What a Provisional Surf Rating Means (and Why Admitting Uncertainty Helps)

What a Provisional Surf Rating Means (and Why Admitting Uncertainty Helps)

A provisional surf rating means we haven't had a person check the spot against reality yet, so the model runs but caps what it will claim and stays out of alerts. It's a badge, not a score change. You still get a number 0–5 for your level, but with a warning: trust it less, and don't set an email alert on it. Honest uncertainty is the point.

Quick version, before the reasoning:

  • Provisional = the spot's model runs, but no human has verified it against real sessions.
  • Provisional spots carry safety caps on what the rating will assert.
  • They're excluded from email alerts — we won't wake you for a number we don't trust.
  • The badge is per spot, and it comes off once the facing, swell window and tide behavior check out in the water.
  • Where we've been wrong before, that drift is published too — the accuracy panel doesn't hide.

Why does a surf forecast need a "we're not sure" state at all?

Because the coast is bigger than the measurements. Portugal's mainland has roughly 943 km of Atlantic-facing coast (Instituto Hidrográfico), and the national in-situ wave network is small against that. The IH Datawell Waverider network runs monitoring nuclei at Faro, Sines and Leixões on the mainland (Instituto Hidrográfico). Most of that 943 km has no buoy directly offshore.

So national wave forecasts are modeled, not measured everywhere. IPMA's wave forecasts run on the HRES-WAM numerical model from ECMWF (IPMA) — a model, not a reading at every point. That's fine. Models are good. But a model at an unreviewed spot is one step further from the water than a model at a spot someone has surfed and corrected. The provisional badge marks that extra step honestly instead of pretending it isn't there.

What actually makes a spot provisional on Peel Watch?

Every spot has a model behind it: a coastline facing measured off OpenStreetMap, a hand-authored swell window with a shadow factor for what wraps in, wind tolerance, shelteredFrom sectors, tide preference and per-level skill gates. That's built for all 120 spots — 100 beach breaks, 12 rivermouths, 4 reefs, 4 points, typed separately because a reef and a beach break don't behave alike.

Provisional means the model exists but hasn't been checked by a person against real sessions at that spot. The facing might be right on paper and wrong in the water because a sandbank moved. When that's the case, the badge stays and the rating is capped. I'd rather show a spot as provisional for a season than hand you a confident 4.2 I can't stand behind.

Why should a moving seabed make you trust static spot data less?

Because the ground under the wave doesn't hold still. About 50% of the low, sandy coastline of mainland Portugal is actively eroding (APA / national coastal data), which reshapes sandbanks and beach profiles over time. A beach break's whole character sits on that sand.

So a swell window authored two years ago can drift as the rebentação shifts down the beach. This is exactly why the provisional badge isn't a permanent label and why review isn't a one-time job. Reefs and points move less; beach breaks move most. When a spot's shape changes and the model hasn't caught up, provisional is the honest state to be in.

One buoy can't speak for the whole coast

Here's the cleanest illustration I have. In the early morning of January 28, 2025, the IH MONIZEE network recorded a maximum wave height of 17 meters at the Nazaré coastal buoy, while other stations recorded very different conditions (Instituto Hidrográfico). Same morning, same country, wildly different water.

If conditions vary that sharply between monitored points, imagine the gap between a monitored point and a spot with no buoy and no human review. That's the whole argument for the badge in one number.

Fact Figure Source
Mainland Atlantic coastline ~943 km Instituto Hidrográfico
Mainland IH monitoring nuclei 3 (Faro, Sines, Leixões) Instituto Hidrográfico
Sandy mainland coast eroding ~50% APA
Nazaré buoy max, Jan 28, 2025 17 m Instituto Hidrográfico

How does admitting uncertainty actually help you plan a session?

It tells you where to spend your trust. A reviewed spot with a 4 for your level is a number you can drive to. A provisional spot with the same 4 is a maybe worth a look — go, but read the water yourself when you arrive.

The alerts follow the same logic. Provisional spots are excluded from email alerts on purpose, so we never send you 45 minutes down the coast on a figure we've flagged as unverified. And the confidence and consistency numbers on each reviewed spot come from historical climatology, not vibes — so you can see how reliable that break tends to be, not just today's score.

One more thing worth saying plainly: forecast drift exists, always. When you're comparing two spots, a small gap in score is noise, not signal. Treat close numbers as a tie and pick the one with the better tide state or the shorter drive.

Same water, three readings — and the badge matters for all three

I surf, I swim distance year-round, and I fish. I read one forecast three ways. A surfer wants period and ondulação direction. A swimmer wants the opposite — calm, low rebentação, no agueiro running. A fisherman wants the tide state and the structure at a specific hour. Peel Watch scores swimming separately on a calmness scale for exactly that reason, and four beaches are marked swim-only.

The provisional badge protects all three readings. If a spot's model isn't verified, I don't want a swimmer trusting a low calmness score any more than a surfer trusting a high surf score. Uncertainty applies across the board, so we flag it across the board — including hazards, which are listed on all 120 spots whether the surf rating is reviewed or not.

Where to go and apply this

Open the map, filter by your region or use the near-me radius, and look at the badges alongside the scores. Reviewed spots you can plan on; provisional ones you check with your own eyes. Peel Watch is free, passwordless, and there's nothing to buy — I built it because the old forecasts wouldn't tell me why. Start on the dashboard and set an alert on a spot you actually know.

Common questions

Does a provisional badge mean the surf rating is wrong?
No. It means no human has verified that spot's model against real sessions yet. The rating still runs, but it carries safety caps and stays out of email alerts. Treat it as a starting point rather than a number to drive to, and read the water yourself when you arrive. The badge comes off once the spot checks out in real conditions.
Why doesn't Peel Watch just show a confident number for every spot?
Because most of Portugal's ~943 km mainland coast has no buoy directly offshore, and forecasts are modeled rather than measured everywhere (IPMA runs the HRES-WAM model). A model at an unreviewed spot is a step further from the water than one at a spot someone has surfed and corrected. Flagging that honestly serves you better than a false 4.2.
Do provisional spots trigger email alerts?
No. Provisional spots are deliberately excluded from email alerts. The idea is simple: we won't send you down the coast for a score we've flagged as unverified. Alerts fire only on reviewed spots, per your chosen level and threshold, one to three days ahead.
Why does a moving seabed affect surf ratings?
Beach breaks sit on sand, and about 50% of mainland Portugal's low, sandy coast is actively eroding, which reshapes sandbanks over time. A swell window authored a couple of seasons ago can drift as the break shifts down the beach. Reefs and points move less; beach breaks move most, which is why review isn't a one-time job.
How big is forecast drift, and how should I read close scores?
Forecast drift is real on every model. When two spots sit within a small margin of each other, treat that gap as noise rather than a meaningful difference. Pick the spot with the better tide state or the shorter drive instead. Per-spot confidence and consistency numbers, built from historical climatology, tell you how reliable a break tends to be.

Get in touch

By submitting, you agree to be contacted about our services. See our privacy policy for how your information is used.