Taiwan’s Taroko National Park

Steep, wet slopes and earthquakes make landslides a common feature in Taiwan’s Taroko National Park.


Click here for free access to our latest coronavirus/COVID-19 research, commentary, and news.

Support nonprofit science journalism

Science’s extensive COVID-19 coverage is free to all readers. To support our nonprofit science journalism, please make a tax-deductible gift today.

A massive experiment in Taiwan aims to reveal landslides’ surprising effect on the climate

TAROKO NATIONAL PARK, TAIWAN—The frequent crackle of tumbling rocks overhead is unnerving, especially when you're picking your way through a pile of jagged debris. "I hate walking down roads like this," says Niels Hovius, a geomorphologist at the GFZ German Research Centre for Geosciences in Potsdam. "I know what can happen here."

Taroko National Park, famous for a precipitous marble gorge that cuts through it, is in a futile fight with gravity. Rockfalls litter the park's serpentine main highway. The scars of at least a dozen landslides punctuate the view in all directions. Maintenance crews are perpetually spraying concrete on slopes in a last-ditch effort to stabilize them. The park gives out safety helmets for free, and strongly encourages visitors to wear them.

For Hovius, all this moving rock and soil makes for a perfect laboratory. For the past 3 years, he and his colleagues have scrambled and rappelled across the park, installing dozens of instruments in what will end up being Taiwan's most comprehensive landscape dynamics observatory. One goal is to monitor landslides and understand their triggers. A bigger aim is to investigate their hidden impact on the climate: As massive chemical reactors, landslides draw carbon dioxide (CO2) out of the sky and sometimes belch it out, too. Understanding their role as both carbon source and sink could help researchers better model the carbon cycle that ultimately controls our planet's climate and habitability.

The recipe for landslides requires three basic ingredients: steep hillslopes, earthquakes to weaken them, and water to make them slick. Taiwan has all three in spades, making it one of the most landslide-prone countries in the world. The island was born 6 million years ago in an ongoing collision of tectonic plates that lifts mountains and generates a drumbeat of earthquakes. And its location in the tropical western Pacific Ocean means typhoons come regularly, occasionally dumping meters of rain in just days. "You can learn a lot by looking at extremes," says Susan Brantley, a geochemist at Pennsylvania State University in State College.

Hovius first visited Taiwan to do fieldwork in the late 1990s, lured by the island's extreme climatic and tectonic forcing and its extensive scientific records. A few years ago, he and GFZ physicist Jens Turowski, working with collaborators at Taiwanese institutions, began to plan something more ambitious. By 2021, their observatory, built with roughly $600,000 in research funds from the Helmholtz Association of German Research Centres, will consist of more than 120 instruments spread across hundreds of square kilometers. Solar powered and autonomous, they will relay data once per hour to servers in Germany and Taiwan.

A technician installs a data-logging station for instruments placed in a stream. Water chemistry holds clues to the effect of landslides on the carbon cycle.


Many of the instruments will work as sensitive landslide detectors. Seismometers will pick up ground shaking from tumbling rocks, and cameras will record fresh landslide deposits and scars. The continuous monitoring is a step up from patchy satellite observations and sporadic reports from tourists and park rangers, Hovius says. He and Turowski plan to share their data with Taroko officials, who can use the near–real-time detections to decide whether to close trails or roads—or search for stranded hikers.

Monitoring landslides will also help the team understand what triggers them and how landscapes recover. In 1999, a magnitude-7.6 earthquake struck central Taiwan, killing thousands of people. A few months later, researchers measured a roughly 20-fold increase in the landslide rate near the earthquake's epicenter. The uptick makes sense: Ground shaking presumably primes the slopes to give way. The surprise was that landslide rates returned to normal after a few years: The landscape somehow knitted itself back together. Rocks settling, fractures filling in, and plant roots binding the soil might all play a role in stabilizing the terrain, Hovius says. To test those ideas, the team might use "crack meters" to monitor how fissures open and close and laser scanners to detect small changes in topography. The observatory's 10-year life span should be long enough to record several large earthquakes, giving the researchers a chance to witness the rise and fall of landslide rates.

The team will also tackle a deeper mystery: the invisible influence of landslides on the atmosphere. The exchange of carbon between the atmosphere, the surface, and the oceans ultimately regulates Earth's habitability. For now, humanity—through industrial and agricultural emissions—is a dominant force in the carbon cycle. But over geologic time, the interaction of water with rocks freshly exposed by erosion—so-called chemical weathering—is another powerful player. And landslides are catalysts that speed up chemical weathering. They exhume fresh rock and grind it down into smaller pieces, creating more surface area for reactions. They also carve depressions in slopes that funnel rainwater into the rock.

Here in Taroko, Hovius and his team hope to learn whether landslide-driven weathering is releasing CO2 into the atmosphere or drawing it down, doing a small part to counter humanity's influence.

It all depends on the rocks and acids available. When atmospheric CO2 dissolves in rainwater, it forms carbonic acid, a weak acid that reacts with silicate rocks—the sandstones, granites, and others that form the majority of Earth's crustal rocks. The reaction liberates bicarbonate ions, which wash down rivers and into the ocean. There, the ions are taken up as calcium carbonate by shell-forming marine organisms. When these die, their skeletons sink to the sea floor, locking up carbon for millions of years in deposits of limestone. Were it not for this steady drawdown of carbon, the carbon emissions from volcanoes would, over the long term, turn Earth into a hothouse, says Jérôme Gaillardet, a geochemist at the Institute of Geophysics in Paris. "Life is possible because we have this process."

When limestone and other carbonate rocks are present, however, they can have the opposite effect, providing sulfide minerals like pyrite also exist nearby. Sulfides react with water and oxygen to form sulfuric acid, which in turn dissolves carbonate rocks to release CO2. About one-quarter of the rocks in Taroko are carbonates.

By analyzing river water and landslide seepage for ions let loose by these two kinds of weathering, scientists can see which one dominates. Hovius and his colleagues plan to sample water from recent landslide scars, slopes that appear to be close to failing, and landscapes that haven't budged in hundreds or even thousands of years. The results, they hope, will show how chemical weathering varies across different settings.

Robert Emberson, a landslide researcher at NASA Goddard Space Flight Center in Greenbelt, Maryland, has studied fresh landslides in Taiwan and shown that they are sources of CO2—right now. He notes that carbonate rocks and sulfuric acid react thousands of times faster than silicate rocks and carbonic acid. But over centuries the reactions can deplete the sulfide minerals or carbonates, paving the way for silicate weathering to take over, although the transition is gradual, says Kuo-Fang Huang, an isotope geochemist at Academia Sinica in Taipei who is involved in the observatory. "It's much more complicated." To see how weathering changes over time, Aaron Bufe, a GFZ earth scientist, plans to gauge the net weathering effect of landslides of different ages—in Taroko and elsewhere—that occurred in similar rock types and climates.

Whether Taroko National Park, and Taiwan as a whole, is a net source or sink of carbon hinges on other factors, too. For instance, landslides mobilize tree trunks and other plant material, kick-starting their journey toward rivers and ultimately the ocean. Rapid burial of this organic matter, before it can decay, removes carbon from the atmosphere. Hovius and his colleagues may monitor this process by recording logs washing downstream after landslides. "The full [carbon] budget remains to be determined," he says.

Hovius is eating a takeout lunch next to the Liwu River, the familiar scarred mountainsides towering overhead. He doesn't mind the prospect of more time in this grand outdoor laboratory. He points to where he lost an underwater acoustic sensor after a storm buried it in 5 meters of sediment. He runs his hand over the rocks he's sitting on and finds evenly spaced holes, relics of a study he did here 20 years ago. Studying this stretch of the river alone could yield 10 papers, he says. But he's got loftier goals. "I really want to understand the whole thing at once."

Katherine Kornei's trip to Taroko National Park was supported by a science journalism fellowship from the European Geosciences Union.