Buoys, M-series and sea state

Met Éireann operates a network of moored buoys designated M1 to M6, positioned in key offshore locations to monitor marine conditions. M1 lies southwest of County Kerry, M2 west of Galway, M3 northwest of Donegal, M4 in the Irish Sea east of Dublin, M5 near the Isle of Man, and M6 southeast of Cork. These buoys transmit data hourly via satellite, forming part of the Global Telecommunication System (GTS) under World Meteorological Organization (WMO) standards. M4 and M6 are equipped with full meteorological and wave sensors; others may report wind and pressure only. The spatial distribution ensures coverage of dominant fetch zones: Atlantic swells enter west of M2 and M3, while the Irish Sea fetch is monitored by M4. Buoys are serviced annually; data gaps occur during maintenance or extreme weather damage, such as M4’s temporary loss during Storm Eunice (18 Feb 2022). Locations are chosen to avoid shipping lanes and represent open-water conditions, minimising coastal distortion. The network complements coastal land stations (e.g., Malin Head, Valentia) by capturing offshore gradients absent on land. For example, during a westerly flow, M2 often records 3–5 m/s stronger winds than Kilkee due to reduced surface roughness. Data latency is typically 10–20 minutes from measurement to GTS availability. The Wind Agent’s obs_vs_model chart aligns buoy timestamps with forecast models, enabling direct comparison.

Buoys measure wind speed and direction using sonic or cup anemometers mounted at 4–5 metres above sea level, pressure via digital barometers, and wave parameters using accelerometers or GPS motion sensors. Wind is sampled at 1 Hz, averaged over 10 minutes; gusts are the maximum 3-second mean within that period. Pressure is recorded at 1 hPa resolution. Wave data include significant wave height (Hs), peak period (Tp), and dominant direction. Hs is computed as the average of the highest one-third of waves in a 20-minute sampling window, reducing noise from individual extremes. Tp is derived from spectral analysis of vertical displacement. For example, if a buoy records wave energies across frequencies, the peak of the spectrum defines Tp. Direction is the mean propagation direction of the dominant swell or wind sea. Data are transmitted in BUFR format, decoded by national meteorological services. The Wind Agent’s sea_state chart displays Hs and Tp alongside wind, allowing users to correlate wave development with local wind forcing. Missing wave data (e.g., sensor fault) are marked as UNKNOWN, not zero. Pressure trends help identify developing lows; a fall exceeding 4 hPa in 3 hours commonly cited by Met Éireann as indicating rapid cyclogenesis. Buoys do not measure sea surface temperature routinely in the Irish network, though M6 has intermittently tested such sensors.

Anemometers on buoys are typically mounted at 4–5 m above sea level, below the 10 m standard for meteorological reporting. To compare with forecasts or land stations, wind speeds must be extrapolated to 10 m using the logarithmic wind profile:

$$ U_{10} = U_z \cdot \frac{\ln(10/z_0)}{\ln(z/z_0)} $$

where $U_z$ is the measured speed at height $z$, and $z_0$ is the roughness length. Over open water, $z_0$ is approximately 0.0002 m. For neutral stability, if a buoy measures 15 m/s at 5 m, the 10 m equivalent is:

$$ U_{10} = 15 \cdot \frac{\ln(10/0.0002)}{\ln(5/0.0002)} = 15 \cdot \frac{\ln(50,000)}{\ln(25,000)} ≈ 15 \cdot \frac{10.82}{10.13} ≈ 15.98 \,\text{m/s} $$

This adjustment adds approximately 0.5–1.0 m/s for typical offshore winds. Stability corrections (e.g., for cold air outbreaks) are rarely applied operationally due to lack of temperature/humidity sensors on most buoys. The Wind Agent applies this log-law correction automatically when displaying buoy data in the glass and obs_vs_model charts, aligning observations with 10 m forecast references. Misreading unadjusted buoy winds as 10 m values underestimates true exposure, particularly in low-wind conditions where shear is stronger. The adjustment is less critical above 15 m/s, where turbulence reduces shear effects.

Significant wave height (Hs) is a statistical measure defined as the mean height of the highest one-third of waves in a given period, typically 20 minutes. It approximates the visual estimate reported by mariners and is expressed in metres. Hs is derived from the zeroth moment of the wave energy spectrum, $H_s = 4\sqrt{m_0}$. For example, if a buoy’s accelerometer records a variance of surface elevation $m_0 = 0.49 \,\text{m}^2$, then $H_s = 4 \times \sqrt{0.49} = 4 \times 0.7 = 2.8 \,\text{m}$. Peak period (Tp) is the wave period with the most energy, commonly 5–15 seconds in Irish waters. During Storm Ophelia (16 Oct 2017), M4 recorded Hs = 14.3 m with Tp = 14.2 s, indicating fully developed seas under prolonged gale-force winds. Hs is not the maximum wave; individual waves can reach 1.6–1.8 × Hs in Gaussian seas. A 2.0 m Hs sea may produce 3.2–3.6 m rogue waves under nonlinear focusing. The Wind Agent’s sea_state chart plots Hs and Tp temporally, allowing users to assess sea state evolution. Wave steepness, $H_s / L_p$ (where $L_p ≈ 1.56 \times T_p^2$), indicates stability: values >1/7 suggest breaking waves. For Tp = 10 s, $L_p ≈ 156$ m; if Hs = 25 m, steepness = 25/156 ≈ 1/6.2, indicating breaking conditions.

Sea state comprises two components: wind sea (locally generated waves) and swell (waves propagated from distant storms). Wind sea has shorter periods (Tp < 8 s), chaotic directionality, and grows with local wind duration and fetch. Swell has longer periods (Tp > 10 s), narrow directional spread, and decays slowly over distance. Spectral analysis separates these: bimodal spectra indicate mixed sea and swell. For example, if M6 records Hs = 3.0 m with dual peaks at 6 s (local westerly wind) and 12 s (north Atlantic swell), the wind sea contributes ~1.8 m and swell ~2.4 m (using $H_s = 4\sqrt{m_0}$ per partition). Swell can dominate Hs even in light winds, misleadingly suggesting local severity. Directional spread is narrower for swell (±15°) than wind sea (±40°). The Wind Agent’s rose_compare chart overlays buoy wave direction with wind direction, helping distinguish local forcing from remote swell. In Irish waters, Atlantic swell from extratropical cyclones often arrives 24–48 hours after peak winds, maintaining Hs > 2 m even after local winds subside. During summer, southerly swells from North African storms occasionally reach M3 with Tp > 13 s, despite calm local conditions. Swell energy attenuates as $e^{-αx}$, where $α ≈ 0.0002 \,\text{m}^{-1}$ and x is distance; a 4 m Hs swell from 2,000 km away may decay to ~2.7 m at M2.

Buoys often record squall peaks 10–30 minutes after coastal land stations due to fetch-limited wave development and sensor response. A sudden wind increase generates waves gradually; the sea state adjusts over time, delaying peak Hs. For example, if wind jumps from 8 to 18 m/s, Hs may take 2–3 hours to reach equilibrium (fetch-limited growth). However, the wind measurement itself should respond instantly. The lag arises because buoys filter high-frequency motion to stabilise readings, applying a 30–60 second damping window to accelerometer data. This smoothing delays peak wind reporting compared to land anemometers with faster response. Additionally, squalls often approach from sea to land; coastal stations (e.g., Malin Head) intercept the leading edge first, while offshore buoys (e.g., M2) are hit later. During Storm Eunice (18 Feb 2022), Malin Head recorded a gust peak at 08:17 UTC, while M2 peaked at 08:42 UTC, a 25-minute delay. The Wind Agent’s diurnal_cycle chart, when applied to marine contexts, reveals such timing offsets. Users relying on buoys for real-time squall detection should consult coastal radar and land stations concurrently. The Agreement Spine feature highlights discrepancies between buoy and model timing, indicating potential propagation delays.

Buoys provide ground truth for validating numerical weather prediction (NWP) models offshore, where satellite and land data are sparse. Forecasters compare observed 10 m wind (height-corrected) and Hs against model outputs (e.g., ECMWF IFS, UKMO UM) using metrics like mean absolute error (MAE) and bias. For instance, if a 5-day forecast predicts 15 m/s at M4 but observation (adjusted to 10 m) is 13.2 m/s, the error is 1.8 m/s. Over a 24-hour period, MAE < 2.0 m/s is typical for modern models in open seas. Hs verification uses the same principle: if ECMWF forecasts Hs = 4.0 m and M6 observes 3.6 m, the model has a +0.4 m bias. The Wind Agent’s obs_vs_model chart overlays buoy observations with ensemble forecasts, showing model spread and bias. A common issue is underprediction of gusts in cold air outbreaks; buoys may record gust factors (gust/sustained) of 1.8–2.0, while models assume 1.5. Buoys also validate wave model physics; WAM and WW3 performance is assessed against Hs and Tp. Persistent overprediction of swell period indicates incorrect source term tuning. The exceedance_curve chart uses buoy climatology to show how often thresholds are breached, informing risk decisions. For example, Hs > 4 m occurs approximately 12% of winter days at M4 based on ERA5 and buoy data.