How AirVeto renders the wind and tracking field.

Wind data

Wind vectors come from Open-Meteo, which blends the ECMWF IFS (European Centre for Medium-Range Weather Forecasts) and NOAA GFS (Global Forecast System) deterministic runs. AirVeto queries Open-Meteo’s gridded hourly product covering an 8-day window: four days back and four days forward. At any given moment, the rendered field is typically on the order of tens of minutes lagged against the underlying model analysis, not the physical atmosphere.

Selectable altitudes: surface, 500 m, 1,500 m, 3,000 m (default), 5,000 m (Pro). Direction and speed are both interpolated from the model grid to the rendered map viewport; no additional post-processing is applied.

Cross-border inflow detection

Border segments on the EU’s eastern frontier with Belarus, Russia, and Kaliningrad are sampled at a configurable stride along the national boundary. Each sample computes the border-normal vector and compares it to the rendered wind vector at that location. A segment is marked orange when the wind vector crosses into the EU within 85° of the border-normal. Border polygons are static GeoJSON, not live-traced.

This is a directional filter, not a forecast and not a trajectory. Orange segments indicate only that the rendered wind is currently oriented to carry airborne objects across that segment — whether anything is airborne is a separate question.

Aircraft tracking (ADS-B)

Aircraft positions come from the public ADS-B receiver network. Coverage is near-complete above roughly 2,000 m across the Baltic ring (Poland, Lithuania, Latvia, Estonia, Sweden, Finland), deteriorates below 1,000 m over open water, and is sparse over Belarus due to low receiver density. Positions are cached region-wide with a time-to-live of approximately 15 seconds.

Vessel tracking (AIS)

Baltic Sea vessel positions come from Finnish Transport Infrastructure Agency’s Digitraffic AIS feed. Coverage is concentrated in the main Baltic shipping lanes; small craft not transmitting AIS do not appear.

Satellite tracking (TLE / SGP4)

Satellite positions are computed client-side from public Two-Line Element sets using the SGP4 orbit propagator (via satellite.js). No external tracking service is queried at render time. TLEs are refreshed on each session load.

Basemap

Vector tiles from OpenFreeMapunder its public licence. OpenStreetMap-derived data © OpenStreetMap contributors, licensed under ODbL.

Validation

Rendered wind direction and speed at the selected altitude can be cross-checked against authoritative aviation meteorological services — ICAO METAR/TAF data from national AMOs covering aerodromes in coverage (EYVI Vilnius, EPWA Warsaw, EVRA Riga, EETN Tallinn, UKBB Kyiv). Material divergence would indicate either a model update lag or a boundary-layer decoupling effect that AirVeto’s gridded rendering cannot capture; in that case, prefer the METAR.

For specific trajectory work, use a proper dispersion model (HYSPLIT, FLEXPART) with appropriate meteorological inputs.

Known limitations

• Altitude sensitivity. Wind at 500 m and at 3,000 m can differ by tens of degrees in direction, especially around frontal passages. Switch altitude for the question you’re answering.
• Model-atmosphere lag. Open-Meteo’s gridded analysis lags the physical atmosphere; expect a smoothing timescale on the order of the model resolution (~9 km for IFS, ~13 km for GFS).
• Cluttered ADS-B below 1,000 m. Below the radar horizon of coastal receivers over open water.
• Border segment density. Inflow detection is sampled along the boundary; very short segments under the sample stride may not surface an orange marker even when conditions are met.

What AirVeto is not for

AirVeto is not for aviation, navigation, or safety-critical decisions. It is a visualization product over public data, intended for research, journalism, and situational awareness of cross-border airflow. Use official authorities for operational decisions.

Publisher

AirVeto is published by AirVeto. See the About page, press kit, and the incident archive for product context.