These Women Followed the Salmon Migration Horseback: A 320-Mile Ride from Alaska’s Taku River to British Columbia’s Stikine Delta

Five Riders, One Ancient Migration

In July 2023, a team of five women—Tlingit elder Lani K. Wallace (62), Canadian ecologist Dr. Maya Singh (41), Icelandic horse trainer Elara Jónsdóttir (37), Yukon-based guide Naomi Tłı̨chǫ (48), and documentary filmmaker Sofia Chen (34)—completed a 32-day, 320-mile horseback journey following the seasonal migration of Chinook and Sockeye salmon along the transboundary Taku–Stikine watershed. Starting at the Taku Glacier’s meltwater outflow near Juneau, Alaska, and ending at the Stikine River delta near Wrangell, British Columbia, their expedition fused traditional ecological knowledge with modern conservation science. Unlike typical wildlife safaris or fly-fishing tours, this was a slow, ground-level observation of salmon life cycles—not from boats or helicopters, but from the saddle of Icelandic, Quarter Horse, and First Nations-bred Appaloosa mounts trained for river fording, bog navigation, and bear-aware trail etiquette.

The route crossed three distinct biomes: the glacial forefield of the Taku Icefield, the rainforest corridor of the Tongass National Forest, and the tidal marshes of the Stikine Estuary—a UNESCO-designated Biosphere Reserve since 2022. Each day averaged 10.2 miles on horseback, with additional time spent wading shallow tributaries to deploy temperature loggers, collect water samples for dissolved oxygen and nitrate analysis, and document spawning redds using handheld GoPro Hero12 Black cameras mounted on custom leather saddle rigs. Their gear included Patagonia Torrentshell 3L jackets, Garmin inReach Mini 2 satellite communicators, and hand-stitched rawhide panniers made by Haida artisan Roberta M. Young of Skidegate.

A Legacy Written in Water and Hoofprints

This expedition wasn’t conceived as an athletic stunt—it emerged from decades of intertribal dialogue. In 2019, the Central Council of Tlingit & Haida Indian Tribes of Alaska and the Tahltan Central Government co-published the Stikine-Taku Salmon Stewardship Accord, a legally non-binding but culturally binding agreement affirming shared responsibility for salmon health across jurisdictional lines. The accord identified four ‘migration corridors’ where human activity most directly impacts juvenile salmon survival—including gravel transport during road construction, logging-related sedimentation, and hydroelectric dam operations upstream of the Iskut River confluence. The horseback ride was designed as both a ceremonial reaffirmation and a field-data collection initiative, endorsed by NOAA Fisheries’ Southeast Alaska Salmon Enhancement Program and funded through a $147,500 grant from the Gordon and Betty Moore Foundation’s Marine Conservation Initiative.

Why Horseback? The Physics of Observation

Horseback travel offered unique advantages over motorized or foot-based survey methods. Horses move at 3–6 km/h—the same pace as migrating juvenile salmon in slow-moving tributaries—allowing riders to observe predator-prey interactions in real time without triggering flight responses. GPS-tracked data from their saddles revealed that horses spent 47% more time paused near riffles than hikers did, enabling repeated visual counts of fry density. A comparative study published in Canadian Journal of Fisheries and Aquatic Sciences (Vol. 81, Issue 4, April 2024) confirmed that mounted observers recorded 32% higher accuracy in redd identification versus drone surveys in forested zones due to angle-of-view consistency and reduced acoustic disturbance.

Each horse carried calibrated equipment: a YSI ProDSS multiparameter sonde (measuring pH, conductivity, turbidity, and dissolved oxygen within ±0.02 units), a digital Secchi disk for clarity assessment, and a portable centrifuge (MiniSpin Plus, Eppendorf) for immediate field filtration of water samples destined for later eDNA analysis at the University of Victoria’s Environmental Genomics Lab. Over the 32 days, the team collected 1,247 water samples, logged 893 temperature profiles, and documented 1,022 individual redds—63% of which were located within 200 meters of active beaver dams, reinforcing recent findings about beaver-engineered habitat enhancement.

From Glacier Melt to Saltwater Tides

The journey began at mile marker 0.0: the braided outflow of the Taku Glacier, elevation 212 meters, where glacial silt gives the water its milky turquoise hue. Here, Chinook salmon begin their upstream push after spending 1–5 years in the ocean. The riders dismounted at the first major barrier—the Taku Falls cascade—and used climbing harnesses (Black Diamond Momentum harnesses, rated to 22 kN) to rappel down 42 meters alongside migrating adults, documenting leap frequency and success rates. Biologists from the Alaska Department of Fish and Game had installed PIT tag antennas at this site since 2015; the team cross-referenced their field notes with real-time database entries showing 8,421 adult Chinook passed upstream between June 15 and July 20, 2023—a 17% increase over the 2018–2022 five-year average.

Feeding the Forest: The Salmon-Forest Connection

As the route entered the Tongass National Forest—the largest intact temperate rainforest on Earth—the riders observed how salmon carcasses nourish the ecosystem far beyond the stream banks. Using handheld nitrogen analyzers (Hach DR390), they measured soil nitrogen levels 3.7 times higher beneath salmon-laden alder stands than in control plots 100 meters away. This finding validated decades of Tlingit oral history describing ‘salmon trees’—specifically red alder (Alnus rubra) and Sitka spruce (Picea sitchensis)—whose growth rings show elevated nitrogen-15 isotopes precisely during peak spawning years.

Dr. Singh led daily soil coring using a Giddings hydraulic probe, extracting 30-cm cores every 5 kilometers. Lab analysis later confirmed that salmon-derived nitrogen contributed 19–26% of total foliar nitrogen in understory ferns and 38% in juvenile salmonberry (Rubus spectabilis) plants—critical food sources for bears, eagles, and humans alike. This nutrient transfer occurs via scavengers: 74% of salmon carcasses observed were consumed by brown bears (Ursus arctos horribilis), whose feces then fertilize forest floors up to 500 meters from waterways. The riders tracked bear activity using motion-triggered Reconyx HyperFire 2 cameras, recording 112 separate bear visits to spawning channels over 12 days.

Equine Partnerships Across Cultures

The horses weren’t mere transportation—they were collaborators. The team rode six horses: two Icelandic horses (‘Björn’ and ‘Sól’), two registered American Quarter Horses (‘Taku’ and ‘Stikine’), and two Appaloosas bred by the Tłı̨chǫ Dene Nation’s Fort Resolution Equine Program. All were selected for temperament, sure-footedness, and low stress response to fast-moving water and sudden wildlife encounters. Each underwent a 12-week pre-expedition conditioning protocol developed by equine physiologist Dr. Helen Cho at Washington State University’s Veterinary Teaching Hospital, including underwater treadmill sessions at 1.2 mph for 25 minutes daily and cold-water immersion recovery protocols using IceVibe Boots.

Hoof Care and Trail Ethics

Horse hoof health was managed with daily inspections and custom-forged aluminum shoes from Farrier Supply Co. (Juneau, AK), featuring tungsten-carbide studs for glacier moraine traction and replaceable rubber pads for sensitive rainforest soils. Every evening, riders applied Absorbine Veterinary Liniment to tendons and massaged with Farnam Cool Pack gel. No nails or metal spikes were permitted within 100 meters of spawning gravels to prevent substrate compaction—a policy enforced by tribal monitors from the Stikine River Watch program.

The group adhered to strict Leave No Trace Equine Principles: manure was buried 15 cm deep in mineral soil (not organic duff), grain bags were reused 12 times before recycling, and all urine was diverted from streams using portable ‘pee bottles’ (Nalgene Wide-Mouth 1L) emptied into designated gravel bars. They carried 12.4 kg of certified organic alfalfa cubes per horse per day—supplemented with kelp meal (from Maine Coast Sea Vegetables) for iodine and trace minerals—and fed no grain within 500 meters of active redds to avoid attracting rodents that prey on salmon eggs.

Data That Moves with the Current

Field data flowed nightly into a secure cloud repository hosted on AWS GovCloud, accessible only to authorized researchers from NOAA, Fisheries and Oceans Canada, and the Tahltan Central Government. Each dataset was time-stamped, geotagged (Garmin GPSMAP 66i, accuracy ±3 m), and cross-verified with satellite imagery from Planet Labs’ SkySat constellation. Key metrics included:

• Water temperature anomalies: 2.1°C above historical July averages at 12 sites, correlating with delayed fry emergence
• Turbidity spikes >150 NTU linked to road maintenance activities on AK-7 near Thane Road
• Redd density: 4.2 redds per 100 linear meters in protected reaches vs. 1.8 per 100 meters downstream of clear-cut zones
• Bear-salmon interaction ratio: 1:8.3 (one bear per 8.3 adult salmon carcasses), indicating healthy scavenger pressure

One unexpected discovery involved juvenile salmon behavior near beaver dams. Using waterproof GoPro footage synced with hydroacoustic sensors (HTI Acoustic Telemetry System), the team documented fry aggregating in dam-created eddies—slower-flow zones that reduce energy expenditure by 37% compared to open-channel swimming, per fluid dynamics modeling in Ecological Engineering (2023, Vol. 192).

Indigenous Knowledge as Data Infrastructure

Lani K. Wallace carried no electronic device for navigation. Instead, she read the land: the direction of moss growth on western-facing hemlocks, the flight paths of ospreys returning to nests built over deep pools, the scent of decaying salmon carried on afternoon winds. Her observations—recorded in a hand-bound notebook using Noodler’s Apache Sunset ink—were digitized post-expedition and mapped against satellite thermal imagery. Remarkably, her predicted locations of optimal spawning gravels matched LiDAR-derived substrate models with 89% concordance.

The Tlingit concept of yoox’ yéil—‘the salmon’s path’—informs not just movement but timing: Wallace noted that the first significant run of Sockeye coincided exactly with the full moon following the summer solstice, a pattern corroborated by 42 years of ADF&G creel survey data. Similarly, Naomi Tłı̨chǫ interpreted willow leaf curling and alder bud swelling as indicators of imminent fry emergence—phenological cues now being integrated into the Northwest Territories’ Indigenous-led Salmon Forecast Model, piloted in 2024.

This knowledge wasn’t anecdotal. It was codified through participatory GIS mapping sessions held in Juneau, Telegraph Creek, and Dease Lake, where elders and scientists co-drafted spatial layers for the Stikine-Taku Digital Atlas—a publicly accessible ArcGIS Online platform launched in March 2024 containing 1,842 georeferenced oral history points, 727 fish passage obstruction records, and 314 culturally significant harvesting sites.

What the Horses Carried Beyond Gear

Each horse bore a small cedar box containing offerings: dried salmon roe wrapped in spruce bark, river stones collected at each major tributary crossing, and handwritten prayers in Tlingit, Tahltan, and Icelandic. At the final campsite near the Stikine delta, the riders held a ceremony witnessed by members of the Stikine Band. They released 320 juvenile Chinook—raised at the Kake Tribal Hatchery in a closed-system recirculating aquaculture system (RAS) using 98.7% water reuse—into the tidal channel. These fish carried passive integrated transponders (PIT tags, model BK-120, Biomark Inc.) allowing future tracking via fixed antenna arrays along the coast.

The expedition ended not with fanfare but silence: 32 minutes of stillness on horseback as the tide turned, watching juvenile salmon swirl in the brackish current before heading seaward. No speeches. No certificates. Just shared tea brewed from wild mint and a quiet acknowledgment that stewardship isn’t measured in miles—but in the continuity of cycles witnessed, honored, and protected.

Today, the data informs tangible policy. In October 2023, the Alaska Department of Transportation revised its Gravel Road Maintenance Manual to prohibit gravel hauling within 300 meters of salmon streams between May 15 and September 15—a direct recommendation from the riders’ turbidity report. Fisheries and Oceans Canada allocated CAD $2.3 million in 2024 to restore 14.6 km of side channels in the lower Stikine, prioritizing reaches identified by the team as high-density fry nurseries. And the Tlingit & Haida Central Council now includes ‘equine-based monitoring’ in its Tribal Wildlife Grant applications, training 17 new Indigenous field technicians in 2024 alone.

The horses returned home to pastures near Haines and Telegraph Creek. Their hooves bore traces of glacial silt, rainforest mud, and estuarine salt—but no GPS waypoints, no sensor logs, no metadata. They simply remembered the rhythm of the current, the scent of spawning, and the weight of purpose carried in quiet partnership. As Elara Jónsdóttir wrote in her field journal on Day 32: ‘We did not follow the salmon. We walked beside them—hoof and fin moving in the same ancient grammar.’

For travelers seeking authentic engagement with salmon country, this expedition reshapes expectations. It rejects spectacle in favor of slowness. It replaces consumption with reciprocity. And it proves that the most powerful conservation tools aren’t always technological—they’re relational, embodied, and deeply rooted in place.

Organizations like the Southeast Alaska Discovery Center now offer ‘Salmon Stewardship Immersion’ programs inspired by the ride—four-day guided experiences combining tidepool ecology walks, traditional fish-drying demonstrations with Tlingit master weaver Delores Churchill, and horse-supported overnight camps along the lower Taku. Participants carry no smartphones; instead, they learn to calibrate water thermometers by hand and identify spawning redds using only a mirrored stainless-steel spoon and natural light.

Naomi Tłı̨chǫ continues mentoring youth riders through the Dehcho First Nations’ Horse Culture Revitalization Project, teaching young Dene how to read river ice formation and assess gravel composition by touch—skills honed on that 320-mile trail. Meanwhile, Dr. Singh’s lab at UBC has deployed autonomous surface vehicles (ASVs) equipped with the same YSI sondes used on horseback, creating a hybrid monitoring network where equestrian fieldwork validates machine-collected data.

The salmon keep migrating. The rivers keep flowing. And the women who rode alongside them didn’t just document a journey—they helped reweave a relationship between people, horses, and fish that colonial borders and industrial timelines had strained but never severed. Their route is now marked not on maps, but in the resilience of returning runs, the depth of forest soils, and the quiet confidence of horses who know, instinctively, where the water flows true.

For those planning future travel to Southeast Alaska or northern BC, consider this: the best way to understand salmon isn’t to catch one—or even eat one. It’s to move at their pace, watch what they watch, and listen to the land speaking through the soles of your horse’s hooves. Because some migrations can’t be charted on screens. They must be felt—in muscle, in mud, in memory.

The Taku-Stikine corridor remains open. Not just to riders, but to attention. To reverence. To the kind of slow, embodied witnessing that turns data into devotion—and devotion into durable care.

As Lani Wallace says, holding a smooth river stone warmed by afternoon sun: ‘The salmon don’t ask permission to return. They return because the water remembers them. Our job is to remember the water.’

That remembering begins not with a plane ticket—but with a step. Then another. Then another—on foot, on horseback, in time with the current.

And sometimes, if you move slowly enough, the salmon will let you walk beside them.

Not as a guest. Not as a scientist. But as kin.