Methodology

Data last computed 2026-10-04. This page describes the exact pipeline.

The window

A tide window is a contiguous interval when the predicted water height at a NOAA station sits below +1.0 ft MLLW — a threshold at which most Pacific tidepool shelves and low-tide beaches become comfortably walkable. Heights come from NOAA CO-OPS harmonic tide predictions (datum MLLW, feet), fetched fresh every day for the next ~400 days.

• Harmonic stations: we fetch the hourly prediction series and linearly estimate the crossing times of +1.0 ft, rounded to 5 minutes.
• Subordinate stations (high/low events only): we use standard cosine interpolation between consecutive extremes to estimate crossings. The difference between hourly and cosine estimates varies by station and tide.

Very shallow dips may be too brief for the hourly samples or five-minute rounding to resolve. We omit intervals that collapse to zero minutes or whose estimated crossings do not surround the predicted low. The complete high/low tide tables still include those NOAA predictions.

Arrive by is 60 minutes before the low — the classic advice for working out to the lowest zone as the water still falls.

The 0–100 score

• Depth (0–50): linear from 0 points at +0.5 ft to 50 points at −2.0 ft (clamped). Lower tide, more exposed seafloor.
• Daylight (0–30): the window's overlap with computed sunrise–sunset, scaling linearly to 30 points at 3 hours. Windows with under 30 minutes of daylight score 0 total and are labeled night tides only when there is no daylight overlap. Sun times are computed with the open-source astronomy-engine library for the station's exact coordinates.
• Timing (0–10): Saturday, Sunday, or US federal holiday +10; Friday +5.
• Season comfort (0–10): a fixed, published table by region and month — a simple heuristic for daylight length and typical shoulder-season comfort. It is deliberately coarse and is not a weather forecast.

Bands: 90–100 Exceptional · 75–89 Great · 60–74 Good · 40–59 Fair · under 40 Skip.

Enrichment data

• Short-range conditions (windows within 7 days only): NWS point forecasts from api.weather.gov, stamped with their fetch date.
• “What's in the pools”: research-grade observation counts within 5 km over the last 60 days from the iNaturalist open API (© contributors, CC BY-NC), filtered to molluscs, echinoderms, cnidarians, and arthropods.

Honest limitations

• Predictions are astronomical. Wind, swell, and pressure routinely shift actual water levels by half a foot or more — check conditions before you go, and never turn your back on the ocean.
• A station's predictions apply exactly at the station. Nearby beaches can lag or lead by tens of minutes; we name the station on every page so you can judge the distance.
• The walkable threshold varies by beach profile; +1.0 ft is a defensible default, not a law of nature.
• A window says when the water is predicted low — it does not verify beach access, seasonal or safety closures, or how long a particular route takes. That information lives with the land manager, not in the tide tables; check their current status before you go.

Automation disclosure

Tidewindow is an automated publication: the data pipeline recomputes every number daily, and the editorial content is produced by an AI system operating under strict rules — every tide statistic must come from the computed dataset (never from a language model's memory), factual claims must cite a source fetched at publish time, and safety guidance is only ever quoted from park services. Full details on the about page. Found an error? Tell us — corrections ship within a day.