The Spread

Surface analysis

Valid now.

GFS's own initialization at 12Z 30 Aug — sea-level isobars every 4 hPa from the model's membrane reduction (the one NCEP's charts contour), highs and lows marked with the field's own values, over the rate at which the model has precipitation falling at that same moment. The frame is the model's full grid, coasts and neighbours included, because an offshore low is still weather. One model's analysis, drawn straight from its grid: no consensus, no correction, and it says so rather than wearing the marker map's roster.

10081008100810081012101210121012101210121012101210121016101610161016101610161016101610201020102010201024 H 1015 L 1005 H 1015 L 1003 H 1016 H 1024 L 1014 H 1023

Unshaded ground is where the model has under 0.01 in/h falling. The heaviest rate on this map is 0.325 in/h, near 33.0N 69.5W — read from the unsmoothed grid, because smoothing fills extremes, so it can sit above the top band the shading actually draws. These are grid-cell averages over about 28 km, which is why they run well under the rates a gauge records: the words “light” and “heavy” are defined for a point and do not transfer, so the key is numbers.

Lines are lightly smoothed for legibility; the number beside each H and L is read from the unsmoothed field. No fronts — the model publishes a field, not an analysis with features drawn on.

Every map on this page is drawn from model output we store ourselves, not from anyone's picture of it. The per-city read of the same day — which cities the models are fighting about, and what the discussion makes of it — is on the national page.