By 2009, NOAA researchers had pieced together circumstantial information about Wyoming's worsening winter ozone events. Ozone spiked when temperature inversions trapped and concentrated pollutants near the valley floor. Extensive snow cover made a difference, too: Sunlight reflecting off it created enough radiation, scientists theorized, to set off the reactions that create ozone. In winter 2010, startlingly high ozone levels were documented in Utah's Uintah Basin under similar conditions.

In both places, booming natural gas fields seemed the source of ozone's primary ingredients: nitrogen oxides from diesel trucks and engines that run compressors and other equipment; and volatile organic compounds, present in the gas itself.

Knowing the source of emissions is enough to devise remedies for many pollution problems. Take sulfate particles, another contributor to smog. Power plants emit sulfur dioxide, which creates sulfate. When plants install scrubbers that capture most sulfur dioxide emissions, you get fewer sulfate particles in the air.

Ozone is much more complicated. "It's a weird beast," says Leonard Herr, the Bureau of Land Management's air-quality specialist for Utah. "You can't just control one thing." Nitrogen oxides (NOx) and volatile organic compounds (VOCs) are the basic building blocks, but you also need reactive atoms or molecules called free radicals -- created when sunlight reacts with things like nitrous acid, a byproduct of NOx, or ozone itself -- to jumpstart the process, the way a spark makes fire out of fuel. But there's no universal recipe; which free radicals react with which VOCs and NOx differs from region to region. Moreover, some environments are sensitive to NOx, meaning that if it's reduced, ozone will be too. Others are sensitive to VOCs. It's even possible to reduce NOx in a VOC sensitive environment, and end up slightly increasing ozone.

So managing ozone requires extraordinary prudence, plus a nuanced understanding of the chemistry that creates it in each locale. Neither Utah nor Wyoming has reached that level of understanding.

The NOAA team intended to change that. For six weeks last winter, they worked out of a sort of man camp in the heart of the gas field. A rental RV -- housing for grad students monitoring a balloon measuring pollution above the surface -- sat alongside trailers and what looked like miniature shipping containers, jammed with instruments measuring pollutants. Wind sensors helped the researchers see how plumes moved. And a team toured the field daily in the van, collecting real-time emissions data from each link of the production chain. If the chemists in the temporary labs could figure out which pollutants were the ozone conspirators, the van could help identify their origins.

Only two things were missing: snow and inversions. The scientists charged with unraveling the ozone riddle and recommending fixes found themselves in the ironic position of wishing for dirty days for the sake of science. But they never materialized. And so, there was no ozone to study.

"We're kind of stuck halfway right now," Roberts says. They did collect some information: They ruled out a nearby coal-fired power plant as a major contributor; although it spews plenty of NOx, its plume travels above the level of inversions. And they believe wintertime ozone has been seen only in the Utah and Wyoming basins because of the way topography restricts air movement during inversions. VOC emissions were found to be especially high from evaporation ponds and during flowback -- the period after a well is hydrofractured when the fracking fluid, gas and other hydrocarbons are regurgitated. The compounds found aren't typically very reactive in small amounts, says Roberts, but their concentrations were very high. NOx levels were sufficient to create ozone, but not particularly high.

"We think there may be some chemistry happening in the snow itself," he says, but it couldn't be studied since there was no snow.

They suspect that this year, with no ozone events, was the anomaly, not the previous two, when snow cover was thick and ozone levels were frequently high. So NOAA, the state and their partners hope to procure funding to study the basin again next year. "The emissions inventory work, gathering fingerprints of sources was successful," says Brock LeBaron, deputy director of air quality for the Utah DEQ, and the state's point person for the study. "Still, the fundamental question is: Should you go after the VOCs, or should you go after NOx?"

Without an answer, it's hard to draft a sure-fire mitigation plan. In the meantime, though, regulators in both Wyoming and Utah are "casting a big net" when it comes to pollution controls, as Wyoming's air quality chief, Steve Dietrich, puts it. "We'd be remiss if we concentrated on VOCs and didn't concentrate on NOx. So we're trying to reduce both of them."

"Green completions," which capture hydrocarbons and VOCs during flowback, are required throughout the Upper Green River Basin, and for new projects in the Uintah Basin, where such projects have been approved only if they install technology such as low-bleed valves to control VOC leaks from well-pad equipment; minimize truck trips; use clean-burning engines at drill rigs; and add any additional controls deemed necessary as regulators' understanding of local ozone chemistry advances. The state and BLM also hope to work with industry to ensure no net increase in emissions, beginning next winter. Wyoming drillers have taken similar steps. Most of these measures aren't legally required, but the industry and states have been trying to avoid becoming 'non-attainment' zones for ozone pollution, an EPA designation that can bring mandatory -- and probably expensive -- regulations, and could limit new energy development.

Since 2009, the industry estimates that it's reduced VOC emissions by 21 percent, and NOx by 17 percent, in the Upper Green River Basin. Yet despite these emissions cuts, the problem persists. The EPA just declared the Upper Green River Basin a non-attainment zone, underscoring the need for better information to shape a strategy. Utah lacks such statistics because it's just begun requiring NOx and VOC controls of big new projects in development.

"The Uintah Basin has a couple years, probably, where we can try some out-of-the-box proactive strategies,"before the EPA enters the fray, says the BLM's Herr. Still, in that short time,  he says, "It's kind of a long shot that we'll be able to solve it."