Méthode & sources des données
Cette page explique d'où viennent les données et comment chaque valeur est calculée. Aucune intelligence artificielle n'est utilisée pour les requêtes météo, les calculs ou les résumés – toutes les valeurs et tous les textes sont calculés à partir des données journalières selon des règles fixes.
1. The comparison period
- For every comparison year exactly the same calendar days are evaluated – both arrival and departure day are included.
- Periods across New Year (e.g. 27 Dec – 5 Jan) belong to the start year and are labelled “2024/25”.
- Leap years: if 29 February lies inside the period it is counted in leap years. If 29 February itself is the arrival or departure day, 28 February is used in non-leap years and the year is marked.
- Only fully completed periods are used: the end of the period plus the data source's publication delay (3–7 days) must lie in the past. The current year only appears once the period is over.
- Monthly averages or climate normals are never used as a substitute for missing days.
2. Calculated values
| Max. moy. le jour | Mean of the daily maximum temperatures (2 m) of all days in the period |
|---|---|
| Min. moy. la nuit | Mean of the daily minimum temperatures – usually reached at night or early morning |
| Température la plus élevée / Température la plus basse | Highest daily maximum and lowest daily minimum in the period, with date |
| Précipitations | Sum of daily precipitation (rain, snow as water equivalent) in mm |
| Jours de pluie | Days with at least 1.0 mm of precipitation (standard meteorological definition) |
| Heures d'ensoleillement | Sunshine duration per day where the source provides it (Open-Meteo: derived from radiation following the WMO definition). Visual Crossing does not provide sunshine duration. |
| Vent | Mean of the daily maximum wind speed (10 m) |
| Température de la mer | Mean of the daily sea surface temperature at the determined sea point (see below) |
Overall values are the mean of the yearly values; every complete year has the same weight. Summary texts are produced from these numbers with fixed templates.
3. Weather icons and prevailing weather
Every day is classified by fixed rules (in this order):
- Precipitation ≥ 1 mm: thunderstorm (if reported by the source), snow (fresh snow ≥ 0.5 cm), rain (≥ 5 mm) or light rain
- Otherwise by sunshine share of day length: sunny from 65 %, partly cloudy from 35 %, otherwise cloudy
- Without sunshine duration by mean cloud cover: sunny up to 30 %, partly cloudy up to 70 %, otherwise cloudy; fog only if reported by the source
The prevailing weather is the most frequent category. In case of a tie the less favourable category is chosen, so the result is never embellished.
4. Missing and incomplete data
- A day is valid if maximum and minimum temperature and precipitation are available and the source marks it as genuine historical data. Statistical normals or forecasts are discarded.
- Complete years (all days valid) are used for the overall values.
- Years with at least 80 % valid days are shown but flagged “incomplete” and excluded from overall values.
- Years with fewer valid days are not evaluated.
- Sunshine, wind and sea temperature are only averaged if they are available for every day – otherwise “n/a” is shown.
5. Sea temperature and sea point
For the sea temperature we search rings of 3 to 20 km around the place for open-sea points, based on the Natural Earth coastline (1:10 m). A point only counts as sea if its surroundings within 2 km are water as well – this excludes the coastline itself, narrow inlets and lakes. If the dataset has no values at that point (e.g. the model's land mask), the next candidate is tried. Without a suitable point no sea temperature is shown.
- NASA JPL MUR SST (default, via NOAA CoastWatch ERDDAP): daily analysis of satellite and in-situ measurements on a 0.01° grid (≈ 1 km), since 2002 – ideal for coastal places. Alternatively NOAA OISST (0.25°, since 1981).
- Copernicus Marine OSTIA: daily analysis of satellite and in-situ measurements on a 0.05° grid (≈ 5.5 km), since 1981. It is the “foundation” temperature a few metres below the surface, without the daytime-warmed top layer.
- Open-Meteo Marine (alternative): Météo-France ocean model, data only from 2022.
6. Types of data
| Station météo (mesure) | Direct measurement at a nearby weather station |
|---|---|
| Stations météo + télédétection | Station measurements complemented by satellite/radar or reanalysis data |
| Réanalyse (modèle météo calé sur des observations) | Reconstruction of the past by a weather model constrained by millions of observations (e.g. ERA5, MERRA-2/NASA POWER). Values represent a grid cell (25–50 km), not a point. |
| Analyse satellite et in situ | Gap-free analysis built from satellite and buoy measurements (OSTIA) |
| Modèle océanique | Output of an ocean model |
7. Attribution
- Weather Data Provided by Visual Crossing
- NASA POWER: Data obtained from the NASA Langley Research Center (LaRC) POWER Project funded through the NASA Earth Science/Applied Science Program
- MUR SST data were provided by JPL under support by NASA MEaSUREs program; NOAA OISST v2.1 (NOAA NCEI); accessed via NOAA CoastWatch ERDDAP
- Weather data by Open-Meteo.com (CC BY 4.0); ERA5: Hersbach et al. (2020), Copernicus Climate Change Service
- Generated using E.U. Copernicus Marine Service Information; OSTIA (doi:10.48670/moi-00168, doi:10.48670/moi-00165)
- Place search: GeoNames (CC BY 4.0) / © OpenStreetMap contributors
- Coastline: Natural Earth (public domain)
- Flags: Twemoji (CC BY 4.0), as a font «Twemoji Country Flags» by TalkJS