Two Stanford researchers have transformed data of a dying Alaskan tree into song.
Quick, bright piano notes open the melody. It’s lively, almost playful; conjuring images of birds chirping through sun-dappled forests.
Then suddenly, a shift. The piano notes quieten; the silence interrupted only by the lament of a lone flute. But soon, other instruments compete for the listener’s attention, the strings and a clarinet jostle to be heard.
The piece becomes a roller coaster of sound and emotion, until finally it ends with a trill of a flute — almost abruptly— much more subdued than how it began.
This is a 100-year-old story captured in a 3-minute song; the story of a dying tree and its changing forest home. It’s a song about life and death. It’s a song about climate change.
Two Stanford researchers collaborated to create the music through an emerging analytical toolknown as data sonification. It captures the ecological transformation of Alexander Archipelago, a 300-mile chain of islands off the southeastern coast of Alaska. There, yellow cedars have been dying in staggering numbers over the past century. Other trees have been moving in to take their place.
The piano in the song represents the dying yellow cedar; violin and viola symbolize mountain hemlock; the clarinet is shore pine; cello and bass are Sitka spruce; and western hemlock, the conifer which has successfully usurped the cedar in some swathes of the archipelago, is played by the mournful flute.
Yellow cedar has long been a culturally, ecologically and economically valuable tree in southeastern Alaska. Many of the trees are ancient, aged between 700 and 1,200 years old. Their leaves and surrounding soils play host to unique bacterial communities; animals rely on the trees for food and shelter, while humans prize them for their durable, closed-grained wood.
For the past 100 years, however, the yellow cedar has been dying off in large numbers across Alaska and British Columbia. As of 2012, 60 to 70 percent of yellow cedar in a 600,000-acre area in the region have died.
Climate change is to blame for this die-off, scientists say. They’ve concluded that though the trees could withstand centuries of bugs, rot and injury, their shallow roots are vulnerable to freezing when the surrounding soil is not protected by a layer of insulating snow.
Alaska has been melting for decades. The average temperature across the state has increased by about 3 degrees Fahrenheit over the past 60 years. That’s more than twice the warming recorded in the rest of the U.S., according to the Environmental Protection Agency.
Interested in these dying cedars and, more specifically, the environmental changes that the tree deaths had triggered, ecologist Lauren Oakes visited the Alexander Archipelago in the summers of 2011 and 2012 to conduct field research. Oakes, then a doctoral student at Stanford’s Emmett Interdisciplinary Program in Environment and Resources, started in the north of the archipelago in Glacier Bay National Park where healthy yellow cedar forests still thrive. She steadily moved south to areas increasingly afflicted by warming temperatures and dying cedars, finally ending her research at the base of Slocum Arm in the remote wilderness of the Tongass National Forest.
Wherever Oakes and her team of field assistants went, they identified trees and other plants in the forest, taking note of measurements including the height and diameter of each specimen. Her data set ended up containing nearly 30 variables for more than 2,000 conifers. Oakes published her findings in an article for the journal Ecosphere in 2014. But as with most academic writing, it was not widely-read beyond a select group of academicians.
That didn’t sit well with Oakes.
“My path in life to date has been at the intersection of environmental research and communications — so I’m always looking for interesting ways to share important stories of environmental change with a broad public,” she shared with The Huffington Post in an email last week.
Oakes double-majored in environmental studies and visual art, specializing in film and photography, at Brown University. She later worked on a series of environmental documentaries, including one for PBS’ Frontline, and wrote for The New York Times Green blog. She recently began work on a book about her Alaskan research for a mainstream audience. It’s due to be published in 2018.
“In any project I tackle, I want to publish the research in journals for the science community, [but also] work to communicate what I learn through other creative means to a broader public,” she said.
So when Stanford University PhD student Nik Sawe emailed her earlier this year about his pioneering work in data sonification, Oakes said she was instantly intrigued and replied immediately. Sawe had been looking for interesting datasets to sonify.
Data sonification is, essentially, the transmutation of data into music. It involves a computer and a composer, who feeds information into a special modeling software that reinterprets data so it’ll make sense to a musical instrument digital interface, or MIDI.
Similar to the more mature tools of data visualization, sonification is meant to help uncover or better explore patterns and trends. These patterns, in turn, could unveil news way to see the world.
The tool is still in its infancy, but many interesting sonification projects have already emerged. Musician Brian Foo has sonified income inequality in New York City and air quality in Beijing. Chris Chafe, director of the Center for Computer Research in Music and Acoustics, Stanford’s computer music research center, and a mentor of Sawe’s, has sonified electrode recordings from epileptic patients.
And then, there’s the climate change music — the result of the sonification of Oakes’ data from Alaska. The music maps Oakes’ journey from the healthy cedar forests of the north to the dying ones of the south. “What I was really looking for was not just a pattern but a narrative,” Sawe said. “The decline of the yellow cedars and the link to climate change was a very clear story in the data. The trick was trying to represent that in audio.”
Then Sawe realized that Oakes’ trip had, in effect, been a journey back in time.
“As Lauren went further south along the Alaskan coast, the forests had been wrestling with climate change for a longer time. So that travel through space, from north to south, became time in the song. We first hear the forests in the north — still largely untouched by climate change — and as we travel south the changing landscape becomes more and more evident,” he said.
Every moment in Sawe’s composition is meaningful. Every note is a tree and every species plays a different instrument. The pitch of each note is related to the height of the tree; how hard the note is hit is linked to the diameter at the base, and the duration correlates to tree health.
“We spend the same amount of time in every place that Lauren visited, so if there are say, 40 yellow cedars in one location, they’ll play very quickly and lively; if there are 3, they’ll be long slow notes,” Sawe said. “But here’s perhaps the most important part of the mapping: dead trees count as silences. So in a place with 3 live trees and 30 dead, we’d hear 3 sporadic notes surrounded by silence. And that’s why you can really hear, about midway through the song, this abrupt change in the liveliness of the yellow cedar. More and more of the trees in the forest are dead.”
In the recording below, the yellow cedar’s piano melody is isolated, allowing for closer listening to the individual tree. Note the silences that begin near the middle of the song:
Sonification of data involves a precise, scientific process — but there’s artistic flair intertwined in it too. In the case of the Alaskan forest composition, for instance, Sawe made two obvious aesthetic choices. One was the use of cello and bass to represent Sitka spruce, which is a tree whose wood is often used to make stringed instruments.
The other was the key. “I picked D minor as it sounded rather poignant,” Sawe said.
The song seems to take on an elegiac quality as a result, a ballad to the dying cedar tree. But a different key could’ve perhaps reflected something else Oakes discovered in her research.
“This is a story of death, yes, but it’s also a story of regeneration, and a new forest community taking shape,” she said. “In the song, you hear yellow cedar trees drop out and then other species regenerate. In the middle, you hear a busy flurry of instruments as various species compete to regenerate. Then over time, in the end, you hear more flute ― Western hemlock dominates.”
Sonification could change the way scientists, and the public, analyze and appreciate data.
“For scientists to interpret their data in a new way, this could be really valuable,” said Sawe, who is currently working on another research project at Stanford to sonify California’s drought. “In terms of how we explore our data, and what questions we ask next as we endeavor to do good science, this is a really promising tool.”
The accessibility and emotiveness of music means that broader audiences can also be reached ― and touched.
“You don’t have to understand statistics or how to read a graph, you don’t need access to an academic journal, you don’t have to dig through the raw data where Lauren painstakingly recorded the attributes of thousands of trees. I can describe what you’re about to hear in 20 seconds and then you’ve got it, and you’re able to hear every single tree that Lauren visited down the Alaskan coast inside of 3 minutes,” said Sawe, adding that the emotional quality of music allows for listeners to experience data “viscerally.”
Next spring, a live orchestra will be playing Sawe’s climate change composition on Stanford’s campus. He and Oakes say they hope to also launch a public exhibition that will showcase the music as well as the photos Oakes captured in the Alaskan wilderness.
“I’d like our work to communicate science in a more accessible way and for it also to help raise awareness about the kinds of impacts occurring as a result of climate change,” said Oakes. “As Nik would tell you, information alone isn’t often enough to motivate action. Feeling motivated is often connected to emotion, and we tend to strip emotion from science.”