Every way to read the wind at one place.

How much stronger the wind is aloft. Each row is a model level (10, 80, 120, 180 m) and each cell is coloured on one speed ramp, so a bright top row over a dark bottom row means strong shear — typical at night when the ground cools and the surface layer decouples. The filament line is your working height; between levels it is interpolated, outside 10–180 m it is extrapolated and marked as such.

Seven national models, each drawn as issued. When the hairlines bunch, the forecast is well constrained; when they fan out, nobody knows yet and a single number would be false comfort. The filament is ECMWF IFS, our primary — it is never averaged with the others. The agreement chip summarises the next 24 hours from the speed spread and the circular spread of direction.

One model run many times with slightly different starting conditions. The inner band holds half the members, the outer band 80 % of them, and the paper line is the median. A plume that widens quickly with lead time means the forecast is losing skill; the member count is stated because a probability from 28 members is not the same as one from 122. Raw member fractions, not calibrated probabilities.

Read it like an insurance table: for any wind speed on the x-axis, the curve gives the chance it is exceeded in the chosen window. The 'peak hour' curve answers 'will it happen at all in the window' (the worst single hour); the 'member-hours' curve answers 'how much of the time'. The curves are reconstructed from the published ensemble percentiles, so the tails beyond P10/P90 are approximate.

Inland, wind usually rises after sunrise as the sun stirs the lower atmosphere and falls away at dusk; on coasts and hills the cycle is weaker or reversed. This compares the coming week's average day with the last 30 days of reanalysis. Thirty days is recent weather, not a climate normal — it tells you what has been typical lately, not what is typical for the season over decades.

Gust divided by mean speed. Over open, flat ground a factor of about 1.3–1.5 is normal; above roughly 1.8 the air is turbulent — rough terrain, showers, convection or simply a light mean wind with occasional puffs. A high factor matters for anything sensitive to peak load (lifts, drones, scaffolding sheets) even when the mean looks benign. Models routinely give 1.8–2.2 in light inland winds, so always read the factor together with the gust value itself. Undefined below 2 mph mean, where the ratio is meaningless.

The forecast translated into the Beaufort scale, the 1805 force scale still used in marine and public forecasts. Each cell is one hour; the force comes from the 10 m mean speed, not the gust. The sea-state descriptions are for open sea well away from the coast — sheltered water and shallow bays look very different at the same force.

Direction against time, wrapped so a shift across north does not draw a false spike. Long steady runs mean the wind is set by a broad weather pattern; frequent large swings mean fronts, showers or a light, wandering breeze. Hollow dots are calm hours where direction has no meaning. Useful for spray drift, smoke, berthing and runway planning, where a steady direction matters as much as speed.

The tips of the wind vectors at 10, 80, 120 and 180 m, joined in height order. A curve turning clockwise with height (veering) is classic warm-air advection; turning anticlockwise (backing) indicates cold advection. A long hodograph means strong shear. Arrows point where the wind blows to. It is a model profile, not a sounding.

Is this week unusual? The band is the 10th–90th percentile of hourly ERA5 10 m wind for ±15 days around today over the last ten years; the filament is the forecast. Hours above the band are windier than nine in ten comparable hours at this point. ERA5 is a ~31 km reanalysis, so it is smoother than any one exposed site — compare shapes, not single values.

The seasonal rhythm of the wind here, from ERA5 daily means since 1990. Bars are the average daily mean speed per calendar month, ticks are the 90th-percentile day, and dots the average daily maximum gust. In Ireland winter months typically run well above summer, which is why calm-weather work is scheduled May–August.

Wind speeds follow a skewed distribution that a two-parameter Weibull curve usually fits well: k describes the shape (higher = steadier), c the scale. Because power rises with the cube of speed, the rare strong hours carry most of the energy, which is what the violet curve shows. Fitted to the last 30 days at 10 m only: this is a teaching instrument, not a wind resource assessment, and hub-height energy is far higher.

The share of hours with less than 3 mph of wind, by hour of day. Calms cluster around dawn inland, when a temperature inversion traps still air near the ground — exactly when spray drift, smoke and frost risks peak. Compares the last 30 days with the coming week.

The nearest offshore weather buoy: significant wave height (the average of the highest third of waves), wave period and wind over the last 24 hours. Period matters as much as height — long-period swell carries far more energy than a short wind chop of the same height. Buoy anemometers sit 3–4 m above the sea, so they read lower than a 10 m mast in the same wind.

Delta T is the dry-bulb minus wet-bulb temperature — a measure of how fast spray droplets evaporate. Agronomy guidance commonly cites 2–8 °C as the preferred band: above about 10 °C droplets shrink fast and drift; below 2 °C conditions often coincide with inversions that carry fine droplets far. Computed from the forecast temperature and humidity (Stull 2011). Always defer to the product label and local rules.

Mean sea-level pressure and how fast it is changing. A fall of more than 3 hPa in three hours is the forecaster's classic warning of an approaching deep low and strengthening wind; a rapid rise behind a front brings strong, gusty post-frontal air. The descriptors follow the WMO tendency wording.

Two wind roses on the same scale: the ERA5 climate for this time of year and the next seven days of forecast. When the forecast rose points somewhere the climate rose rarely does, the week is unusual for direction — which matters for exposed sites, sheltered harbours and runway choice even when the speed is ordinary.

The decision chart. The violet fan is the ensemble at your height, the filament the primary model, and the dashed line your own limit. Ember ticks mark hours where at least 40 % of members exceed the limit. It does not say whether to work — it says how often the model world breaks your number.

The Shear Glass: the wind at four heights in one cylinder, looked at from the side. Each filament's length is speed and its angle is where the wind blows to. The ring is your working height — solid if interpolated between model levels, dashed if extrapolated. A fog wedge appears when the ensemble disagrees.

Route exposure lives in the instrument's route mode: draw a line, set a departure and speed, and each segment is checked against the forecast for the time you will actually be there — headwind, crosswind and the chance of crosswind over your limit. This card links to it.

How well the model is doing right now. The nearest station's readings today are plotted against the model's 10 m wind for the same hours. A consistent offset (bias) is common — stations are points, models are grid averages — and knowing it lets you mentally correct the forecast. Large, growing errors are a signal to trust the forecast less today.

What actually happened over the last seven days, from ERA5 reanalysis where it exists and short-range model analysis for the most recent days (ERA5 runs about five days behind). It is the best gridded estimate of the past, not a station record — useful for checking a claim, a delivery delay or an incident against the weather.

Ten years of ERA5 hours for ±15 days around today, sorted by the direction the wind came from. Wedge length is how often, colour is how strong on average. The prevailing direction in most of Ireland is south-west to west, and that is also where the strongest winds come from.

One cell per hour, two rows. The top row is the ensemble chance of breaking your limit (ember when ≥ 40 %); the bottom row is whether the seven deterministic models agree on speed and direction. An hour that is ember on top and violet below is a genuine toss-up; ember on top and green below is a well-agreed problem.