How it works

Every status here is a best guess from data. None of it says a slope is safe or in shape. Late-season snow means hard ice, moats, rockfall and running water.

Satellite

Sentinel-2 passes over Colorado every few days. For each snowfield we read a small window of 20 m pixels (0.04 ha each) and count the ones that look like snow: bright, much brighter in green than in shortwave infrared, and not dark in near infrared (that's how water gives itself away). ESA's cloud mask decides whether a pass counts.

• Cloudy is not zero. A pass that is cloudy over a snowfield, or only partly covers it, is counted as a “no observation”. You can see the pass, but it doesn't update the status.
• One zero isn't gone. North-facing cirques sit in shadow by September. A snowfield reads Gone only after two clear zeros in a row. The tricky part here is that often, the satellite can't “see” it because it's covered in rock and debris by late summer. I mark these as best I can, but the best option is to go have a look!
• New snow is tricky. When many snowfields jump at once, that date is flagged as a likely storm and status falls back to the passes before it.
• Small things are invisible. A skiable strip narrower than about 40 m can read zero. This is common! Some very skiable (small) patches just don't show up.
• Nothing goes live until it's published. I usually review each scan run before the site shows it.

Ground reports

Members can post what they saw or skied. A report on a tracked snowfield is shown first for 10 days, then the snowfield falls back to the satellite. Reports always say who and when.

New snowfields

A sweep of the whole range looks for snow patches that aren't tracked yet. I checks each one before it's added.