⚠️ Now in Early Access — RTurbSim is live for public testing. Expect bugs, downtime, and data resets. Join the test

Real-time turbulence detection

See turbulence
before you fly through it.

Crowdsourced. Live. Network-wide. Built for MSFS 2020/2024, X-Plane 11/12, P3D, and FSX.

scroll

Without a turbulence network

Turbulence is invisible.

METARs don't show it. Forecasts smooth it over. Your weather engine guesses. You only know it's there when your coffee is on the ceiling.

With RTurbSim

Until now.

Every pilot on the network reports turbulence the moment they hit it. The result is a live, color-coded map of where the air is rough, built from the people actually flying through it.

How it works

Measured from real motion.

No black boxes. Your simulator feeds real aircraft motion to the network, and we quantify it the same way the aviation industry quantifies turbulence. Open any part to see exactly how, including what it deliberately doesn't do.

#1 How your flight data is captured & transmitted

A lightweight Windows app reads your simulator through SimConnect, FSUIPC, or XPUIPC on MSFS, X-Plane, P3D, or FSX. You sign in once; there's nothing else to configure.

While you fly, it streams a compact reading roughly every five seconds: vertical g-load, vertical speed, bank and pitch, altitude, heading, airspeed, position, and your aircraft type. That stream is the entire raw material of the network: real motion from a real flight, never a forecast.

#2 How turbulence is measured

Turbulence is rapidly fluctuating vertical acceleration, so we measure exactly that: the standard deviation of vertical g over a ~30-second rolling window, the same statistical basis ICAO and airline turbulence systems use. Not the single biggest jolt, which any input can cause; how rough the air is across the whole window.

  • Each g-value is bank-corrected (1 / cos(bank) removed) so a steep turn never reads as turbulence.
  • The bar to count as turbulence widens during climb, descent and approach, where thrust, trim and gear add motion that isn't weather, and stays tight in cruise.
  • Every reading must be backed by at least six high-rate samples; a throttled, one-sample-a-minute session is discarded as smooth rather than published as a false spike.
#3 The severity scale

Roughness maps to the standard light / moderate / severe scale pilots already use. Reference thresholds (vertical-g standard deviation, after correction) are shown. We publish them on purpose.

SmoothAir you wouldn't comment on. Not plotted.< 0.02 g
LightNoticeable chop; drinks ripple.0.02 to 0.05 g
ModerateDefinite bumps; seatbelt-sign air.0.05 to 0.12 g
SevereSharp movement against the belts.0.12 g +

The output is a normalized 0 to 1 index aligned to this scale, a faithful proxy, not a physical EDR measurement (see #6 below).

#4 Calibration across aircraft

The same pocket of air throws a Cessna around far more than a 737. So we normalize each report toward an airliner baseline by aircraft class, so reports stay comparable on one map. We only damp sensitive airframes toward the baseline; we never amplify heavier ones, which would just magnify autopilot jitter.

ClassNormalizationExamples
Light single0.25×C172, PA28, Cirrus
Light twin / turboprop0.45×King Air, TBM
Regional0.70×CRJ, E-Jet, Q400
Narrow-body (ref)1.00×737, A320
Wide-body / heavy1.00×777, A350, 747
Helicopter0.30×Rotorcraft

Empirically tuned from network data, not first principles, and refined as the dataset grows.

#5 From a reading to the live map

Each reading is placed on a hexagonal grid (about 2 nm per cell). Same-severity neighbours merge into the continuous ribbon that follows each route instead of scattered dots. A cell's colour reflects its most recent qualifying pass; report counts and altitudes accumulate so you can see how well-supported it is. The live map shows the last 30 minutes by default and rewinds up to 48 hours.

#6 What it is and what it isn't

We'd rather you trust a smaller claim completely than a bigger one halfway:

What you can rely on

  • Every reading is a real aircraft's real response in the sim. Measured, never guessed.
  • Severity uses the same statistical basis the industry uses for turbulence.
  • Aircraft, phase and maneuvers are corrected for, so the number reflects the air, not the flying.
  • Low-confidence data is discarded, not smoothed over.

What it is not

  • Not real-world weather. No METARs, PIREPs or radar. It reflects what your sim renders.
  • Not certified EDR. Shares EDR's method, not its calibrated units; don't compare it to real-world EDR.
  • Not a discrete-jolt detector. A rolling window can smooth over a rare sub-second bump.
  • No altitude climatology. The same bands apply low and high.
Website map RTurbSim live map showing turbulence hex trail
Android app RTurbSim Android companion app
iOS app Coming soon

Hexagonal binning.

Every reading is dropped into a uniform hexagonal grid cell, so coverage is even, severity stacks cleanly, and the map renders fast at any zoom.

  • Severity-coded hex overlay (light → severe)
  • Smooth flight-path trail by altitude
  • Live aircraft icons + callsigns
  • RainViewer 512 radar tile overlay

Where we are

Built in public. Roadmap is the receipt.

Pipeline status 5 of 7 milestones shipped
71% complete
  1. Windows .exe Shipped
  2. Android app Shipped
  3. Website Shipped
  4. Server infra Shipped · ongoing
  5. Early access Live now
  6. iOS app In development
  7. Full release Q4 2026
Early Access: Active

Help us build the turbulence network.

Free to join. You get full access to the Windows app, live map, and Android companion. In return, your flight data tunes our detection and proves coverage at scale.

Status Active
Full release Q4 2026
Cost Free during testing

What to expect: bugs, missing features, occasional downtime, data resets. Your feedback is what makes the network production-ready.

Join Early Access

Sim compatibility

If you fly it, we map it.

  • MSFS 2020 / 2024 via SimConnect
  • X-Plane 11 / 12 via XPUIPC
  • Prepar3D all versions, via FSUIPC
  • FSX via FSUIPC