The Biological Pump: How the Ocean Buries Carbon

The reaction that starts it all: photosynthesis in the sunlit ocean

The pump begins where light and nutrients overlap — the euphotic zone, the top ~100 m of ocean where roughly 1% of surface sunlight still penetrates. Here phytoplankton run oxygenic photosynthesis, converting dissolved CO₂ into biomass. In its simplest balanced form:

• 6 CO₂ + 6 H₂O → C₆H₁₂O₆ + 6 O₂   (ΔG° ≈ +2870 kJ/mol, driven by photons)

But real plankton aren't made of glucose. Alfred Redfield showed in 1934 that marine organic matter has a remarkably fixed elemental recipe, the Redfield ratio of 106 C : 16 N : 1 P. Writing the full 'Redfield equation' with nitrate and phosphate as the nutrient sources gives a balanced net reaction:

• 106 CO₂ + 16 NO₃⁻ + HPO₄²⁻ + 122 H₂O + 18 H⁺ → C₁₀₆H₂₆₃O₁₁₀N₁₆P + 138 O₂

That single equation is the heart of the pump. It removes CO₂ from surface seawater (lowering the partial pressure pCO₂ and letting the sea inhale more from the air), it consumes nitrate and phosphate, and it releases O₂. The enzyme RuBisCO catalyses the CO₂-fixing step, using the Calvin cycle. Crucially, the reaction as written is reversible — run it backwards and you get respiration, which is exactly what happens on the way down.

Marine snow: how carbon actually leaves the surface

Fixing carbon isn't enough — most of it (>70%) is grazed and respired back to CO₂ within the euphotic zone in days. The pump only works for the fraction that sinks before it can be eaten. That fraction rides down as marine snow: aggregates of dead cells, mineral ballast, gelatinous 'transparent exopolymer particles,' and above all zooplankton fecal pellets, which can sink at 100–1000 m per day.

• Aggregation: sticky diatom blooms clump into millimetre-to-centimetre flakes that fall fast (Stokes' law: sinking speed ∝ radius²).
• Ballasting: dense mineral grains — CaCO₃ and opal (SiO₂·nH₂O) from shells — weight the aggregates, which is why ballasted particles reach the seafloor most reliably.
• Active transport: zooplankton that feed at the surface at night and descend by day (diel vertical migration) physically inject carbon 200–1000 m down and respire it there.

The amount crossing the 100 m 'export horizon' is the export production, ~5–11 Gt C/yr. Only a slice of that survives to real depth: about 2.7 ± 0.6 Gt C/yr is sequestered for centuries, and a mere ~0.2 Gt C/yr — roughly 0.4% of what was fixed — is finally buried in seafloor sediment. Over geological time, that tiny leak became every oil field, coal seam and limestone cliff on Earth.

Remineralization and the Martin curve: the carbon rains back into CO₂

As particles sink, bacteria and animals eat them and remineralize the organic carbon — running the Redfield reaction in reverse and consuming dissolved O₂:

• C₁₀₆H₂₆₃O₁₁₀N₁₆P + 138 O₂ → 106 CO₂ + 16 NO₃⁻ + HPO₄²⁻ + 122 H₂O + 18 H⁺

This is why the ocean's interior is rich in CO₂ and nutrients but poor in oxygen — the pump literally strips O₂ from the 'twilight zone' (200–1000 m) and creates the world's oxygen minimum zones. Where O₂ runs out, microbes switch electron acceptors: first denitrification (using NO₃⁻ → N₂), then sulfate reduction (SO₄²⁻ → H₂S). The remineralized CO₂ acidifies deep water; combined with pressure, this drives dissolution of falling CaCO₃ below the calcite compensation depth (~4–5 km).

The depth at which carbon is respired is everything. John Martin's 1987 sediment-trap data gave the famous Martin curve, a power law for flux F with depth z:

• F(z) = F₁₀₀ × (z / 100)⁻ᵇ, with b ≈ 0.86 on average

A larger b means faster shallow remineralization and a weaker pump; a smaller b means carbon reaches depth and stays down longer. Because deep water may not touch the atmosphere again for ~1000 years, whether a carbon atom is respired at 300 m or 1500 m can decide its fate for a millennium.

The carbonate counter-pump: why building shells releases CO₂

Not all sinking carbon lowers atmospheric CO₂. Calcifying plankton — coccolithophores, foraminifera and pteropods — build shells of calcium carbonate, and the precipitation reaction is counterintuitive:

• Ca²⁺ + 2 HCO₃⁻ → CaCO₃(s) + CO₂ + H₂O

Making a shell produces CO₂ at the surface and removes two units of alkalinity (the 2 HCO₃⁻) for every one carbon locked in mineral. This shifts the carbonate equilibrium CO₂ + CO₃²⁻ + H₂O ⇌ 2 HCO₃⁻ and actually raises surface pCO₂. That's the carbonate counter-pump: it exports carbon downward but pushes CO₂ back toward the atmosphere. Because the organic (soft-tissue) pump is roughly an order of magnitude larger at 100 m, the net effect is still a powerful CO₂ sink — but the counter-pump means the two biological pumps are, chemically, at war.

The rain ratio — the ratio of inorganic (CaCO₃) to organic carbon leaving the surface, typically ~0.05–0.25 — sets the balance. A more calcifying ocean pumps less net CO₂ down; a diatom-dominated (opal, no counter-pump) ocean pumps more. Ice-age climate swings may have partly ridden on shifts in this single ratio.

What controls the pump: nutrients, iron, and the limiting variables

The pump's strength is set by whatever runs out first. In most of the ocean that's nitrogen (as NO₃⁻) or phosphorus (as PO₄³⁻), delivered by upwelling and rivers. But across huge regions — the Southern Ocean, the equatorial and subarctic Pacific — nitrate and phosphate are abundant yet plankton stay sparse. These are the HNLC (High-Nutrient, Low-Chlorophyll) zones, and the missing ingredient is iron.

• Iron limitation: Fe is a cofactor in nitrate reductase and photosynthetic electron transport. In oxic seawater it exists as insoluble Fe³⁺ oxyhydroxides at vanishingly low concentration (~0.1–1 nmol/L), so plankton starve amid plenty. Most bioavailable iron arrives as windblown desert dust.
• Light and mixing: deep winter mixing drags cells below the euphotic zone; a shallow, stratified 'mixed layer' in spring triggers blooms.
• Temperature: warming stratifies the surface, throttling the upward nutrient supply — a key reason models project the pump may weaken as the ocean warms.

The iron leverage is enormous: because the Redfield-equivalent ratio to iron is roughly 106,000 C : 1 Fe, one atom of iron can, in principle, escort ~100,000 atoms of carbon into biomass. That arithmetic is what launched the idea of ocean iron fertilization as geoengineering.

Why it matters: climate, oxygen, and the deep-time carbon budget

The biological pump is a load-bearing wall of the planet's climate. By continually stripping CO₂ from the surface and stashing it at depth, it maintains a vertical carbon gradient that keeps atmospheric CO₂ roughly 150–200 ppm lower than a lifeless ocean would allow. That's the difference between the world we have and a much hotter one. Today atmospheric CO₂ is ~420 ppm and surface ocean pH has fallen ~0.1 units (from ≈8.2 to ≈8.1) since preindustrial times as the sea absorbs our emissions.

The pump also runs the ocean's oxygen budget and its fertility. Every nutrient upwelling that feeds a fishery is carbon that the pump previously carried down and remineralized. And over hundreds of millions of years, the ~0.2 Gt C/yr that escapes into sediment is the source rock of the entire fossil-fuel economy — the slow carbon cycle we are now reversing in a few centuries.

Climate change threatens the pump from several directions at once:

• Stratification reduces nutrient resupply, shrinking primary production.
• Warming speeds bacterial respiration, remineralizing carbon shallower (a bigger Martin b) so it returns to the air sooner.
• Acidification dissolves the CaCO₃ ballast and threatens pteropods and coccolithophores, altering the rain ratio.
• Deoxygenation expands oxygen minimum zones.

Interventions being tested — iron fertilization, macroalgae sinking, and ocean alkalinity enhancement (adding crushed silicate/carbonate to boost the carbonate buffer) — all try to lean on the same chemistry the plankton have run for two billion years. The catch, as sediment traps keep showing, is that most fertilized carbon is respired long before it reaches the millennium-safe deep, making verified, durable sequestration far harder than the back-of-envelope 106,000:1 ratio suggests.

The two biological pumps run in opposite directions on atmospheric CO₂. The soft-tissue (organic) pump pulls CO₂ down; the carbonate pump partly pushes it back — the 'counter-pump.'

Property	Soft-tissue (organic) pump	Carbonate pump
What sinks	Organic matter C₁₀₆H₂₆₃O₁₁₀N₁₆P	CaCO₃ shells (calcite/aragonite)
Key organisms	Diatoms, cyanobacteria, all phytoplankton	Coccolithophores, foraminifera, pteropods
Effect on surface CO₂	Lowers it (draws CO₂ down)	Raises it (releases CO₂)
Effect on alkalinity	Small	Removes 2 mol alkalinity per mol C
Export at 100 m	~5–11 Gt C / yr	~0.5–1.5 Gt C / yr (≈10× smaller)
Net climate role	Cools (sequesters carbon)	Warms slightly (counter-pump)