OPOpenPithMake curiosity visibleGeography lab / Seawater density and circulation

About this lab

Five lessons, one thread: density differences move water up and down, wind moves surface water sideways, and together they make ocean circulation.

Seawater density depends on temperature, salinity and pressure. Cooling or adding salt makes water denser, so it sinks; warming or freshening makes it lighter, so it floats. Surface water is also pushed by wind, deflected by Earth’s rotation and blocked by continents. Deep water feels almost no wind and is driven mainly by density.

The sinking in the glass tank is a teaching model: blob speed comes from the density difference, mixing and diffusion are ignored, the constants are not calibrated to measurements, and only relative changes between runs mean anything; a tank cannot explain the whole ocean circulation. The weighing-station equipment is about twice life size, so its scale differs from the tank. The pale blue and yellow in the tank are teaching dyes: colour shows salinity, not the true colour of seawater. Current lines and the conveyor on the globe are schematic paths, not observed flow fields; the conveyor depth is exaggerated about 40 times.

Five conclusions

Five conclusions

  • Salinity and density: 3.5 g of salt made up to 100 g with water gives a brine of 1.0246 g/mL at 20 °C against 0.9982 g/mL for fresh water, about 2.6% denser. The denominator of the recipe is the total mass, not the mass of water. With coarse gear (0.1 g balance, 1 mL cylinder divisions) the smallest mass fraction you can tell apart is 2.76%; with fine gear (0.01 g, 0.2 mL) it is 0.54%. When the density difference is smaller than the sum of the two uncertainties, the honest result is “cannot tell apart”.
  • Temperature–salt layering: the density difference decides whether water sinks or rises and where it stops. Fresh water is densest at 3.98 °C, so 4 °C fresh water sinks into 10 °C fresh water; above a salinity of about 24.6 g/kg the density maximum falls below the freezing point and water keeps getting denser until it freezes. For water of 35 g/kg at 2 °C, density rises from 1027.97 to 1070.38 kg/m³ between the surface and 10000 dbar.
  • Evaporation, rain and freezing: take 1 kg of seawater (35 g of salt, 965 g of water). Evaporating 96.5 g of water raises the salinity to 38.74 g/kg; adding 96.5 g of rain lowers it to 31.92; adding 200 g of river water lowers it to 29.17; freezing out 100 g of ice (ice salinity taken as 4 g/kg) raises what is left to 38.44. The salt stays, the water changes, and salinity and density follow.
  • Surface currents: driven by wind, deflected by rotation and constrained by continents. The five subtropical gyres turn clockwise in the Northern Hemisphere and anticlockwise in the Southern. Warm and cold currents are defined by relative water temperature: sea-surface temperature along the path minus the mean for the same ocean at the same latitude. The Brazil and East Australian warm currents flow south, while the Peru and Benguela cold currents flow north. A warm current does not always flow north; the correct statement is that warm currents run from low to high latitude and cold currents from high to low.
  • The conveyor and upwelling: warm, salty surface water in the North Atlantic flows north, is cooled to a higher density at high latitude and sinks, flows south in the deep, circles Antarctica, rises in the North Pacific and returns to the surface; NOAA estimates that one parcel of water takes about 1,000 years to complete the circuit. When wind blows along a coast, Ekman transport pushes surface water away from the shore and deep water rises to replace it: an 8 m/s wind at 36.5°N moves 1.14 m²/s and lifts water at 3.95 m/day; nearer the equator the Coriolis parameter is smaller, so the same wind at 14.5°S moves 2.71 m²/s. What rises is cold, nutrient-rich water, not water created from nothing.

Model and sources

Model and sources

  • Density uses the UNESCO 1983 international equation of state, EOS-80 (a pure-water polynomial plus salt terms, with the secant bulk modulus for pressure). All four UNESCO check values are reproduced; it agrees with the GSW-C 3.05 seawater check values within 0.11 kg/m³ and with the Tanaka et al. pure-water density within 0.001 kg/m³. Mass fraction is converted to Absolute Salinity as SA = 1000 w g/kg and divided by 1.004715 to get practical salinity; a sodium chloride solution in a lab differs from sea salt, so this is a teaching approximation and the difference has not been checked. The 75-term TEOS-10 polynomial is not used: it is valid only inside the oceanographic funnel and is 0.3 kg/m³ off for pure water at 25 °C.
  • The weighing-station readings are simulated by the model, not measurements: the three repeats have preset meniscus offsets of 0, +0.4 and −0.4 half-divisions, and masses are rounded to the balance’s last digit. In the tank the blob speed is v = 0.25·√(|g′|·d), where g′ is reduced gravity and d = 2 cm; on reaching the bottom it spreads like a gravity current. These constants are teaching values.
  • Sea temperature and salinity on the globe are the 1° annual mean fields of the NOAA World Ocean Atlas 2018 (WOA18); salinity is practical salinity, and so is the 3000 m level. Land and missing cells are filled with the nearest valid value for colour, while readings use only valid cells. Current lines are drawn by hand from names and positions; warm and cold are computed from WOA18 sea-surface temperature, not entered by hand. The conveyor path is schematic: sinking sites, branches and depths are not drawn one by one. The Atlantic section is at 30.5°W.
  • Upwelling: wind stress τ = ρa·Cd·U² (ρa = 1.22 kg/m³, Cd = 1.3×10⁻³), Ekman mass transport M = τ ÷ |f|, volume transport Tv = M ÷ 1025, upwelling velocity w = Tv ÷ L with L = 25 km, and the rising water comes from 100 m depth; these constants are teaching values. Coastal surface temperature and nitrate come from WOA18 1° profiles (units µmol/kg): for California at 36.5°N and Peru at 14.5°S, the nearest coastal grid cell and a reference cell 10° of longitude to the west. The grid is coarse and does not represent the true upwelling core. The picture is stretched vertically, 1 unit = 60 m, while 1 unit horizontally = 7 km. Physically supplied nitrate is not productivity, and the page does not convert it into fish catch.
Back to the geography index ↑