Manhattan Is the Wrong Unit

Manhattan is the wrong unit. Not imprecise — wrong. Area is not the quantity that determines whether this event matters, and we keep reporting it as though it were.

Start with the thing that sounds like a correction and isn't quite one. The Petermann ice tongue was already floating. It had already displaced its water, in the bathtub sense, for as long as it had been afloat. Its departure raises global sea level by approximately nothing. Archimedes worked this out in a tub in the third century BC and the arithmetic has not been repealed. So the iceberg itself — the photogenic part, the part with the area — is causally inert.

Which is why the standard next move is to invoke buttressing. The floating tongue is wedged into a fjord; it drags along the rock walls; that shear pushes back up the glacier and holds the grounded ice behind it. Grounded ice is the part that counts, because it is currently resting on bedrock above sea level and would very much like not to be. Remove the cork, the argument goes, and the bottle empties.

Except Petermann already ran that experiment. Calving events in 2010 and 2012 took roughly 50 to 60 percent of the floating tongue between them. If the cork model were right, the glacier should have lurched. It barely moved — modeled speedups of a few tens of meters a year out on the tongue, nothing propagating inland, the entire response smaller than the seasonal variation the glacier does on its own every year.

So the honest quantity is neither area nor total tongue length. Modeling of this specific glacier points to something less quotable: distance from the grounding line — the hinge where ice stops resting on rock and starts floating. The thin, soft ice more than twelve kilometers out front does almost no buttressing work, which is exactly why the big calvings barely registered: discharge across the grounding line went from 10.12 gigatons a year to 10.26 after 2010, and 10.47 after 2012. Take ice from within twelve kilometers of the hinge and the picture changes — inland flow speeds jump by something like 900 meters a year, and stripping the tongue entirely drives discharge toward 25 gigatons a year, roughly two and a half times baseline. That is the range with sea level in it.

Say plainly what that threshold is, though: a model output. The 2010 and 2012 events killed the cork story by observation — they happened, the glacier shrugged, argument over. Nothing has ever calved inside twelve kilometers of Petermann's hinge, so nothing has ever tested the number that replaced it. That is not a reason to discount it. It is the reason to watch the distance.

So I went looking for the distance. Here is what that search was actually like. It is not in the announcement. It is not in the wire copy, the aggregator rewrites, or the photo captions — all of which found room for Manhattan. It is in a Science Advances paper on Petermann published this month, in a subordinate clause: once the calved ice clears, the glacier's terminus will sit within 40 kilometers of the grounding zone, its farthest upstream position observed since 1923.

Forty kilometers, against a threshold of twelve. Which makes the honest reading of this event the deflating one. The iceberg did not matter. The two sections queued behind it — 94 square kilometers and 84 square kilometers, both already fractured and expected to go soon, together with the one that left amounting to about 22 percent of the remaining tongue — will not matter either. Even after all three, the front stays well outside the zone where calving reaches grounded ice. I would rather say that than borrow an urgency I can't price.

But farthest upstream since 1923 is the part with a direction in it. A century of observation, and the front has never been this far back. Each departure moves it closer to the hinge, and the hinge itself has been moving the other way: grounding-line retreat and fresh fracturing were documented at Petermann beginning after 2016, with the grounding zone migrating kilometers back and forth with the tides. Both ends of the measurement are walking toward each other.

That is not an event. It is a ratchet — slower than the coverage implies, and pointed somewhere the coverage never mentions.

None of the fractures were a surprise. Sentinel-1, the European Space Agency's radar mission, has passed over this fjord every few days for a decade, seeing through cloud and indifferent to weather, and the centerline cracks have been widening in the returns for years. Imagery from August 3 showed clear deterioration. Then it went.

The way our species found out was this: Adam Garbo, a PhD student in glaciology at the University of Ottawa, was looking at the pictures. On purpose. As a habit. Working with colleagues at Stirling, Lancaster, Leeds, and Environment and Climate Change Canada, he identified the largest Arctic calving event since 2020 by noticing it in grayscale radar returns from a place almost nobody will ever stand.

There is no alarm on the fjord. Nothing rings. Northwest Greenland will keep doing this in front of an audience of exactly whoever bothers to watch.

But watching turns out not to be the bottleneck, and that is the thing I had wrong walking in. Somebody did measure the distance to the hinge. It is sitting in a journal, published the same month as the event. It simply did not travel — and the question worth asking is why Manhattan travels and forty kilometers doesn't.

Because area is the only quantity in this story that is visible in the photograph. You can see 76.4 square kilometers. It has an outline, it detaches, it drifts, it looks like something, and it fits in a caption. Distance-from-the-grounding-line has no image at all: the hinge sits under roughly 600 meters of ice, inferred from how the surface flexes with the tide in radar interferometry, and it slides back and forth twice a day. It cannot be photographed, so it cannot be captioned, so it does not enter the public record of what happened.

The wrong unit isn't chosen. It's the one that survives compression. And the arrangement runs well past this fjord: at nearly every ice margin on Earth, the causally inert thing is the photogenic one and the causally decisive thing is invisible. What can be pictured gets reported, what gets reported gets funded, what gets funded gets measured. That ordering is backwards, and nothing in the fjord is going to correct it.

The island broke free. Fine. The question is how close to the hinge the next one goes — and now you know why nobody will lead with the answer.