The Yakutat microplate is stuck in tectonic traffic.

Along much of the southern edge of Alaska, the Pacific plate dives northwestward beneath North America. But where the state turns the corner to fling its eastern arm down part of western Canada, there’s a different plate heading down the hatch — the Yakutat microplate.

The Yakutat microplate is a peculiar piece of plate, its provenance contentious. Some say it’s part of neither the Pacific nor North America proper. It’s a bit of curiously placed, somewhat buoyant oceanic plateau that’s stuffing itself into the subduction zone at Alaska’s crook. The subduction zone is, in the parlance of scientists, congested.

One way to see what’s happening below the surface — how the microplate slips and shoves below North America — requires earthquakes. Earthquakes illuminate subsurface structures in several ways. For example, their locations accentuate active structures. In other instances, as seismic waves travel through different layers, the waves change speed and angle, highlighting changes in composition. Detecting even extremely small earthquakes can help scientists “see” under our feet.

In new research published in The Seismic Record, a team of scientists led by Meghan Miller of the Australian National University used machine learning to uncover thousands of previously unknown tiny temblors. These quakes trace the razor-sharp eastern edge of the Yakutat microplate, much farther east than previously thought, as the microplate journeys northward into Earth’s mantle.

Volcanoes, mountains, faults

Alaska’s landscape is shaped in part by its geology. Head toward the bend in Alaska’s coastline by boat, and navigate through Cook Inlet to the capital city of Anchorage. To the west, the long line of active Aleutian volcanoes begins, with Mount Spurr, one of the easternmost volcanos, visible from Anchorage on clear days. The approximately 1,100 mile-long Aleutian Range proceeds along the Alaska Peninsula before hopping offshore and continuing as the Aleutian Islands, reaching for Russia. These volcanoes, their eruptions, and megathrust earthquakes along the chain are the above-mentioned products of Pacific subduction.

Once you dock at Anchorage, go ashore and make your way through the city. Glance to the east; you’ll see the jagged peaks of the Chugach Mountains looming, the range shaped by and draped with glaciers. Several summits pierce the sky at over 10,000 feet. Northeast of the Chugach range and out of sight of the city, the volcanic peaks of the Wrangell Mountains punctuate the landscape toward Canada. Beyond the Wrangells, the non-volcanic Saint Elias Mountains span the border, heading southeastward into Canada.

Around where the Saint Elias Mountains and the Chugach range intersect, Alaska’s eastern appendage turns the corner, stiff-arming its way down the west coast of North America. The offshore, plate-bounding, right-lateral strike-slip Queen Charlotte Fault parallels this northwest-southeast trending limb, coming onshore as it heads north, becoming the Fairweather Fault. This system of faults, like the mountain ranges, turns the corner in Alaska’s armpit (arguably the most spectacular underarm in the world).

The Queen Charlotte-Fairweather fault system gets complicated at the curve. One of the complexities: The system connects to the largely east-west trending Denali fault system via the Totschunda Fault.

The Alaska range, the Denali Fault

If you head north from Anchorage, you can catch the Parks Highway toward Fairbanks — more than 300 miles away. You’ll pass the Talkeetna Range’s snowy crags, which rise less than 8,900 feet above sea level.

Continue onward and you’ll approach the Alaska Range, its pinnacles soaring in Denali National Park. The arcuate range, a mere 84 miles south of Fairbanks, extends into both Wrangell-St. Elias National Park to the east, and Lake Clark National Park to the west. Denali, North America’s highest peak, crowns the range. The north face of the mount, Wickersham Wall, is shaped by the right-lateral Denali Fault. In fact, a bend in the Denali Fault works to raise the range’s high topography.

The Denali fault system is like its southeastern sister — the Queen Charlotte Fault. Both are right-lateral continental strike-slip faults that cut across hundreds of miles. But the Denali Fault does not delineate a plate boundary. Instead, it’s a fault cutting across Central Alaska, across a tectonic plate. It’s an intracontinental fault. Its easternmost extent swings southeast, paralleling the Queen Charlotte Fault. To the west, it ends in the Bering Sea, well north of the Aleutian subduction zone.

Major quake

On the afternoon of November 3, 2002, Seattle’s Lake Union began to slosh. Further afield, pools and ponds as far as Texas slopped from side to side as seismic waves triggered standing waves called “seiches” in enclosed water bodies.

A magnitude 7.9 earthquake had struck central Alaska on the Denali fault system. The earthquake not only sliced through nearby streams and roads, but also littered fault scarps across tundras and glaciers. Yet, the rupture nucleated elsewhere, on an unstudied fault.

Earthquake scientists responsible for mapping the surface rupture in the quake’s aftermath found some structures signaling relatively shallow thrust faulting — not strike-slip. Studies showed that the mainshock consisted of three subevents — three distinct, sequential components of the rupture.

The first subevent occurred on the previously unknown Susitna Glacier Fault, in which thrust motion — the shoving of one side over the other — transitioned to the Denali Fault, where motion switched to right-lateral strike-slip. The second subevent began about 80 kilometers east of the epicenter, ripping eastward, exclusively on the Denali Fault. A third subevent, about 190 kilometers east of the epicenter, caused another burst of right-lateral slip on the Denali Fault. The rupture jumped south, continuing on part of the Totschunda Fault. That third and final subevent thankfully ended there, instead of continuing onto the Queen Charlotte fault system. The easternmost extent of the Denali Fault remained unruptured.

Complex tectonic confluence

Alaska’s bend is a confluence of complexity from ocean to continental interior. The Yakutat microplate awkwardly subducts at a relatively shallow angle, with the Pacific plate subducting nearby. The Yakutat, says Miller, is both thick and buoyant compared to the steeply dipping, more dense Pacific plate.

Along the bend, the expected subduction-driven arc is missing. Instead, scientists note a marked lack of volcanism in what is known as the Denali Volcanic Gap — which spans the space between the Aleutians and the Wrangells.

Typically, the locations of subduction zone earthquakes allow scientists to trace the downgoing slab. Called the Wadati-Benioff zone, it’s hard to track without detecting lots of local earthquakes. And at the purported eastern edge of the Yakutat microplate as described in previous studies, this paucity of earthquakes detected to the east means that the microplate seems to vanish.

Moreover, a weird kind of seismic signal has been identified here — tectonic tremor, found in subduction zones (and other major faults) around the world. Tectonic tremor is a low-amplitude seismic signal of extended duration, and lacks the higher frequencies of regular earthquakes. Its presence tells us about slow slip, and here, it seems to be concentrated just underneath the Alaska range, where we might expect to find subducted Yakutat.

Other oddities persist. The volcanoes of the Wrangell field, punching through continental crust, form an outlandish sort of arc — part of a volcanic arc-transform junction that began to erupt nearly 30 million years ago. Today, only one volcano, Mount Wrangell, is active, while several mud volcanoes persist.

To explore the lithosphere of this enigmatic area, Miller and colleagues crafted a three-pronged attack using waveform data available from the NSF NGF data archive. They first created a new machine-learning based earthquake catalog. They used data from more than 200 stations installed near the central section of the Denali Fault. The data spanned January of 2018 to December of 2021. The team’s final catalog included 64,870 events.

In the second analysis, the team looked at earthquakes that occurred at distances between 30° and 98° — thousands of miles away. The seismic waves produced by these events, with magnitudes greater than 6.0, sample Earth’s structures along their path. As P-waves cross structural boundaries, they produce additional seismic waves, including S-waves. In this kind of analysis, scientists measure the time difference between the arrival of these two waves at a single station. That difference tells us about the depth at which the conversion happened, which in turn tells us the depth of the structure. Do this for lots of earthquakes at lots of stations, and you can start to see the depth to structures like the Mohorovičić discontinuity, or Moho, which marks the transition from crust to mantle. 

The final analysis the team undertook involved using vertical-component ambient noise data (think of this as the background noise that seismic stations are always recording). The data, collected between August of 2017 to the end of July 2021, allowed the team to generate a three-dimensional shear-wave velocity model of the crust and uppermost mantle, a depth of about 50 kilometers.

Patterns of the Yakutat plate

The machine-learning based earthquake catalog allowed the team to identify several structures. For instance, earthquakes illuminate what is likely subduction of the Yakutat microplate beneath the Denali Volcanic Gap.

To the east the researchers see earthquakes setting alight the main strand of the Denali Fault, as well as its connection to the Totschunda Fault. The westernmost volcano of the Wrangell volcanic field, the inactive Mt. Drum, also hosted clusters of quakes, termed the Glenallen cluster in previous work.

However, the authors also observed a new pattern: a line of earthquakes trending north-northwest, tracing what’s likely to be the eastern edge of the Yakutat microplate. East of this line, beneath the Wrangell volcanic field, there’s almost no seismicity at any depth.

This streak of earthquakes stretches about 250 kilometers and includes about 1,750 events south of the Denali Fault. Moreover, the linear feature also correlates with the northeastern extent of tectonic tremor in the region, interpreted to mark the interface between the Yakutat and North American plates. As tremor ceases, the earthquakes continue to a depth of about 120 kilometers — beneath and beyond the Alaska range.

Another key observation: The magnitude 7.9 Denali earthquake’s epicenter and its aftershocks sit along this line. The epicenter also resides at the apex of the curve of the Alaska range and Denali Fault. Miller and colleagues argue that the combination of thrust and strike-slip faulting during the 2002 earthquake may result from a combination of stress loading caused by plate motion plus the curving of the Denali fault. It’s also possible that the sharp edge of the Yakutat microplate — directly below the epicenter and curved fault — could have “influenced the nucleation of the 2002 event,” they write.

The three-dimensional velocity model of the crust — above the Yakutat microplate — also emphasizes the linear seismicity zone; the edge of a fast S-wave velocity zone in the lower crust aligns itself along the same boundary. Fast seismic velocities represent cold North American crust, Miller explains. A low-velocity zone east of the line — below the Wrangell volcanic field — indicates an absence of cold North American crust at that depth.

Mantle wedge?

The analyses of distant earthquakes map out the subducting Yakutat microplate in greater detail, supporting the hypothesis that the edge of the plate corresponds to the line of earthquakes. Moreover, the complete analysis supports previous studies that conclude the subducting Yakutat plate is in contact with North American lithosphere — without an asthenospheric wedge.

The Yakutat microplate’s lower density than the Pacific explains its relatively shallow subduction angle. There’s simply no space for a mantle wedge, which in turn explains the Denali Volcanic Gap’s existence.

Volcanism in the Denali Volcanic Gap that began about 1 million years ago might indicate that a mantle wedge may be establishing itself, according to Miller and colleagues. They note that active volcanism just beyond the edge of the subducted Yakutat slab — Buzzard Creek and Jumbo Dome volcanoes north of the microplate, and the Maclaren River volcanic field to the east — further support an encroaching mantle wedge.

Putting the pieces together

The zone of tectonic tremor suggests the Yaktutat microplate incrementally slides beneath the continent, as opposed to moving in the fits and starts that would produce earthquakes. But at the eastern edge of the slab, after tremor lessens, a different deformation mechanism seems to be taking over; stick-slip earthquakes indicate an increase in fault strength, according to Miller and colleagues. And so, the rock tends to break, instead of sliding.

The “razor-sharp line,” the authors argue, “is the locus of changing slab morphology as it cuts through the lithosphere.” In this way, the edge of the Yakutat controls the seismicity in the overriding plate — perhaps also including the nucleation of the 2002 magnitude 7.9 earthquake. It controls the curious curve of the Alaska Range. And it controls the recent volcanism along its edge.

“Knowing more precisely where past earthquakes, even small magnitude ones, occur can help seismologists and geoscientists understand where others may more likely occur in the future,” Miller says. Small earthquakes can inform seismic hazard models. Plus, finding more of them could further illuminate the congested tectonic traffic jam in and around Alaska’s exceptional bend.

This study utilized seismic data from the NSF National Geophysical Facility data archive.