DC Forecast Discussion

The stage the forecast plays out on: a model surface analysis rebuilt live in your browser — air masses, fronts, highs and lows, wind. Scrub the timeline to watch the pattern evolve; hover anywhere for numbers. The note under the map reads the pattern and says why it is (or isn't) raining at DC.

The story LWX is telling, the atmosphere behind it, and the hour-by-hour at the field. Their full discussion is at the bottom — dotted terms carry plain-English tooltips.

A forecast never stands still — each new discussion quietly rewrites the last. The drift card puts today in the middle: three days back, what KDCA actually recorded against what was called for; three days ahead, how the numbers have moved since the earlier forecast. Below it, word-level diffs (added / removed) show the reasoning itself being revised — open an issuance to read one. New issuances are picked up every 10 minutes.

Forecast windows close at different times, so each is checked when it closes and the rest stay listed as open — at noon that's last night's low and the morning's ceilings, not a verdict on storms that haven't fired yet. Day rows are the DC forecast against KDCA's METARs; field rows are the KANP hourly grid against KNAK, the field's own sensor (KDCA for days before that stream existed, and for thunder always). Ceilings are checked by height band, not just flight category — a 3,000 ft VFR deck and a clear sky are different forecasts. When something closed and missed, the model's own hindcast says where it went sideways. The ‹ › arrows step back through earlier days, every window closed.

KDCA's pressure trace and the weather it brought — rain, storms, fog and gusts marked where the METARs recorded them, LWX warnings along the top. Falling pressure means a front or low approaching; rising means high pressure building in. The amber area is hourly GFS CAPE at the field, on its own scale, so a ⚡ over a tall amber peak is fuel spent, and a peak with no ⚡ is a lid that held. All read from this site's own archive.