Skip to content

Wind Waves, Wave Spectra, Forecasting, Tides and Tidal Currents

CSIR NET Ocean Sciences: linear wave theory, dispersion, wave statistics, PM and JONSWAP spectra, forecasting, tidal constituents, amphidromes, tidal currents.

  • 21 min read
  • Updated
On this page · 15 sections
  1. Prerequisites
  2. The Hook: Swell From a Storm You Never Saw
  3. Linear (Airy) Wave Theory
  4. Generation and Growth of Wind Waves
  5. Describing a Random Sea
  6. Wave Forecasting
  7. Tides
  8. Tidal Currents
  9. Memory Tricks
  10. Don’t Confuse
  11. Common Mistakes
  12. Practice
  13. Quick Revision Table
  14. Key Takeaways
  15. What’s Next

Prerequisites

The Hook: Swell From a Storm You Never Saw

Linear (Airy) Wave Theory

The dispersion relation

For small-amplitude waves on water of depth \( h \), with wavenumber \( k = 2\pi/L \) and angular frequency \( \omega = 2\pi/T \):

$$\boxed{\omega^2 = gk\tanh(kh)}$$
Deep water (\( h > L/2 \))IntermediateShallow water (\( h < L/20 \))
\( \tanh kh \)\( \approx 1 \)full expression\( \approx kh \)
Phase speed \( c \)\( \frac{gT}{2\pi} \approx 1.56\,T \) m s\(^{-1}\)\( \sqrt{\frac{g}{k}\tanh kh} \)\( \sqrt{gh} \)
Wavelength \( L \)\( \frac{gT^2}{2\pi} \approx 1.56\,T^2 \) msolve iteratively\( T\sqrt{gh} \)
Group velocity \( c_g \)\( c/2 \)\( nc \), \( n = \frac{1}{2}\left(1 + \frac{2kh}{\sinh 2kh}\right) \)\( c \)
Dispersive?yes: long waves fasteryesno: all periods at \( \sqrt{gh} \)
Particle orbitscircles, radius decaying as \( e^{kz} \)ellipsesflat ellipses, same horizontal motion at all depths

Waves “feel the bottom” once \( h < L/2 \): below \( z = -L/2 \) the orbital motion has fallen to \( e^{-\pi} \approx 4\% \) of its surface value.

Energy and power

Energy per unit surface area (half kinetic, half potential):

$$\boxed{E = \frac{1}{8}\rho g H^2}$$

Energy travels at the group velocity, so the energy flux (wave power) per metre of crest is \( P = E c_g \). For regular deep-water waves \( P = \rho g^2 H^2 T/(32\pi) \); for a random sea with significant height \( H_s \) and energy period \( T_e \), \( P = \rho g^2 H_s^2 T_e/(64\pi) \), about \( 0.49\,H_s^2 T_e \) kW m\(^{-1}\).

A regular 2 m, 10 s deep-water wave carries \( E = 5.0 \) kJ m\(^{-2}\) and about 39 kW per metre of crest.

Capillary and gravity waves

Surface tension is the restoring force for the shortest waves. The minimum phase speed, about 0.23 m s\(^{-1}\), occurs at a wavelength of about 1.7 cm; shorter waves are capillary waves (ripples), longer ones gravity waves. Capillary ripples are the first response to wind and give the sea surface the roughness that radar scatterometers use to measure wind.

Steepness and breaking

  • Deep-water waves break when steepness \( H/L \) reaches about \( 1/7 \) (crest angle 120°).
  • In shallow water they break when \( H \) is about 0.78 of the depth.
  • In practice wind seas break (whitecap) well below the theoretical limit.

Generation and Growth of Wind Waves

Wind waves grow by transfer of momentum from the air:

  1. Resonance (Phillips, 1957): turbulent pressure fluctuations moving at the speed of a free wave excite it; important for initial growth.
  2. Shear-flow instability (Miles, 1957): the airflow over existing waves produces pressure out of phase with the surface; growth is exponential.
  3. Nonlinear wave–wave interactions: resonant quartets transfer energy within the spectrum, shifting the peak to lower frequencies as the sea develops.
  4. Dissipation: whitecapping, breaking, bottom friction.

The state of the sea depends on three factors:

FactorLimit
Wind speed \( U \)sets the ultimate size
Fetch \( F \) (distance over which the wind blows)fetch-limited sea
Duration \( t \)duration-limited sea

If both fetch and duration are long enough, the sea stops growing: a fully developed sea, in which input from the wind is balanced by dissipation. Waves that leave the generating area become swell: longer-crested, more regular, sorted by period.

Wind seaSwell
Locationinside the generating areafar from it
Steepnesshighlow
Crestsshort, irregularlong, regular
Spectrumbroad, peak near the wind speed’s frequencynarrow, low frequency

Describing a Random Sea

Wave statistics

A record of the sea surface contains waves of many heights. Defined heights:

SymbolDefinition
\( H_{1/3} \) (\( H_s \))significant wave height: mean of the highest one-third of waves; close to what an observer estimates by eye
\( H_{rms} \)root-mean-square height
\( \bar{H} \)mean height
\( H_{1/10} \)mean of the highest tenth
\( H_{\max} \)highest wave in the record

For a narrow-band sea, wave heights follow the Rayleigh distribution:

$$P(H > H^*) = \exp\left[-\left(\frac{H^*}{H_{rms}}\right)^2\right] = \exp\left[-2\left(\frac{H^*}{H_s}\right)^2\right]$$

Ratios that follow:

Ratio to \( H_s \)Value
\( \bar{H} \)0.63
\( H_{rms} \)0.71
\( H_{1/10} \)1.27
\( H_{\max} \) in about 1000 wavesabout 1.86 (it grows slowly, as \( \sqrt{\ln N} \))

The wave spectrum

Treat the surface elevation as a sum of sinusoids with random phases. The variance spectrum \( S(f) \) gives the energy (per \( \rho g \)) at each frequency. Its moments are

$$m_n = \int_0^\infty f^n S(f)\,df$$
QuantityFrom the spectrum
Variance of the surface\( m_0 \)
Spectral significant height\( H_{m0} = 4\sqrt{m_0} \) (≈ \( H_{1/3} \) for a narrow-band sea)
Mean period\( T_{m01} = m_0/m_1 \)
Zero-crossing period\( T_z = \sqrt{m_0/m_2} \)
Peak period \( T_p \)\( 1/f_p \), where \( S \) is largest

With the directional spreading function \( D(f, \theta) \) the full description is the directional spectrum \( S(f, \theta) = S(f)D(f, \theta) \).

Standard spectral forms

Pierson–Moskowitz (1964), fully developed sea:

$$S(\omega) = \frac{\alpha g^2}{\omega^5}\exp\left[-\beta\left(\frac{\omega_0}{\omega}\right)^4\right], \quad \alpha = 8.1\times 10^{-3},\ \beta = 0.74,\ \omega_0 = g/U_{19.5}$$

where \( U_{19.5} \) is the wind speed at 19.5 m. It gives \( H_s \approx 0.21\,U_{19.5}^2/g \) and a peak at \( \omega_p \approx 0.88\,g/U_{19.5} \).

JONSWAP (Joint North Sea Wave Project, 1973), fetch-limited, growing sea: the PM shape multiplied by a peak enhancement factor \( \gamma \) (mean value about 3.3), giving a sharper, higher peak. The JONSWAP experiment showed that nonlinear wave–wave interactions control the growth and the downshift of the peak.

Both have a high-frequency tail falling as \( f^{-5} \) (the saturation range).

Wave Forecasting

Methods, in historical order

MethodBasisNotes
Significant-wave (SMB) methodSverdrup and Munk’s work for wartime amphibious landings, later revised by Bretschneider; empirical curves of \( H_s \) and period against wind speed, fetch and durationintroduced the significant wave height; still used for quick estimates
Spectral (PNJ) methodPierson, Neumann and James (1955): forecast the spectrum, then derive heightsfirst spectral approach
Numerical spectral modelssolve the energy (action) balance equation on a gridfirst, second and third generation; operational today

The energy balance equation

$$\frac{\partial E}{\partial t} + \mathbf{c}_g\cdot\nabla E = S_{in} + S_{nl} + S_{ds}$$

for the directional spectrum \( E(f, \theta; \mathbf{x}, t) \): wind input, nonlinear transfer, dissipation (whitecapping, bottom friction, breaking). In coastal models refraction and depth-induced breaking are added.

GenerationTreatment of \( S_{nl} \)Example
Firstnone; spectrum grows to a prescribed limitearly models
Secondparameterised spectral shape—
Thirdexplicit (Discrete Interaction Approximation)WAM, WAVEWATCH III; SWAN for nearshore

In India, INCOIS (Hyderabad) runs operational ocean-state forecasts (waves, swell, currents) for the Indian coast and fishing community.

Empirical growth laws (fetch-limited)

Nondimensional JONSWAP-type relations, with \( U \) the wind speed at 10 m:

$$\frac{gH_{m0}}{U^2} \approx 1.6\times 10^{-3}\left(\frac{gF}{U^2}\right)^{1/2}, \qquad \frac{gT_p}{U} \approx 0.286\left(\frac{gF}{U^2}\right)^{1/3}$$

Height grows as \( \sqrt{F} \), period as \( F^{1/3} \), until full development.

Locating a distant storm from swell

Deep-water swell of frequency \( f \) travels at \( c_g = g/(4\pi f) \). Waves of all frequencies leave a storm at distance \( x \) at the same time \( t_0 \), so frequency \( f \) arrives at

$$t - t_0 = \frac{x}{c_g} = \frac{4\pi f x}{g} \quad\Rightarrow\quad \boxed{x = \frac{g}{4\pi\,(df/dt)}}$$

A record in which the peak frequency rises linearly with time gives \( x \) from the slope, and extrapolating the line back to \( f = 0 \) gives \( t_0 \).

Tides

The tide-generating force

The Moon pulls every point of the Earth, but by different amounts. The difference between the pull at a point and the pull at the Earth’s centre is the tide-generating force. It varies as

$$F_{\text{tide}} \propto \frac{M}{d^3}$$

which is why the Moon dominates despite the Sun’s much greater mass:

$$\frac{F_{\text{Moon}}}{F_{\text{Sun}}} = \frac{M_{\text{Moon}}}{M_{\text{Sun}}}\left(\frac{d_{\text{Sun}}}{d_{\text{Moon}}}\right)^3 \approx 2.2$$

The horizontal component (tractive force) moves water toward the points directly under and opposite the Moon, producing two bulges.

Equilibrium theory

Newton’s equilibrium theory imagines an ocean covering the Earth and responding instantly.

  • Two bulges, so two high tides per lunar day of 24 h 50 min: the semidiurnal tide has period 12 h 25 min.
  • When the Moon is north or south of the equator (declination), an observer passes through the bulges unevenly: successive high waters differ, the diurnal inequality. At high declination the tide becomes partly diurnal; at zero declination it is purely semidiurnal.
  • Spring tides at syzygy (new and full moon): solar and lunar bulges align, range large. Neap tides at quadrature (first and third quarter): range small. The spring–neap cycle is the beat between M2 and S2: \( 1/(1/12.00 - 1/12.42) \) h \( \approx 14.77 \) days.
  • Perigean tides (Moon closest) are larger; the anomalistic month is about 27.55 days.

Equilibrium theory gets the periods right and the amplitudes and timing wrong: a long wave cannot travel round the Earth fast enough to keep under the Moon (it would need about 450 m s\(^{-1}\), while \( \sqrt{gh} \) for a 4 km ocean is about 200 m s\(^{-1}\)), and the continents block it.

Dynamic theory

Laplace treated the tide as the forced response of real ocean basins on a rotating Earth. Each basin responds to the periodic forcing with its own shallow-water waves, strongly modified by rotation and resonance:

  • The tide propagates as a Kelvin wave round each basin, with the coast on its right in the NH.
  • The result is an amphidromic system: a point of zero tidal range (the amphidromic point) round which high water rotates once per tidal period; anticlockwise in the NH for Kelvin-wave systems, clockwise in the SH (with exceptions where basin geometry dominates).
  • Cotidal lines (same time of high water) radiate from the amphidrome; co-range lines (same range) are roughly concentric, the range increasing outward.
  • Resonance: a gulf closed at one end resonates when its length is about a quarter wavelength of the tide, natural period \( T \approx 4L/\sqrt{gh} \). The very large ranges in the Bay of Fundy and the Gulf of Khambhat come from near-resonance and funnelling.

Tidal constituents

Harmonic analysis represents the tide at a place as

$$\eta(t) = Z_0 + \sum_n f_n H_n\cos(\omega_n t + (V_0 + u)_n - g_n)$$

with amplitude \( H_n \) and phase lag \( g_n \) for each constituent (the \( f \) and \( u \) terms are nodal corrections). The main constituents:

SymbolNamePeriodEquilibrium amplitude relative to M2
M2principal lunar semidiurnal12.42 h1.00
S2principal solar semidiurnal12.00 h0.47
N2larger lunar elliptic12.66 h0.19
K2lunisolar semidiurnal11.97 h0.13
K1lunisolar diurnal23.93 h0.58
O1principal lunar diurnal25.82 h0.42
P1principal solar diurnal24.07 h0.19
M4shallow-water overtide of M26.21 h— (generated in shallow water)
Mf, Mm, Salong-period: fortnightly, monthly, annual13.66 d, 27.55 d, 1 yrsmall

Rayleigh criterion: to separate two constituents of frequencies \( f_1 \) and \( f_2 \), the record must be at least \( 1/|f_1 - f_2| \) long. M2 and S2 need about 15 days; K1 and O1 about 14 days; S2 and K2 about half a year. A 29-day record (one lunar month) is the standard minimum for the main constituents; a year resolves most of them.

Tide types: the form factor

$$\boxed{F = \frac{K_1 + O_1}{M_2 + S_2}}$$

(using amplitudes).

\( F \)TypeCharacter
< 0.25semidiurnaltwo nearly equal highs and lows a day
0.25–1.5mixed, mainly semidiurnaltwo highs and lows of unequal height
1.5–3.0mixed, mainly diurnalsometimes only one high a day
> 3.0diurnalone high and one low a day

Most of the Indian coast has semidiurnal or mixed, mainly semidiurnal tides. Diurnal tides occur in parts of the Gulf of Mexico and the South China Sea, where the basin responds strongly at diurnal periods.

By range: microtidal (< 2 m), mesotidal (2–4 m), macrotidal (> 4 m). The Gulfs of Khambhat and Kachchh are macrotidal; much of the Indian east coast is microtidal.

Tidal Currents

Rotary and reversing currents

TypeWhereBehaviour
Rotaryopen sea, shelvesdirection rotates through 360° each cycle (clockwise in the NH in many open-shelf settings); the tip of the current vector traces a tidal ellipse
Reversing (rectilinear)channels, estuaries, inletsflows one way (flood, landward) then the other (ebb, seaward), with slack water at the turn

Progressive versus standing tide

The phase relation between level and current tells you what kind of wave the tide is.

Progressive tidal wave (open coast, long channel)Standing tidal wave (closed basin near resonance)
Maximum flood currentat high watermidway between low and high water
Slack waterat mid-tideat high and low water
Current speed\( u = a\sqrt{g/h} \) (shallow-water wave of amplitude \( a \))maximum at the node, zero at the closed end

Real estuaries are usually in between, with slack occurring some time after high and low water.

Shallow-water effects and asymmetry

  • In shallow water the crest of the tide travels faster (\( \sqrt{g(h + \eta)} \)) than the trough, so the rising tide becomes shorter than the falling tide: flood dominance, with stronger flood currents. This distortion appears as the M4 overtide. Flood-dominant estuaries import sediment.
  • Extreme steepening forms a tidal bore, a breaking front moving upriver (the Hooghly has one).
  • Tidal currents dissipate energy by bottom friction; on shelves this mixes the water column. The Simpson–Hunter criterion \( h/u^3 \) separates mixed (small \( h/u^3 \)) from stratified (large) shelf seas; tidal mixing fronts form at the boundary and are productive.
  • Globally, tidal friction slows the Earth’s rotation and transfers angular momentum to the Moon, which recedes by about 3.8 cm per year (lunar laser ranging).

Memory Tricks

Don’t Confuse

ThisvsThat
Phase speed \( c \)Group (energy) speed \( c_g \): half of \( c \) in deep water
Wind sea: steep, irregular, broad spectrumSwell: regular, narrow spectrum, dispersed by period
\( H_{1/3} \) (from wave-by-wave analysis)\( H_{m0} = 4\sqrt{m_0} \) (from the spectrum)
Pierson–Moskowitz: fully developedJONSWAP: fetch-limited, peak enhancement \( \gamma \approx 3.3 \)
Spring/neap: phase of the Moon, 14.77 dPerigean/apogean: distance of the Moon, about 27.55 d
Cotidal lines: same phaseCo-range lines: same range
Amphidromic point: zero rangeNode of a seiche: zero range, but no rotation
Rotary current: open shelfReversing current: channels and estuaries

Common Mistakes

Practice

All questions below are original practice questions written for this page. They are not from any CSIR NET paper.

Level 1: Foundation

  1. What is the deep-water wavelength of a 6 s wave?
  2. Name the two constituents whose beat produces the spring–neap cycle.
  3. At what depth, relative to wavelength, does a wave start to feel the bottom?
Answers
  1. About 56 m: \( 1.56\times 36 = 56.2 \) m.
  2. M2 and S2.
  3. About half a wavelength (\( h < L/2 \)).

Level 2: Part B Style

Practice Q1. A wave spectrum has zeroth moment \( m_0 = 0.36 \) m\(^2\). The spectral significant wave height is:

(a) 0.6 m (b) 1.2 m (c) 2.4 m (d) 4.8 m

Answer: (c) 2.4 m

\( H_{m0} = 4\sqrt{0.36} = 4\times 0.6 = 2.4 \) m.

Practice Q2. Tidal harmonic amplitudes at a port are M2 = 0.40 m, S2 = 0.15 m, K1 = 0.35 m, O1 = 0.20 m. The tide is:

(a) semidiurnal (b) mixed, mainly semidiurnal (c) mixed, mainly diurnal (d) diurnal

Answer: (b)

\( F = (0.35 + 0.20)/(0.40 + 0.15) = 0.55/0.55 = 1.0 \), in the range 0.25–1.5.

Practice Q3. A tsunami-like long wave and a 12 s swell both travel over the 4000 m deep ocean. The ratio of the long wave’s speed to the swell’s group velocity is about:

(a) 2 (b) 10 (c) 21 (d) 42

Answer: (c) about 21

Long wave: \( \sqrt{9.81\times 4000} = 198 \) m s\(^{-1}\). Swell: \( c_g = gT/(4\pi) = 9.81\times 12/(4\pi) = 9.37 \) m s\(^{-1}\). Ratio \( = 21.1 \).

Practice Q4. In a narrow-band sea with \( H_s = 3 \) m, the probability that an individual wave exceeds 4.5 m is about:

(a) 0.011 (b) 0.05 (c) 0.11 (d) 0.22

Answer: (a) 0.011

\( P = \exp[-2(4.5/3)^2] = \exp(-4.5) = 0.0111 \).

Level 3: Part C Style

Practice Q5. A wave-buoy record on the west coast of India shows the peak frequency of arriving swell rising steadily from 0.050 Hz to 0.070 Hz over 48 hours. Assume deep water along the path and linear dispersion. Which statements are correct?

  1. The storm was about 6700 km away.
  2. Waves of 0.050 Hz travelled with a group velocity of about 15.6 m s\(^{-1}\).
  3. The 0.050 Hz waves had been travelling for about 5 days when they arrived.
  4. If the storm were twice as far away, the rate of frequency increase would double.

(a) 1, 2 and 3 only (b) 1 and 4 only (c) 2 and 3 only (d) 1, 2, 3 and 4

Answer: (a) 1, 2 and 3

\( df/dt = 0.020/(48\times 3600) = 1.157\times 10^{-7} \) Hz s\(^{-1}\).

  1. \( x = g/(4\pi\,df/dt) = 9.81/(4\pi\times 1.157\times 10^{-7}) = 6.74\times 10^6 \) m ≈ 6700 km. Correct.
  2. \( c_g = g/(4\pi f) = 9.81/(4\pi\times 0.05) = 15.6 \) m s\(^{-1}\). Correct.
  3. \( t = x/c_g = 6.74\times 10^6/15.6 = 4.32\times 10^5 \) s = 5.0 days. Correct.
  4. \( df/dt = g/(4\pi x) \): doubling \( x \) halves the rate. Wrong.

Practice Q6. An idealised gulf, closed at its head, is 270 km long and 70 m deep. Which statements are correct?

  1. Its fundamental natural period, \( 4L/\sqrt{gh} \), is about 11.4 h.
  2. It is close to resonance with the M2 tide, so a large tidal range at its head is expected.
  3. At its head, the tidal current is strongest at high water.
  4. Near-resonant amplification would be stronger for the K1 tide than for M2.

(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2 and 4 only (d) 1, 3 and 4 only

Answer: (a) 1 and 2
  1. \( \sqrt{9.81\times 70} = 26.2 \) m s\(^{-1}\); \( T = 4\times 2.7\times 10^5/26.2 = 4.12\times 10^4 \) s = 11.4 h. Correct.
  2. 11.4 h is close to 12.42 h. Correct.
  3. Wrong: at the closed head of a standing wave the current is zero; slack water there coincides with high and low water.
  4. Wrong: K1 (23.93 h) is far from 11.4 h.

Practice Q7. A wind of 15 m s\(^{-1}\) blows offshore over a fetch-limited sea. Using the JONSWAP-type growth laws on this page, if the fetch increases from 100 km to 400 km:

(a) \( H_{m0} \) doubles and \( T_p \) increases by about 59% (b) \( H_{m0} \) quadruples and \( T_p \) doubles (c) \( H_{m0} \) doubles and \( T_p \) doubles (d) both increase by about 41%

Answer: (a)

\( H_{m0}\propto F^{1/2} \): \( 4^{1/2} = 2 \). \( T_p\propto F^{1/3} \): \( 4^{1/3} = 1.587 \), an increase of about 59%. From the worked example, \( H_{m0} \) goes from 2.4 m to 4.8 m and \( T_p \) from 7.1 s to 11.3 s. This assumes the sea is still fetch-limited. At 15 m s\(^{-1}\) and 400 km it is approaching full development, so in reality growth would begin to level off.

Quick Revision Table

IdeaFormula or fact
Dispersion\( \omega^2 = gk\tanh kh \)
Deep water\( c = gT/2\pi \), \( L = gT^2/2\pi \), \( c_g = c/2 \)
Shallow water\( c = c_g = \sqrt{gh} \), non-dispersive
Energy\( E = \rho gH^2/8 \); \( P = Ec_g \)
Breaking\( H/L \approx 1/7 \) deep; \( H \approx 0.78h \) shallow
Rayleigh\( P(H > H^*) = e^{-2(H^*/H_s)^2} \)
Spectral height\( H_{m0} = 4\sqrt{m_0} \)
PMfully developed; \( H_s \approx 0.21U^2/g \)
JONSWAPfetch-limited; \( \gamma \approx 3.3 \); \( H\propto F^{1/2} \), \( T\propto F^{1/3} \)
ForecastingSMB → PNJ → spectral models (WAM, WAVEWATCH III, SWAN)
Storm distance\( x = g/(4\pi\,df/dt) \)
Tidal force\( \propto M/d^3 \); Moon/Sun ≈ 2.2
Spring–neap14.77 d (M2–S2 beat)
Form factor\( (K_1 + O_1)/(M_2 + S_2) \): 0.25 / 1.5 / 3
AmphidromesKelvin waves; anticlockwise in NH
Resonance\( T = 4L/\sqrt{gh} \) for a gulf closed at one end
Tidal currentsprogressive: max at HW; standing: slack at HW

Key Takeaways

What’s Next