San Francisco, United States · America/Los_Angeles
Fort Point
A cobble ledge that only works in a real swell, and only with enough tide to keep it from grinding.
15:00 local · 2026-10-04
26% — the read for this hour
Swell
0.48 m
Period
7 s
Swell from
252°
Wind
3.3 m/s
Wind from
258°
Air
20°C
Stay home. At Fort Point on this hour the read is 26%. What earns it: peel speed at 20.8 km/h. What costs it: tide window at 3.42 m over the take-off.
2 hours worth driving for today: 00:00, 01:00.
Tide window for a reef: 0.6–1.4 m of water over the take-off; below 0.40 m it wants the reef just covered matters and the reef is exposed.
Dashed line is the tide above MLLW, solid line is the swell height. Column height is the engine score for that hour. Faded columns are gated. The hour you select is in the URL, so this page is shareable as-is.
Tide station
- Station
- San Francisco
- Datum
- MLLW
- Range
- 1.82 m
- Mean
- 0.97 m
Signature interaction
Drag the tide and watch the reef change its mind
The cross-section is drawn from the break's own reef slope. Every change re-runs the shared engine on the server and returns a fresh seal — this control cannot work without the API.
The dashed arrow is the peel, at c = √(g·depth). More water over the take-off means a faster peel, which is why the same swell behaves differently at low and mid tide.
Re-solved by swellread-engine/2026.10.1
- Water over take-off
- 0.37 m
- Break height
- 0.28 m
- Peel speed
- 20.8 km/h
- Wave power
- 1.0 kW/m
Stay home. At Fort Point on this hour the read is 26%. What earns it: peel speed at 20.8 km/h. What costs it: tide window at 3.42 m over the take-off.
Every factor, re-weighted for this tide
- Swell power1.0 kW/m · 1.0%
- Peel speed20.8 km/h · 19.0%
- Wind quality3.3 m/s straight onshore · 0.1%
- Tide window3.42 m over the take-off · 0.0%
- Direction match50° off the 302° peel · 4.2%
- Period cleanliness7.0 s · 1.7%
Show your working
Every factor, its weight, and the arithmetic behind it
Weights are renormalised over the factors that have real data behind them. A factor with no data is dropped and said so, never quietly scored at zero.
- Swell power1.0 kW/m · 1.0%
weight 0.220
- Peel speed20.8 km/h · 19.0%
weight 0.200
- Wind quality3.3 m/s straight onshore · 0.1%
weight 0.200
- Tide window3.42 m over the take-off · 0.0%
weight 0.180
- Direction match50° off the 302° peel · 4.2%
weight 0.120
- Period cleanliness7.0 s · 1.7%
weight 0.080
engine seal 7b904740e578327dfe7732a982560d871bd835e04f98a6c918100c2ff99d760c275ba7abda668db6677aef667aa8d339
Swell power
P = 1/16 · rho · g · Hs² · Tp = 0.628 · 0.48² · 7 = 1.0 kW/m, against a 22 kW/m "serious day" reference.
Peel speed
c = sqrt(g · depth) = sqrt(9.81 · 3.42 m) = 20.8 km/h over a 3.42 m take-off, with 0.28 m standing up. Full marks 8.5–19 km/h: slower is mush, faster than ~25 km/h is unridable.
Wind quality
Wind is straight onshore at 3.3 m/s. Alignment = 1 − (1 + cos 6°)/2 = 0.00; speed envelope at 3.3 m/s = 0.95. Product 0.00.
Tide window
reef wants the reef just covered. Ideal band 0.6–1.4 m of water over the take-off; below 0.40 m the reef is exposed.
Direction match
Swell from 252°, this break peels cleanly from 302°. 50° of error: inside 12° is full marks, past 75° the swell wraps the headland and does not arrive.
Period cleanliness
Tp = 7.0 s. Energy organises between 6 s and 9.5 s; beyond 14 s the swell stands up steeper than a shallow reef can hold, so the score is eased off by up to 35%.
Provenance
Every number above came from one of these
- liveOpen-Meteo Marine APICC BY 4.0, non-commercial attribution requiredfetched 2026-10-04 20:28 UTC168 hourly rows, swell and wind wave.
- liveOpen-Meteo Forecast APICC BY 4.0, non-commercial attribution requiredfetched 2026-10-04 20:28 UTC168 hourly rows, wind and air.
- liveNOAA CO-OPS tide predictionsU.S. Government public domainfetched 2026-10-04 20:28 UTC144 hourly predictions, MLLW datum, station 9414290 (San Francisco).
Snapshot generated 2026-10-04T20:28:09.517Z. Break orientation, reef slope and take-off depth come from this project's catalogue and are labelled as estimates, not survey data.