As the methane leak from a pipeline in Alaska’s Cook Inlet continues unabated, the first water samples show that the water is being impacted, though the extent is still unknown.

The samples showed low levels of oxygen in the water in the area around the leak, and levels of methane high enough to be dangerous to fish. The results hinted at troubling impacts to the water, but few samples were taken and they were not very close to the leak, so they provided an incomplete picture

“Clearly there is some sort of a signal,” said Chris Sabine, a chemical oceanographer with the National Oceanic and Atmospheric Administration. “We need to do a better job of assessing the real potential there.”

An underwater natural gas pipeline has been leaking almost pure methane since late December, and can’t be fixed until ice in the inlet melts, which could still be weeks away. This has raised concerns for regulators and environmentalists, particularly because the area is home to an endangered group of beluga whales.

Hilcorp Alaska, the company that owns the pipeline and the four oil platforms that the pipeline supplies with gas, conducted the sampling on March 18 and 19 by sending four buoys with sensors attached into the area near the leak. It was the first time that any samples have been taken near the leak since it was first reported on February 7.

Hilcorp has been ordered by the federal Pipeline and Hazardous Materials Administration (PHMSA) to stop the leak or shut the pipeline down by May 1. The agency also has ordered Hilcorp to inspect and report on an oil pipeline that is adjacent to the leaking gas pipeline, and which could also be vulnerable to a leak. Both are 8-inch pipelines that are 52 years old.

Regulators and environmentalists are concerned about dissolved oxygen and methane concentrations. Changes in either could pose a problem for the inlet whales, as well as their prey.

Hilcorp reported dissolved oxygen measurements as low as 7.8 milligrams per liter, about one third less than the levels elsewhere in the inlet. And the true level could actually be worse than that, according to Graham Wood, a program manager with the Alaska Department of Environmental Conservation (ADEC).

Because of the methods the company used, the locations of the samples and the small number of samples collected, Hilcorp’s finding “doesn’t represent maximum most probable concentrations from the bubble field,” Wood wrote in a letter to the company.

Sabine was not involved in the sampling but reviewed Hilcorp’s report and ADEC’s response. He said the dissolved oxygen levels are truly dangerous below 2 milligrams per liter. “So it’s above that,” he said. “But clearly there’s something going on if they have measureable decreases in oxygen.” And without having measurements closer to the source of the leak, it’s impossible to know just how much oxygen is in the water.

Sabine said he was ultimately left with a lot of questions. The four buoys that Hilcorp used to take the measurements did not appear to get measurements from close to the leak. He said they can serve only as a snapshot and not a three-dimensional picture of how the plume might be spreading.

The methane readings were also worrisome. In its report to ADEC, Hilcorp wrote that it found “low dissolved methane concentrations consistent with the initial modeling estimates” the company had provided ADEC back in February. But in fact, the maximum levels they reported were three times as high as the model had predicted. The level found, Wood wrote in his letter to the company, “is above concentrations shown to cause adverse responses to fish … based on published data previously shared with Hilcorp.”

One of the few studies that has examined the impact of a methane leak on fish found that the gas can enter through the gills and have almost immediate—and potentially deadly—effect.

In the case of both measurements, the most problematic readings came at roughly 22 feet underwater, not at lower depths. Wood wrote about his concern regarding this in his letter and to Sabine echoed that in his review of the results. Presumably, the highest concentration of methane should be closest to the source of the leak, which is occurring 60 feet deeper on the inlet floor. And yet Hilcorp’s deeper readings were less abnormal.

Sabine pointed out that the samples are just a first effort and subsequent samples could provide a more complete picture. Hilcorp has been taking more samples and will be filing weekly reports with ADEC.

“For a first try, I think it was a reasonable effort,” he said.