Formulary

Key Equations and Numbers

The key equations of the masterclass, with the numbers an oceanographer carries in their head, collected unit by unit. Each one is derived or motivated in its unit, with worked examples from real measurements.

Oceanography is a quantitative science with a small set of relations that come back everywhere. A balance of pressure against rotation explains the Gulf Stream, the eddies, and the slope of the sea surface that satellites measure. One wave speed explains why a tsunami crosses the Pacific at the speed of an airliner and why swell arrives sorted by period. One definition of transport, the sverdrup, lets you compare a river, a current, and the overturning of a whole ocean in a single breath. A mass balance for salt, heat, or an isotope turns a measurement into a history. This page collects those relations, in the order of the units, each with its symbols explained and a typical value given.

Next to the equations stand the numbers an oceanographer carries in their head and uses to judge every other number: about 1 atmosphere for every 10 metres of depth, a mean depth of about 3,700 m, about 17 Sv across the Atlantic at 26.5° N, 3.3 mm of sea level rise per year since 1993, 180 ppm of carbon dioxide in glacial air. Each card states when the relation applies, and where a textbook value is a simplification, the unit says so and gives the real range. The page grows as units are published, so an equation you cannot find yet belongs to a unit that is still to come.

The Oceanographer's Toolkit

The sverdrup

\[ 1\unit{Sv} = 10^{6}\unit{m^{3}\,s^{-1}} \] The volume transport of a current is the volume of water that passes through a cross-section of it each second: width times depth times speed, averaged over the cross-section. One sverdrup is a million cubic metres per second. All the rivers of the world together deliver about \(1.2\unit{Sv}\) to the sea; the Amazon alone carries about \(0.2\unit{Sv}\).

The nautical mile and the length of a degree

\[ 1\unit{nautical\ mile} = 1{,}852\unit{m} = \text{one minute of latitude}, \qquad 1^{\circ}\ \text{of latitude} \approx 111\unit{km} \] A degree of longitude is \(111\unit{km} \times \cos\varphi\), because the parallels shrink towards the poles: 111 km at the equator, 96 km at 30°, 55.5 km at 60°, and zero at the poles. One knot is one nautical mile per hour, which is \(0.514\unit{m\,s^{-1}}\).

Depth from an echo

\[ H = \frac{c\,t}{2} , \qquad c \approx 1{,}500\unit{m\,s^{-1}} \] An echo sounder measures the time \(t\) between a sound pulse and its echo. The sound travels down and back, so the depth \(H\) is half the travel time multiplied by the speed of sound \(c\) in seawater.

Hydrostatic pressure

\[ p(z) = p_{\mathrm{atm}} + \rho g z \] The pressure at depth \(z\) is the pressure of the atmosphere plus the weight of the water column above each square metre. This is the hydrostatic equation. With \(\rho = 1{,}025\unit{kg\,m^{-3}}\) and \(g = 9.81\unit{m\,s^{-2}}\) the water adds \(1.01\unit{bar}\), very nearly one atmosphere, for every 10 m of depth.

Heat

\[ Q = m\, c_p\, \Delta T \] \(Q\) is the heat in joules that raises the temperature of the mass \(m\) by \(\Delta T\), and \(c_p\) is the specific heat capacity of the substance at constant pressure. For seawater \(c_p \approx 3{,}990\unit{J\,kg^{-1}\,K^{-1}}\), written "about 4,000" in text, against about 1,000 for air and about 800 for rock and sand.

Ocean Basins and the Seafloor

The depths of an ocean basin

\begin{align*} \text{shelf break} &\approx 130\unit{m}, & \text{ridge crest} &\approx 2{,}500\unit{m}, \\ \text{mean depth} &\approx 3{,}700\unit{m}, & \text{abyssal plains} &\approx 4{,}000 \text{ to } 6{,}000\unit{m}, \\ \text{Challenger Deep} &= 10{,}935 \pm 6\unit{m}. && \end{align*} These are the five numbers that place any sounding, as a single measurement of depth is called. The shelf break is the edge of the continent, the ridge crest is the youngest floor, the plains are the oldest, and the trenches are where the floor leaves the surface.

Spreading rate, distance, and age

\[ v = \frac{d}{t}, \qquad t = \frac{d}{v} \] A magnetic stripe of age \(t\) found at the distance \(d\) from the ridge axis gives the half spreading rate \(v\), the speed of one flank away from the axis; the two plates separate at the full spreading rate \(2v\). Read the other way round, the distance of any point from the axis divided by the half rate is the age of the crust there. A convenient conversion is that 1 cm per year is 10 km per million years, because a million centimetres make 10 km.

Depth of the ocean floor against its age

\[ d \approx 2{,}500\unit{m} + 350\unit{m}\,\sqrt{\frac{t}{1\unit{Myr}}} \qquad (t < 70\unit{Myr}) \] The formula that Barry Parsons and John Sclater (1977) fitted to the measured mean depth \(d\) of ocean floor of age \(t\). The crest lies at about 2,500 m, and the floor lies 350 m deeper for each unit of \(\sqrt{t}\) with \(t\) in millions of years, so 350 m deeper at 1 million years, 700 m at 4 million years, and 1,050 m at 9 million years. For older floor the depth levels off near 6,000 m. Individual soundings scatter by several hundred metres about the relation because of seamounts, sediment, and hot spots.

Sedimentation rate and the time in a core

\[ t = \frac{h}{r} \] A layer of thickness \(h\) deposited at the rate \(r\) represents the time \(t\). With \(r\) in millimetres per thousand years and \(h\) in millimetres, \(t\) comes out in thousands of years. Typical rates are 1 to 5 mm per thousand years for red clay, a few millimetres for siliceous ooze, and up to a few centimetres for calcareous ooze.

The width of a multibeam swath

\[ W \approx 3H \text{ to } 4H, \qquad \text{footprint of one beam} \approx H\,\theta \] A deep-water multibeam sounder maps a strip of width \(W\) about three to four times the water depth \(H\) on every pass, and each beam of angular width \(\theta\) (in radians) resolves a patch of about \(H\theta\) across. A ship therefore maps a wide swath in deep water and a narrow one on the shelf, and the finest detail it can see grows with the depth.

Salt, Heat, and the Density of Seawater

Practical salinity

\[ S = 35 \quad\Longleftrightarrow\quad \text{about } 35\unit{g} \text{ of dissolved salt per kilogram of seawater} \] Salinity \(S\) on the Practical Salinity Scale of 1978 is a conductivity ratio and carries no unit of measurement. The open ocean lies between about 33 and 37, with a mean of 34.7, and the symbol \(S\) is used for it throughout this program.

Potential temperature

\[ \theta = T - \Delta T_{\mathrm{ad}} , \qquad \Delta T_{\mathrm{ad}} \approx 0.1\,^{\circ}\mathrm{C} \text{ per } 1{,}000\unit{m} \] The potential temperature \(\theta\) is the temperature a parcel would have if it were raised to the surface adiabatically. \(T\) is the measured, or in situ (Latin for “in place”), temperature and \(\Delta T_{\mathrm{ad}}\) the adiabatic warming between the surface and the depth of the parcel. The rate of 0.1 ° C per 1,000 m is a rule of thumb; the exact value grows with pressure and with temperature.

The linear equation of state

\[ \rho \approx \rho_0\left[\,1 - \alpha\,(T - T_0) + \beta\,(S - S_0)\,\right] \] \(\rho_0\) is the density at a reference temperature \(T_0\) and salinity \(S_0\), for example \(1{,}027\unit{kg\,m^{-3}}\) at 10 ° C and \(S = 35\). The thermal expansion coefficient \(\alpha\) is about \(2\times10^{-4}\unit{K^{-1}}\) at mid-latitude temperatures, about four times smaller at 0 ° C, seven times smaller at the freezing point, and about one and a half times larger in the tropics. The haline contraction coefficient \(\beta\) is about \(7.6\times10^{-4}\) for each increase of 1 in salinity and varies little. The relation holds at a fixed pressure.

The freezing point of seawater

\[ T_{\mathrm{f}} \approx -0.054\,S \;\;{}^{\circ}\mathrm{C} \quad\text{at the surface, so } T_{\mathrm{f}} = -1.9\,^{\circ}\mathrm{C} \text{ at } S = 35 \] Each increase of 1 in salinity lowers the freezing point by a little more than a twentieth of a degree. Pressure lowers it further, by about 0.075 ° C for every 100 m of depth, so that water in contact with the base of an ice shelf 1,000 m down can be as cold as \(-2.7\) ° C and still liquid.

The buoyancy frequency

\[ N^2 = \frac{g}{\rho}\,\frac{\Delta\rho}{\Delta z} \] \(N\) is the buoyancy frequency (also called the Brunt-Väisälä frequency), the angular frequency with which a displaced parcel oscillates about its resting depth, with \(g = 9.81\unit{m\,s^{-2}}\) and the density increase \(\Delta\rho\) over the depth increase \(\Delta z\). \(N^2 > 0\) means a stable stratification. Typical values of \(N\) are \(10^{-2}\unit{s^{-1}}\) in a seasonal thermocline, a few times \(10^{-3}\unit{s^{-1}}\) in the main thermocline, and \(10^{-3}\unit{s^{-1}}\) or less in the deep ocean.

The Ocean and the Atmosphere

The surface heat budget

\[ Q_{\mathrm{net}} = Q_{\mathrm{sw}} - Q_{\mathrm{lw}} - Q_{\mathrm{lat}} - Q_{\mathrm{sens}} \] The net heat gain \(Q_{\mathrm{net}}\) of the sea surface, in watts per square metre, is the absorbed sunlight minus the net long-wave radiation, the latent heat of evaporation, and the sensible heat given to the air. The global ocean averages are about \(170\), \(53\), \(100\), and \(16\unit{W\,m^{-2}}\) (Wild et al. 2015), so the four terms nearly cancel and \(Q_{\mathrm{net}}\) is close to zero.

Latent heat flux and evaporation

\[ Q_{\mathrm{lat}} = L_v\,E, \qquad L_v \approx 2.5\times10^{6}\unit{J\,kg^{-1}} \] The latent heat flux \(Q_{\mathrm{lat}}\) in \(\mathrm{W\,m^{-2}}\) equals the latent heat of vaporization \(L_v\) times the evaporation rate \(E\) in kilograms of water per square metre per second. One kilogram of water is one litre, and a litre spread over a square metre is a layer one millimetre thick, so \(E\) is also an evaporation depth per unit of time. \(L_v\) falls slightly with temperature, from \(2.50\times10^{6}\unit{J\,kg^{-1}}\) at 0 °C to \(2.43\times10^{6}\) at 30 °C.

Warming of a mixed layer

\[ \Delta T = \frac{Q\,\Delta t}{\rho\,c_p\,h} \] A net surface heat flux \(Q\) acting for the time \(\Delta t\) warms (or, if negative, cools) a mixed layer of depth \(h\) by \(\Delta T\), where \(\rho \approx 1{,}025\unit{kg\,m^{-3}}\) is the density and \(c_p \approx 3{,}990\unit{J\,kg^{-1}\,K^{-1}}\) the specific heat capacity of seawater. The relation assumes that the heat stays in the layer and is not carried away by currents.

Ocean heat content

\[ \mathrm{OHC} = \sum_{\text{layers}} \rho\,c_p\,\Delta T\,\Delta V \] The ocean heat content anomaly is the sum, over all layers and regions of the ocean, of the density \(\rho\) times the specific heat capacity \(c_p\) times the temperature change \(\Delta T\) of that volume \(\Delta V\) relative to a reference period. It is quoted in zettajoules, \(1\unit{ZJ} = 10^{21}\unit{J}\), usually for the top \(700\unit{m}\) or \(2{,}000\unit{m}\).

Wind-Driven Currents and the Great Gyres

Wind stress

\[ \tau = \rho_{\mathrm{air}}\,C_D\,U_{10}^{2} \] The wind stress \(\tau\) on the sea surface grows with the square of the wind speed \(U_{10}\) measured 10 m above the water, where \(\rho_{\mathrm{air}} \approx 1.22\unit{kg\,m^{-3}}\) is the density of the air and \(C_D\) is the drag coefficient, a pure number without a unit, about \(1.3\times10^{-3}\) for moderate winds. The relation is called the bulk formula for the wind stress; \(C_D\) is a value found by measurement rather than a constant of nature, and it rises to about \(2\times10^{-3}\) in gales, when the sea is rough.

The Coriolis parameter

\[ f = 2\,\Omega\,\sin\varphi , \qquad a_{\mathrm{C}} = f\,v , \qquad T_{\mathrm{i}} = \frac{2\pi}{|f|} \] The Coriolis parameter \(f\) at latitude \(\varphi\) is twice the rotation rate of the Earth, \(\Omega = 7.29\times10^{-5}\unit{rad\,s^{-1}}\) (one turn in 23 hours and 56 minutes), multiplied by the sine of the latitude. A parcel moving at speed \(v\) is deflected sideways with the acceleration \(a_{\mathrm{C}} = f v\), to the right of its motion where \(f\) is positive (Northern Hemisphere) and to the left where it is negative (Southern Hemisphere). Left alone, the parcel goes round a full circle in the inertial period \(T_{\mathrm{i}}\), so called because nothing but the parcel’s own inertia and the turning Earth is involved. At the equator \(f = 0\), at 43° of latitude \(f = 1.0\times10^{-4}\unit{s^{-1}}\), and at the poles \(1.46\times10^{-4}\unit{s^{-1}}\).

The Ekman transport

\[ M_E = \frac{\tau}{\rho\,f} \] The volume of water carried by the wind-driven layer across each metre of a line at right angles to its motion, in \(\mathrm{m^{2}\,s^{-1}}\), equals the wind stress \(\tau\) divided by the density of seawater \(\rho\) and the Coriolis parameter \(f\). The transport is directed 90° to the right of the wind in the Northern Hemisphere and 90° to the left in the Southern Hemisphere. It holds for a steady wind over deep water away from the equator, where \(f\) is too small for the balance to form; multiplied by the length of a line in metres it gives a transport in cubic metres per second.

The Sverdrup transport

\[ M_y = \frac{1}{\rho\,\beta}\,\mathrm{curl}\,\vec{\tau} , \qquad \mathrm{curl}\,\vec{\tau} \approx -\frac{\Delta\tau_x}{\Delta y} \] The north-south volume transport \(M_y\) of the whole water column per metre of east-west width, in \(\mathrm{m^{2}\,s^{-1}}\), is the curl of the wind stress divided by the density of seawater \(\rho\) and by \(\beta\), the rate at which the Coriolis parameter increases northward. The curl of the wind stress measures how much the wind twists the sea surface; for a wind that blows east or west and changes only with latitude it is the northward change of the eastward stress \(\tau_x\) with the sign reversed, which is the second form. Between the trades and the westerlies the curl is negative in the Northern Hemisphere, so the interior transport is toward the equator. This relation is called the Sverdrup balance; it holds for the steady interior of a basin, away from the boundary currents that carry the water back.

Waves and Tides

The dispersion relation of water waves

\[ \omega^2 = g\,k\,\tanh(kH), \qquad k = \frac{2\pi}{L}, \quad \omega = \frac{2\pi}{T_{\mathrm{w}}}, \quad c = \frac{\omega}{k} \] The wavenumber \(k\) counts how many radians of the wave cycle fit into one metre, and the angular frequency \(\omega\) how many radians pass a fixed point in one second, so their ratio \(c = \omega/k\) is the familiar \(L/T_{\mathrm{w}}\). Here \(g = 9.81\unit{m\,s^{-2}}\) is the acceleration of gravity, and \(H\) is the water depth. The first paragraph of the unit used the same letter for the wave height, as the literature does, so wherever a formula contains \(H\) the unit says in words which of the two is meant. The function \(\tanh\) is the hyperbolic tangent, which rises like its argument near zero and levels off at \(1\). Because \(\tanh(kH)\) is close to \(kH\) for small \(kH\) and close to \(1\) for \(kH\) above about \(3\), which with \(kH = 2\pi H/L\) means a depth of about half a wavelength, there are two limits: \[ c = \sqrt{\frac{gL}{2\pi}} \approx 1.25\sqrt{L} \quad (\text{deep water}, H > L/2), \qquad c = \sqrt{gH} \quad (\text{shallow water}, H < L/20), \] with \(c\) in metres per second and \(L\) and \(H\) in metres.

Group speed

\[ c_g = \tfrac{1}{2}\,c \quad (\text{deep water}), \qquad c_g = c = \sqrt{gH} \quad (\text{shallow water}) \] The group speed \(c_g\) is the speed of a wave packet and of its energy: half the phase speed in deep water, so that swell crosses an ocean at half the speed of the crests seen inside it, and equal to the phase speed in shallow water, so that a tsunami arrives as fast as its first crest.

Wave energy and wave power

\[ E = \tfrac{1}{8}\,\rho\,g\,H^2, \qquad P = E\,c_g \] The energy \(E\) of a wave per square metre of sea surface grows with the square of the wave height \(H\) (the height, not the depth, in this formula), where \(\rho \approx 1{,}025\unit{kg\,m^{-3}}\) is the density of seawater. The power \(P\) passing through each metre of a line drawn across the waves, in watts per metre of crest, is the energy times the group speed \(c_g\), because the energy travels with the group. For example, a regular swell \(3\unit{m}\) high with a period of \(10\unit{s}\) has \(E = 1.13\times10^{4}\unit{J\,m^{-2}}\) and \(c_g = 7.8\unit{m\,s^{-1}}\), half of the phase speed \(1.56\times10 = 15.6\unit{m\,s^{-1}}\), so it carries about \(88\unit{kW}\) per metre of crest, the average electricity use of some 200 European homes. A real sea with \(H_s = 3\unit{m}\) holds about half as much, because \(H_s\) describes only its highest third and most of its waves are lower.

Significant wave height and the Rayleigh law

\[ H_s \approx 4\,\sigma_\eta, \qquad P(H > h) = \exp\!\left(-\frac{2h^2}{H_s^2}\right) \] The significant wave height \(H_s\) is four times the standard deviation \(\sigma_\eta\) of the sea surface height, and the fraction of individual waves higher than a given height \(h\) follows the Rayleigh law: \(e^{-2} = 13.5\) per cent of waves are higher than \(H_s\), one in 3,000 is higher than \(2H_s\), and one in about 70 million is higher than \(3H_s\). The law assumes many small, independent waves of similar period.

The tide-generating force

\[ a_{\mathrm{tide}} \approx \frac{2\,G\,M\,r}{d^3} \] The tide-generating acceleration at the surface of a planet of radius \(r\), due to a body of mass \(M\) at distance \(d\), falls with the cube of the distance, one power faster than gravity itself, because it is the difference of two inverse-square attractions. It points away from the centre of the planet on the near and the far side alike, which is why there are two bulges and two high tides a day.

The resonant period of a bay

\[ T_{\mathrm{res}} = \frac{4L}{\sqrt{gH}} \] A bay of length \(L\) and mean depth \(H\) that is open at one end has a natural period \(T_{\mathrm{res}}\) equal to the time a shallow-water wave takes to travel four times its length. If the tidal period is close to \(T_{\mathrm{res}}\), the tidal range at the head of the bay is amplified; a bay much shorter or deeper than this fills and empties with the ocean outside and gains nothing from resonance.

Seawater Chemistry and the Ocean Carbon Cycle

Apparent oxygen utilization

\[ \mathrm{AOU} = \mathrm{O_2^{sat}}(\theta, S) - \mathrm{O_2} \] AOU is the difference between the oxygen concentration \(\mathrm{O_2^{sat}}\) that the water would hold at equilibrium with the air at its potential temperature \(\theta\) (Unit 3) and salinity \(S\), and the oxygen \(\mathrm{O_2}\) actually measured. It is the oxygen consumed since the water left the surface, on the assumption that it left in equilibrium, and divided by the respiration rate along the parcel’s path it gives the parcel’s age.

The carbonate equilibria

\[ \mathrm{CO_2} + \mathrm{H_2O} \rightleftharpoons \mathrm{H_2CO_3} \rightleftharpoons \mathrm{H^+} + \mathrm{HCO_3^-} \rightleftharpoons 2\,\mathrm{H^+} + \mathrm{CO_3^{2-}} \] Dissolved carbon dioxide and carbonic acid together (written \(\mathrm{CO_2}\) for short, since the acid is less than a thousandth of the pair) exchange hydrogen ions with bicarbonate and carbonate. Which form dominates depends on the hydrogen ion concentration, and in seawater at a pH of 8.1 the dissolved inorganic carbon is about 88 per cent bicarbonate, 12 per cent carbonate, and only half a per cent dissolved \(\mathrm{CO_2}\) (at 25 ° C; in cold water the carbonate share is smaller).

Dissolved inorganic carbon and total alkalinity

\[ \mathrm{DIC} = [\mathrm{CO_2}] + [\mathrm{HCO_3^-}] + [\mathrm{CO_3^{2-}}] , \qquad \mathrm{TA} \approx [\mathrm{HCO_3^-}] + 2\,[\mathrm{CO_3^{2-}}] + [\mathrm{B(OH)_4^-}] + \ldots \] DIC is the total dissolved carbon, about \(2{,}000\unit{\mu mol\,kg^{-1}}\) in warm surface water and \(2{,}350\unit{\mu mol\,kg^{-1}}\) in the deep Pacific; total alkalinity TA, about \(2{,}300\) to \(2{,}400\unit{\mu mol\,kg^{-1}}\), counts the charge that weak bases can neutralize. From the two together, with temperature, salinity, and pressure, the pH, the partial pressure of \(\mathrm{CO_2}\), and the carbonate ion concentration follow.

Hydrogen ions from pH

\[ [\mathrm{H^+}] = 10^{-\mathrm{pH}} , \qquad \frac{[\mathrm{H^+}]_2}{[\mathrm{H^+}]_1} = 10^{\,\mathrm{pH}_1 - \mathrm{pH}_2} \] Because the pH scale is logarithmic, a fall of 0.1 in pH is a rise of \(10^{0.1} = 1.26\) in the hydrogen ion concentration, that is, 26 per cent more hydrogen ions; a fall of 0.3 doubles them, and a fall of 1.0 multiplies them by ten.

The saturation state

\[ \Omega = \frac{[\mathrm{Ca^{2+}}][\mathrm{CO_3^{2-}}]}{K_{\mathrm{sp}}} \] \(\Omega\) (capital omega) is the saturation state of seawater with respect to calcite or aragonite, the two crystal forms of calcium carbonate, each with its own solubility product \(K_{\mathrm{sp}}\). Water with \(\Omega > 1\) is supersaturated and shells are stable in it; water with \(\Omega < 1\) is undersaturated and shells dissolve. Since the calcium concentration of seawater is nearly constant at \(10.3\unit{mmol\,kg^{-1}}\) for a salinity of 35, \(\Omega\) is in effect set by the carbonate ion concentration, and the water is saturated where \([\mathrm{CO_3^{2-}}]\) equals \(K_{\mathrm{sp}}/[\mathrm{Ca^{2+}}]\): about \(42\unit{\mu mol\,kg^{-1}}\) for calcite and \(66\unit{\mu mol\,kg^{-1}}\) for aragonite in cold water at the surface (a little less in warm water), and more at depth.

The delta notation

\[ \delta^{18}\mathrm{O} = \left(\frac{R_{\mathrm{sample}}}{R_{\mathrm{standard}}} - 1\right)\times 1{,}000\unit{\text{\textperthousand}} , \qquad R = \frac{{}^{18}\mathrm{O}}{{}^{16}\mathrm{O}} \] \(\delta^{18}\mathrm{O}\) (“delta oxygen-18”) is the relative departure of the sample’s isotope ratio \(R\) from that of a standard, in per mil (‰, parts per thousand); \(\delta^{13}\mathrm{C}\) is defined in the same way with \(R = {}^{13}\mathrm{C}/{}^{12}\mathrm{C}\). The standards are Vienna Standard Mean Ocean Water (VSMOW) for the oxygen of water and Vienna Pee Dee Belemnite (VPDB) for carbonates; a positive value means the sample is richer in the heavy isotope.

The Global Overturning Circulation

Density gain from cooling and from brine

\[ \Delta\rho \approx \rho\left(-\alpha\,\Delta T + \beta\,\Delta S\right), \qquad \alpha \approx 2.5\times10^{-5}\unit{K^{-1}},\;\; \beta \approx 7.9\times10^{-4} \text{ at } -1.9\,^{\circ}\mathrm{C} \] The change of density \(\Delta\rho\) caused by a change of temperature \(\Delta T\) and a change of salinity \(\Delta S\), from the linear equation of state of Unit 3 with \(\rho \approx 1{,}028\unit{kg\,m^{-3}}\). The coefficients given are those of Weddell Sea surface water at its freezing point: there, a change of 1 in salinity changes the density as much as a change of 30 ° C would, and since the water cannot be cooled below its freezing point, salt is the only lever left.

The radiocarbon age of a water parcel

\[ t = -8{,}033\,\ln\!\left(\frac{A}{A_0}\right) \text{ years} \] \(A_0\) is the \(^{14}\mathrm{C}\) content of the water when it left the surface and \(A\) its content now; 8,033 years is the conventional mean life of radiocarbon (from the Libby half-life of 5,568 years). Two corrections are always needed: surface water starts with a deficit of about \(-50\) ‰ against the atmosphere, equivalent to about 400 years, because the surface ocean never catches up with the air, and mixing with younger water lowers the apparent age of a sample below the true average age of its parts.

The residence time of the deep water

\[ t = \frac{V}{Q} \] The average time \(t\) that water spends in a reservoir of volume \(V\) that is refilled at the rate \(Q\). For the deep ocean below the thermocline, \(V\) is about \(1.0\times10^{18}\unit{m^{3}}\) and \(Q\) the rate at which new deep water is made, about 30 Sv, which gives about a thousand years. The formula assumes a steady state, with the inflow equal to the outflow.

Munk’s balance of upwelling and mixing

\[ -w\,\frac{\partial T}{\partial z} = \kappa\,\frac{\partial^{2} T}{\partial z^{2}} \qquad\Longrightarrow\qquad T(z) = T_{\mathrm{deep}} + \left(T_{0} - T_{\mathrm{deep}}\right) e^{-z/h}, \quad h = \frac{\kappa}{w} \] In the deep ocean, the slow rise of cold water at the upward speed \(w\) is balanced by heat mixing down with the mixing coefficient \(\kappa\). The symbol \(\partial T/\partial z\) is the rate at which temperature changes with depth \(z\) (the slope of the profile, in the sense of Unit 1) and \(\partial^{2} T/\partial z^{2}\) the rate at which that slope changes; the minus sign is there because \(z\) is counted downwards while the water moves up. The balance is satisfied by a temperature profile that decays exponentially with depth towards the deep value \(T_{\mathrm{deep}}\) with the scale depth \(h = \kappa/w\), \(T_{0}\) being its extrapolated surface value; measured deep profiles have \(h \approx 1\unit{km}\).

The transport through a section

\[ Q = \sum_{i} v_i\,A_i \] The volume transport \(Q\) through a vertical section is the sum, over all the pieces \(A_i\) into which the section is divided, of the velocity \(v_i\) through each piece (positive northwards) times its area, expressed in sverdrups. The meridional overturning circulation at a latitude is the net northward transport summed from the surface down to the depth at which the flow turns southward, which in the Atlantic at 26.5° N lies at about 1,100 m; the same amount flows south below that depth.

The heat transport of the overturning

\[ \mathrm{HT} = \rho\,c_p\,Q\,\Delta T \] The heat carried across a latitude by an overturning of transport \(Q\) whose northward and southward branches differ in temperature by \(\Delta T\), with \(\rho \approx 1{,}027\unit{kg\,m^{-3}}\) the density and \(c_p \approx 4{,}000\unit{J\,kg^{-1}\,K^{-1}}\) the specific heat capacity of seawater (Unit 1). The product \(\rho c_p\) is about \(4\times10^{6}\) joules per cubic metre and per degree. At 26.5° N the Atlantic carries about 1.2 PW northwards, where \(1\unit{PW} = 10^{15}\unit{W}\).

Rotation, Geostrophy, and Ocean Eddies

The Coriolis acceleration

\[ \vec{a}_{C} = -2\,\vec{\Omega}\times\vec{u}, \qquad \text{horizontally}\quad a_x = f v, \quad a_y = -f u, \quad f = 2\Omega\sin\varphi \] A parcel moving with velocity \(\vec{u}\) relative to the Earth, which rotates with \(\vec{\Omega}\), appears to accelerate at right angles to its motion. Horizontally the acceleration has the magnitude \(f\) times the speed and points to the right of the motion where \(f > 0\) (Northern Hemisphere) and to the left where \(f < 0\) (Southern Hemisphere). \(f\) is the Coriolis parameter of Unit 5, about \(10^{-4}\unit{s^{-1}}\) at mid-latitudes and zero at the equator.

The Rossby number

\[ Ro = \frac{U}{f L} \] The ratio of the acceleration of the water to the Coriolis acceleration for a flow of speed \(U\) and horizontal scale \(L\) at a place where the Coriolis parameter is \(f\). When \(Ro \ll 1\) the rotation of the Earth controls the motion and the geostrophic balance of Section 2 holds; when \(Ro \gg 1\) the rotation can be ignored. It is named after Carl-Gustaf Rossby (1898-1957), whom Section 5 introduces.

The geostrophic relation

\[ f v = \frac{1}{\rho}\,\frac{\partial p}{\partial x}, \qquad f u = -\frac{1}{\rho}\,\frac{\partial p}{\partial y}, \qquad\text{and at the surface}\qquad v = \frac{g}{f}\,\frac{\partial \eta}{\partial x}, \quad u = -\frac{g}{f}\,\frac{\partial \eta}{\partial y} \] In a steady, frictionless flow with a small Rossby number, the Coriolis acceleration balances the pressure gradient force. The current \((u, v)\) runs along the lines of constant pressure, the isobars, with the high pressure on its right in the Northern Hemisphere and on its left in the Southern Hemisphere, and its speed is the pressure gradient divided by \(\rho f\). Near the surface the pressure gradient comes from the slope of the sea surface \(\eta\) (hydrostatic pressure, \(p = p_{\mathrm{atm}} + \rho g\,(\eta - z)\) with \(z\) upward), which gives the second pair, already used in Unit 5. The relation fails within a few degrees of the equator, where \(f\) goes to zero.

The thermal wind relation

\[ f\,\frac{\partial v}{\partial z} = -\frac{g}{\rho_0}\,\frac{\partial \rho}{\partial x}, \qquad f\,\frac{\partial u}{\partial z} = \frac{g}{\rho_0}\,\frac{\partial \rho}{\partial y} \] The vertical shear of the geostrophic current (its change per metre of height, \(z\) upward) is set by the horizontal gradient of density. In the Northern Hemisphere the current increases upward with the light water on its right; a current flowing eastward with cold, dense water on its northern side, like the Gulf Stream, is therefore strongest at the surface and weakens with depth. The name comes from meteorology, where the same relation links the westerly winds aloft to the temperature difference between the equator and the poles.

The Rossby radius of deformation

\[ L_R = \frac{\sqrt{gH}}{|f|} \quad\text{(barotropic)}, \qquad L_R = \frac{c_1}{|f|} \approx \frac{N H}{\pi\,|f|} \quad\text{(baroclinic)} \] The horizontal scale on which the rotation of the Earth takes over from gravity. For motions of the whole water column of depth \(H\) the relevant wave is the long surface gravity wave of Unit 6 with speed \(\sqrt{gH}\), which gives about \(2{,}000\unit{km}\) at mid-latitudes. For motions of the stratified upper ocean it is the fastest internal wave, with speed \(c_1\). For a stratification that is uniform over a depth \(H\) that speed is \(NH/\pi\), with \(N\) the buoyancy frequency of Unit 3, and textbooks often write the radius as the rounder scale \(NH/f\); in the real ocean, where \(N\) is large in the thermocline and small below it, \(c_1\) is computed from measured profiles as \(\frac{1}{\pi}\int N\,\dd z\) over the whole water column and comes to \(2.5\) to \(3\unit{m\,s^{-1}}\) in the tropics and about \(2\unit{m\,s^{-1}}\) at mid-latitudes, which gives tens of kilometres. The baroclinic radius sets the size of ocean eddies.

The speed of long Rossby waves

\[ c = -\beta\,L_R^{2} \] A Rossby wave much longer than the Rossby radius \(L_R\) travels westward (the minus sign) at the speed \(\beta L_R^{2}\). For the baroclinic waves of the thermocline this is centimetres per second at mid-latitudes and tens of centimetres per second in the tropics, because \(L_R\) grows toward the equator while \(\beta\) changes little. Shorter waves travel more slowly, and the fastest barotropic Rossby waves, with \(L_R\) of thousands of kilometres, cross an ocean in a few weeks.

Coasts, Estuaries, and Upwelling

The flushing time of an estuary

\[ t_f = \frac{V_{\mathrm{fresh}}}{R}, \qquad V_{\mathrm{fresh}} = V\left(1 - \frac{\bar{S}}{S_{\mathrm{sea}}}\right) \] The flushing time \(t_f\) is the time the river flow \(R\) needs to replace the fresh water held in the estuary, \(V_{\mathrm{fresh}}\), which is the volume \(V\) multiplied by the fresh water fraction; \(\bar{S}\) is the mean salinity of the estuary and \(S_{\mathrm{sea}}\) the salinity of the sea it opens into. It is the time a dissolved substance that entered with the river stays in the estuary on average, hours to days for a small tidal inlet and decades for an enclosed sea.

Upwelling transport and upwelling speed

\[ M_E = \frac{\tau_{\parallel}}{\rho\,|f|}, \qquad w = \frac{M_E}{L_{\mathrm{up}}} \] A wind stress \(\tau_{\parallel}\) along the coast drives an Ekman transport \(M_E\) offshore, in cubic metres per second per metre of coastline (which is square metres per second), where \(\rho\) is the density of seawater and \(f\) the Coriolis parameter; only the component of the stress along the coast counts. The water rises at the mean speed \(w\) across a strip of width \(L_{\mathrm{up}}\), about one baroclinic Rossby radius, 10 to 50 km. The transport per \(100\unit{m}\) of coastline is the upwelling index of Andrew Bakun (1939-2025), who tabulated it for the North American coast from atmospheric pressure fields in 1973; values of 100 to 300 cubic metres per second per 100 m of coast are typical of summer off northern California.

The height of a storm surge

\[ \eta_{\mathrm{surge}} \approx \frac{\tau}{\rho g H}\,L \;+\; \frac{\Delta p}{\rho g} \] The storm surge at a coast is the wind set-up, which is the slope \(\tau/(\rho g H)\) multiplied by the length \(L\) of shelf over which the wind stress \(\tau\) blows toward the coast over water of mean depth \(H\), plus the inverted barometer rise, about 1 cm per hectopascal that the pressure \(\Delta p\) falls below normal. The relation gives the right order of magnitude and the right dependence on depth and fetch; the surges that engineers design against are computed with the full equations of motion on a model of the real coast, whose shape, tide, and waves change the answer by a factor of two.

The sea-level budget

\[ \frac{\Delta h}{\Delta t} = \underbrace{1.2}_{\text{thermal}} + \underbrace{0.7}_{\text{glaciers}} + \underbrace{0.65}_{\text{Greenland}} + \underbrace{0.3}_{\text{Antarctica}} + \underbrace{0.3}_{\text{land water}} \approx 3.2\unit{mm\,yr^{-1}} \quad (1993\text{ to }2018) \] The observed rise of the global mean sea level is the sum of the thermal expansion of the warming ocean and of the water added by glaciers, by the two ice sheets, and by changes in the water stored on land, each measured independently (ocean temperatures, glacier surveys, satellite gravity and altimetry over the ice, hydrological models). When the sum matches the observed rise within the uncertainties, the budget is said to close, and for 1993 to 2018 it closes to within about 0.2 mm a year.

El Niño and Ocean Climate Variability

The tilt of the thermocline under the trade winds

\[ |\tau_x|\,L = \rho\,g'\,H\,\Delta h , \qquad \Delta\eta = \frac{\Delta\rho}{\rho}\,\Delta h \] A westward wind stress of magnitude \(|\tau_x|\) acting over a basin of length \(L\) thickens a warm layer of mean thickness \(H\) and reduced gravity \(g' = g\,\Delta\rho/\rho\) by \(\Delta h\) toward the west; the sea surface rises in the same direction by the much smaller amount \(\Delta\eta\). The balance holds for the mean state; a change of the wind is followed by the adjustment of Section 3, which takes months.

The speed of an equatorial Kelvin wave

\[ c = \sqrt{g'\,H_1} , \qquad g' = g\,\frac{\Delta\rho}{\rho} \] A long wave on the thermocline travels at the speed \(c\), where \(H_1\) is the thickness of the warm layer and \(g'\) the reduced gravity of the density step \(\Delta\rho\) across the thermocline. For the Pacific, with \(H_1 \approx 150\unit{m}\) and \(\Delta\rho \approx 4\unit{kg\,m^{-3}}\), \(c\) is about \(2.4\unit{m\,s^{-1}}\); observed Kelvin waves travel at \(2\) to \(3\unit{m\,s^{-1}}\). An equatorial Kelvin wave moves eastward only.

The equatorial Rossby radius

\[ L_{\mathrm{eq}} = \sqrt{\frac{c}{\beta}} , \qquad \beta = \frac{\dd f}{\dd y} = 2.3\times10^{-11}\unit{m^{-1}\,s^{-1}} \text{ at the equator} \] The distance from the equator over which the Coriolis parameter becomes large enough to hold a wave of speed \(c\) in place. Equatorial Kelvin and Rossby waves and the equatorial currents are confined to a band of a few \(L_{\mathrm{eq}}\) on either side of the equator.

The recharge oscillator

\[ \frac{\dd T_E}{\dd t} = R\,T_E + F_1\,h_W , \qquad \frac{\dd h_W}{\dd t} = -r\,h_W - F_2\,T_E \] \(T_E\) is the sea surface temperature anomaly of the eastern Pacific and \(h_W\) the anomaly of the warm layer thickness in the west, a measure of the warm water stored along the equator. The first equation is the Bjerknes feedback: the east warms at a rate set by its own anomaly (through \(R\), which is positive when the feedback is strong) and by the stored warm water (\(F_1\)). The second is the discharge: a warm east drives the winds that drain the warm water away (\(F_2\)), while \(r\) is the slow damping. The two together oscillate with a period of a few years.

The Niño 3.4 index and the Oceanic Niño Index

\begin{gather*} \text{Niño 3.4} = \overline{T'}\ \text{averaged over } 5\,\text{\textdegree N to }5\,\text{\textdegree S}\text{ and }170\,\text{\textdegree W to }120\,\text{\textdegree W}, \\ \mathrm{ONI} = \text{three-month running mean of Niño 3.4} \end{gather*} The Niño 3.4 index is the sea surface temperature anomaly \(T'\), the departure from the long-term mean of the same month, averaged over the region. The Oceanic Niño Index of the NOAA Climate Prediction Center is its mean over three consecutive months, with the anomaly computed against a 30-year base period that is moved forward every five years so that the warming of the ocean itself does not count as El Niño. An El Niño episode is declared when the ONI stays at or above \(+0.5\,\text{\textdegree C}\) for at least five consecutive overlapping three-month seasons, and a La Niña when it stays at or below \(-0.5\,\text{\textdegree C}\) for as long.

The Southern Oscillation Index

\[ \mathrm{SOI} = \frac{\Delta p - \overline{\Delta p}}{\sigma_{\Delta p}} , \qquad \Delta p = p_{\mathrm{Tahiti}} - p_{\mathrm{Darwin}} \] The monthly pressure difference between Tahiti and Darwin minus its long-term mean for that calendar month, divided by its standard deviation, so that the index is a pure number of typical size one. The forecast centres differ in detail: the NOAA Climate Prediction Center standardizes the two stations separately before taking the difference, and the Australian Bureau of Meteorology multiplies its index by ten, so that its values run to about \(\pm30\). Sustained values below about \(-1\) (or \(-10\) on the Australian scale) go with El Niño.

Polar Oceans, Sea Ice, and Ice Shelves

The freezing point of seawater and the salt that freezing leaves behind

\[ T_{\mathrm{f}} \approx -0.054\,S \;\;{}^{\circ}\mathrm{C}, \qquad \Delta S \approx \frac{\rho_{\mathrm{i}}\,h}{\rho_{\mathrm{w}}\,H}\,(S_{\mathrm{w}} - S_{\mathrm{i}}) \] Seawater of salinity \(S\) at the surface freezes at \(T_{\mathrm{f}}\), which is \(-1.9\) ° C for \(S = 35\). When a layer of ice of thickness \(h\), density \(\rho_{\mathrm{i}} = 917\unit{kg\,m^{-3}}\), and salinity \(S_{\mathrm{i}}\) forms from seawater of salinity \(S_{\mathrm{w}}\), the rejected salt raises the salinity of the water column of depth \(H\) and density \(\rho_{\mathrm{w}}\) below it by \(\Delta S\), if nothing carries the brine away (exactly, the denominator is \(\rho_{\mathrm{w}} H - \rho_{\mathrm{i}} h\), a correction of about 1 per cent). This brine rejection is the third way of making dense water. Unit 8 worked the numbers for a 100 m column on the Weddell Sea shelf, where one metre of new ice raises the salinity by about 0.24; over the 50 m column of a coastal polynya the same ice gives a rise of about 0.5, enough to make the shelf water denser than the deep water offshore and start it down the slope as Antarctic Bottom Water.

The thickness of sea ice after a given cold

\[ h = \sqrt{\frac{2\,k_{\mathrm{i}}}{\rho_{\mathrm{i}}\,L_{\mathrm{f}}}\,\mathrm{FDD}} \approx 3.5\,\sqrt{\mathrm{FDD}}\;\text{cm}, \qquad h_{\mathrm{Lebedev}} \approx 1.33\,\mathrm{FDD}^{0.58}\;\text{cm} \] The thickness \(h\) of snow-free sea ice grows as the square root of the accumulated freezing degree days (in ° C days), with the numerical value for \(k_{\mathrm{i}} = 2.2\unit{W\,m^{-1}\,K^{-1}}\), \(\rho_{\mathrm{i}} = 917\unit{kg\,m^{-3}}\), and \(L_{\mathrm{f}} = 334\unit{kJ\,kg^{-1}}\). The first formula is Stefan’s upper bound; the second is the empirical fit that Lebedev published in 1938 for first-year ice measured at Soviet stations in the Arctic, snow and ocean heat flux included, and it gives about two thirds to three quarters of Stefan’s value, the fraction rising slowly with the length of the winter.

Absorbed sunlight and the albedo

\[ Q_{\mathrm{abs}} = (1 - \alpha)\,Q_{\mathrm{sun}}, \qquad \alpha \approx 0.06 \text{ (water)}, \quad 0.5 \text{ to } 0.7 \text{ (bare ice)}, \quad 0.8 \text{ to } 0.9 \text{ (snow)} \] A surface of albedo \(\alpha\) absorbs the fraction \(1 - \alpha\) of the incoming solar flux \(Q_{\mathrm{sun}}\). Melting ice therefore sets off the ice-albedo feedback: when ice gives way to water, the surface absorbs two to nine times more sunlight (snow at 0.9 absorbs a tenth of it, water 94 per cent), the water warms, more ice melts from the side and from below, and the warming grows. The values are those measured by Donald Perovich and colleagues on Arctic ice; melt ponds lie in between, at 0.2 to 0.4.

The transport of the Antarctic Circumpolar Current

\[ Q_{\mathrm{ACC}} = \int\!\!\int u\;\dd z\,\dd y \approx 130\text{ to }170\unit{Sv} \] The transport is the eastward velocity \(u\) added up over the depth and the width of the passage, \(173 \pm 10\unit{Sv}\) as measured in Drake Passage in 2007 to 2011. Most of it is set by the thermal wind of Unit 9: the water is denser on the southern side by about \(0.5\unit{kg\,m^{-3}}\) across the 800 km of the passage, so the eastward speed grows upward through the water column, by about 0.12 m/s between 2,500 m and the surface at the latitude of Drake Passage, and a current of that modest speed over a cross-section 800 km wide and 2,500 m deep integrates to about 120 to 130 Sv. The remaining 40 to 50 Sv is a flow of the whole water column that the thermal wind cannot see and that only the moored current meters measured; the current is slow, since 0.1 m/s is a slow walk, but it is wide and it reaches the seafloor.

The freezing point at depth

\[ T_{\mathrm{f}}(S, p) \approx -0.054\,S - 7.5\times10^{-4}\,p \;\;{}^{\circ}\mathrm{C} \] With the pressure \(p\) in decibars, the freezing point of seawater of salinity \(S\) falls by about 0.75 ° C for every 1,000 m of depth, so water that is at the freezing point at the surface, \(-1.9\) ° C, is 0.75 ° C above the freezing point at a grounding line 1,000 m deep and can melt ice there. The deepest parts of an ice shelf are therefore the first to melt, and the freezing point rises again as the meltwater ascends.

The sea-level equivalent of an ice mass

\[ \Delta\mathrm{SL} = \frac{M}{\rho_{\mathrm{w}}\,A_{\mathrm{ocean}}}, \qquad 362\unit{Gt}\;\text{of ice} = 1\unit{mm}\;\text{of global sea level} \] A mass \(M\) of ice that melts and runs into the sea raises the global mean sea level by \(\Delta\mathrm{SL}\), with \(\rho_{\mathrm{w}} = 1{,}000\unit{kg\,m^{-3}}\) for the fresh meltwater and \(A_{\mathrm{ocean}}\) the area of the ocean, 361 to 362 million square kilometres depending on which marginal seas are counted; 362 Gt per millimetre is the conventional figure. Averaged over the GRACE record from 2002 to 2025, NASA gives the loss of Antarctica as about 135 Gt per year and that of Greenland as about 264 Gt per year, 400 Gt per year together, so the two ice sheets raise the sea by \(400/362 = 1.1\unit{mm}\) per year, roughly a third of the 3.3 to 4 mm per year that the altimeters have measured in recent years (Unit 10); the rest comes from the thermal expansion of the warming water and from mountain glaciers. Greenland holds about 7 m of sea-level equivalent, West Antarctica about 3 to 5 m, and East Antarctica about 50 m; the ice sheets dominate the long-term projections because their losses are accelerating and because the Amundsen Sea sector alone could add more than a metre over the coming centuries.

Reading the Seafloor Archive

Age, depth, and accumulation rate

\[ t = \frac{h}{r} , \qquad \Delta t_{\mathrm{sample}} = \frac{\Delta h}{r} \] A layer of thickness \(h\) deposited at the rate \(r\) represents the time \(t\), and a sample of thickness \(\Delta h\) averages over the time \(\Delta t_{\mathrm{sample}}\). With \(r\) in centimetres per thousand years and \(h\) in centimetres, \(t\) comes out in thousands of years. The relation assumes that the rate did not change within the layer, which is the main assumption behind every simple age model.

Radioactive decay and the radiocarbon age

\[ N = N_0\,2^{-t/t_{1/2}} , \qquad t_{1/2} = 5{,}730\ \text{years}; \qquad t_{\mathrm{conv}} = -8{,}033\ \text{years}\times\ln\frac{A}{A_0} \] \(N_0\) atoms of a radioactive isotope with the half-life \(t_{1/2}\) leave \(N\) after the time \(t\). The conventional radiocarbon age \(t_{\mathrm{conv}}\) is computed from the measured activity \(A\) relative to the modern standard \(A_0\) with the mean life of 8,033 years that belongs to the Libby half-life of 5,568 years, and is quoted in radiocarbon years before present, where “present” is the year 1950. It becomes a calendar age only after calibration.

The oxygen isotope palaeotemperature equation

\[ T \approx 16.9 - 4.38\,(\delta_c - \delta_w) + 0.10\,(\delta_c - \delta_w)^2 \quad(\text{in }^{\circ}\mathrm{C}) \] \(T\) is the temperature at which a calcite shell grew, \(\delta_c\) the \(\delta^{18}\mathrm{O}\) of the calcite against VPDB, and \(\delta_w\) the \(\delta^{18}\mathrm{O}\) of the water it grew in, converted to the same scale by subtracting 0.27 ‰ from its VSMOW value. Near 15° C the slope is about \(-4.4\)° C per ‰, that is about \(-0.23\) ‰ per degree, which the text rounds to the customary \(-0.22\) ‰. Shackleton’s 1974 calibration is given; others agree within a few tenths of a degree.

The ice volume effect

\begin{gather*} \Delta\delta_w \approx 0.010\unit{\text{\textperthousand}}\ \text{per metre of sea-level fall} \\ \Delta\delta_w \approx 1.0\ \text{to}\ 1.2\unit{\text{\textperthousand}}\ \text{for the glacial fall of about 120 m} \end{gather*} Locking light water into ice sheets with a mean \(\delta^{18}\mathrm{O}\) of about \(-30\) to \(-35\) ‰ raises the \(\delta^{18}\mathrm{O}\) of the whole ocean by about one hundredth of a per mil for every metre that sea level falls. The glacial value is measured directly in the pore water of deep-sea sediments, which still holds the glacial ocean’s signature, and lies between 1.0 and 1.2 ‰; the value depends on the mean composition of the vanished ice, which is why a range is given.

The Mg/Ca thermometer

\[ \frac{\mathrm{Mg}}{\mathrm{Ca}} = B\,e^{A\,T} \qquad\Longleftrightarrow\qquad T = \frac{1}{A}\ln\frac{\mathrm{Mg/Ca}}{B} \] The Mg/Ca ratio of foraminiferal calcite, in millimoles per mole, grows exponentially with the temperature \(T\) at which the shell formed. The widely used calibration of Anand, Elderfield, and Conte from sediment traps in the Sargasso Sea gives \(A = 0.090\) per ° C and \(B = 0.38\) mmol/mol for several planktic species together; other species and other oceans have their own constants, and \(A\) lies between about 0.08 and 0.10 for all of them.

The Ocean in the Ice Ages

The daily insolation at the top of the atmosphere

\[ \bar{Q} = \frac{S_0}{\pi}\left(\frac{\bar{d}}{d}\right)^{2}\bigl(H_0\sin\varphi\sin\delta + \cos\varphi\cos\delta\sin H_0\bigr), \qquad \frac{d}{\bar{d}} = \frac{1-\varepsilon^2}{1+\varepsilon\cos(\lambda - \tilde{\omega})} \] \(\bar{Q}\) is the daily mean insolation in \(\mathrm{W\,m^{-2}}\) at latitude \(\varphi\), \(S_0 = 1{,}361\unit{W\,m^{-2}}\) is the solar constant (the flux at the distance \(\bar{d}\), the semi-major axis \(a\) of the orbit), \(d\) the actual distance, \(\delta\) the declination of the Sun (its latitude on the sky, \(\sin\delta = \sin\varepsilon_{\mathrm{obl}}\sin\lambda\)), \(\lambda\) the Sun’s longitude along the orbit (0° at the March equinox, 90° at the June solstice), and \(H_0\) the half length of the day as an angle, from \(\cos H_0 = -\tan\varphi\tan\delta\) (\(H_0 = \pi\) in the polar day, 0 in the polar night). The bracket contains the geometry of the day, the distance factor the orbit.

The precession index

\[ P = \varepsilon\sin\tilde{\omega}, \qquad \frac{\Delta \bar{Q}_{\mathrm{June}}}{\bar{Q}_{\mathrm{June}}} \approx 2\varepsilon\sin\tilde{\omega} \] The precession index \(P\) is positive when perihelion falls in the northern summer and negative when it falls in the northern winter, as today (\(P = -0.016\)). The fractional change of the June solstice insolation relative to a circular orbit is about \(2P\), so the precession can change the northern summer sunshine by up to about 12 per cent when the eccentricity is large, and by almost nothing when it is small. Figure 2 shows its beating envelope.

Ice volume and sea level

\[ V_{\mathrm{ice}} = A_{\mathrm{ocean}}\,\Delta h\,\frac{\rho_{\mathrm{w}}}{\rho_{\mathrm{ice}}}, \qquad 362\unit{Gt\ of\ ice} = 1\unit{mm\ of\ sea\ level} \] A fall of the sea level by \(\Delta h\) over the ocean area \(A_{\mathrm{ocean}} = 361\) million \(\mathrm{km^2}\) corresponds to the ice volume \(V_{\mathrm{ice}}\), with the densities of water, \(\rho_{\mathrm{w}} = 1{,}000\unit{kg\,m^{-3}}\), and of glacier ice, \(\rho_{\mathrm{ice}} = 917\unit{kg\,m^{-3}}\). The second form is the one of Unit 12. The relation neglects the shrinking of the ocean area as the shelves emerge, which the full reconstructions include. The same ice raises the \(\delta^{18}\mathrm{O}\) of the ocean by about 0.01 ‰ per metre (Unit 13).

The carbon budget of the glacial atmosphere

\[ \Delta M_{\mathrm{atm}} = 2.12\unit{Pg\,C\ per\ ppm}\times\Delta(\mathrm{CO_2}) \] A change of the atmospheric carbon dioxide by \(\Delta(\mathrm{CO_2})\) parts per million is a change of the atmosphere’s carbon by \(\Delta M_{\mathrm{atm}}\). The factor is the mass of the atmosphere converted to moles of air and multiplied by the molar mass of carbon, and it is the same factor as in Unit 7.

The solubility effect on the carbon dioxide of the air

\[ \frac{\Delta p\mathrm{CO_2}}{p\mathrm{CO_2}} \approx 0.04\,\Delta T \quad \text{per }^{\circ}\mathrm{C} \] At fixed carbon content and alkalinity, the partial pressure of carbon dioxide over seawater rises by about 4 per cent for each degree of warming and falls by the same fraction for each degree of cooling (the measured coefficient is 4.2 per cent per degree). Applied to the whole ocean surface, it gives the change of the atmospheric concentration that a uniform change of sea surface temperature produces once ocean and atmosphere are in balance.

Abrupt Climate Changes and the Ocean

The protactinium-thorium ratio as a rate meter

\[ \left(\frac{^{231}\mathrm{Pa}}{^{230}\mathrm{Th}}\right)_{\mathrm{sediment}} = 0.093\,(1 - f_{\mathrm{exp}}) \] Both isotopes are produced in seawater in the ratio 0.093, and \(f_{\mathrm{exp}}\) is the fraction of the protactinium that the overturning carries out of the Atlantic before it settles. A sediment ratio at the production value means that nothing was exported and the overturning was off; the Holocene value of the Bermuda Rise, 0.055, means that about 40 per cent was exported. The relation between \(f_{\mathrm{exp}}\) and the transport in sverdrups is only roughly proportional, because the particle flux also matters.

The thermal bipolar seesaw

\[ \frac{\dd T_{\mathrm{S}}}{\dd t} = -\,\frac{T_{\mathrm{N}} + T_{\mathrm{S}}}{\tau} \] \(T_{\mathrm{N}}\) is the temperature anomaly of the North Atlantic, \(T_{\mathrm{S}}\) that of the Southern Ocean reservoir, and \(\tau\) the time the reservoir needs to adjust. The reservoir is pushed toward the mirror image of the north, \(-T_{\mathrm{N}}\), but only at the rate \(1/\tau\); Stocker and Johnsen found that \(\tau\) of about 1,000 to 1,500 years, best 1,120 years, turns the GRIP record into the Byrd record.

Stommel’s two-box model

\[ q = k\,(\alpha\,\Delta T - \beta\,\Delta S), \qquad |q|\,\Delta S = S_0\,F_{\mathrm{w}} \ \ \text{in equilibrium} \] The flow \(q\) is proportional to the density difference between the boxes, with \(k\) a constant of the pipes; \(\alpha\) and \(\beta\) are the expansion coefficients of Unit 3 and \(\Delta T\) is held fixed by the fast exchange of heat with the air. Fresh water enters the cold box at the rate \(F_{\mathrm{w}}\) and is evaporated from the warm one, which is the same as a salt flux \(S_0 F_{\mathrm{w}}\) in the opposite direction, and in equilibrium the flow carries exactly that much salt back, \(|q|\,\Delta S\). Temperature drives the flow and salt opposes it; \(q > 0\) is the mode in which cold water sinks in the north.

Fresh water, volume, and sea level

\[ V = F_{\mathrm{w}}\,t, \qquad \Delta h = \frac{V}{A_{\mathrm{ocean}}}, \qquad \text{rate} = \frac{\Delta h}{\Delta t} \] A freshwater flux \(F_{\mathrm{w}}\) in cubic metres per second kept up for a time \(t\) delivers the volume \(V\); spread over the ocean’s area \(A_{\mathrm{ocean}} = 3.61\times10^{14}\unit{m^{2}}\) it raises sea level by \(\Delta h\) if it came from ice on land. A flux of 0.1 Sv for a century is \(3.2\times10^{14}\unit{m^{3}}\), about 320,000 km³, and 0.9 m of sea level.

Detecting a trend in a noisy record

\[ \sigma_{b} = \sigma\sqrt{\frac{12}{n\,(n^{2} - 1)}} \] A trend \(b\) fitted by a straight line to \(n\) yearly means that scatter by the standard deviation \(\sigma\) about the line has itself the standard deviation \(\sigma_{b}\). The trend counts as detected when \(b\) exceeds about \(2\sigma_{b}\), so the number of years needed grows as the two-thirds power of the noise-to-trend ratio. The formula assumes that the years scatter independently; when they do not, more years are needed.

Warm Worlds of the Past

Carbon isotope mass balance

\[ \Delta\delta^{13}\mathrm{C} \approx \frac{M_{\mathrm{added}}\,(\delta_{\mathrm{added}} - \delta_{\mathrm{res}})}{M_{\mathrm{res}} + M_{\mathrm{added}}} \qquad\Longleftrightarrow\qquad M_{\mathrm{added}} = \frac{M_{\mathrm{res}}\,\Delta\delta^{13}\mathrm{C}}{\delta_{\mathrm{added}} - \delta_{\mathrm{res}} - \Delta\delta^{13}\mathrm{C}} \] A mass \(M_{\mathrm{added}}\) of carbon with the isotope ratio \(\delta_{\mathrm{added}}\) added to a reservoir of mass \(M_{\mathrm{res}}\) and ratio \(\delta_{\mathrm{res}}\) shifts the ratio of the mixture by \(\Delta\delta^{13}\mathrm{C}\). The relation assumes that the added carbon mixes through the whole reservoir before any of it is removed, so it gives a lower bound for a release that lasted thousands of years.

The weathering time scale

\[ \tau_{\mathrm{w}} \approx \frac{M_{\mathrm{added}}}{\Delta F_{\mathrm{w}}} \] The time \(\tau_{\mathrm{w}}\) to remove an added mass \(M_{\mathrm{added}}\) of carbon by weathering is the mass divided by the extra weathering flux \(\Delta F_{\mathrm{w}}\) that the warmer climate drives. Today’s silicate weathering consumes carbon at about 0.1 to 0.2 Pg per year, and a warming of several degrees raises it by a few hundredths of a petagram per year, so a pulse of a few thousand petagrams needs about 100,000 years, the recovery time seen in Figure 1 for the PETM and in the carbonate of Figure 2.

Climate sensitivity

\[ \Delta T = S\,\log_2\!\left(\frac{C}{C_0}\right) \] The equilibrium warming \(\Delta T\) for a carbon dioxide concentration \(C\) relative to a reference \(C_0\) (280 ppm for the pre-industrial atmosphere) is the sensitivity \(S\), in degrees per doubling, times the number of doublings. The Sixth Assessment Report of the IPCC gives \(S\) a best estimate of 3 °C and a likely range of 2.5 to 4 °C. The relation holds for the fast feedbacks only; with ice sheets and vegetation free to change, the Earth system sensitivity is larger, and the Pliocene is the measure of how much.

Sea-level equivalent of an ice mass

\[ \Delta h = \frac{m_{\mathrm{ice}}}{\rho_{\mathrm{w}}\,A_{\mathrm{ocean}}} , \qquad 362\unit{Gt} \;\approx\; 1\unit{mm} \] An ice mass \(m_{\mathrm{ice}}\) that melts into an ocean of area \(A_{\mathrm{ocean}}\) raises sea level by \(\Delta h\), with \(\rho_{\mathrm{w}}\) the density of the melt water. A layer of 1 mm over the ocean’s \(3.61\times10^{14}\unit{m^{2}}\) is \(3.61\times10^{11}\unit{m^{3}}\) of water, which at \(1{,}000\unit{kg\,m^{-3}}\) is 361 gigatonnes, rounded in the literature to 362 as in Unit 12. Greenland holds about 7 m, West Antarctica about 3 to 5 m, and East Antarctica about 50 m of sea level. The relation ignores the small corrections for the ice that already floats or rests below sea level and for the Earth’s response to the shifted load.