Nine Years of Nothing, Then a Signal

Tromsø —

The instruments went down in 2015. They were recovered, serviced, and redeployed every eighteen months, at considerable expense, in weather that permits the operation about four months a year.

For nine years the array recorded, in the words of its principal investigator, “exactly what we expected, which is the worst possible outcome for a grant renewal.”

In the autumn of 2024, it recorded something else. The finding has since been replicated by two other groups and has changed a modeling assumption used across the field. It could not have been detected in under a decade of continuous observation, and the project came within one funding cycle of not existing that long.

The structural problem with null results

Nothing about this is unusual, which is the point. Long-baseline observation is the only way to detect changes that occur slowly, and it is systematically disadvantaged by every mechanism used to allocate research funding.

• Grant cycles are typically three to five years; the phenomena of interest are not.
• Interim reports on a monitoring project describe continued operation, which reads as an absence of progress.
• Publication incentives reward findings, and nine years of stable measurements produce very few.
• Instrument maintenance is capital-like spending that most schemes classify as operational and therefore discourage.

How the project survived

Not gracefully. Veyra reviewed the project’s funding history across eleven years. It was sustained by four separate grants from three bodies, two periods of institutional bridging finance, and one year in which the ship time was donated by another project that had spare capacity on a transit leg.

The 2021 gap is the one everyone remembers. Funding lapsed in March and did not resume until the following January. The instruments stayed on the seabed, unserviced, past their design interval. Four of the eleven nodes were lost. The data from that period carries a caveat in every publication that uses it.

Had the recovery vessel not been available in early 2022, the entire nine-year baseline would have been broken, and the 2024 signal would have been uninterpretable — a change measured against nothing.

What the field is trying

Several funders have introduced longer instruments — ten- and fifteen-year awards specifically for observational infrastructure. They remain a small fraction of total allocation, and researchers we spoke to described a second-order problem: a fifteen-year award concentrates risk, because losing one is catastrophic rather than inconvenient.

The alternative proposal, favored by most of the people who actually maintain arrays, is duller and probably better. Fund the maintenance separately from the science. Treat a monitoring array the way a country treats a tide gauge: as infrastructure that exists, with a budget line that does not have to justify itself with a discovery every three years.

“We built something that can see slowly,” the investigator said. “Then we asked it to prove itself quickly. That is not the instrument’s fault.”