Live monitoring of a slow process

Is the tail being realised?

The downstream package puts a number on the worst case for downtown Miami: under IPCC AR6 SSP5-8.5 low confidence, sea level at Virginia Key reaches 3.43 ft at 2100, with a 95th percentile of 8.39 ft. That p95 is not thermal expansion and it is not Greenland. It is almost entirely marine ice sheet instability in West Antarctica. This page watches the upstream end of that number.

01
The tail
p95 of 8.39 ft at Virginia Key in 2100. Published, sourced, and already in the dataset.
02
The mechanism
Warm Circumpolar Deep Water thinning Amundsen Sea ice shelves, removing buttressing from grounded ice.
03
The watch
Stated tripwires, set in advance, with what each would and would not mean.
04
The consequence
What that stage does to Brickell, the CBD, and Edgewater — building by building.
Not a specialist? This page assumes some background. The plain-language guide explains sea level rise, ice shelves, and El Niño from scratch, with the common misunderstandings corrected, and takes about five minutes.
The trap this page is designed around. There is a latency-versus-significance inversion in cryosphere data. What updates fastest here — sea ice extent, daily — contributes nothing directly to sea level, because sea ice is already floating. What matters most — ice sheet mass — is monthly, and is currently unavailable to this monitor at all. A wall that put those side by side without saying so would be lying by layout. Every tile therefore carries its observation date, its lag, and whether it raises sea level.

Indicators

Fetched on a six-hourly schedule. The lag shown is the gap between what the instrument observed and now — not how recently this page was refreshed.

loading…

Tripwires

These are not consensus scientific thresholds. They are this project's pre-registered positions on what would count as evidence, published in advance so they cannot be quietly moved afterwards. Each states what it would mean and — equally important — what it would not.

loading…

El Niño, and why the sign is a trap

ENSO does not act on Antarctica with a single sign. During a strong El Niño, Amundsen Sea ice shelves tend to gain height from increased snowfall while losing mass to warmer Circumpolar Deep Water at the base. The two signals oppose each other.

A monitoring dashboard built on satellite altimetry height would therefore show those shelves thickening during exactly the episodes that are thinning them. This is the clearest example on the whole site of why a confident number can be worse than no number. The result traces to Paolo and colleagues, Nature Geoscience, 2018 — cited here from recollection and flagged for verification before it is relied on.

Visual observation

Imagery is where most people actually check on the ice, so it belongs here — with its limits stated.

MODIS Terra true-colour satellite image of the Jakobshavn Isbrae outlet region, west Greenland
Jakobshavn Isbræ, west Greenland. MODIS Terra true colour. Northern summer, so optical imagery works and is refreshed daily. Cloud is common; a white frame is usually weather, not ice.
Not available today
It is polar night over the Amundsen Sea. True-colour imagery of Thwaites and Pine Island returns a black frame from roughly April to September.
Thwaites & Pine Island, West Antarctica. Seeing through polar night needs radar or passive microwave. I could not get a working embedded microwave layer from the snapshot API for this region, so this is a stated gap rather than a filled panel — open it directly in NASA Worldview instead: Worldview, Antarctic view ↗

Source health

What this monitor can and cannot see, stated as a count. A source we cannot reach is published as a gap, never dropped.

Raw snapshot: /api/ice · history: /api/ice/history