Solano weather glossary

A Solano-specific indicator measuring the historical precision of a weather model on your exact geographic point. It is computed by comparing past forecasts (J-1 to J-15) to the ERA5T reanalysis, via the MAE (mean absolute error).

This score answers the question: has this model, on this exact anchorage or col, been accurate in the past? It is computed separately per variable (wind, gusts, direction, rain, temperature, cloud, pressure) and per lead time (J+1 to J+7). Shown in the Accuracy tab of the Consistency panel.

A complement to the stability score: stability says whether the model has changed its mind between runs; accuracy says whether it was right.

A measure of the energy available for vertical air motion. The higher the CAPE, the faster a parcel of air can rise if it is lifted, with an increased risk of deep convection (thunderstorms, cumulonimbus).

Interpretation reference points: CAPE < 300 J/kg = stable atmosphere, weak convection; 300–1000 J/kg = moderate convection, thunderstorms possible if triggered; > 1000 J/kg = unstable atmosphere, thunderstorms probable and possibly violent; > 3000 J/kg = extreme conditions, tornadoes possible (rare at our latitudes).

CAPE alone isn't enough — a triggering mechanism (front, terrain, convergence) is needed to activate convection, and it is always read with the CIN and the Lifted Index. Solano combines them in a thunderstorm risk index.

The energy (in J/kg) that prevents a parcel of air from rising to the level where convection runs away. It is the "cap" of the atmosphere: as long as it isn't overcome (by daytime heating, terrain, a front…), the storm doesn't trigger — even with an enormous CAPE.

Reference points: CIN < 25 J/kg = open cap, easy triggering; 50–150 = needs a forcing; > 300 = strong cap, convection suppressed. A large CAPE under a strong CIN is a "loaded gun": unlikely, but explosive if the cap breaks.

A model whose grid is fine enough (≈ 1 to 3 km) to compute the updrafts of a storm itself, rather than estimating their average effect with a formula. Coarse-grid models (13 to 25 km: GFS, ECMWF, ARPEGE) don't "see" a cell a few kilometres across: they parameterise it. A convection-permitting model, on the other hand, draws it — it places cells, with a time and a place.

In Solano: AROME HD, ICON-2I, ICON-D2 and ICON-CH1/CH2. Their trade-off is a short horizon (34 h to 5 days) and a regional footprint — and a caveat in reading: the exact placement of a cell in a deterministic run at 2 km moves by tens of kilometres from one run to the next. They answer "where and when it may fall", not "is the atmosphere charged" — that is the role of the thunderstorm risk.

The process by which a numerical model folds in real weather observations to build its initial state — the most faithful possible snapshot of the atmosphere at launch time. These observations come from very diverse sources: surface weather stations, radiosondes, satellites, marine buoys, aircraft measurements in flight…

Assimilation reconciles this heterogeneous — sometimes contradictory — data with a first guess of the atmospheric state (from the previous run) by minimising the overall gap. The result is the initial condition from which each run starts. The quality of the assimilation directly conditions the accuracy of short-lead forecasts.

The European Centre for Medium-Range Weather Forecasts, based in Reading (United Kingdom). An intergovernmental organisation founded in 1975 and supported by 35 member states. Recognised as one of the world references in numerical weather prediction.

ECMWF operates the IFS model (Integrated Forecasting System) — available in Solano under the name ECMWF IFS — and produces the ERA5 reanalysis, used as the reference for the accuracy score. ECMWF also publishes the SEAS5 seasonal forecasts used in Solano's Climatology section.

ERA5 is the atmospheric reanalysis from ECMWF covering the period from 1940 to today. It is not a direct field measurement: it is a coherent reconstruction of the state of the atmosphere hour by hour, obtained by data assimilation of sparse observations — surface weather stations, radiosondes, satellites, marine buoys — into ECMWF's IFS model.

Why "sparse data"? Weather stations and buoys only cover a fraction of the globe's surface (the oceans in particular). ERA5 fills these empty zones coherently, using the physics of the atmosphere to interpolate between observation points. The result is a best estimate of what happened — not an absolute truth.

ERA5T is the "real-time" version of this reanalysis, available with roughly a 1-to-2-day delay. Solano uses it as the reference for computing the MAE (accuracy score) and for feeding the 📡 Verif. tab of the Consistency panel.

The altitude at which the air temperature is 0 °C. It is a key thermal boundary: below it, precipitation falls as rain; above it, as snow. In winter or in disturbed weather, the freezing level can drop down to the coastal zone.

Useful for mountain activities (hiking, skiing), for estimating the risk of ice at altitude, or for understanding why it snowed at 800 m but not at 400 m. In Solano, the freezing level is available via the Snow/Freezing level layer.

Note: ECMWF IFS and ARPEGE Europe don't transmit this variable via Open-Meteo — the layer may be absent for these models.

A line joining points of equal atmospheric pressure on a map. Isobars let you visualise the structure of the pressure field at a glance and identify weather systems:

• Closed, tight isobars with values decreasing towards the centre = a low (low pressure, disturbed weather)
• Closed isobars with values increasing towards the centre = a high (high pressure, stable weather)
• Isobar spacing = the strength of the pressure gradient → strong wind if tight, weak if spread out

In Solano, isobars are available as a map overlay (MSL pressure parameter), drawn every 4 hPa.

A unit of speed used in marine meteorology, aviation and navigation. 1 knot = 1 nautical mile per hour = 1.852 km/h. Solano expresses wind speeds and currents in knots.

Correspondence with the Beaufort scale:

The rate at which temperature falls with altitude, in °C/km. Solano uses the mid-tropospheric 700→500 hPa lapse rate: the steeper it is, the more unstable the atmosphere and the more vigorous the convection for the same CAPE.

Reference points: < 5.5 °C/km = soft profile; > 7 °C/km = steep profile (explosive convection). This is one of the reasons why "a CAPE of 2,000 isn't the same" depending on the season and the situation.

The interval between the moment a forecast is issued and the moment forecast. J+0 denotes the run's issue time (an "instantaneous" forecast), J+3 a 3-day forecast, J+7 a 7-day one, etc.

Uncertainty grows with lead time: models are generally reliable up to J+3 to J+5, and their skill decreases rapidly beyond that. That is why Solano's accuracy score is measured separately for each lead time — a model can be excellent at J+2 on your spot and mediocre at J+5.

An instability index (dimensionless): a parcel of low-level air is lifted to 500 hPa and its temperature is compared to that of the environment. Negative = unstable (the parcel is warmer, it keeps rising); positive = stable.

Reference points: LI > 0 = stable, no storm; 0 to −2 = marginal; −2 to −4 = unstable, thunderstorms probable; < −6 = very unstable, violent thunderstorms. More robust than CAPE alone for judging instability, because it is less sensitive to the season.

A statistical measure of the gap between a model's forecasts and a reference. In Solano, the reference is the ERA5T reanalysis. The formula:

This calculation is carried out separately for each model, each variable (wind, gusts, direction, precipitation, temperature, cloud, pressure) and each lead time (J+1 to J+7). A wind MAE of 2 knots at J-3 means the model was on average 2 knots off over your last 15 days when forecasting three days ahead.

A computer program that solves the equations of fluid dynamics (Navier-Stokes equations, thermodynamics, cloud microphysics…) on a three-dimensional grid of the atmosphere. The world is divided into cells of a few kilometres to a few tens of kilometres — each cell exchanges mass, heat and momentum with its neighbours at every time step.

The model is initialised from an "initial state" (see data assimilation) and computes the probable evolution of the atmosphere hour by hour up to several days. It starts from scratch at each run. The horizontal resolution varies: AROME France HD = 1.5 km, ICON-EU = 7 km, ECMWF IFS = ~9 km, GFS = ~13 km.

The models available in Solano: ARPEGE Europe, AROME France HD, ECMWF IFS, GFS NOAA, ICON-EU, ICON seamless, UKMO (UKV + global), GEM GDPS/RDPS/HRDPS (Canada), ICON-2I (Mediterranean).

A weather observation made by a weather balloon fitted with a radiosonde — a box of a few hundred grams that captures temperature, pressure, humidity and wind speed as it climbs through the atmosphere, up to about 30 km altitude.

About 900 stations worldwide launch two radiosondes a day, at the synoptic hours of 00Z and 12Z. These vertical measurements are one of the most valuable sources for data assimilation: they provide a full profile of the state of the atmosphere where satellites only see integrated layers.

The balloon generally bursts between 20 and 35 km (expansion from depressurisation), and the radiosonde comes back down under a parachute. Some are recovered and reused.

A satellite product (NWC SAF, EUMETSAT) that automatically detects and tracks thunderstorm cells from their development phase, using geostationary imagery. It spots nascent storms even before the first lightning, and — being satellite-based — also covers the open sea, where ground radars see nothing.

Each cell is outlined by a contour and labelled with its minimum top temperature (°C): the colder it is, the higher the top and the more powerful the cell. Reference points: > −40 °C = weak (cumulus/congestus) · −40 to −55 °C = moderate · −55 to −65 °C = strong · < −65 °C = very intense (overshooting tops).

In Solano, RDT is an overlay layer of the satellite imagery, alongside lightning and infrared. The "See the sky now" button of the thunderstorm risk activates it automatically to move from forecast to observation.

A coherent, homogeneous reconstruction of past atmospheric states, obtained by passing archives of historical observations through a modern numerical model via data assimilation. Unlike a direct measurement, a reanalysis fills the zones without stations and produces a global state at every hour across the whole globe.

Advantages over raw observations: spatial and temporal consistency, global coverage (including oceans and sparsely observed regions), derived variables available (boundary layer, fluxes…). Limits: accuracy depends on the density of observations available at each period — recent decades are better analysed than the 1940–1960 years.

Solano's reference reanalysis is ERA5 (ECMWF). Other reanalyses exist: MERRA-2 (NASA), JRA-55 (JMA), CFSR (NOAA).

One launch of a forecast model at a precise moment, called the "run validity time". Operational models run at regular intervals: most produce 4 runs a day at the synoptic hours of 00Z, 06Z, 12Z and 18Z (UTC). Some high-resolution models run every hour or every 3 hours.

Each run starts from scratch with the most recent available assimilation and produces forecasts up to its time horizon (from 2 to 16 days depending on the model). Two successive runs of the same model can give appreciably different forecasts for the same target day — it is this variability that Solano's stability score measures.

The delay between a run's validity and its public availability (processing delay) varies: AROME France HD is available ~1h30 after its validity time, ECMWF IFS in ~3h, GFS in ~4h.

A Solano-specific indicator, scored from 0 to 10, measuring the consistency of a single model's successive runs. A high score (8–10) means the model "hasn't changed its mind" between its recent runs for the target date — the sign of a well-established, predictable atmospheric situation. A low score (0–4) means significant divergence between runs, revealing strong uncertainty.

Recent runs count more than older ones: the weighting decreases exponentially (yesterday's run = high weight, a run 4 days old = low weight). The score measures the internal consistency of the forecasts, not their accuracy — a model can be very stable and very wrong if the atmospheric situation is poorly initialised.

Complement: the accuracy score (ERA5T MAE) answers the accuracy question.

A Solano-specific categorical index — Low · Moderate · Marked · Severe — computed hour by hour to go beyond reading CAPE alone. It combines the fuel (CAPE sharpened by the Lifted Index and the lapse rate) and a cap factor (the CIN): a lot of energy under a strong cap gives a low risk but a "loaded gun" flag ⚠️. The shear then refines the category — "Severe" is reserved for organised storms.

The index is anchored on the GFS model, the only one to provide every ingredient — so it describes the environment at the basin scale (point brought back to ≈ 25 km) and doesn't localise the cells, which is the role of triggering. It is accompanied by a measure of agreement between models (CAPE spread): when they diverge, the risk is flagged as uncertain. It describes a favourable environment, never a thunderstorm at a precise time.

The second reading of thunderstorms in Solano, alongside the thunderstorm risk. Where the risk describes an environment at the basin scale (is it charged? is the cap holding? would it organise?), triggering answers a different question: does a convection-permitting model place a cell on this point, and at what time?

It reads from the total precipitation of the fine model, reinforced by a simultaneous gust. Counter-intuitive but measured: when a model resolves convection, the shower is in the total rain and not in the "showers" field, where it is now just a residue. Three levels: shower, formed cell, vigorous cell.

⚠️ It never enters the /10 score of the thunderstorm risk, and for a physical reason: a model that triggers a cell consumes the CAPE. A low CAPE at 5 pm may mean "storm in progress", not "calm". The two readings complement each other, they don't average.

A meteorological symbol representing the direction and speed of the wind simultaneously. The shaft points in the direction the wind comes from (a symbol pointing up indicates a north wind). The ticks on the shaft (the "barbs") encode the speed:

• Short tick = 5 knots
• Long tick = 10 knots
• Pennant (filled triangle) = 50 knots
• Empty circle = calm (< 2 knots)

A barb with a pennant and a long tick = 60 knots. This convention is universal in meteorology and aviation.

The variation of the wind (speed and direction) between the surface and altitude, measured here as the vector difference between the wind at 10 m and the wind at 500 hPa (≈ 5.5 km), in m/s. Shear doesn't create the storm but decides its organisation.

Reference points: < 10 m/s = isolated, short-lived cells (showers); 15–20 = organised systems, supercell possible; > 20 = strongly organised (squall lines, hail, downbursts). It's what distinguishes a thundery shower from a dangerous, lasting system.