HOW THE FORECAST WORKS
The Quota 4000 dropzone weather forecast starts from the best-match model auto-selected for the area (typically ICON-EU over continental Italy, AROME France HD over the north-west), and in parallel queries 4 numerical weather prediction (NWP) models — ICON-EU, ECMWF IFS, AROME France HD, GFS — to compute their agreement on wind, gust and precipitation. Inter-model spread becomes the hour-by-hour confidence indicator. On top of that we apply skydiving-specific operational thresholds instead of the usual "sunny / cloudy" labels.
The 4 models we combine
The primary hourly data you see on the dropzone detail (temperature, low/mid/high cloud, visibility, CAPE, wind at 10/80/120 m, gusts, etc.) comes from the best-match model auto-selected by Open-Meteo for the area — typically ICON-EU over continental Italy, AROME France HD over the north-west. In parallel we query the 4 models below on surface wind, gust and precipitation and compute σ across them: model agreement becomes our hour-by-hour confidence indicator.
Best for central-north Italy — high spatial resolution over the Po valley and eastern Alps.
The "gold standard" of global forecasting. Lower resolution than ICON-EU but excellent physics on synoptic-scale dynamics (fronts, depressions).
Finest resolution available at 1.3 km. Convection-permitting model: it explicitly represents cumulus and storm cells rather than parametrising them. Best over alpine orography.
Global reference baseline, free and frequently updated. On its own it never beats ECMWF, but in an ensemble it acts as an "independent vote" that catches scenarios the other 3 miss.
All 4 models are queried via Open-Meteo (open-source / free tier for non-commercial use), which acts as the aggregator: a single HTTP call returns the 4 models' outputs on the 3 confidence variables (wind, gust, precipitation), aligned on the same hourly grid as the best-match.
Confidence indicator
For every hour in the next 24h we compute the standard deviation (σ) of wind, gust and precipitation across the 4 models. If models agree, σ is low: the forecast is "robust", confidence is high. If models diverge, σ is high: scenario likely ill-defined, exercise caution.
Jumpability score · how it's computed
The score starts at 100 and applies non-linear penalties for each condition outside operating limits. Logic mirrors a jumpmaster, not a meteorologist: surface wind and gusts dominate, clouds are weighted against the dropzone-specific exit altitude, storms and rain in the next 2h are hard stops.
| Condition | Penalty |
|---|---|
| Gust > 13 m/s | −100 · Hard stop everyone |
| Gust 10–13 m/s | −35 · No students |
| Surface wind > 12 m/s | −80 · No-go everyone |
| Surface wind 8–12 m/s | −30 · No students |
| Upper wind 80m > 18 m/s | −25 · Heavy drift on opening |
| Cloud base < 600 m | −100 · Below VFR minimums |
| Cloud base < DZ exit altitude | −60 · No visual exit |
| Cloud base < exit + 500 m | −20 · Tight margin |
| CAPE > 1500 J/kg or rain 2h > 1 mm | −100 · Storms or incoming rain |
| Rain probability > 60% | −40 · Rain likely |
| Visibility < 5 km | −80 · Low-vis landing |
The final score is clamped to 0–100. Green ≥ 70 (jumpable), yellow 40–69 (borderline), red < 40 (no-go). Cloud base is derived locally from the Espy formula: cloudBaseM ≈ 125 × (T − Td), where T is temperature and Td is dewpoint — Open-Meteo doesn't expose cloud base directly, but the estimate is conservative when low clouds actually exist.
List score vs detail score
The list shows a single number: a weighted average of the score across the next 36 operating hours (08:00–19:00, weight 0 for night hours). When you open a DZ detail, you see the score hour-by-hour. A DZ can appear yellow (e.g. 55/100) in the list and then show a 0/100 hour in the detail: it means some hours are outside range but others over the next 1–2 days recover. Open the detail to see where the usable window falls.
Known limits
- Complex orography: narrow alpine valleys, the Adriatic coast, inland Sicily can produce local winds (foehn, sea breeze) that 3–13 km models don't resolve. AROME at 1.3 km helps in the NW, but the "miss" risk remains — always verify with the DZ.
- Cloud base derived from Espy: rough estimate, not direct sensing. If low cloud is < 50% coverage we show "—" (no penalty).
- Open-Meteo refreshes ICON-EU every 6h. Our cron pre-warms the cache every 3h; worst-case you see data 3 hours old — the "fetched X min ago" under the toggle is always truthful.
- The forecast never replaces the on-site jumpmaster's briefing. What you see here is a pre-decision tool (is it worth driving?). The final go/no-go is always taken by the DZ with the actual ground wind reading.
Sources and attribution
- Open-Meteo · API aggregator for the models, free non-commercial tier. open-meteo.com ↗
- DWD · ICON-EU model, CC BY 4.0 licence. dwd.de ↗
- ECMWF · IFS model, ECMWF open data licence. ecmwf.int ↗
- Météo-France · AROME France HD model, Etalab Open Licence 2.0. meteo.data.gouv.fr ↗
- NOAA · NCEP · GFS model, public domain. noaa.gov ↗