Weather & Climate FAWP Detection
Detect the Information-Control Exclusion Principle in atmospheric forecast systems.
The core question
"Is there a regime where the atmosphere remains forecastable, but weather-modifying interventions have already lost their effect?"
That is FAWP in the weather domain: prediction persists, control has collapsed.
Install
Three entry points
1. From ERA5 reanalysis data (no API key)
from fawp_index.weather import fawp_from_open_meteo
result = fawp_from_open_meteo(
latitude = 51.5, # London
longitude = -0.1,
variable = "temperature_2m",
start_date = "2015-01-01",
end_date = "2024-12-31",
horizon_days = 7, # 7-day forecast horizon
)
print(result.summary())
2. Multi-location grid scan
from fawp_index.weather import scan_weather_grid
cities = [
{"lat": 51.5, "lon": -0.1, "name": "London"},
{"lat": 48.9, "lon": 2.4, "name": "Paris"},
{"lat": 40.7, "lon": -74.0, "name": "New York"},
]
results = scan_weather_grid(cities, variable="temperature_2m",
start_date="2015-01-01", end_date="2024-12-31")
for r in results:
flag = "🔴 FAWP" if r.fawp_found else "—"
print(f"{r.location:<20} {flag} gap={r.peak_gap_bits:.4f}b")
3. From your own NWP forecast arrays
from fawp_index.weather import fawp_from_forecast
result = fawp_from_forecast(
forecast = nwp_output, # model forecast values
observed = verification, # actual measurements
intervention = model_nudge, # forecast adjustment / ensemble spread
horizon_days = 5,
variable = "temperature_2m",
location = "50.0N 0.0E",
)
print(result.summary())
Supported variables (ERA5 via Open-Meteo)
| Variable | Description |
|---|---|
temperature_2m |
2m air temperature (°C) |
precipitation_sum |
Daily precipitation (mm) |
wind_speed_10m |
10m wind speed (m/s) |
surface_pressure |
Surface pressure (hPa) |
cloud_cover |
Cloud cover fraction (%) |
et0_fao_evapotranspiration |
Reference ET (mm) |
shortwave_radiation |
Solar radiation (W/m²) |
Interpreting results
| Field | Meaning |
|---|---|
fawp_found |
FAWP regime detected — prediction persists but control collapsed |
peak_gap_bits |
Maximum leverage gap (bits) — larger = stronger FAWP |
odw_start/end |
Operational Detection Window (Ï„ range) |
tau_h_plus |
Post-zero agency horizon — where control first vanishes |
tau_f |
Failure cliff — where the system becomes fully uncontrollable |
Physical interpretations
- High
peak_gap_bits: The forecast model retains skill at lags where interventions (nudges, corrections, warnings) have already lost effect. - Narrow ODW: FAWP window is short — small lead time available before cliff.
- FAWP in precipitation: Predictability persists into a regime where cloud-seeding or model re-initialization no longer changes outcomes.
- FAWP in wind energy: Grid operator can forecast output but can no longer route or curtail fast enough to affect the outcome.
Papers
- E1–E7: doi:10.5281/zenodo.18663547
- E8: doi:10.5281/zenodo.18673949
- E9 (SPHERE_15): Experiment 9 confirmation suite