raw × 0.514444
The input was scaled twice.
IGRA already stores wind in tenths of meters per second. Treating that integer as knots made the starting field about 5.14× too fast.
Loading the latest balloon launches
The wind,
five kilometers up.
A living reconstruction of the 500 hPa wind field over the United States, built from NOAA radiosondes and corrected for the unit and time-step errors in my original graduate model.
Mean wind—mph
Strongest—mph
Latest sounding—Across — stations
01 / LIVE FIELD
500 hPa winds,
in motion.
0–15 m/s 15–30 m/s 30+ m/s
Interpolated from radiosondesPressure altitude ≈ 5.5 km / 18,000 ft
Particle motion is illustrative; station arrows and reported speeds come directly from the latest valid NOAA 500 hPa observations.
02 / THE CORRECTION
Why the first model
made impossible wind.
u × du/dt × Δt
The update multiplied state.
Wind was multiplied by its tendency on every step, turning normal upper-air flow into values measured in millions of meters per second.
u + du/dt × Δt
The corrected relationship.
This page uses properly scaled observations. Future forecast evolution will use an additive, stability-limited time step.
03 / SOUNDING NETWORK
What the balloons
actually measured.
Twice a day, weather balloons climb through the atmosphere carrying instruments that report pressure, temperature, humidity, wind direction, and wind speed.
| Station | Observed | Direction | Wind | Temperature | Height |
|---|---|---|---|---|---|
| Loading NOAA observations… | |||||
04 / SOURCE + CADENCE
Measured in the sky.
Updated automatically.
The updater runs after both standard global sounding cycles and publishes the newest valid 500 hPa observation for each station. NOAA notes that final archive availability can trail launch time.
Open the NOAA IGRA archivePage data generated —