data sources and data aggregators often don’t contain enough information about how the data was collected. Rigorous research typically cannot proceed without this metadata. Are there really more sloths in this location-time, or just more survey effort? Camera traps and acoustic monitoring offer a lower-cost alternative to terrestrial transect surveys. Like the surveys, cameras and microphones can be placed to enable representative population estimates of a study site.

Ten years ago, only humans classified the animals photographed by camera traps or the bird songs recorded by microphones. Machine learning classification models make this task dramatically easier. Today, scientists can collect biodiversity outcome data using camera traps and, with human oversight, process the images with an “AI” tool like MegaDetector. For example, Jay Qi from DrivenData discussed how Zamba can be used to automate some of this pipeline. Electronic monitoring — cameras that some governments require on their fishing vessels — could be similarly used to measure catches of marine species.

Open Data versus Indigenous Data Sovereignty?

Open, publicly-available data enables research by making data access inexpensive. It also builds trust in research by facilitating the replication and verification of published analyses. If researchers had access to more public, high-quality panel data, they would be able to devote more time to conducting rigorous research about what works for global conservation efforts. Despite the collective benefit such data provide, many governments, NGOs, and research organizations limit access. The lack of incentives for individual researchers and institutions to pool resources results in its systematic underprovision and duplication of efforts.

Lily Xu, assistant professor at Columbia University, emphasized the gains possible simply by data sharing and open data efforts. In an example from East Africa, Xu explained that dozens of NGOs and ecologists work independently to understand the health of lion communities. “Frustratingly, so few of them actually share data, that there are actually lions in national parks who have two GPS collars because different organizations refuse to share data,” said Xu. “So, yes, there’s a shortage of data and we need to collect more of it — but we also need to de-duplicate efforts.”

Emilio Tripp and Jessica Camarena of the Karuk Tribe Wildlife Program shared concerns about opening tribal lands and data to unregulated research. Instead, they collaborate closely with researchers to ensure collective benefit and control access to data. Given the centuries of violence against Indigenous peoples, trusting relationships with non-Indigenous researchers are paramount. McKalee Steen, a citizen of the Cherokee Nation, and a PhD Candidate at UC Berkeley’s Environmental Science, Policy, and Management Department, highlighted how addressing capacity constraints can unlock more productive collaborations. “Any of the work [to] engage Tribes should build Tribal capacity [to use] these tools,” said Steen. Building capacity to integrate novel Privacy Enhancing Technologies (PETs), in particular, could address both researchers’ desire for open data and honor communities’ rights to privacy and data sovereignty.

Building trusting relationships with partners offers an additional benefit: it increases the likelihood that the evidence produced supports decision-making. Indigenous scientists emphasized that real empowerment requires giving them ownership over the data. Researchers often overlook this step because their focus is on publication. Genuine co-production ensures that research aligns with the evidence needs of local collaborators and is useful for policy and practice. Indeed, this approach is generalizable beyond biodiversity research.

If you were unable to attend, you can watch the conference presentation here and visit the MeasureDev 2025 event page to explore select slide decks. We particularly recommend the two keynotes, Laura Pollock and David Kroodsma, who survey the existing and emerging biodiversity measurement universe in an accessible and inspiring manner. You can also watch MeasureDev attendees reflect on the conference’s insights in the above video.