BERKELEY, Calif.—It is (there is no better way to put this) a dead-plant library.

Not a library of books about dead plants—though there are books here, and photos and oil paintings—but of plants themselves. They lie in wait on big, stiff sheets of yellowed paper, wrapped gingerly in envelopes and stacked on top of each other, before being inserted into dozens and dozens of cabinets. This is the fifth-largest dead-plant archive in the country, and you can turn a corner and see yards and yards of these cabinets. It is a menagerie of dead plants.

This is a herbarium. There is a network of them stretched across the country: one in St. Louis and one in New York, one in Chicago and one in Fort Worth. Each holds more than a million dead plants. This, the University and Jepson Herbaria at the University of California, Berkeley, contains more than 2.3 million specimens.

Herbaria are physical spaces—but they are also, increasingly, digital resources. For the past half-decade, the National Science Foundation has helped fund the processing of turning these dead-plants—many of which come with a species title, a collection data, and a geotag—into a database. Now, researchers at Princeton University and elsewhere have figured out how to use this database to reveal how climate change is mangling California—and the world’s—ecosystems.  

As my colleague Ed Yong has detailed, biologists have lately taken a renewed interest in natural-history museums. Their collections contain a physical record of a biodiversity now greatly denuded. Some of the most important classification and genetic-identification work is happening, not with new specimens harvested from the field, but with century-old ones called up from collections.

“There’s a tendency to see research on collections as inherently old-fashioned—a misconception of dustiness, or looking towards the past,” Kris Helgen, the head curator for mammals at the Smithsonian Museum of Natural History, told him. “But it’s inherently forward-looking.”

A similar process is now happening with herbaria. “People will look in 100 years and say isn’t it crazy that they didn’t know all the species that were in here,” Brent Mishler, the director of the Berkeley herbaria, says. “This [herbarium] only gets more valuable.”

The California herbarium, when it was located in a private house in Berkeley (Robinson Meyer / The Atlantic)

The standard herbarium specimen consists of a dried plant, pasted to a board. It also contains information on when it was collected, who collected it, where it was, and often a reference to a page of a field book. These field book entries—which the NSF has also been digitizing—often detail what the plant was near and how it was growing.  

The geographical information has become one of the most important parts of the database. Recent research has focused on a database constructed from 35 herbaria across California. It is the most complete database of plant biodiversity for any place in the country.

“There’s a lot you can tell about the history of science by looking at this data,” Jon Christensen, a history professor at the University of California, Los Angeles, said of the archive. “You can see the path [a particular botanist] took around a particular mountain collecting specimens on a particular day.”

There’s only one problem: That kind of human story is useless to science.

“All of the stuff which is of interest to a historian like me is bias,” said Christensen, pausing dramatically. That’s especially true “when you’re looking for the signal that actually the plants themselves are giving off about their habits and their range,” he said.

Thousands of cabinets like this one hold the Herbaria’s more than 2 million specimens. (Robinson Meyer / The Atlantic)

Last month, a team of researchers that included Christensen figured out how to peel away that bias. Adam Wolf, a statistician at Princeton, normalized the data in such a way that it stripped away all the noise of collection: the bio-blitzes, the well-documented townships, the individual botanists traveling a certain path 150 years hence.

What that study found was disconcerting. Many plants are moving north or uphill in response to a warmed climate, the study found. But native plants (those that are found elsewhere in the American West) are responding to global warming much faster than endemic plants, which are found only in certain parts of California.

Worse still, invasive plants are moving around most of all. While this raises the specter that invasive plants could be most effective in a climate-altered world, it may also mean that invasive plants are just generally conducive to moving around. “These are species on the move already—in a sense they have not settled down. They are still adapting, they are still finding their niche and their range,” said Christensen.

The results made sense to Mishler, familiar both with working in the field and with the University and Jepson’s collection. “I always tell people, the flora of California is best understood inside these walls. What’s outside is a shadow now of what used to be,” he told me. “There are as many weeds as native plants now.”

The herbarium is one of the oldest types of scientific institutions. Its aims are pleasurably, confidently Aristotelian—it wants to contain one of every plant, that has ever lived, anywhere, ever—and the fact that it not only still exists but also proves so useful is like finding a little living, breathing chunk of 19th-century empiricism. But understanding its aims in terms of biodiversity, as Christensen and Mishler do, makes it seem startlingly contemporary. Before there was Big Data, there was a cavernous archive—an herbarium.

“Some biologists think these kinds of collections are old fashioned. They ask, aren’t you done yet?,” said Mishler. The answer, he said, was hardly. Twelve new species are discovered every year in this state alone: “California is probably the best-known place in the U.S., and people are still going out there and saying, what is this?”

“Even within the U.S., there’s lots and lots of species that we don’t know about,” he told me. “The job of finding all the plants in the world is not done.”