Wave Buoy Spectra, Significant Height and Direction

The sea state behind one buoy number

A wave buoy can report significant wave height, peak wave period, mean wave period and dominant wave direction in the same record. Those four numbers are not interchangeable descriptions of a single wave. They summarise different features of many waves passing an instrument during a measurement interval. To translate the record correctly, the reader must distinguish vertical sea-surface displacement from its frequency spectrum, recognise the directional convention, and keep the observation period attached to every conclusion.

An ocean wave spectrum distributes measured surface-elevation variance across wave frequencies. A high peak on the spectrum is not itself a wave height in metres, and a long peak period is not a promise that every crest will arrive exactly that many seconds apart. Observations combine locally generated wind sea, swell arriving from elsewhere, and sometimes multiple swell trains. The spectral picture is useful because it preserves distinctions that disappear when a complex sea is reduced to one height.

This guide explains how wave-buoy motion becomes spectral measurements, why significant wave height is calculated from the integrated spectrum, how peak and average periods differ, why wave direction is commonly described as coming from a bearing, and where the uncertainty enters. It provides dimensional checks, worked calculations, diagnosis of misleading displays, practice questions with answers, and first-party NOAA methods. All numerical situations are illustrative teaching examples, not live conditions at a named station.

Before trusting a number, look for the station identifier, the observation timestamp and time zone, the data product and whether it is provisional or checked. A buoy snapshot and a forecast for the same patch of water may legitimately differ because they describe different things. A measured twenty-minute interval cannot establish the maximum crest of the next six hours. A map drawn with a large coloured circle may imply spatial coverage, but the buoy records motion at its local position.

The most important reading habit is to ask which definition a named variable follows. Significant wave height can be described conversationally as a characteristic height of the highest third of waves, while modern spectral systems typically report an estimate obtained from the total spectral variance. Peak period is drawn from a spectral maximum. Average period depends on a specified weighting or algorithm. Their labels can look familiar while their exact computations differ across providers.

From buoy motion to a frequency spectrum

The sea surface goes up and down as waves pass. Buoys capture this motion using accelerometers, displacement sensors or other motion estimates, along with orientation information when directional waves are requested. A processor converts a sampled time series into frequency-domain information, commonly through Fourier-based analysis, with corrections for buoy dynamics and instrument response. NOAA’s National Data Buoy Center explains that its published spectral products are derived from processed buoy motion, not a photograph or a direct count of breaking crests.

A simplified surface-elevation time series can be imagined as a succession of numbers measured at regular intervals: positive above a local reference, negative below it. A repeating smooth oscillation at 0.10 cycles per second has a ten-second period. Real seas contain many frequencies at once, and individual waves need not repeat regularly. The spectrum answers a different question from an individual crest trace: how much of the elevation variance is associated with each frequency band during the sampled interval?

The horizontal axis of a conventional nondirectional spectrum is frequency in hertz, or cycles per second. Its vertical axis is often variance spectral density in square metres per hertz. Multiply density in a band by the band’s frequency width and you obtain the approximate contribution of that band to elevation variance, in square metres. Add the band contributions and you obtain the zeroth spectral moment, conventionally written m0. This is the area under the spectrum, not the height of its tallest peak.

Frequency and period are inverses. A frequency of 0.05 hertz means one cycle in twenty seconds; 0.10 hertz corresponds to ten seconds; 0.20 hertz corresponds to five seconds. Period increases as frequency decreases, so a spectrum with its long-period swell on the left can look counterintuitive to someone accustomed to an ordinary time graph. Always read the axis before calling the left-hand peak the ‘fast’ waves.

Fourier analysis and windowing introduce choices. How long a record is analysed, whether segments overlap, the spectral estimator, smoothing, high- and low-frequency cut-offs, sensor calibration and response corrections all affect the final curve. A narrow, tall spectral spike can represent a concentrated wave component but may also be influenced by limited frequency resolution. Two instruments processing different record lengths need not have the same-looking spectral peak even under broadly similar sea conditions.

Wave motion can be distorted by buoy hull dynamics. A sensor attached to the buoy feels the behaviour of a floating object rather than a mathematical point on an undisturbed surface. The instrument applies a response model to estimate wave energy, but there are frequency ranges where this correction is less reliable. Mooring forces, a damaged hull, sensor saturation, tilt and magnetic interference can add uncertainty. Responsible translation explains the chain from motion to processed spectrum instead of presenting the estimate as direct, assumption-free truth.

Record the station’s sampling period. A twenty-minute buoy spectrum represents the measured waves within that interval, not every minute of the hour or the entire day. The timestamp may identify the start, middle or end of processing, according to the provider. A sudden squall between scheduled samples can change the sea state faster than the reported update cadence. A comparison with a ship report should align time and position before interpreting disagreement.

Directional spectra add an angular dimension: energy as a function of both frequency and arrival direction. In practice a directional buoy estimates wave direction from combinations of vertical motion and horizontal or rotational response, using system-specific techniques. It does not identify a single exact path for each crest. Direction estimates have uncertainty, especially when energy at a frequency is weak, multiple systems overlap or the platform’s compass is wrong.

Significant wave height: area under the curve

Significant wave height is often abbreviated Hs, Hm0, WVHT or a similar provider-specific name. A common spectral calculation is Hm0 = 4 times the square root of m0, where m0 is the integrated surface-elevation variance in square metres. The square root converts square metres to metres, and the factor four is part of the conventional relationship between variance and characteristic wave height. It is not a direct instruction to multiply the height of a graph peak by four.

Suppose a teaching spectrum has integrated variance m0 = 0.25 square metre. The square root is 0.50 metre; four times that value is a spectral significant height of 2.00 metres. If another record has m0 = 0.5625 square metre, its Hm0 is 3.00 metres. Squaring a significant height ratio shows how sensitive energy-like variance is to height: increasing Hm0 from two to three metres corresponds to m0 increasing by a factor of 2.25.

Two spectral peaks do not usually contribute their significant heights by simple addition. Suppose an idealised low-frequency swell band contributes m0 = 0.09 square metre and a higher-frequency wind-sea band contributes 0.16. Their separate Hm0 values are 1.2 and 1.6 metres, while combined m0 is 0.25 and combined significant height is 2.0 metres, not 2.8 metres. Variances add for disjoint spectral bands; characteristic heights are recovered from their sum.

That example exposes a common report-writing error. A chart with swell height 1.2 metres and wind-wave height 1.6 metres cannot be summarised as ‘2.8-metre waves’ by arithmetic addition of component heights. The combined characteristic height depends on how the provider partitions and integrates the spectrum. With independent, nonoverlapping energy partitions in the standard formulation, the square-root-of-sum-of-squares relationship is the relevant check. Avoid assuming a provider’s rounded components will reconstruct its total exactly.

Significant wave height is not maximum individual wave height. A high individual crest-to-trough event can exceed Hs during the measurement interval, sometimes substantially. How much higher depends on the sea state, the number of waves, spectral shape, nonlinear effects and sampling, rather than on one universal multiplier that guarantees safety. A navigation or coastal-risk explanation should distinguish a representative sea-state statistic from extremes and breaking conditions.

The traditional highest-one-third concept and spectral Hm0 are related, but they are not mathematically identical for every real sea. Spectral bandwidth, statistical assumptions and wave definition can produce differences. NOAA’s wave analysis guidance explains that the spectral estimate is a useful practical measure, while exact comparisons depend on the wave field. Readers should not manufacture precision by claiming the number is literally the arithmetic average of individually identified wave crests unless that is the algorithm used.

Variance spectral density may be presented on a linear or logarithmic axis, and frequency bands can have unequal widths. A tall bar with a narrow width can contribute less m0 than a shorter but much wider band. When integrating discrete spectral data, multiply each spectral density by that particular band’s bandwidth. Adding raw density numbers without weighting by frequency width can give a nonsensical result with the wrong units.

A quick dimensional audit catches many mistakes. The units of S(f) are metres squared per hertz; multiply by delta-frequency in hertz to obtain square metres. Sum to get m0 in square metres. Take the square root to obtain metres, then multiply by a dimensionless four. If a spreadsheet produces wave height in square metres or reports negative m0 despite nonnegative spectral density, investigate the input interpretation and numerical formula before drawing conclusions.

Peak period, average period and the word dominant

Peak period is the reciprocal of the frequency at which spectral density is maximal. If the maximum occurs at 0.10 hertz, peak period is ten seconds. If it moves to 0.08 hertz, peak period becomes 12.5 seconds. A longer peak period need not mean a larger significant height; the overall variance can increase, decrease or remain similar independently of where its biggest spectral peak sits.

An awkward feature of peak period is that it can jump between two neighbouring peaks. Suppose long swell at 0.08 hertz and shorter wind sea at 0.17 hertz have nearly equal spectral density. Small measurement or smoothing changes may make one peak temporarily larger. The reported dominant period can switch between roughly 12.5 and 5.9 seconds even though the underlying sea changed gradually. Inspect the full spectrum before interpreting a sudden peak-period jump as a dramatic physical event.

Average period is not one universally computed quantity. The NOAA NDBC wave calculation page describes an average period derived as the square root of m0 divided by m2, where m2 is the second frequency moment of the spectrum. Another provider may use a different spectral moment ratio or zero-crossing definition under a similar label. A correct article names the precise source and formula, rather than declaring every published ‘average wave period’ equivalent.

The second spectral moment weights each frequency band’s variance contribution by frequency squared. In symbols, m2 is the sum of frequency squared times spectral density times the band’s width. If a hypothetical spectrum has m0 = 0.25 square metre and m2 = 0.010 square metre per second squared, the NDBC-style expression gives the square root of 25 seconds squared, or five seconds. This example is arithmetic, not a claim about a particular ocean buoy.

Peak period is controlled by the highest individual density peak; a moment-derived period responds to the distribution of energy over many frequencies. This is why a long-period swell can determine peak period while short-period waves influence the mean period. Neither summary alone describes wave grouping, directional spread or maximum crest height. The complete spectrum contains information that is deliberately compressed when displayed as one number.

A period value must be interpreted in the context of how it was sampled and reported. If a system rounds period to whole seconds, an apparent difference from 10 to 11 may partly reflect resolution. If a dataset uses a missing-value code, a zero could mean no valid result rather than waves with zero-second periods. Before making a trend chart, check the provider’s codebook, update interval, quality flags and instrument changes.

Longer-period waves often originate from storms farther away, while locally forced waves can have shorter characteristic periods, but this distinction is not absolute. Fetch, storm duration, wind strength, ocean depth and propagation all shape the spectrum. It is reasonable to propose swell and wind-sea interpretations from separate spectral peaks, yet an observational record needs direction, wind data and regional context before claims about a specific distant storm can be made.

Directional convention and what the arrow actually says

Wave products commonly express the direction from which waves arrive. In NOAA NDBC’s detailed-wave product, mean wave direction at the dominant period is measured clockwise from true north and describes where the waves come from. A direction of 270 degrees is therefore waves arriving from the west and generally propagating eastward, not a westbound wave current. This is the opposite verbal convention from an eastward ocean-current vector, which describes motion toward east.

A wave-direction arrow can be drawn as an arrow pointing toward propagation or as a barb indicating the source direction. The plot legend decides which. If a diagram labels an incoming northwesterly swell, that can mean waves originate in the northwest and travel toward the southeast. Do not trust an unlabeled arrow to establish the convention. Read the data provider’s definition, especially when combining forecasts, buoy observations and satellite products.

True north and magnetic north must not be blended. Directional buoy measurements require orientation information; a compass problem can rotate every reported bearing. Separate a genuine change in incoming wave direction from sensor heading error, particularly after equipment service or unusual magnetic conditions. When comparing two stations, local shoreline orientation, diffraction and refraction can also change the directions that arrive at each site.

A single mean direction can be actively misleading when two wave trains approach from different bearings at similar frequencies. Their directional distribution may be broad or multi-peaked, while an average angle suggests a tidy compromise that no dominant train actually follows. Inspect directional spread, energy by frequency and the plotted directional spectrum. If the source contains only nondirectional measurements, do not infer directional structure from the presence of a height and period field.

Direction uncertainty can become large in low-energy frequency bins. A small numerical change in a weak angular signal may be uninterpretable even when the height estimate remains useful. Direction also wraps at north: 359 degrees and 1 degree differ by two degrees, not 358. Ordinary arithmetic averaging of bearings can make a northerly sea look southerly; appropriate circular statistics or vector methods are necessary.

Sea state, sampling and the temptation to overpredict

A wave record says what was observed locally in its acquisition interval, subject to processing and quality controls. A forecast predicts future conditions from a model and its initial data. A hindcast reconstructs past conditions with a model. These products can be compared, but the comparison needs matching location, depth, forecast valid time, analysis period, variable definition and uncertainty. A buoy cannot validate an offshore grid cell solely because both points lie in the same coloured map square.

Wave transformation near shore complicates comparisons. As waves approach shallow water, refraction can rotate their path, shoaling can change height, and breaking can reduce wave energy. Harbour walls, reefs, channels and bathymetric features can make coastal measurements different from offshore buoy measurements even over modest distances. A large difference may be physical rather than instrumental, but it demands a geographical explanation, not an assumption that one device must be wrong.

The time scale of a wave observation matters for safety. Significant height says something about a sea state, not a guaranteed upper bound for individual waves, and real seas evolve while a ship travels. A responsible public explanation distinguishes what the latest observation measured from what is forecast later, cites the timestamp and avoids promising safe passage from one buoy number. Use operational guidance appropriate to the maritime activity rather than converting an educational interpretation into navigation instructions.

A numerical workshop for spectral wave interpretation

Worked example one begins with a simplified measured spectrum. There are three frequency bands with densities 2.0, 3.0 and 5.0 square metres per hertz, each with bandwidth 0.01 hertz. Their m0 contributions are 0.02, 0.03 and 0.05 square metre, summing to 0.10 square metre. Spectral significant height is four times the square root of 0.10, approximately 1.265 metres. The calculation shows why the integral must retain bandwidth, and it deliberately does not suggest these are typical ocean spectral values.

Worked example two changes only the band widths. Suppose the same three density values are assigned widths 0.02, 0.01 and 0.005 hertz. Their contributions become 0.04, 0.03 and 0.025 square metre; m0 is 0.095, producing Hm0 approximately 1.233 metres. The tallest density is still five, but the largest variance contribution is now the first band. A reader who merely sums density values would miss both the correct units and the correct ranking of energetic contributions.

Worked example three reverses the usual calculation. If a record reports a spectral significant height of 2.4 metres, the implied m0 under Hm0 = 4 sqrt(m0) is (2.4 / 4) squared, equal to 0.36 square metre. If another interval reports 1.2 metres, its implied variance is 0.09 square metre. The height ratio is two, but the variance ratio is four. This matters when comparing the strength of sea states without exaggerating the numerical change in height.

Worked example four considers two wave systems. A lower-frequency swell component has spectral height 1.5 metres and a wind-sea component has height 2.0 metres. Under the additive, nonoverlapping variance assumption, the total is the square root of 1.5 squared plus 2.0 squared, equal to 2.5 metres. The answer is not 3.5 metres. If reported component totals fail this check slightly, investigate rounding and the provider’s exact partition before assuming a mistake.

Worked example five calculates a peak period. If the highest spectral density occurs at 0.125 hertz, Tp is 1 / 0.125, or eight seconds. A nearby peak at 0.08 hertz represents 12.5 seconds but is not the dominant peak unless its density exceeds the first. A 12.5-second swell can still contribute substantial integrated energy without setting the reported Tp. The phrase ‘the waves are eight seconds apart’ incorrectly turns a peak statistic into a clockwork promise.

Worked example six reads two successive spectra with equal m0 but different peak locations. Both spectra integrate to 0.25 square metre and therefore both give Hm0 = 2.0 metres. One peaks at 0.10 hertz, the other at 0.20. Their Tp values are ten and five seconds. The sea states can have the same spectral height while concentrating energy at different periods; wave response of vessels and coastal structures may differ substantially even with identical Hs.

Worked example seven audits an average period. A simplified dataset has m0 = 0.36 square metre and m2 = 0.0144 square metre per second squared. Under the NDBC-style formula, average period is the square root of 0.36 / 0.0144, equal to five seconds. Check the ratio’s units: seconds squared under the square root. Calling this a peak period would be wrong unless independently shown that the spectral maximum corresponds to 0.20 hertz.

Worked example eight translates direction. An NDBC-style MWD value of 225 degrees indicates waves arriving from the southwest, generally propagating northeast. A current reported as 225 degrees toward would move southwest. The numeric coincidence does not imply the two processes travel together. The final sentence should say ‘waves were arriving from the southwest’ rather than the ambiguous ‘waves headed southwest’, which would invert the physical narrative.

Worked example nine handles directional wrap. Two consecutive incoming-wave bearings are 358 and 2 degrees. Their physically relevant separation is four degrees across north, not 356 degrees. The ordinary arithmetic average of 358 and 2 is 180 degrees, which would falsely suggest a southerly source. A circular or vector-based average can recover a mean close to north. This is a general lesson in translating bearings, applying to winds and currents with their own source-direction conventions.

Worked example ten distinguishes an observation from a forecast. A buoy’s latest twenty-minute sample gives Hm0 = 1.8 metres at 10:00 UTC, while a model predicts 2.5 metres for 13:00 UTC at a nearby grid location. These are different times, locations and products. The only defensible immediate statement is that the available observation was lower than that future model estimate; it is not evidence that the forecast has failed. Verification requires observations valid at the forecast time.

How to recognise mistaken readings before they spread

Misreading one: a headline says that the highest peak on the frequency plot measures four metres. The peak’s vertical axis may be square metres per hertz, not metres. Repair the statement by computing integrated variance and then spectral height. If the underlying frequency-band widths are unavailable, the chart alone may not support an accurate reconstruction. A responsible writer says what cannot be calculated, rather than estimating area by eye without scales.

Misreading two: an analyst adds swell height and wind-wave height. The arithmetic produces an exaggerated combined significant height. Repair by asking whether the components are partitioned spectral energies and, if they are, combining their squared heights and taking the square root. Preserve the provider’s total if available because component boundaries and reporting precision may differ. The goal is to explain the measurement’s mathematical structure, not outvote a validated product with a homemade number.

Misreading three: a large peak period is announced as evidence of towering waves. Peak period identifies the frequency of maximum density, not the integrated energy. A weak long swell can dominate peak period during otherwise modest conditions. Repair by reading Hm0 and the full spectrum together, and by discussing directional context if needed. Long-period swell can be operationally consequential without requiring exaggerated claims about the wave height recorded at that instant.

Misreading four: a direction of 90 degrees is described as waves travelling east. In many wave-buoy summaries, 90 degrees means waves coming from east and travelling broadly west. Repair by checking the provider’s direction field definition and stating ‘from east’ explicitly. If another source uses a toward convention, document the conversion instead of mixing arrow styles. Always retain true-versus-magnetic orientation where relevant.

Misreading five: a smooth mean wave period is assumed to be a universal average of individual crest gaps. Some average periods are spectral-moment quantities, and formulas differ among programmes. Repair by identifying the variable definition from the dataset documentation. If comparing multiple networks, either recompute a common measure from comparable spectra or explain that the averages are not directly interchangeable. An elegant multi-station graph is not a substitute for a common definition.

Misreading six: a rapid change in Tp is treated as a sudden storm arrival. Two competing spectral peaks may cross in rank with only a small change in energy. Repair by plotting the spectra before and after, computing the contribution of each system, and checking wind and direction data. The real ocean may have changed, but the scalar peak statistic also introduces a switching mechanism. Separating those explanations is a valuable diagnostic practice.

Misreading seven: a buoy is treated as if it measures every crest along an entire shipping route. Waves vary spatially with bathymetry, shelter, currents and storm geometry. Repair by describing the buoy’s actual observing location and interval. A route-scale inference needs appropriate observations or models across the route, and operational decisions require official maritime guidance. A local buoy is a sensor, not a guarantee for unobserved locations.

Misreading eight: a chart’s blank band is converted into a zero-energy band. Some gaps represent unavailable or rejected frequency measurements, not physical absence of waves. Repair by distinguishing missing-value flags, quality controls, frequency limits and actual low measured energy. Keep the raw spectrum or provider status codes when possible. A plot that paints missing cells white without a legend can hide precisely the uncertainty that matters most.

Quality control, uncertainty and a useful comparison protocol

Begin quality review with station metadata: location, deployment depth, buoy type, wave sensor, acquisition length, processing version, quality status and maintenance history. If a station has been redeployed or upgraded, discontinuities may reflect changed hardware or algorithms. Compare time series around such events and avoid treating an instrument change as a climatic trend. Record time in UTC or another explicit time zone consistently; local daylight-saving conversions can scramble regional comparisons.

Next, inspect raw or minimally processed spectral shapes where accessible. Is there a physically plausible distribution of energy? Is there an unexpected narrow spike at a frequency associated with instrument response? Are the high-frequency values truncated, the low-frequency bands missing or the directional estimates implausibly stable despite weak energy? The presence of a smooth Hs time series does not settle these questions; scalar summaries can conceal processing artefacts.

Separate three kinds of uncertainty. Measurement uncertainty comes from sensors, platform response and sampling. Statistical uncertainty arises because a finite interval contains a limited realisation of a variable sea. Representativeness uncertainty arises when the buoy’s particular position and time are used to describe a larger sea area or future interval. These are not interchangeable and should not all be reduced to one plus-minus number without explicit assumptions.

Very small significant heights require special care. As sea-surface motion approaches noise or instrument-resolution limits, direction and period fields may become unstable or be flagged. A high-frequency spectral tail can also be sensitive to sensor and processing characteristics. The correct conclusion may be that the broad sea state is calm but a particular direction or peak period is unreliable. It is entirely legitimate to give different confidence levels to the different fields from the same observation.

Comparing buoys is most meaningful when location, depth, exposure and processing are considered together. An offshore open-ocean buoy and a harbour buoy need not have matching height because waves transform before reaching the harbour. A difference in frequency distribution may reveal sheltering or shallow-water effects. A directional difference could be refraction, different waves or a convention mismatch. List plausible mechanisms and seek independent evidence rather than choosing the most dramatic story.

Comparing observations with models also requires the correct target. A model may publish significant height, peak period and mean direction for a grid cell, while a buoy reports point-scale spectral observations. A fair verification aligns the valid time, identifies whether wind sea and swell are combined, and notes spatial mismatch. A model can have small height error and serious direction error, or vice versa. Treat verification as a multi-variable exercise.

Precision should reflect limitations. If an observed height is displayed to two decimal places, that does not automatically establish two-centimetre accuracy. If a mean direction changes by two degrees while sensor uncertainty is much larger, the comparison may not support an actual directional shift. State the unit, computation and quality class, and avoid decorating a narrative with decimals that cannot be defended.

A good public summary contains three layers: what was observed, how it was calculated, and why it matters under stated conditions. For instance: ‘The buoy’s spectral significant wave height increased during the sampled intervals, meaning integrated wave-elevation variance increased. The peak period remained near the long-period swell band, so the change in height did not require a shift in dominant frequency. These observations describe the buoy site and are not a route-wide forecast.’ The explanation is specific without pretending to know the storm’s whole history.

Transfer questions with answers that can be checked independently

Question one provides m0 = 0.16 square metre. Find Hm0. The answer is 4 × 0.40 = 1.60 metres. The correct intermediate unit is the square root of square metres. If a learner obtains 0.64 metres by multiplying m0 itself by four, the error is omission of the square root, not a matter of rounding.

Question two provides two independent spectral components with heights 0.9 and 1.2 metres. Under additive variance, find the combined height. The answer is sqrt(0.9 squared + 1.2 squared) = sqrt(2.25) = 1.5 metres. Adding the components to produce 2.1 metres is not the correct spectral total. The independence and nonoverlap assumption should be stated.

Question three gives a peak frequency of 0.0625 hertz. The peak period is sixteen seconds. A learner who reports 0.0625 seconds has confused frequency with period. A second question asks which is longer-period: 0.0625 or 0.125 hertz? The first, because the inverse relation makes smaller frequency correspond to longer time.

Question four gives mean wave direction 315 degrees under a from-direction convention. The source lies northwest and propagation is broadly southeast. The check is geometric: add 180 degrees modulo 360 to convert incoming bearing to approximate outgoing propagation direction. One should not carry this conversion into a current vector whose data definition is already toward.

Question five provides m0 = 0.49 square metre and m2 = 0.0196 square metre per second squared, with the NDBC-style average period. The answer is sqrt(0.49 / 0.0196) = sqrt(25) = five seconds. The peak period remains unknown without a spectral density maximum. This problem tests whether readers will invent an absent value.

Question six asks whether Hs = 2 metres means every wave is below two metres. The answer is no: Hs is a sea-state statistic, not a bound on individual waves. The size and frequency of exceptionally large waves need separate statistical and physical analysis. A safety claim based on the incorrect bound would be irresponsible.

Question seven describes a sudden peak-period jump from six to twelve seconds but almost unchanged integrated m0. The first interpretation to test is whether a pair of spectral peaks exchanged rank. Plot both spectra, check frequency resolution and quality flags, and compare wind and directional data. Do not infer that wave energy necessarily doubled or that the sea instantly became twice as high.

Question eight asks a student to compare 23:00 UTC observations with midnight local observations. The first step is to convert both timestamps into the same zone and verify whether the records refer to starting, midpoint or ending time. Without that alignment, an apparent one-hour lag may be entirely manufactured. The answer is a procedure, not a numeric conclusion.

Question nine asks what to do with a blank directional field while Hm0 is valid. The answer is to preserve the valid height and report direction as unavailable. Fill it neither with zero degrees, which is a real north bearing, nor with a nearby station’s direction without an explicit interpolation model. Different variables can have different quality statuses.

Question ten asks for a concise explanation of a dual-peaked spectrum. An acceptable answer is: ‘The buoy recorded wave energy in two distinguishable frequency bands, consistent with more than one wave system. The integrated energy determines the characteristic height, while the tallest density peak controls the reported peak period. Directional and meteorological evidence is needed before attributing either band to a particular source storm.’ This connects the measurement to its limits.

Frequently asked questions and a concise reporting standard

Is Hs the height of the biggest observed wave? No; it is a characteristic statistic and, in common spectral reporting, a function of total elevation variance. Does a longer peak period mean rougher seas? Not necessarily, because total wave energy and the location of the spectral peak are different. Can wind-sea and swell heights be added as ordinary lengths? Not to obtain their combined spectral significant height under the standard component-energy method.

Are wave and current directions interchangeable? No. Wave directions in NDBC-style products refer to where waves come from; a current-vector direction convention typically describes where water flows toward. Can two buoys disagree without either being broken? Yes; they may sample different water depths, shelter, geometry, wave systems and times. Confirm definitions and metadata before claiming a contradiction.

What is the shortest defensible summary? Give station, time and quality, the significant wave height and its spectral meaning, the peak period and what it represents, the incoming-wave direction if valid, and the main limitation. When uncertainty is consequential, show the spectrum rather than overinterpreting a compact number. Scientific translation is successful when the words enable a reader to reconstruct the inference and see its boundaries.