A meteorologist can track a storm today using satellite or pulse-Doppler radar data. A weather historian can track a storm from a century ago using records—watching it crawl across the plains through the precipitation totals of yellowed farm journals and log books.

Now, two scientists may have found a way to track a storm—or, at least, track the average of all storms across the season—325 years in the past.

In a paper published this month in Science Advances, Erika Wise and Matthew Dannenberg, both paleoclimatologists at the University of North Carolina, teased three centuries of local climate history out of the ponderosa pine trees that dot the Pacific Northwest. Their finding will resonate beyond the Cascades: Because that region is so important to how moisture enters and fans across North America, the study helps answer important questions about how climate change will roil the United States.

At the center of the paper is an especially sensitive species nestled across an especially ecologically diverse region. From every October to March, storms roll in from the Pacific and break across the Cascades mountain range in Washington and Oregon. The direction of a storm’s path makes a big difference in the precipitation each valley and ridge in the region receives. And ponderosa pines scattered across the two states absorb the moisture, thriving in wet years and adding thin growth layers in dry years.

The paper relies on the relationship between the Cascades,  the Pacific, and thousands of ponderosa pines. All three create and respond to the annual storm track: the average of all the paths taken by winter storms within a year.

How so? “The storm track really affects the water on the eastern side of the Cascades,” says Wise. “If storms come west to east, there’s a big rain shadow. But if they come at a slight angle, there’s less of a shadow, and the eastern side is a little bit wetter compared to the west side.”

Ponderosa pines are nestled across this crucial region. They form stands in valleys, on ridge lines, and on the lee sides of mountains. Every year, like all trees, a pine inscribes the outside perimeter of its bark into its heartwood as a ring. Like many other trees, too, ponderosa pines mark wetter years—which often boost their growth—as a thicker splotch.

Six years ago, Wise began going out in the field, collecting data from dozens of stands of trees across the Pacific Northwest. From each stand, she or her colleagues would sample 25 or 30 different trees, taking two cores per tree. Once enough samples have been collected per stand, then a kind of average tree ring history emerges for the site. Every ring in this mean tree is associated with a year, and researchers can tell whether the year was wet or dry, depending on the thickness of the ring.

They also merged this new data with a broader, older library of tree-ring chronologies recorded across the Pacific Northwest.

Then, the authors collated this tree-ring data with the past 50 years of observational measurements for the region. People have been recording storm tracks in the region since the 1950s. By matching this data to the geocoded tree rings, Wise and Dannenberg could learn what kinds of storms soak a certain valley and what kinds leave them dry. (They also focused their analysis on storm tracks measured during the October to March rainy period. This is when low-pressure storms tumble out of the Pacific, delivering Oregon and Washington most of their annual rainfall.)

Then they went a step further. Using the correlations observed from those 50 years of known meteorological data matched to tree rings, they were able to back-track. They estimated years of storm tracks from the additional 275 years of data in the tree rings. And using other paleoclimate records—such as coral skeletons from the Pacific or ice cores from the Andes—they could augment their local tree ring data with knowledge of when a broader El Niño or La Niña event was convulsing the Pacific.

These are some of the storm tracks produced by Wise and Dannenberg’s study. Each line represents the average storm track for that year, and lines are grouped based on what regions they left especially dry or wet. In all the storm tracks in A, for instance, both the east and western sides of the Cascades were unusually dry. In B, the west side was wet, while the east side was dry. In C, the eastern side of the mountains were wetter than usual. In D, both sides received more rain than usual. (Wise and Dannenberg / Science Advances)

Wise and Dannenberg’s data also cleared up a mystery about the last half century of weather records. Basically since modern meteorological records began in the 1950s, storms seemed to be getting more and more intense. Since the 1980s, the average storm track had also been drifting further north.

Some computer models suggest that this northern drift is a result of human-caused climate change. If it were to continue, eventually it would push some important precipitation northward—which would bring farmland north with it, pulling it out of the United States and into Canada.

The tree-ring data puts these modern observations into context.  It strongly suggests that the mid-20th century was an unusually weak period for storms in the Pacific Northwest. They also trended more southernly than the historical average.

You can see both these trends in the two charts below, which measure storm landfall and storm intensity since 1693. Since scientists have begun directly observing storms around 1950, they have gotten significantly stronger. Since 1980, they have also arrived further up the coast, though that data is noisier.

These two charts reveal how storms have changed over the Pacific Northwest since 1693. The top chart measures the latitude at which the storm made landfall (a higher number means it arrived to the north), and the bottom chart measures storm intensity. The tree-ring data is black and the observational data is either orange or teal. (Wise and Dannenberg / Science Advances)

So if storms now seem to be more intense and more northernly, it’s partly a regression to the mean. But something else is going on, too. “If you look at end of the intensity record, that line keeps going up, up, up,” says Wise, of the chart above. At least according to the tree-ring data, rainy-season storms in the Pacific Northwest are now as intense as they have been in the last 325 years—a major finding.

Will this trend hold? As scientists collect more data in the years to come, they’ll now be able to see whether storms keep growing more intense—outside the bounds of what’s historically normal—or whether they return to the baseline.

Overall, the paper confirms what many scientists expect for the region. “This doesn’t change the projections of what we expect to happen in the future with storm tracks [in the Pacific Northwest,]” Wise told me. “They are expected to shift north and get stronger—this doesn’t disprove that at all.”

“Particularly for the Pacific Northwest, that will really change where the rain and snow usually goes. It might mean that Seattle’s wetter but Boise is drier. That’s why these shifts make a big difference—they are what brings water to the west,” she said.

But this study is important beyond just providing a long, long, long-term forecast for Portland and points east. It represents a step forward for tree ring science.

“This is a great example of where the field has been trying to go for a long time,” says Andy Bunn, an environmental scientist and tree-ring researcher at Western Washington University. “It’s an important, super novel paper.”

For the past four decades, climate scientists have been using the basic tree-ring mechanism to piece together the climate history of an area. Take enough samples of tree rings, and you can see whether the past 100 years of meteorological data represent typical weather for a region or something stranger.

Tree rings support other findings that El Niño can cause big changes in how storms hit the Pacific Northwest. In La Niña years on the left, the average winter storm makes landfall in the Olympic Peninsula (C), and the entire region is wetter as measured by the Standardized Precipitation Index (A). On the right, which showed El Niño years, storms make landfall at the Columbia River (D) and land east of the Cascades is drier than normal (B). (Wise and Dannenberg / Science Advances)

Wise’s research goes further, says Bunn: It uses tree ring data from across the northwest to suss out the enigmatic history of the climate across the Americas. Instead of writing a climate history of the Pacific Northwest, she pokes at the mysteries of how climate works in the first place.

“It’s moving the field of dendrochronology away from simple statistical studies that look at correlations, to looking at the atompshere as a dynamic system,” he said. (Dendrochronology—from the Greek dendron, meaning tree, and chronology—is the study of tree rings across time.)

“Dendrochronology has been interested in paleoclimate, but the application has been telling very cool stories about individual spots. What this does, which I think is so important and so neat, is it looks at the engine of the atmosphere and the oceans in a way that is more physically realistic. It’s more consistent with a larger understanding of the way the climate system and the Earth system works,” he told me. “It moves away from saying: Okay, here’s the wiggle of one record, match it to a wiggle of another record, and then let’s speculate about it.”

If the paper has weaknesses, they arise from the ambiguity of trees as a long-term historical record, he said.

“Trees are not thermometers or rain gauges. They are biological creatures—and when we look at tree rings, we’re looking at one expression of their biology that leaves a record in time. As we move from trying to document a pattern to trying to understand a mechanism, it means we’re moving from the physics of atmosphere to the biology of the species. And biology is just really complicated,” he said.

Wise and Dannenberg’s paper did a good job of dealing with many of those questions by working from a large sample size, he said. Wise also told me that they had to discard some data because it was older than the rest of their sample size: Some of the ponderosa pine cores she collected dated back to the 14th century. (“If you have two trees that go back to the 1300s, that’s great and interesting but you can’t really use it,” she told me.)

And there is one more deeper premise that runs beneath much of dendrochronology, Bunn said. “The bedrock assumption we make is: Whatever created the pattern we’re observing in the past, the same mechanism will create a mechanism in the present. We assume the same processes are creating that pattern. And that’s a dicey assumption in a world that is changing so quickly.”