This time of year, the Middle Fork of the Flathead River is ice-bound and quiet – except when a freight train rumbles past.

The transcontinental “Hi-Line” follows the river along Glacier National Park’s southern border for about 30 miles between Bear Creek and West Glacier. Each week, several trains pass through this scenic corridor carrying oil from North Dakota’s Bakken fields.

BNSF Railway owns these tracks and insists they are safe and well-maintained. But in recent years, as trains carrying Bakken crude have derailed, spilled and exploded in other locations across North America, conservationists have grown worried.

American Rivers placed the Middle Fork on its 2017 list of “America’s Most Endangered Rivers,” warning that a spill “could permanently degrade the river and downstream waters, harming communities and the economy.”

This possibility isn’t lost on the railway, which has prepared a geographic response plan for its staff and local first-responders to follow if a spill occurs. Flathead National Forest District Ranger Bob Davies said in an email that “since I started working on the Flathead in 2012, BNSF has contacted me annually to review [the response plan], update contact info and review emergency procedures.”

But these plans may rest on insufficient data, say scientists who study the Middle Fork. In interviews and documents shared with the Daily Inter Lake, they’ve called for more research on the complex hydrology around the river – and explained how BNSF Railway rejected a proposal to test it.

Their interests focus on the waters around and beneath the Flathead. Most rivers are “not like a pure gutter,” explained Bob Hall, a professor at the University of Montana’s Flathead Lake Biological Station. In some stretches, the Flathead squeezes through well-sealed bedrock canyons. But it also flows through, broad, gravel-lined valleys.

Hall says that one of these valleys, the Nyack floodplain, “is really, really un-gutter like.”

Standing on the river’s snow-covered bank, the ecologist likened it to a different plumbing fixture. “If you dug a hole where I’m standing, and poured a bucket of water, it’d be like flushing a toilet, and the water would just go ‘fmm,’ running right out into the aquifer.”

Beneath the Nyack’s soil lies a layer of pebbles and fist-sized “cobbles” a mile wide, 6 miles long, and 70 feet thick. Hall explained that, when river water enters the valley, it’s forced down between these stones. It resurfaces about 6 miles downstream, the “upwelling” end of the aquifer. About a dozen sensor-equipped wells around the valley record the water’s temperature, depth and dissolved oxygen concentration for scientists.

“You think of groundwater as just this dead water, but nothing could be further from the truth,” Hall explained, standing over one of the wells.

“There’s a whole assemblage of animals that live down there” he said, spacing his fingers an inch apart to show the size of stoneflies that have crawled out of the well. Thanks to bacteria living there, “It’s the site of lots of transformation of nutrients” carried underneath by the river.

Other scientists are also watching how water moves through the area. According to Davies and Flathead County Office of Emergency Services manager Rick Sacca, U.S. Geological Survey gauges helped generate the information, such as river water travel times, that guide BNSF’s geographic response plan.

But Jack Stanford, professor emeritus at the Flathead Lake Biological Station, worried that the plans didn’t account for all the complex piping beneath the river.

The station’s director for 36 years, Stanford wanted to better model how oil would move through the area. In a 2016 research proposal, he joined the station’s assistant director, Tom Bansak, and Montana State University’s Geoffrey Poole in arguing that “current models of contaminant transport in stream networks do not account for the effects of vertical and lateral dimensions associated with hyporheic hydrology” – the up-down and side-to-side flows of water around and beneath the river.

“We know that estimates provided by conventional modeling approaches will be wrong,” they continued, “but we don’t know the expected magnitude of the error in model results.” But based on existing knowledge, “we expect the model error to be substantive.”

To craft a new method, they explained, they would need to collect physical data by releasing dye into the Flathead River. The compounds contained in and derived from crude oil, Stanford said, are “all going to move through those systems differently, but the dye test would give us a baseline.”

Hall made clear that “the dye would not necessarily say, ‘this is what’s going to happen to the oil’...But it would allow us to build a hydrologic model that we could then couple to an oil-spill model to predict how oil is transported.”

To reach that goal, the scientists would need more than coloring. In a 2016 letter to his colleagues on the proposal, Stanford argued that simply “dumping dye into the river...will not provide the data needed to produce a predictive model of short and long term consequences of a spill.”

To generate that data, he continued, the dye would have to be injected at several points along the Middle Fork, and include a salt that sensors placed in the area could detect.

“You have to put sensors in the river system and in the groundwater in order to capture when the [dye] plume moves through,” Stanford explained in a phone interview, “and those things are not cheap.”

In his letter, he estimated that designing a test would cost $18,000. Carrying it out and building the model would take about $300,000 – to be funded, they hoped, by the railway. “We propose to work with BNSF, GNP [Glacier National Park], and other stakeholders” to study the issue, the grant proposal began.

But BNSF showed little interest.

“We had proposed that the railroad ante up enough money for us to do that kind of release and we actually wrote a pre-proposal for them to consider,” Stanford remembers, “but I never heard anything more about it.”

The railway eventually addressed the proposal in a May 16, 2016 letter to the leaders of another water-quality group, the Flathead Lakers. BNSF Assistant Vice President for Community Affairs Andrew Johnsen argued that dye couldn’t replicate oil.

“While a soluble dye tracer is useful for time of travel studies, mixture and retention analyses, and calibrating stream-flow models, it does not simulate oil spills well,” he wrote to Lakers president Greg McCormick and executive director Robin Steinkraus. “A miscible dye” – one that fully dissolves in water – “does not replicate how a non-miscible product such as crude oil would act in water.”

But the Biological Station researchers expect that some oil compounds would get churned into the river.

“Some of it will get suspended in the river as little droplets, especially when it hits rapids,” Hall explained. “It’s like making mayonnaise...some of it will travel underwater. Not all of it, but some of it.”

That fraction of the oil, and its components, will seep into aquifers like Nyack’s. Poole, a professor in MSU’s Department of Land Resources and Environmental Sciences, said the dye test would help them understand where those compounds would be carried.

“Given what we do know today, any water soluble chemical (or the water soluble fraction of any other spill like fuel or oil) is likely to penetrate deep into the gravel aquifers along the river and persist for perhaps months to years,” he wrote in an email. “Dye tests would give us a better idea of the movement and persistence of water – and therefore water soluble chemicals – in these aquifers.”

Hall phrased it more simply. “If we add 10 pounds of dye up here, and only 5 pounds leaves the river, that means 5 pounds is underneath our feet somewhere.”

The information gathered, Stanford said, could “provide response teams perspective because in the end, a spill with the complex compounds and chemicals involved will be worse than the dye may indicate.”

According to Erin Sexton, senior research scientist with the biological station, “the dye test is still a priority, but funding has not been found yet to fill this important information gap.”

If and when the test gets done, it may show a problem with no solution.

When oil spills into water, first responders focus on the surface, trying to corral it with snake-like floating booms.

While the railway’s response plan isn’t publicly available, the Forest Service’s Rob Davies and Flathead County’s Rick Sacca have access to the plan. Both say that BNSF stores its booms and other equipment, like skimmers and storage tanks, at Whitefish and West Glacier, and that it has seven pre-planned boom deployment sites along the Middle Fork. The Nyack Flats are hemmed in by two of these sites, at the Moccasin Creek and Cascadilla boat launches.

But the fraction of oil that gets mixed into the river poses more of a challenge. Davies said that the Forest Service and Glacier Park had suggested the railway look at ways to increase protection of the Nyack, including possibly “a strategy using deflectors that maybe could deflect or steer oil-laden water away from known locations where water infiltrates into the gravels.” Sacca said he had no further information on groundwater protection.

“I frankly don’t know how the hell you’re gonna mitigate [a spill in] a high-volume groundwater-surface water interaction like we have at Nyack,” Stanford said. “The volume of movement is so fast that the contamination of the entire thing within a few hours is really possible.”

If that happens, he warned, “it’ll bleed out of there for years and years and years.” Exactly how many years isn’t known, but both he and Hall noted that, more than two decades after the Exxon Valdez spilled its crude into Alaska’s Prince William Sound, oil was still being found along area beaches.

Oil in the aquifers would bode ill for the nutrient processing that goes on down there, and the ecosystem it supports, Hall explained. “Where any oil goes, it will change the bacteria assemblage...It would for sure affect the animals, and it would definitely affect the bacteria. How, I’m not exactly sure.”

Before these questions move from theoretical to practical, Stanford wants the stakeholders to ”spend a little money on consultation, get heads together,”

“Don’t [let it] be the railroad on one side and us on the other side. Let’s sit down around the table and understand how they might respond and where their deficiency might be.”

In four decades doing research on the Flathead, he reflects, “it has always puzzled me why it becomes sort of an us-versus-them scenario” with the railway.

In an emailed statement, BNSF spokesperson Ross Lane emphasized the measures the company has taken to keep a spill from happening.

“BNSF has invested over $600 million in Montana to improve our physical infrastructure, further reduce risk, and leverage new technology.”

Improvements along the Middle Fork include Positive Train Controls, which use GPS and electronic signals to keep trains running at safe speeds. Data from the Federal Railroad Administration show that in 1997 there were nine railroad accidents in Flathead County. Last year through November there were three, none of them derailments.

And the route’s traffic rises and falls with the oil fields’ fortunes. According to the Montana Public Service Commission, 10 to 18 trains per week were taking this route in 2015. By February 2017, when BNSF submitted its most recent report, it gave a rate of five to 10 per week.

But even if the odds of a crash are down, some researchers and conservationists say that the calculus has shifted now that the Bakken’s boom has turned the Middle Fork into an oil corridor. “We don’t insure based on the likelihood of something going wrong,” said Michael Jamison, senior program manager with the National Parks Conservation Association, “we insure based on the value of the risk.”

For Stanford, the value of understanding what’s at stake can’t be measured.

“Glacier Park and the Bob Marshall Wilderness–Great Bear Wilderness complex is priceless, and Flathead Lake is priceless. It’s the cleanest lake really in the world that has people living around it, and a railroad spill, even of one of those big tank cars, is unthinkable.”

Reporter Patrick Reilly can be reached at preilly@dailyinterlake.com, or at 758-4407.