The conditions that night were a perfect storm -- foggy, low cloud cover, an early fall evening that was right for flight.
How many birds flew past the flare on a recent September night at the Canaport LNG plant, north of Maine in New Brunswick, Canada, isn't known.
But managers at the liquefied natural gas import facility estimate that about 7,500 migrating birds, mostly small songbirds, were drawn into the flare and died from its heat.
The issue of wind energy-related bird deaths has received wide publicity with extensive news reports and some studies stemming from the realization in recent years that deadly bird strikes at wind turbine farms can add up. But there is little tracking of avian mortality at natural gas facilities.
In the Canaport case, officials said that despite regular natural gas flaring at the facility to keep equipment pressures in a safe range, the mass bird kill was unprecedented.
"This is the first occurrence of its kind at Canaport LNG; we have never had this happen before," spokeswoman Kate Shannon Marie said. "Once a root cause of this incident has been determined, we will be better able to mitigate the risks."
Shannon Marie said the company has just brought online a two-year, $45 million upgrade to reduce flaring at the terminal by adding two compressors, a retrofit that was not yet in place in mid-September when the incident took place.
"These compressors will allow our facility to send the excess gas to the pipeline instead of to the flare, reducing the visible flare and its impact on the environment, significantly preventing the risk of this occurring again," she said.
An initial assessment of the singed carcasses strewn around the facility afterward found mostly songbirds -- vireos, several types of warblers, a few thrushes and some grosbeaks, the head of zoology at the New Brunswick Museum, where many were taken, told CBC News -- but did not immediately identify any threatened or endangered birds among the casualties.
Andrew Farnsworth, a research associate at Cornell University's ornithology laboratory, said there has been little study of bird mortality from oil and gas development, despite the fact that rigs and other facilities are, under certain conditions, a predictable source of harm to birds.
"People understand very well from experience on their own that moths are attracted to light. ... It's not that surprising that other animals are, too," Farnsworth explained. "It's a very ancestral response to light. Light is a beacon; light is an attractive force."
Farnsworth said that a natural gas flare stack -- or a communication tower, skyscraper, or any other facility that emits light at night -- can draw birds in and lead to collisions or dangerous disorientation. That is especially likely when other navigational cues they use, such as sight and sound, are obscured by bad weather conditions.
"When there's poor visibility, there are fewer things available to them to help them stay on track," he said. "Especially when there's light combined with poor visibility, whether it's lights from natural gas, or lights from a wind turbine, or lights on the Empire State Building, or from a city, that's a very dangerous situation for birds," he added. "There [have been] some pretty horribly gruesome incidents."
Farnsworth said an incident with 7,500 birds killed ranks as significant but far from the largest mass mortality ever seen. Anecdotal reports suggest some events killing off tens of thousands to a hundred thousand birds, even, though details are shoddy.
Pointing a finger
The extent to which such incidents matter, whether measured in single numbers or thousands of birds, is a subject of debate, Farnsworth conceded.
The biggest concern would be for birds that are endangered or threatened already, for which a single event of the magnitude seen in New Brunswick "could, conceivably, have population-level effects," he said. "In that sense, it's possible that an event like this could be really devastating."
At Canaport, the most abundant of the species killed appear to have been birds with healthy populations.
Farnsworth said in that case, the question becomes one of how often such events are taking place, and whether they cluster along, say, a particular species' major migration route.
"It's definitely still something, in my mind, that requires extremely careful attention and diligence and monitoring," he said. "Things could get out of control quickly."
In the context of human-linked bird mortality, such light-related dangers are one hazard among many. A 2005 study by the environmental consulting group Western EcoSystems Technology Inc. estimated that between 500 million and a billion birds are killed annually in the United States from "anthropogenic sources," including collisions with vehicles, buildings and windows, power lines, communication towers, and wind turbines; electrocution; oil spills and other contamination; pesticide exposure; cat predation; and commercial fishing bycatch.
"Many of those deaths from these sources would be considered unlawful" under federal laws, including the Endangered Species Act, the Migratory Bird Treaty Act and the Bald and Golden Eagle Protection Act, the authors wrote.
Farnsworth said it's important to keep those overall figures, and mortality sources, in mind when considering events like the one at Canaport.
"I don't think it's right in this case to point a finger at any specific industry, because the problem isn't with the industry itself. It's with anything that's lit up at night," he said. "The issue is light and birds. It's not natural gas and birds, or wind and birds."
Contacted for this story, some major bird-related interest groups had little knowledge of how natural gas facilities might affect avian populations.
Given the lack of systematic data collection on the issue, that may not come as a surprise. Farnsworth said there have been some accounts of bird-kill events from Midwestern and Rocky Mountain West states with oil and gas activities, but no comprehensive analysis has been published.
The limited literature on oil- and gas-related bird kills points to offshore installations as a particular hazard for migratory birds. Mass burn casualties have been sporadically documented for decades at offshore oil and gas platforms on the North Sea and on the Gulf Coast. Emerging technologies allowing extraction in more challenging offshore environments could increase the number of such encounters.
Onshore, new oil and gas development activities popping up in Ohio, Pennsylvania, North Dakota and elsewhere could have notable impacts, too, with around-the-clock drilling schedules leading to new light sources in rural and semi-rural landscapes, and high flaring rates in areas where natural gas transportation infrastructure fails to keep up with production.
Proposed new LNG export facilities could raise similar concerns. Canaport's case, in which the addition of new compressors to the facility is expected to significantly reduce flaring, underlines the extent to which the equipment configurations at individual terminals could affect those outcomes.
Daniel Donovan, a spokesman for Dominion Resources Inc., which owns the Dominion Cove Point LNG import terminal near Lusby, Md., said that as currently configured, the plant does not flare natural gas. In the case of an emergency, natural gas would be released by venting it, without a flame, allowing the lighter-than-air gas to escape upward, he said.
Dominion is proposing to redevelop Cove Point for LNG exports, a complex and expensive undertaking requiring extensive new construction as well as new safety analyses. Donovan said that during the liquefaction process before export, other hydrocarbons are present that are heavier than air. In an emergency release, those would be burned as a flare to prevent pooling at ground level, he said. Normal shutdown operations could also require some flaring of those hydrocarbons, Donovan acknowledged, though he said the extent of it would be limited.
The Federal Energy Regulatory Commission is tracking four proposed or potential LNG import terminals in the United States and 21 proposed or potential export terminals in the lower 48 states. While no one expects all of those facilities to be financed and constructed, any that are ultimately built could be occasional sources of natural gas flaring.
The point that Cornell's Farnsworth makes, that the oil and gas industries should not be singled out for their avian mortality impacts, is important to keep in mind. Feral cats, he points out, are potentially a vastly more significant source of bird deaths.
But he suggests that more complete study is needed on the issue, as well as efforts to bring science, policy, conservation and economics together around questions of how and at what cost businesses can reduce avian impacts under circumstances in which mass bird-kill events are particularly likely.
"Whatever the number is [of annual bird deaths] linked to natural gas pales in comparison with what feral cats do," Farnsworth said. "Does that mean we should just forget about it, because it's just a drop in the bucket? I don't think so. Does it have the legs to be a huge problem, like cats or other sources? I don't know."