In 2025, the Amazon rainforest shrank by exactly 1.63 million hectares. But nobody can say how much wildlife vanished with it.

This illustrates one of the biggest challenges confronting scientists and others working to stem the loss of biodiversity. While satellites offer a sweeping birds-eye view of the planet, it can seem like ecologists tracking the ebb and flow of species in a place like the Amazon are studying a distant galaxy through a pinhole.

“If a tree has fallen almost anywhere in the world we know it almost in a matter of days. But I can tell you almost nothing about how biodiversity is faring below the canopy,” said Amy Rosenthal, who heads conservation initiatives at Planet Labs, a U.S. company that specializes in producing satellite imagery for various purposes.

But technology has reached the point where scientists and more than two dozen organizations including Planet are betting that they can pierce the forest canopy and create a near real-time view of trends in wildlife across the vast Amazon.

Last week, they announced the project Amazon.ia (the “i.a.” is an acronym for “inteligencia ambiental” or “environmental intelligence”). The plan calls for a network of between 10 and 30 sites across the Amazon, equipped with wildlife-tracking instruments. Their output can then be fed into computer models. Other sites could be added on a temporary basis. The end goal is a high-resolution picture of the state of biodiversity across the region, updated continuously as new data flows in.

While the project is still in its infancy, and observers caution that it may not be able to make good on all of the grand promises, there is also a sense that tools developed in the last decade are now at the point where it’s reasonable to launch such an effort.

“We’ve been talking like ‘Oh, we need to take an ecosystem and just do it,’” said Tanya Berger-Wolf, an Ohio State University computational ecologist who isn’t part of the initiative. “We need to monitor everything everywhere all at once and make sense of it all, often in real time. … This is the holy grail.”

In recent years, scientists have been working to devise ways to improve wildlife monitoring by marrying new tech with powerful artificial intelligence. Data from cheap, durable audio recorders and motion-activated cameras is crunched by computer programs capable of identifying individual species and discerning patterns. Increasingly, automated field labs can collect and sequence wildlife DNA from insect traps. Satellites equipped with hyperspectral imaging can see a broader light spectrum than visible light, enabling researchers to detect things like tree stress brought on by drought. Computer models can combine such data streams with a detailed map of a region’s geography to estimate what’s happening in places well beyond a specific research station.

Some of these new approaches were highlighted in a recent XPrize competition in which competing teams used drones, rapid DNA analysis, audio recorders and other technology to document hundreds of species in a patch of Amazon rainforest in 24 hours.

The new initiative would equip core sites with camera and audio recorders as well as insect traps and links to computers capable of making sense of the data. The organizers include several South American research groups already running field stations that will be among the first to feed into this network, including ones in Peru, Ecuador and Brazil.

The new initiative should help in part by enabling scientists to discern patterns across the entire region, rather than noticing them in one place and wondering what’s happening elsewhere, says Corine Vriesendorp, an ecologist and head of science for the Peruvian organization Conservación Amazónica, which operates a station on the Los Amigos River that’s in the new program.

One example of such blindspots is recent research showing that populations of birds in intact forests at a field station in Ecuador and another in Brazil have decreased. Based on scattered existing data, it’s not clear if it’s part of a broader trend. Nor do scientists know if insect populations – a food source for many of the birds – have also fallen.

Such region-wide insights could also help galvanize political action, she said. “Being able to say something about what’s going on in all of these places will be super important politically and also from a science research perspective.”

It remains to be seen whether Amazon.ia can live up to that hope. The organizers say they need $100 million to run the program for 10 years, with much of the expense in equipment, computers, staffing and training people in the various countries. So far, the group has raised around $20 million, said Rosenthal, with some $14 million of it from the philanthropic Bezos Earth Fund. They have recruited five locations to act as data-collecting points, including the Los Amigos site, leaving between five and 25 to be brought on board.

There are technical limitations that will almost certainly emerge. For instance, acoustic data can currently help scientists know whether a particular bird species has passed nearby. But it’s not very good at telling you how big that bird population is, says Justin Kitzes, a quantitative ecologist at the University of Pittsburgh who specializes in bioacoustics.

Then there is the challenge of producing a reliable, detailed view of things happening across a river basin spanning 7 million square kilometers, roughly the same size as the lower 48 U.S. states combined.

Even with advanced computer models and dozens of observation posts, Kitzes expects efforts to understand how biodiversity is changing across the Amazon will run up against a lack of data. “With 30 stations spread across the Amazon with that type of technology, will you know everything going on in the Amazon? Of course not,” said Kitzes. “Will you know a lot more than you used to? Probably.”