Method & data

How it works

From about 20 years of hourly ERA5 reanalysis, we chain randomly chosen 10-day blocks (kept in season) into thousands of weather scenarios, count the hours when wind and waves are below your limits, and measure how many calendar days it takes to reach the planned workable days (P50/P90). Wind, waves and rain are taken together from the same block, so their real co-occurrence is preserved.

w(t) = 1[t ∉ suspension] · 1[U(t) < U_th] · 1[H_s(t) < H_s,th]
D = min {t : Σ w(s)·Δt ≥ T_plan × 24 h} · P50 = Q₀.₅(D), P_x = Q_x(D) (x = 70–95, default 90)

Climate change & typhoons

For start dates 5 or more years ahead, monthly change factors from CMIP6 high-resolution models are applied to wind and waves (wave height approximated as wind to the power 1.5). SSP2-4.5 is approximated by scaling the SSP5-8.5 signal by the ratio of global warming. The stochastic typhoon model is in beta and its rates are provisional.

Limitations

• It does not predict the weather on a specific day; it samples a probability distribution.
• The ERA5 wave grid is coarse and has gaps near the coast.
• Non-weather factors such as vessel availability, port closures and equipment failures are not included.
• Earthquake and tsunami effects are not included.
• Typhoon rates are provisional and not yet calibrated.

Data sources and licences

• Open-Meteo — CC BY 4.0 (non-commercial use on the free API)
• ERA5 reanalysis — Copernicus Climate Change Service — Licence to use Copernicus products
• CMIP6 HighResMIP via Open-Meteo Climate API — CC BY 4.0
• 地理院タイル (GSI Japan) — Government Standard Terms of Use v2.0
• OpenStreetMap — ODbL

Weather data: Open-Meteo.com (CC BY 4.0); ERA5: Copernicus Climate Change Service. Map: Geospatial Information Authority of Japan tiles.