Turbulence Analysis — GTGN Nowcast + Multi-Model Wind Shear
PlaneWX uses the FAA/NCAR Graphical Turbulence Guidance Nowcast (GTGN) as the primary turbulence assessment, updated every 15 minutes with EDR-based severity across CONUS. Multi-model wind shear analysis from HRRR, GFS, and ECMWF provides the meteorological mechanism — identifying jet streams, mountain waves, and frontal boundaries that explain what GTGN detects. Together, they give you the most complete turbulence picture available in general aviation.
Why Multi-Model Turbulence Matters
Traditional turbulence information comes from two sources: G-AIRMET Tango (large geographic areas with blanket severity ratings) and PIREPs (subjective reports from other pilots). Both have significant limitations for pre-flight planning:
G-AIRMET Tango: Too Broad
- • Cover hundreds of thousands of square miles
- • Single severity for the entire area and altitude range
- • No information about which specific altitudes are roughest
- • Don't indicate the physical mechanism causing turbulence
- • Updated every 3 hours — jet stream patterns shift faster
PIREPs: Subjective & Sparse
- • Severity is subjective — varies by aircraft type and pilot experience
- • "Moderate" in a 737 feels very different than in a Cirrus
- • Sporadic coverage, especially at night and in rural areas
- • Often hours old by the time you see them
- • Absence of PIREPs doesn't mean smooth air
PlaneWX's approach: We use the FAA GTGN nowcast as the primary turbulence assessment — it provides EDR-based severity sampled directly along your route, updated every 15 minutes. Multi-model wind shear analysis (HRRR, GFS, ECMWF) provides the meteorological context: vertical wind shear and Richardson number at twenty-four pressure levels identify the physical mechanisms driving turbulence. When GTGN says "moderate," the models tell you whether it's jet stream shear, mountain wave propagation, or frontal activity — so you can make an informed decision about altitude changes or route deviations.
FAA GTGN Nowcast — The Primary Assessment
The Graphical Turbulence Guidance Nowcast (GTGN) is an operational turbulence product produced by the FAA and NCAR (National Center for Atmospheric Research). It is the same EDR-based turbulence assessment used by airline dispatchers and air traffic control. PlaneWX ingests GTGN data directly and uses it as the primary turbulence severity assessment for your briefing.
Updated Every 15 Minutes
GTGN refreshes four times per hour across the entire CONUS domain, providing near-real-time turbulence conditions rather than a forecast that may be hours old.
EDR-Based Severity
Uses Eddy Dissipation Rate (EDR) — a universal, objective measure of turbulence intensity that doesn't depend on aircraft type or pilot perception. EDR is the ICAO standard for turbulence reporting.
52 Vertical Levels
From the surface to FL500, on a 0.25° grid (∼15 NM resolution). PlaneWX samples this grid at your actual climb, cruise, and descent altitudes using bilinear interpolation.
What GTGN Fuses
GTGN is not a single-source product — it fuses multiple data streams into a unified EDR field:
- • Automated in-situ EDR from equipped airline and business jet aircraft
- • PIREPs converted to calibrated EDR values
- • Satellite-derived turbulence indicators
- • Radar convective turbulence signatures
- • NWP model guidance (RAP/HRRR) for background fields
EDR-to-Severity Thresholds (FAA/NCAR)
GTGN EDR values are mapped to severity categories using FAA/NCAR thresholds that account for aircraft weight class. Lighter aircraft experience turbulence more intensely at the same EDR:
Light Aircraft (<15,500 lbs — most GA)
Medium Aircraft (15,500–300,000 lbs)
PlaneWX classifies your aircraft based on service ceiling: piston and turboprop aircraft use Light thresholds, jets (service ceiling >35,000 ft) use Medium thresholds.
How GTGN and NWP Work Together
GTGN tells you “what” — the EDR severity at each point along your route, based on fused observational and model data. NWP wind shear analysis tells you “why” — whether the turbulence is caused by jet stream shear, mountain wave propagation, frontal boundaries, or low-level wind shear. PlaneWX automatically selects the right source for each point along your route using a per-waypoint temporal relevance model.
Because GTGN is a nowcast (current conditions, not a forecast), it becomes less reliable the further into the future you look. For each point along your route, PlaneWX calculates the effective age — how old the GTGN data will be by the time you actually reach that waypoint. This accounts for both the current staleness of the GTGN data (how long since the last update) and the estimated time of arrival at each point (using ground speed that accounts for headwind/tailwind from NWP models).
GTGN Zone
Effective age < 1 hour
GTGN severity is authoritative. The EDR nowcast is fresh enough to trust for this waypoint.
Blend Zone
Effective age 1–2 hours
Conservative blend: the worse of GTGN and NWP is used. Prevents cliff-edge transitions.
NWP Zone
Effective age ≥ 2 hours
NWP forecast models provide the assessment. GTGN data is too stale for this waypoint.
This means a near-term flight departing in 15 minutes will primarily use GTGN along most of the route, while a flight departing tomorrow will use NWP models entirely. For flights departing in the next hour or two, you may see different segments of the same route using different sources — the first half might be GTGN-authoritative while the latter half transitions to NWP as the effective age increases.
Canonical Turbulence Phases
Your briefing’s phase summary tiles (Climb, Cruise, Descent) show a single, authoritative severity that consolidates all available turbulence sources:
- • GTGN/NWP hybrid — per-waypoint severity from the temporal blending model above
- • G-AIRMET Tango — FAA turbulence advisories intersecting your route
- • PIREPs — pilot reports along or near your route (GA-normalized)
The worst severity from any source wins for each phase. The turbulence route bar also respects these consolidated severities — if a G-AIRMET Tango or PIREP indicates higher severity than the model data alone, the bar will reflect the elevated severity for that phase.
Hover tooltips: Each segment of the turbulence route bar shows a card-style popup with the EDR value, severity, data source (GTGN / NWP / blend), and temporal zone — so you can see exactly why each waypoint received the severity it did.
GTGN Limitations
- • CONUS only — no coverage for Hawaii, Alaska, Caribbean, or international routes. For non-CONUS flights, NWP models provide the full turbulence assessment automatically.
- • Terrain masking — EDR data near high-elevation airports may be unreliable below terrain level; PlaneWX clamps sampling to airport elevation + 500 ft AGL to avoid false readings
- • 15-minute cadence — rapidly developing convective turbulence may not appear until the next update cycle
- • Nowcast, not forecast — GTGN describes current conditions. The per-waypoint hybrid system handles this automatically: for near-term waypoints GTGN is authoritative, for waypoints further out PlaneWX smoothly transitions to NWP forecast models via the conservative blend described above
- • GTGN data gaps — if GTGN has no EDR data at a specific waypoint (e.g., near domain edges), PlaneWX falls back to the NWP severity for that point rather than defaulting to “none” — ensuring no false all-clear assessments
The Three NWP Models — Wind Shear Analysis
PlaneWX derives turbulence mechanisms from the same model soundings used for icing — HRRR, GFS, and ECMWF. Each model resolves the atmosphere differently, and their wind field differences are themselves a signal about forecast uncertainty.
High-Resolution Rapid Refresh
NOAA's highest-resolution hourly model — best for capturing terrain-induced turbulence and mesoscale wind features that coarser models miss.
3 km
Every hour
0–18 hours
Mountain wave, low-level wind shear
Global Forecast System
NOAA's primary global model — good for large-scale wind patterns, jet stream positioning, and upper-level turbulence forecasts.
~13 km
4 times daily
0–16 days
Jet stream CAT, upper-level shear
European Centre for Medium-Range Weather Forecasts
The world's leading global forecast model — provides an independent European perspective on upper-level dynamics and wind patterns.
~9 km
4 times daily
0–10 days
Jet stream dynamics, moisture-driven instability
Why Three Models?
Each model resolves wind fields differently due to different grid resolutions, physics parameterizations, and initial conditions. When all three models agree on turbulence at a given altitude, confidence is HIGH. When they disagree, the atmosphere is uncertain — and that disagreement is itself valuable intelligence for your decision-making.
The Physics: Wind Shear & Richardson Number
PlaneWX derives turbulence from two fundamental atmospheric quantities: vertical wind shear (the engine that creates turbulence) and the Richardson number (the gate that determines whether the atmosphere allows it).
Vertical Wind Shear
Wind shear is the change in wind speed between two altitude levels, expressed in knots per 1,000 feet. When wind speed changes rapidly with altitude, the resulting shear can generate turbulence — especially near jet streams, frontal zones, and temperature inversions.
Example calculation:
18,000 ft: 47 kt from 251°
14,000 ft: 31 kt from 278°
Speed difference: 16 kt over 4,000 ft
Shear = 16 / 4 = 4.0 kt/1000ft
Note: PlaneWX also accounts for directional shear (wind direction changes between levels), which can produce turbulence even when speed shear alone appears modest.
Richardson Number (Ri)
The Richardson number is a dimensionless ratio that measures the balance between thermal stability (which suppresses turbulence) and wind shear(which generates it). It answers: "Is the atmosphere stable enough to resist the shear trying to break it up?"
Ri = (g / θ) × (dθ/dz) / (dU/dz)²
Where g = gravity, θ = potential temperature, dθ/dz = static stability, dU/dz = wind shear
Important Nuance: Ri Is a Gate, Not a Severity Scale
A low Ri tells you the atmosphere can support turbulence — it does NOT tell you how severe that turbulence will be. Ri < 0.25 is common in model data near jet streams and doesn't automatically mean severe turbulence. Think of Ri as a gate: when Ri ≥ 0.5, severity is capped at LIGHT; when Ri < 0.25 and shear ≥ 4 kt/1000ft above 5,000 ft AGL, severity is boosted one level (KHI instability). The KHI boost is suppressed in the boundary layer (≤ 5,000 ft AGL) where low Ri is normal mechanical mixing. How intense the turbulence gets depends on the wind shear magnitude, not on Ri alone.
Severity Classification
PlaneWX classifies turbulence severity using a multi-step process. First, wind shear magnitude determines the baseline severity. Then, the Richardson number acts as amodifier — capping severity in stable conditions and boosting it when dynamic instability is detected.
Step 1: Shear-Based Severity
Wind shear magnitude is evaluated first using GA-calibrated thresholds (see Step 2 below). The Richardson number then modifies this baseline.
Step 2: GA-Calibrated Shear Thresholds
Wind shear magnitude determines the baseline severity:
< 6 kt/1000ft
SMOOTH
No significant turbulence detected by models. This does not guarantee smooth air — localized turbulence from convection, terrain, or wake is always possible and model resolution cannot capture all sources.
≥ 6 kt/1000ft
LIGHT
Slight, erratic changes in altitude and/or attitude. GA pilots report light chop at energy levels well below what transport-category aircraft pilots notice.
≥ 12 kt/1000ft
MODERATE
Changes in altitude and/or attitude occur. The low-moderate threshold where GA structural stress begins. Occupants feel definite strains against seat belts.
≥ 20 kt/1000ft
SEVERE
Large, abrupt changes in altitude and/or attitude. Model-derived shear ≥ 20 across a wide layer is almost always a severe event for light GA aircraft (3,000–6,000 lb).
GA-Calibrated Thresholds
PlaneWX thresholds are specifically calibrated for General Aviation aircraft (3,000–6,000 lb). GA aircraft experience turbulence at much lower energy states than transport-category aircraft — a “Light” report from a Boeing 737 corresponds roughly to “Moderate” for a Cessna 172 or DA40. Additionally, model data at 2,000–6,000 ft vertical resolution smooths peak shear by 40–60%, so a measured 8 kt/1000ft across a bulk layer often masks a 15+ kt/1000ft shear in a 500 ft sub-layer. The 6/12/20 kt thresholds account for both GA airframe sensitivity and model smoothing.
Stability Cap & KHI Boost
When Ri ≥ 0.5 (stable or transitionally stable), severity is capped at LIGHT — never overridden to SMOOTH. Bulk Richardson numbers from NWP model layers (2,000–6,000 ft thick) can miss thin shear layers near fronts, low-level jets, and mountain waves where turbulence co-exists with a stable bulk layer (Sharman et al. 2006). Capping at LIGHT instead of clearing to SMOOTH prevents dangerous false clears.
Conversely, when Ri drops below 0.25 (dynamic instability / Kelvin-Helmholtz onset) above 5,000 ft AGL and shear is at least 4 kt/1000ft, severity is boosted by one level. This captures turbulence from KHI billows in environments near the tropopause or convective boundaries — a source of false negatives in shear-only models.
Boundary layer suppression: The KHI boost is suppressed below 5,000 ft AGL. Low Richardson numbers in the atmospheric boundary layer are routine mechanical mixing from surface friction and gusty winds — not the free-atmosphere dynamic instability that KHI detection targets. This prevents strong surface gusts (e.g., 40 kt) from triggering false SEVERE classifications in the lowest flight levels during approach and departure.
Severity band downgrading: Separately from the KHI boost suppression, any SEVERE turbulence band that lies entirely below 5,000 ft AGL is automatically downgraded to MODERATE and flagged as low-level wind shear. Strong surface gusts during approach and departure can produce high model-derived shear in the lowest layers, but this is a surface phenomenon — not the kind of en-route severe turbulence that warrants a SEVERE classification in your briefing. The 5,000 ft threshold is computed from airport elevation when available, so it works correctly at high-elevation airports like Denver (5,431 ft MSL → cap at ~10,431 ft MSL).
Understanding “SMOOTH” Values in the Table
When you look at the turbulence table in the Models tab, you’ll often see many levels showing SMOOTH across all models. This means no significant turbulence signal was detected by the models — not a guarantee of smooth air. Here’s why:
Low Shear (< 6 kt/1000ft)
Wind shear below the GA detection threshold at model resolution. You might see Ri = 0.41 with shear = 4.5 showing SMOOTH — the atmosphere is marginally unstable, but the shear isn’t strong enough for models to predict perceptible turbulence. Localized turbulence from convection, terrain, or wake is always possible.
Stable Layers Show LIGHT, Not SMOOTH
When Ri ≥ 0.5 (stable atmosphere) and shear is above the light threshold, severity is capped at LIGHT rather than cleared to SMOOTH. This conservative approach prevents dangerous false clears — bulk Ri from NWP model layers (2,000–6,000 ft thick) can miss thin shear layers near fronts, low-level jets, and mountain waves.
What Highlighted Rows Mean
Levels with turbulence (LIGHT, MODERATE, SEVERE) are highlighted with an amber background. These are the altitudes where wind shear is strong enough to generate bumps. A table with mostly SMOOTH and a few highlighted rows tells you where the models detect turbulence — but remember that SMOOTH means “no significant modeled signal,” not “guaranteed smooth air.”
Reading the Color-Coded Ri Column
The Ri column in the turbulence table uses color coding to help you quickly assess atmospheric stability:
Red
Ri < 0.25
Unstable
Amber
Ri 0.25 – 0.5
Transitional
Green
Ri ≥ 0.5
Stable (cap at LIGHT)
Multi-Model Consensus
Each model independently computes wind fields, which means each has its own shear and Ri values. PlaneWX combines them into a consensus assessment using a conservative-max approach.
Conservative Max Consensus
For safety-critical aviation, PlaneWX uses a conservative max consensus: the highest severity predicted by any model is used, unless that model is an outlier by two or more categories (in which case it is tempered down by one level).
- Safety first: If the high-resolution HRRR detects severe turbulence that GFS and ECMWF smooth away, a median-based approach would report “Light” — a dangerous false clear that could lead a pilot into a KHI zone or rotor.
- Outlier protection: If one model says MODERATE and two say SMOOTH (a 2-category gap), the consensus is tempered to LIGHT rather than accepting the outlier value at face value.
- Per-model columns: You can still see each model's individual assessment in the table. If models disagree, the per-model columns make that visible so you can factor it into your decision.
HIGH
Unanimous
All models agree on turbulence severity at this altitude. The forecast is reliable.
MODERATE
Majority
Most models agree. One model sees different conditions. The majority view is likely correct, but consider the outlier model's perspective.
LOW
Split
Models disagree significantly. This happens near jet stream boundaries where small position differences produce big shear differences. Plan conservatively.
Route Sampling
PlaneWX samples 3 to 7 points along your route (depending on route length) and queries all three models at each point. This captures how turbulence conditions change along your path — critical for long flights that cross jet stream boundaries or frontal zones.
At Each Sample Point, We Compute:
- Wind speed and direction at 24 pressure levels (1000–100 hPa) from each model
- Vertical wind shear between each pair of adjacent levels
- Richardson number using temperature gradient and shear
- Per-model severity assessment at each level
- Multi-model consensus severity and agreement
- Turbulence layer boundaries (base and top altitudes)
What We Detect (and What We Don't)
PlaneWX's model-derived analysis is designed to detect Clear Air Turbulence (CAT) — turbulence caused by wind shear in clear air, typically near jet streams, frontal zones, and temperature inversions. This is the most common type of turbulence at cruise altitudes.
What Our Model Analysis Detects
- • Clear Air Turbulence (jet stream CAT)
- • Upper-level wind shear zones
- • Frontal zone turbulence
- • Temperature inversion shear layers
- • Tropopause-level turbulence
- • Mountain wave turbulence (cross-barrier flow detection)
Covered by PIREPs & G-AIRMETs
- • Convective turbulence (thunderstorms)
- • Mechanical turbulence (terrain/buildings)
- • Wake turbulence
- • Low-level wind shear (LLWS)
PlaneWX includes PIREP-reported turbulence and G-AIRMET Tango advisories alongside the model analysis to provide complete coverage of all turbulence types.
Mountain Wave Detection
Mountain wave turbulence is a leading cause of GA accidents, and standard vertical wind shear models can miss it entirely if the shear is low but terrain-forced gravity waves are present. PlaneWX includes a cross-barrier flow check that detects conditions conducive to mountain wave turbulence independently of the shear/Ri analysis.
How It Works
- PlaneWX fetches a detailed terrain elevation profile along your route using a dedicated elevation API.
- Ridge orientation is estimated dynamically from the actual terrain gradient, with confidence predicates that assess gradient magnitude and bearing coherence across neighboring points. Low-confidence bearings trigger conservative fallbacks.
- Model sounding winds are interpolated at ridge level and decomposed into perpendicular-to-ridge and parallel-to-ridge components.
- Relief-scaled thresholds: Wind speed thresholds are dynamically adjusted based on terrain relief. Lower-relief terrain (Appalachians, coastal ranges) uses lower thresholds to avoid under-detection, while major barriers (Rockies, Sierra) retain higher thresholds. The formula is: base = 18 + (relief_ft / 1,000) × 4 kt, with SEVERE = base + 10 kt, MODERATE = base + 5 kt, LIGHT = base.
- Advisory tier: 15+ kt perpendicular flow over terrain with ≥2,000 ft relief generates a marginal mountain wave advisory in your briefing. This carries no WX Score penalty but alerts you to conditions where light to moderate mountain wave turbulence is possible near ridgelines.
- Sub-2,000 ft terrain promotion: Lower-relief terrain like the Appalachians, Ozarks, and coastal ranges (<2,000 ft crest-to-valley relief) can still produce significant turbulence when flow is blocked. When the Froude number indicates blocked flow (Fr < 0.60) and cross-ridge winds are ≥25 kt, severity is promoted by one level even over modest ridges. This prevents under-detection of real hazards in lower-relief terrain.
- Low-confidence ridge fallback: When terrain gradient is weak or ridge orientation cannot be determined with confidence, PlaneWX uses 70% of the total wind speed as a conservative estimate of the perpendicular component (the RMS value of |sin(θ)| over all possible ridge orientations). This avoids worst-case 100% assumptions that would over-alert, while still flagging genuine risk.
- Results feed into pre-flight checks, WX Score deductions, and the LLM briefing prompt as terrain-specific hazards.
Data availability: Mountain wave and rotor detection require model wind data at ridge level. If model soundings are unavailable (e.g., fetch timeout), PlaneWX will prominently note that wind-terrain interaction analysis was not performed and only terrain clearance was assessed. The briefing will explicitly warn when mountain wave analysis is unavailable over mountainous routes.
Mountain wave turbulence can occur in perfectly clear skies with no G-AIRMET coverage. Rotors and lee-side downdrafts near ridgelines are particularly hazardous for light GA aircraft with limited climb performance. When PlaneWX flags mountain wave potential, treat it seriously.
Shear Guard Protection
NWP models report winds on thick pressure layers (2,000–7,500 ft apart) that can smooth out real-world shear. The bulk Richardson number (Ri) — which PlaneWX normally uses to cap severity at LIGHT when Ri ≥ 0.5 — can therefore mask significant shear signals. The Shear Guard prevents this false-quiet failure mode.
Altitude-Banded Thresholds
When resolved shear exceeds the guard threshold for that altitude band, the Ri cap is bypassed and the shear-based severity stands:
Low Level
1000–750 hPa
≥ 18 kt/1000ft
Mid Level
749–500 hPa
≥ 14 kt/1000ft
Upper Level
< 500 hPa
≥ 12 kt/1000ft
Lower-altitude layers are relatively thinner and resolve shear more faithfully, so a higher threshold is appropriate. Upper layers are much thicker (~2,500–7,500 ft between pressure levels), heavily smoothing shear — a lower threshold protects weaker signals that likely represent stronger real-world shear.
These thresholds were validated against 1,034 GA-specific PIREPs and calibrated to keep false alarms within the FAA-accepted ~10% tolerance. The methodology was independently reviewed and stress-tested before deployment. Read the full methodology review
Mountain Rotor Detection
Rotors are violently turbulent circulations that form on the lee side of mountain ridges when airflow is partially or fully blocked. They are among the most dangerous phenomena for GA aircraft, capable of producing updrafts and downdrafts exceeding aircraft performance. PlaneWX uses Froude number analysis to detect rotor risk.
Froude Number (Fr)
Fr = Uperp / (N × h), where all quantities are in SI units:
- Uperp — perpendicular wind component across the ridge in m/s (converted internally from knots; 1 kt = 0.514 m/s)
- N — Brunt-Väisälä frequency (atmospheric stability, s-1), computed from potential temperature lapse between valley and crest
- h — effective ridge height (meters), max-min elevation within ±2.7 NM window, requiring ≥500 ft relief
SEVERE Rotor
Fr < 0.60 and Uperp ≥ 25 kt (12.9 m/s)
Blocked flow — hydraulic jump
WX Score: NO-GO
MODERATE Rotor
Fr 0.60–0.75 and Uperp ≥ 20 kt (10.3 m/s)
Partial blocking — trapped lee waves
WX Score: −30 points
Rotor turbulence occurs below ridge crest level and is invisible to conventional turbulence forecasts. When PlaneWX flags rotor risk, avoid flying on the lee side of the ridge below crest altitude. Consider routing upwind of the ridge or crossing well above ridge crest level. Note: Froude-number rotor detection requires terrain relief ≥ 500 ft and a high-confidence ridge orientation (gradient ≥ 100 ft/NM with coherent neighboring bearings). When ridge geometry is uncertain, the system skips rotor analysis and falls back to the cross-barrier wind mountain wave heuristics described above, using a conservative wind estimate.
Scorer Parameter Gate
Not all blocked-flow conditions produce rotors. Rotor turbulence requires trapped lee waves, which only form when the atmospheric profile supports wave trapping. PlaneWX computes the Scorer parameter (l²) from model soundings between ridge crest and ~4 km above crest to verify this condition:
l² = (N² / U²) − (1/U)(d²U/dz²)
- When min(l²) ≥ 0.25 between crest and crest+4 km, wave trapping is supported and the rotor alert stands.
- When min(l²) < 0.25, waves propagate vertically rather than trap — rotor formation is suppressed and the alert is removed. This prevents false alarms in common winter jet-stream scenarios where strong upper-level winds prevent wave trapping.
- When sounding data is insufficient to compute the Scorer parameter, the system falls through to the Froude-number analysis alone (conservative).
Complex Terrain Modifier
Simple 2D Froude analysis can over-predict rotor severity in complex 3D terrain where multiple ridge orientations exist (sinuous valleys, overlapping ridges). PlaneWX computes the circular variance of ridge-crest orientations within the analysis window. When variance exceeds 0.35 (indicating highly complex terrain), the Froude number is inflated by 25% to account for 3D flow-splitting effects that weaken rotors. Deeply blocked flows (Fr < 0.35) are exempt from this adjustment because 3D splitting can concentrate downslope flow in confined valleys.
Methodology Review
PlaneWX’s mountain wave and rotor detection algorithms are validated against published mountain meteorology literature (FAA AC 00-57, Durran 1990, Vosper 2004, Reinecke & Durran 2008) and continuously verified through automated regression testing against known mountain routes including Rocky Mountain, Sierra Nevada, Appalachian, and Colorado Front Range crossings.
PIREP GA Normalization
A “Moderate” turbulence report from a Boeing 747 (~875,000 lb) is a vastly different experience than “Moderate” in a Cessna 172 (2,300 lb). PlaneWX uses a three-tier weight classification to normalize PIREP turbulence intensity for GA aircraft sensitivity.
Weight Categories
Light
< 15,500 lbs
C172, DA40, SR22
No adjustment
Medium
15,500 – 300,000 lbs
B737, A320, CRJ, E-jets
Half-step bump
Heavy
> 300,000 lbs
B747, B777, A380, A330
Full-category bump
Normalization Examples
- Heavy aircraft (B747, B777, A330, A380): “Light” becomes “Moderate” and “Moderate” becomes “Severe” in your briefing.
- Medium aircraft (B737, A320, CRJ, E-jets, biz jets): “Light” becomes “Light-Moderate” and “Moderate” becomes “Moderate-Severe.”
- “Smooth” and “Negative” reports are not adjusted — smooth air is smooth regardless of aircraft weight.
- Reports from Light GA aircraft are not adjusted — they are already calibrated for your experience.
- Unknown aircraft types are flagged with a low-confidence note in the briefing. The intensity is displayed as-reported (not normalized), and the briefing explicitly notes that the aircraft type could not be identified. This helps you weigh the report appropriately.
Correlation with PIREPs
Your turbulence briefing includes both model analysis and recent PIREPs. PlaneWX correlates the two to give you a complete picture:
Models and PIREPs Agree
When model-predicted turbulence at an altitude matches PIREP-reported turbulence nearby, confidence in the forecast increases. Your briefing will note this correlation.
PIREPs Exceed Model Predictions
Pilots reporting worse turbulence than models predict is common — models can underpredict due to coarse resolution. Your briefing will flag this discrepancy and prioritize the PIREPs as ground truth.
Models Show Turbulence, No PIREPs
Absence of PIREPs doesn't mean smooth air — it may mean no one has filed a report. Model-predicted turbulence without PIREP confirmation should still be taken seriously, especially for future flights where PIREPs don't exist yet.
Turbulence Exposure Time
Severity alone doesn't tell the full story. A brief encounter with moderate turbulence is very different from 40 minutes in it. PlaneWX estimates how many minutes of moderate-or-worse turbulence you can expect during each phase of flight — climb, cruise, and descent — based on the hybrid GTGN + NWP analysis and your estimated ground speed.
Where to Find It
- • Turbulence grid tiles — Each phase tile (Climb / Cruise / Descent) shows a sublabel like ~14 min when exposure is significant. The Max Severity tile shows the total across all phases.
- • Turbulence section bullets — When GTGN data is active and exposure exceeds 1 minute, a bullet appears: "Estimated moderate-or-worse turbulence exposure: ~14 min cruise, ~7 min descent = ~21 min total."
- • Briefing summary — If turbulence is driving a score reduction, exposure is noted alongside the severity.
- • Pre-flight checks — The "Why the score reduction?" callout includes per-phase and total exposure.
How It's Calculated
For each segment between waypoints, PlaneWX:
- Determines the hybrid severity (GTGN / NWP / blend) at the midpoint of that segment
- If severity is Moderate or worse, calculates the segment's flight time using your estimated ground speed at that waypoint
- Assigns the segment to its flight phase (Climb / Cruise / Descent) based on distance from departure
- Sums the exposure minutes per phase
Per-phase values are independently rounded before summing, so the displayed per-phase numbers always add up to the total shown — no rounding surprises.
GTGN required for exposure estimates. Exposure time is only computed when GTGN nowcast data is available and active for your route (departure within approximately 2 hours). For later departures, the phase tiles show peak NWP severity only, without exposure estimates — because NWP models provide a forecast of where turbulence may exist but cannot reliably estimate how long you'll be in it at the per-waypoint level.
Sustained vs. Peak Severity
A single severe waypoint in the middle of an otherwise smooth cruise phase doesn't mean the entire cruise is severe. PlaneWX applies exposure-aware severity smoothing to distinguish between brief spikes and sustained conditions — while still showing you the peak so you know the worst-case reading.
Sustained Severity (Phase Badge)
The severity shown on each phase tile (Climb / Cruise / Descent) is the sustained severity:
- • A severity level is sustained if it appears at 2 or more consecutive waypoints, or at ≥20% of waypoints in that phase
- • If neither threshold is met, the phase is classified at one level below the peak
- • This prevents a single anomalous GTGN spike from driving the entire phase badge to SEVERE
Peak Severity (Always Visible)
The highest single-waypoint severity is always shown alongside the sustained level:
- • The Max Severity tile shows the highest severity detected at any waypoint across all phases
- • Tooltip text on the route bar identifies the exact waypoint and source that drove the peak reading
- • The briefing narrative calls out peak severity separately when it differs from the sustained level
Why This Matters
Traditional tools show a single severity for the whole phase, which can either over-alarm (one spike = entire cruise shown as SEVERE) or under-alarm (the spike is averaged away). PlaneWX's sustained/peak split gives you both: the practical assessment you need for your go/no-go, and the worst-case data point for situational awareness.
GTGN SEVERE: Strong Caution, Not Hard No-Go
When GTGN detects turbulence that exceeds your hard limit (e.g., SEVERE turbulence when your maximum is MODERATE), PlaneWX handles it differently depending on the data source.
PIREP/AIRMET-Sourced Hard Exceedance
When corroborated PIREPs or active SIGMETs report turbulence beyond your hard limit, this triggers a hard violation — the pre-flight check fails and the WX Score reflects a no-go condition. PIREPs and SIGMETs are confirmed reports, not model estimates.
GTGN-Sourced Exceedance
When GTGN is the primary source of a hard limit exceedance, PlaneWX applies a caution with a score reduction of approximately 30 points rather than a hard no-go — typically bringing the WX Score to around 70%. The pre-flight check flags it as a caution, not a violation, and recommends verifying with other sources before departure.
Why the Distinction?
GTGN is a single-cycle nowcast — a snapshot of current conditions. Consecutive-cycle confirmation (two or more GTGN updates both showing severe conditions) provides much stronger confidence than a single cycle. Until multi-cycle confirmation is implemented, PlaneWX treats GTGN SEVERE as a "proceed with awareness" rather than an automatic hard no-go. A score around 70% means the conditions are worth taking seriously, but they don't justify grounding the flight the way a confirmed SIGMET or corroborated PIREP would. The final call remains with you as PIC per 14 CFR 91.3.
How Turbulence Affects Your WX Score
Turbulence deductions use your personal minimums — the same soft/hard limit system used for icing, winds, and visibility.
Soft / Hard Limit System
Comfort (Soft Limit)
The turbulence severity you're comfortable with. Exceeding this puts you in the caution zone.
Example: Soft = Light. If moderate turbulence is forecast, you'll see a caution deduction.
Limit (Hard Limit)
The maximum turbulence you'll accept. Exceeding this is unfavorable.
Example: Hard = Moderate. Severe turbulence forecast = unfavorable.
Turbulence vs. Other Hazards — Proportional Scoring
En-route turbulence (outside of mountain wave and convective contexts) is primarily a comfort and airframe stress issue, not a lethality issue. Icing, convection, and IMC below minimums carry materially higher risk of fatal outcome. PlaneWX reflects this in how penalties are weighted:
Icing / Convection
Higher penalty weight — confirmed severe can reach 0%
Turbulence (GTGN SEVERE)
~30 pt caution — score typically ~70%
Turbulence (Moderate)
10–20 pt caution — score typically 80–90%
Mountain wave and rotor turbulence are treated more conservatively, as severe mountain wave has a documented history of catastrophic structural failure. PIREP- or SIGMET-confirmed severe turbulence (any type) triggers a hard violation rather than a caution.
Reading the Turbulence Table
The turbulence table in the Models tab shows all the raw data behind the analysis. Here's what each column means:
| Column | What It Shows |
|---|---|
| Alt (ft) | Altitude of the upper level in each adjacent pair |
| Wind | Consensus wind speed at this altitude (average across models) |
| Dir | Consensus wind direction (degrees true) |
| Shear | Vertical wind shear in kt/1000ft between this level and the one below. Color-coded: red ≥ 9, amber ≥ 6, yellow ≥ 3 |
| Ri | Richardson number. Color-coded: red < 0.25 (unstable, KHI boost above 5,000 ft AGL), amber 0.25–0.5 (transitional), green ≥ 0.5 (stable, capped at LIGHT) |
| HRRR / GFS / ECMWF | Per-model severity assessment (SMO = smooth, LIG = light, MOD = moderate, SEV = severe) |
| Severity | Multi-model consensus severity (conservative max with outlier tempering) |
| Agreement | How well models agree: unanimous, majority, or split |
Limitations & Important Notes
Coarse Vertical Resolution
Weather models provide data at 24 pressure levels (1000–100 hPa), with layer thickness ranging from ~700 ft at low altitudes to ~7,500 ft at upper levels. Sharp shear layers that exist between these levels can be smoothed out in the data. Real turbulence can be more localized than what the model resolves. PIREPs from pilots at specific altitudes remain valuable for filling these gaps.
CAT Focus Only
The shear/Ri model analysis primarily detects Clear Air Turbulence (wind shear-driven). Mountain-wave risk is evaluated separately via the terrain-based cross-barrier flow detection described above. Convective turbulence (thunderstorms) and mechanical turbulence are covered by PIREPs, G-AIRMETs, and SIGMETs — all of which PlaneWX includes separately in your briefing.
Forecast, Not Observation (NWP); Observation, Not Forecast (GTGN)
NWP model-derived turbulence is a forecast of where conditions will support turbulence, not a real-time measurement. GTGN is the opposite: a near-real-time observation that decays in relevance over time. PlaneWX’s per-waypoint hybrid system automatically selects the best source for each segment of your route based on when you’ll actually be there. For immediate flights, GTGN’s EDR nowcast provides ground-truth conditions; for flights hours or days out, NWP models provide the forward-looking assessment. Always cross-reference with current PIREPs regardless of departure time.
Not a Substitute for Official Briefing
PlaneWX's multi-model turbulence analysis supplements — but does not replace — official weather products. Always cross-reference with current G-AIRMET Tango, SIGMETs, and PIREPs. The pilot in command is always responsible for the final go/no-go decision per 14 CFR 91.3.
About This Documentation
PlaneWX’s turbulence analysis combines the FAA/NCAR GTGN EDR nowcast (updated every 15 minutes via direct LDM ingest) with multi-model wind shear analysis from HRRR, GFS, and ECMWF soundings accessed via the Open-Meteo API. A per-waypoint temporal relevance model automatically selects GTGN, NWP, or a conservative blend for each segment of your route based on the effective age of the GTGN data at your estimated time of arrival. Canonical turbulence phases consolidate GTGN/NWP hybrid data with G-AIRMET Tango advisories and GA-normalized PIREPs for a complete assessment. All calculations are performed server-side in real-time for each briefing.
Research
- Iterative AI Validation of Turbulence Detection — 8 debates, 3 days, shear guard & rotor detection
- Adversarial AI Validation of Mountain Wave Detection — Scorer parameter, relief-scaled thresholds, 6 improvements