Turbulence Prediction — The Science Behind Forecasts

Turbulence prediction has advanced dramatically over the past decade. Here's how modern forecasts work — and what their accuracy limits are.

The three layers of turbulence prediction

Modern turbulence forecasting uses three complementary data streams. Numerical Weather Models compute the atmosphere's state using physics equations solved across a 3D grid. Models like NOAA's GFS and ECMWF calculate wind shear, jet stream position, and stability indices at every grid point every 6–12 hours. These give 24–72 hour forecasts of where turbulence is likely but cannot resolve small-scale eddies. SIGMET (Significant Meteorological Information) alerts are issued by meteorological watch offices when severe turbulence is forecast for a specific area. These cover large geographic regions and altitudes, valid for 4–6 hours. PIREPs (Pilot Reports) are real-time observations from crews who encountered turbulence. A PIREP includes location, altitude, aircraft type, and intensity rating. These are the most accurate local data — but they only tell you where turbulence was, not necessarily where it will be.

EDR — the standard unit of turbulence measurement

Eddy Dissipation Rate (EDR) is the modern standard for quantifying turbulence intensity. Unlike older subjective ratings (light/moderate/severe based on pilot feel), EDR is an objective physical measurement derived from the aircraft's accelerometers and flight data. Modern aircraft automatically compute and report EDR continuously. An EDR < 0.1 is smooth. 0.1–0.2 is light. 0.2–0.4 is moderate. > 0.4 is severe. > 0.7 is extreme (very rare). EDR-based reporting removes the subjectivity of individual pilot perception — a seasoned military pilot and a nervous passenger would rate the same bump very differently, but the EDR is the same.

Machine learning in turbulence prediction

Airlines and weather services now use machine learning models trained on historical EDR data to predict turbulence from atmospheric variables. These models learn correlations that physics equations miss — for example, which combinations of jet stream curvature, wind shear, and atmospheric stability produce turbulence versus which don't. GTG (Graphical Turbulence Guidance) from NOAA is the most widely used operational product, updated hourly. Airline dispatch systems ingest GTG data alongside PIREPs and model output to generate route-specific turbulence forecasts 12–24 hours ahead.

Accuracy and limitations of turbulence prediction

Current turbulence forecasts are approximately 70–85% accurate at predicting moderate-or-above turbulence within ±100km and ±3 hours. Clear-air turbulence is harder to predict than convective turbulence — it has no visual signature and occurs in apparently clear air with no radar return. False positives (predicting turbulence that doesn't occur) are more common than false negatives (missing turbulence that does occur) because forecasters correctly err on the side of caution. Within 6–12 hours of a flight, PIREP-based nowcasting significantly improves accuracy for specific altitudes and routes.

Most Turbulent Routes

Ranked by historical turbulence score — click any route for details

Heavy
Heavy
Heavy
Heavy
Heavy
Heavy
Heavy
Heavy
Heavy
Heavy
Moderate
Moderate
Browse all routes →

Frequently Asked Questions

How far in advance can turbulence be predicted?
72-hour forecasts identify the general areas and altitudes likely to have turbulence with moderate confidence. 24-hour forecasts significantly improve accuracy as weather models update with more observation data. 6-hour forecasts combine model output with real-time PIREPs and can achieve 80–90% accuracy for specific route segments. Turbulence that begins suddenly (convective outflow, breaking mountain waves) can be predicted only minutes ahead even with the best systems.
What is the GTG turbulence forecast?
GTG (Graphical Turbulence Guidance) is a product developed by NOAA's Aviation Weather Center that provides a graphical turbulence forecast updated every hour. It covers altitudes from 10,000 feet to 45,000 feet and predicts turbulence intensity on the EDR scale. GTG is used by airline dispatch centers and is the basis for the turbulence layer shown in major aviation weather apps. It is freely available at aviationweather.gov.
Do pilots get turbulence predictions before the flight?
Yes — turbulence is a key component of every pre-flight briefing. Pilots receive a SigWx (Significant Weather) chart covering the entire route, SIGMET alerts, PIREPs from previous flights on the same route, and dispatch output from the airline's meteorology team. For transatlantic and transpacific flights, airline dispatchers specifically coordinate altitude changes with ATC to avoid areas with active SIGMET advisories.
Can passengers access the same turbulence data as pilots?
Passengers can access SIGMET alerts through aviationweather.gov (FAA) or the Meteorological Watch Offices worldwide. PIREPs are also publicly available. Apps like TurboTrack aggregate and interpret this data into passenger-friendly route forecasts. The main difference is that pilots also have access to company meteorology reports and real-time ACARS turbulence data from the airline's fleet — these are not public.
Real-time turbulence in your pocket
Live PIREPs, SIGMETs, AI forecast & best seat guide — free on iOS
Download Free on iOS