Convective Watch — Thunderstorm Scoring
How PlaneWX assesses thunderstorm risk along your route using multi-model atmospheric analysis, composite scoring, and horizon-scaled penalties.
The Big Picture
Before every flight, PlaneWX samples 9–15 points along your route and pulls atmospheric data from up to five weather models chosen by geography (for example HRRR, HRDPS, ICON-EU, ICON-D2, AROME on western European routes, GFS, and ECMWF). At each point, it measures four atmospheric properties that together determine whether thunderstorms are likely, and if so, how strong they’d be.
These measurements are combined into a single Threat Score (0–100) that appears on your Convective Watch card. The higher the score, the greater the thunderstorm risk along your route.
0–15
None
No thunderstorm risk
15–35
Isolated
Scattered pop-ups possible
35–60
Scattered
Multi-cell storms likely
60–100
Widespread
Severe-capable instability
These four labels measure instability, not sky coverage. They come entirely from the sounding — CAPE, lifted index, K-index and cap strength. “Widespread” means the environment supports storms along the whole route, not that storms fill the sky. Whether that environment can actually build the continuous, unavoidable convection the heaviest deductions are priced for depends on wind shear, and that is checked separately — see storm organization below.
What makes this different? Most weather apps show CAPE or a single instability number. PlaneWX combines four atmospheric indices, weights them by importance, checks whether the models agree, and factors in atmospheric caps that might suppress storms. The result is a score that reflects true risk — not just raw energy.
Reading the Convective Watch Card
When conditions support thunderstorms along your route, a Convective Watch card appears in your briefing. Here’s what each metric means:
Threat Score (0–100)
The composite score combining all four atmospheric indices, model agreement, and suppression factors. This is the single best number for gauging overall thunderstorm risk along your route.
Instability / CAPE (J/kg)
The maximum CAPE (Convective Available Potential Energy) found along your route — the atmosphere’s “fuel tank” for storm updrafts. Higher means stronger potential storms.
Affected Segments
How many of the sampled route segments show thunderstorm potential. “2 / 15” means 2 of 15 points along your route have conditions that support storms — the threat is localized, not route-wide.
Two different counts matter, and they are not the same number. This one counts every segment carrying any convective signal, however weak. When PlaneWX decides whether you have gaps to steer through, it counts only the segments at scattered coverage or heavier — see Counting the gaps.
Storm Probability / K-Index
Predicts the probability of air-mass (pop-up) thunderstorms based on temperature and moisture through the atmosphere. Above 35 is high probability.
Stability / Lifted Index
Measures whether air parcels will rise or sink. Negative values mean the atmosphere is unstable. Below −4 indicates severe weather potential.
Cap Strength / CIN (J/kg)
Convective Inhibition — the atmospheric “cap” that suppresses storm development even when energy is high. Low CIN (<25 J/kg) means the cap is essentially gone and storms can fire with no trigger. High CIN (>100 J/kg) means the cap is strong and convection is unlikely despite high CAPE. The gauge is inverted: red is dangerous (weak cap), green is suppressed (strong cap).
Storm Trigger / LFC (ft)
The Level of Free Convection — the altitude at which a rising air parcel becomes warmer than the surrounding atmosphere and starts rising on its own, creating a self-sustaining storm. A low LFC (<6,500 ft) means minimal lift is needed to initiate and sustain a storm, and the atmosphere becomes self-sustaining very quickly once triggered. SPC research links low LFC environments to elevated tornado probability when supercell shear is also present.
Where the Data Comes From
PlaneWX draws on numerical weather models, WPC discussions, and TAF phrases. No single weather product captures the full convective picture, so combining them gives you the most reliable assessment.
Numerical Weather Models
Regional and global models (for example HRRR 3 km over CONUS, HRDPS over Canada, ICON-EU and ICON-D2 over Europe, AROME 2.5 km on western European routes, plus GFS ~13 km and ECMWF ~9 km globally) provide atmospheric profiles at multiple pressure levels. PlaneWX extracts CAPE, K-Index, Lifted Index, and CIN from each model at each sample point.
WPC Discussions (CONUS only)
For US flights, the Weather Prediction Center’s short-range and extended discussions are analyzed for convective language. This adds regional context that models alone may miss. WPC data is geographically filtered so you only see language relevant to your route — not weather in distant regions.
TAF Convective Scanning
Terminal Aerodrome Forecasts at departure, arrival, and en-route waypoint airports are checked for thunderstorm indicators including TS, TSRA, VCTS, CB, TCU, and PROB30/40 convective groups. En-route TAF scanning catches convective weather forecasts that only appear at airports along your corridor, not just at your endpoints.
How Cumulonimbus Is Graded
A CB in the forecast is not automatically a no-go. How much of the sky it covers changes what you can actually do about it, so the sky-coverage prefix sets the severity:
Caution. Isolated or scattered cells you can normally see and fly around. The score is reduced, not floored.
No-go. Broken or overcast CB means embedded convection, where visual deviation is unreliable and you can fly into a cell you never saw.
One storm is charged once. A single forecast period often names the same storm several ways at once — TSRA, VCTS and SCT025CB can all appear together. That is one event, so it produces one line in your score breakdown, listing every token the TAF carried: [VCTS, SCT025CB]. You see the full picture without being penalised three times for one thunderstorm.
METAR Convective Scanning
Real-time METARs at departure, arrival, nearby (30 NM radius), and en-route corridor airports are scanned for live convective indicators. This catches storms that are happening now, which models may not have predicted.
Potential Max Tops (Equilibrium Level)
When convective conditions are detected, PlaneWX computes the Equilibrium Level (EL) — the theoretical maximum altitude a thunderstorm updraft can reach based on the atmospheric sounding. This is displayed as “Potential Max Tops” alongside your cruise altitude to give you immediate vertical situational awareness.
Important: The EL represents a theoretical maximum, not a guarantee. Actual storm tops can exceed the EL due to overshooting, and significant turbulence and hail can exist thousands of feet above the visible cloud top. PlaneWX never suppresses convective warnings based on altitude — a pilot at FL450 still needs to know about a CB below them because storms can grow at 6,000 ft/min, and hail can be ejected into clear air downstream.
How Potential Max Tops Are Calculated
The Lifted Parcel Method with Most Unstable (MU) parcel fallback.
How Potential Max Tops Are Calculated
PlaneWX uses the Lifted Parcel Method — the same approach used by the National Weather Service. A parcel of air is virtually “lifted” from the surface through the atmospheric profile:
- Below the LCL (cloud base), the parcel cools at the dry adiabatic rate
- Above the LCL, the parcel cools at the moist adiabatic rate (slower, because condensation releases latent heat)
- The parcel rises until it becomes cooler than the surrounding air — this is the Equilibrium Level
If the surface parcel doesn’t find an EL (e.g., a surface inversion blocks it), the algorithm tries the Most Unstable parcel — the level with the highest equivalent potential temperature in the lowest 300 hPa. This catches elevated convection that can form above inversions, especially at night.
How the Threat Score Is Calculated
The Threat Score uses a weighted composite of four atmospheric indices, adjusted for model agreement and atmospheric suppression. Here’s the breakdown.
Step 1: Convert Each Index to a Signal Score (0–1)
Each atmospheric index is mapped to a 0–1 signal based on meteorologically verified ranges.
Step 1: Convert Each Index to a Signal Score (0–1)
CAPE Signal (Weight: 50%)
Convective Available Potential Energy — the energy available to fuel storm updrafts.
Lifted Index Signal (Weight: 30%)
Atmospheric stability — negative values mean unstable air that wants to rise.
K-Index Signal (Weight: 20%)
Thunderstorm probability from temperature and moisture profiles. Automatically excluded at airports above 5,000 ft where the formula is unreliable.
Step 2: Agreement Multiplier & CIN Suppression
The raw score is adjusted based on whether indices agree and whether an atmospheric cap is holding storms down.
Step 2: Agreement Multiplier & CIN Suppression
Agreement Multiplier
If CAPE, Lifted Index, and K-Index all agree on high risk, confidence is high (multiplier near 1.0). If they disagree — say CAPE is high but Lifted Index is stable — the score is reduced. Missing data is treated as neutral, not as disagreement.
If K-Index is unavailable (e.g. high elevation), its weight is redistributed to CAPE and Lifted Index proportionally.
CIN (Convective Inhibition) Suppression
CIN measures the atmospheric “cap” holding storms down. High CIN means storms are unlikely to fire even if energy (CAPE) is high. The system uses CIN as a multiplier that reduces the score when a strong cap is present.
“Loaded Gun” Safety Floor
When extreme instability is present (very high CAPE + indices in strong agreement), PlaneWX will not let CIN suppress the score below 40%, even if the cap is strong. This is a “loaded gun” scenario — the atmosphere is primed. If a cold front, dryline, or other trigger breaks the cap, explosive storms will develop rapidly. You need to know about this risk.
Step 3: The CAPE Gate
A hard minimum that prevents false alarms when there simply isn’t enough energy for storms.
Step 3: The CAPE Gate
If the maximum CAPE across all models at a sample point is below 100 J/kg, the threat is automatically set to “None” regardless of what the other indices say. Without sufficient energy, thunderstorms simply cannot form.
This prevents false positives at high-elevation airports (like Mexico City at 7,300 ft) where K-Index and Lifted Index might look threatening but there isn’t enough atmospheric energy to produce storms.
Step 4: Multi-Model Consensus
How the models queried for your route are reconciled into a single assessment.
Step 4: Multi-Model Consensus
The regional and global models queried for your route often report different CAPE values because they use different parcel types (surface-based, mixed-layer, most-unstable). To handle this:
- •Median CAPE is used for the actual score — this resists outliers from any single model
- •Maximum CAPE is used for the safety gate — if any model shows energy, we don’t dismiss it
This “safety-first” approach prevents a single model from over-warning you (the median keeps the score grounded) while also preventing all models from silencing a real threat (the max keeps the gate honest).
Final Composite Score
Threat Score = (Weighted Signal Score) × CIN Suppression × Agreement Multiplier
Weighted Signal = (CAPE Signal × 50%) + (LI Signal × 30%) + (K-Index Signal × 20%)
Horizon-Scaled Penalties
Convective forecasts become dramatically more accurate as departure approaches. A thunderstorm forecast 36 hours out is far less certain than the same forecast 3 hours before departure. PlaneWX accounts for this by scaling convective WX Score penalties by the forecast horizon.
Horizon Multiplier
Built-In Safety Guards
CAPE Gate (100 J/kg)
Below 100 J/kg of CAPE, thunderstorms cannot form. The system short-circuits to “None” regardless of other indices, preventing false alarms at high-elevation or cold-weather airports.
Elevation Guard (K-Index)
The K-Index formula uses temperature at 850 hPa and 700 hPa. At airports above 5,000 ft, these pressure levels can be below the surface, making K-Index mathematically invalid. PlaneWX automatically excludes K-Index at high-elevation points and relies on CAPE and Lifted Index instead.
“Loaded Gun” Floor
When extreme energy is present and multiple indices agree, CIN suppression is capped at 0.40× to ensure pilots are warned about “capped but dangerous” environments. If the cap breaks, explosive convection follows.
WPC Geographic Filter
WPC text is filtered to your route’s geographic region. A flight from California won’t see convective language about the Southeast. International flights outside the US bypass WPC entirely and rely solely on model data and TAFs.
Model-Aware METAR Suppression
Some automated weather stations (especially AO2 sensors in the Southeast) routinely report distant lightning (“LTG DSNT”) even when no convective weather threatens your route. A single station’s distant lightning report can create a false-positive score deduction on an otherwise perfect VFR day.
To prevent this, PlaneWX cross-references METAR convective signals against the NWP models queried for your route. When every model unanimously confirms zero convective potential along your route (CAPE = 0, no threatened segments), low-severity signals like distant lightning are suppressed from the WX Score calculation. The observation still appears in your briefing text so you’re aware of it.
Active thunderstorm observations (TS, VCTS, tornado, hail) are never suppressed, and neither are signals from three or more stations — widespread reports always count regardless of what models say.
Time-of-Day Awareness
Surface-based convection is heavily driven by solar heating. Thunderstorms are far more likely during peak afternoon heating (15–19 local) than early morning. PlaneWX applies a diurnal multiplier to the composite threat score based on local solar time at your departure, reducing scores for flights outside the peak convective window without ever fully suppressing warnings.
Storm Type & Bulk Wind Shear
PlaneWX computes 0–6 km bulk wind shear from model sounding data to determine whether convective conditions favor short-lived pulse storms or dangerous organized severe weather. This context appears in the Convective Watch card narrative, and it sets the ceiling on your en-route convective deduction.
Why organization sets the deduction ceiling
An unstable airmass is not the same thing as a line of storms. Without shear to tilt and sustain an updraft, even a very unstable atmosphere produces pulse cells that live 20–40 minutes and leave room to fly between them. With shear, the same instability organizes into clusters and lines that can close a corridor. The heaviest en-route convective deductions — 20, 35 and 40 points — are priced for the second case, so PlaneWX now requires evidence the second case exists before charging them.
Any one of these is enough to keep the full ladder in force:
- •Deep-layer bulk shear of 20 kt or more — anything above the pulse threshold in the table above
- •An active Convective SIGMET on your corridor — a meteorologist has already judged the convection organized
- •Radar showing scattered-or-worse coverage in the 10 NM corridor you actually fly — observation beats forecast
- •Embedded or line storm mode — cells you cannot see to avoid, which always carries the full 40
When none of them is present, the en-route route-coverage deduction caps at 10 points — the same value isolated convection has always carried, for the same reason: individually identifiable cells with navigable gaps. Terminal thunderstorm deductions, SIGMET deductions, METAR and TAF convective findings are unaffected and score exactly as before.
Missing data does not buy relief. If shear cannot be determined for your route, the storm mode is treated as unknown and the full ladder stands. The cap only applies when there is positive evidence of disorganization, never merely an absence of evidence.
Counting the gaps: which segments hide one
Once organization is established and your route lands on the scattered rung, one question decides the deduction: are there at least two segments you could steer through? Two or more clear segments is 20 points; fewer is 35.
That count used to treat any segment with a convective signal as blocked, including segments carrying nothing more than a single isolated cell. On a summer airmass day that is nearly every segment on the route, so the gap count collapsed to zero and the route was billed 35 points even where the actual coverage was thin. A route with scattered coverage on 3 of 15 segments read the same as one covered end to end.
Gaps are now counted from the segments at scattered coverage or heavier. An isolated cell no longer hides a gap, for the same reason an isolated route tier has always capped at 10: individual cells are widely spaced and you can see them and go around.
This moves one rung of one tier. A scattered, organized route that genuinely has gaps now bills 20 instead of 35. Nothing else changes — isolated, widespread, embedded and line routes, unorganized routes already capped at 10, and every terminal, SIGMET, METAR and TAF deduction all score exactly as before. The correction can only ever lower a deduction, never raise one, and if the coverage detail is unavailable for your route the stricter old count still applies.
Sounding Profile Badges & Threat Brief
When the atmospheric sounding shows high instability combined with a weak or absent cap, PlaneWX displays a sounding profile badge and a Threat Brief paragraph inside the Convective Watch card. These describe how close the atmosphere is to producing thunderstorms — and how little it might take to trigger them.
High energy building under a weakening cap — storms possible if triggered
CAPE ≥ 1,000 J/kg with CIN between 50–150 J/kg. The atmosphere has significant energy but it’s being held down by a cap. If a trigger arrives — peak afternoon heating, a cold front, a dryline, or orographic lift — storms can develop. Monitor carefully through the afternoon hours.
Cap is eroding — weak heating or a minor boundary can fire storms
CAPE ≥ 1,000 J/kg with CIN between 25–50 J/kg. The cap is weakening and the atmosphere needs only a weak trigger to produce convection. A sea breeze, a terrain upslope, or afternoon heating may be enough. Convective development can be rapid once the cap is breached.
Cap is essentially gone — rapid storm development possible at any time
CAPE ≥ 2,500 J/kg with CIN < 50 J/kg, or any CAPE ≥ 1,000 J/kg with CIN < 25 J/kg. The suppressing cap has essentially eroded. Convective development can occur with little or no external trigger — storms may initiate on their own once minimal lift is present. Expect rapid intensification and limited warning time.
What is the Threat Brief?
When a sounding profile badge appears, a Threat Brief paragraph is generated directly below it. It translates the model sounding numbers into plain English, covering:
- •What the sounding tier means and why it matters for convective timing
- •The CAPE value and cap status (CIN) in plain terms — e.g. “cap of only 3 J/kg (essentially gone)”
- •LFC altitude and what it means for how quickly storms become self-sustaining
- •Wind shear implications — whether supercell structure or tornado potential is present
- •Storm tops vs. cruise altitude when potential tops exceed your planned flight level
The Threat Brief is also surfaced in the Gotchas section and Briefing Summary so the risk is visible at every level of the briefing without requiring you to scroll to the Convective Watch card.
Important: The sounding profile badge describes atmospheric priming, not active storms. A loaded-gun or explosive-gun environment can exist in the morning with no clouds and clear skies — the danger is that conditions are set for rapid development later in the day. These badges do not affect the WX Score directly; they are informational overlays to help you understand the sounding environment behind the convective threat.
Convective SIGMETs & SPC Day 1 Outlook
Convective SIGMETs
Active Convective SIGMETs (WST) that intersect your search corridor are parsed from SIGMET data and shown in the Convective Watch card with red indicators. Display and scoring are not the same step: a polygon in the corridor is always worth seeing, but the WX Score asks whether you will actually encounter that convection.
- • Time-aware encounter: Storm motion, your ground track, and when you reach the area scale the deduction. Flying away from a SIGMET behind departure, or missing the polygon entirely in time, reduces or removes the en-route penalty versus a head-on sustained encounter.
- • Tops & deviation: Cloud tops (especially FL350+) and how hard a lateral end-run looks still drive severity when an encounter is real.
- • Validity: Score lines note SIGMET valid times; refreshes re-evaluate after expiry.
One Storm, One Deduction (Ledger Dedup)
The same convective event can show up as a Convective SIGMET, a departure TAFVCTS/TSRA group, and model CAPE language. PlaneWX keeps a server-side convective ledger so those products do not stack full penalties for one event. Distinct hazards still score separately — for example, en-route SIGMET convection versus a different terminal thunderstorm period that is not the same cell.
Low CAPE Does Not Erase Active Products
Model instability along the corridor can look modest while a Convective SIGMET or TAF thunderstorm group is still active. When the server scores those products, the SUMMARY and synoptic narrative are required to name them — they must not claim “no significant convective threat” while the score breakdown still shows a SIGMET or TS deduction.
SPC Day 1 Categorical Outlook
For CONUS flights, PlaneWX fetches the Storm Prediction Center Day 1 categorical outlook and checks whether your route passes through any risk area. The maximum risk level encountered is displayed as a metric cell on the Convective Watch card.
TAF Time-Window Filtering
En-route TAFs are now filtered against your actual flight window. Convective hazards in TAF groups that don’t overlap with your departure-to-arrival window are excluded, preventing false alarms from storms forecast hours before or after your flight.
Live Radar Corridor
For near-term flights, Convective Watch also shows live MRMS radar along your route — dual Model / Radar badges, a dBZ-colored segment bar, and a status strip with data age.
When radar has genuinely looked at your corridor and found it quieter than the models forecast, the route-coverage thunderstorm deduction above is reduced, and the breakdown line says so along with the age of the radar image. Radar only ever moves that one deduction, and only downward — SIGMETs, TAF and METAR thunderstorm reports, probability groups, and hard limits are never softened.
Read the Live Radar Corridor guide →When the Air Has Fuel but Nothing Is Forecast to Fire
The Threat Score is instability — CAPE and friends. On a quiet summer day that can still look like scattered storms.
NOAA’s National Blend of Models also publishes a thunderstorm chance at each point on your route, at the time you would be there. If that chance stays low, PlaneWX can ease the route thunderstorm line — the one billed from CAPE. It never adds points. It does not touch a TAF thunderstorm, a SIGMET, or a METAR.
NBM only moves that line inside 72 hours of departure. Farther out, the score stays on the model thunderstorm haircut.
Live radar still has a say for the first couple of hours. If a cell is already on the corridor, we keep at least an isolated deduction. Quiet radar, or no radar yet, lets a low chance zero that line.
In Score Breakdown it looks like this:
−10 Scattered thunderstorms along the route [NBM occurrence along route caps coverage as isolated]
Open Sources on the Convective Watch card to see the max chance, the NBM tier, and whether the score moved.