SPOTLIGHT TOPIC

Beyond the instrumental record:
Paleoceanographic reconstructions from marine sediment cores

How do scientists know how the ocean and climate has varied in the past? Historical measurements of the ocean can take us back 70 years or so, but how do we find out how the ocean behaved in the longer-term past? The answer lies in paleoceanographic reconstructions; records of past ocean conditions preserved in the mud at the bottom of the ocean.

Oceanographers are obsessed with measuring all facets of the ocean, from hydrographic properties such as temperature and salinity, to more dynamic processes such as the strength and direction of different ocean currents. Oceanographers even measure how much whale poop is in the ocean; worth doing as this is a significant carbon sink. Their toolkit is expansive: enormous research vessels traversing the world’s oceans, high-tech moorings spanning ocean basins, and robotic underwater gliders reaching normally inaccessible depths. In fact, one major goal of EPOC is to design an observing system to measure components of the Atlantic Meridional Overturning Circulation (AMOC).

The limits of the observational record

Yet despite all this cutting-edge technology, oceanographic data collection is still constrained by a fundamental limitation: our instrumental records only span the last 30 to 70 years. This makes it difficult to contextualise individual observations; do they reflect a long-term trend? Or are they just natural variability? The AMOC has been monitored continuously since 2004 by the RAPID mooring array, spanning the Atlantic Ocean at 26°N. This record reveals a long-term decline in AMOC strength up until approximately 2016, followed by a modest recovery. But while these observations yielded new insights into the AMOC on short timescales, the extent to which these changes are part of a long-term AMOC trend is unclear. The record is just too short.

So what can we do to solve this conundrum (apart from sitting around for 100 years)? The answer is hidden at the bottom of the ocean — in mud, or to use the correct paleoceanographic terminology, marine sediments.

Paleoceanographic reconstructions

Just like oceanographers, paleoceanographers study all facets of the ocean — except they focus on the ocean in the past. Unable to directly observe or measure the ancient ocean, we paleoceanographers rely on proxy evidence preserved in marine sediments. For example, researchers estimate past ocean temperatures by measuring the ratio of magnesium to calcium preserved in foraminifera shells — tiny single-celled organisms that once inhabited the ocean before sinking to the seafloor.

Researchers also determine past deep-ocean current speeds by measuring sediment grain size. Larger grains indicate faster flow; finer grains indicate slower flow — a concept reminiscent of fast-flowing mountain streams littered with large rocks versus slow-flowing lowland riverbeds covered in fine-grained silts. By deciphering when these proxies were living or deposited, often using radiocarbon dating, paleoceanographers assign an age to each measurement, establishing the conditions of a particular ocean region at a precise point in the past.

Sediment cores: retrieving the record

But while individual measurements are useful, what paleoceanographers are especially interested in is longer records made up of multiple measurements, ideally spanning the observational era and extending further back in time. To achieve this, they rely on marine sediment cores, i.e., long cylinders of mud, retrieved from the bottom of the ocean. Researchers lower a plastic tube to the ocean floor from the side of a ship (not quite as high-tech as an autonomous glider!). A dropped weight or triggered mechanism then drives the tube into the underlying marine sediments before the crew hauls it back aboard.

The law of superposition states that in undisturbed sequences of sediments deposited in layers (such as those on the ocean floor), the youngest layers are at the top and the oldest layers are at the bottom. Paleoceanographers develop a long-term record of past ocean temperature or flow speed by measuring proxies at regular intervals throughout a sediment core. Dating the core also dates the proxy records, allowing researchers to infer past changes in ocean temperature and flow speed.

An imperfect science

The paleoceanographic reconstruction process sounds relatively straightforward. In practice, it comes with significant caveats.

Sediment cores are inherently messy (they are mud, after all!) and can be disturbed or ‘reworked’ by various factors, from burrowing worms to submarine landslides. The 1929 Grand Banks earthquake, for example, appears in numerous cores from around Newfoundland. Reworking leads to younger sediments underlying older ones and, in extreme cases, complete mixing of sedimentary layers. This complicates interpretation, but researchers can account for some disturbances through numerical modelling or by omitting sections where disturbance is evident.

Paleoceanographic proxies present their own challenges. Oxygen isotope ratios in foraminifera respond to temperature, but salinity also influences them, potentially obscuring the temperature signal. Researchers can account for this by reconstructing salinity independently using other proxies. It is also important to verify that a proxy record actually reflects the variable it is supposed to measure. The standard approach is to compare proxy data from the top of a sediment core with modern observations from the same location. In cores with very high sedimentation rates, researchers can go further and compare proxy records directly against longer instrumental records; an approach considered the gold standard.

EPOC paleoceanographic reconstructions

At UCL, we have employed these paleoceanographic techniques to reconstruct deep-ocean flow speeds at multiple sites along the flow path of the Deep Western Boundary Current. Our focus spans the last 1,000 years, with particular attention to the most recent 150 years, during which comparison with instrumental records becomes possible.

We are now working to interpret these records, assessing whether they can help contextualise the observational data of our non-paleoceanographic colleagues. Alongside this, we have compiled paleoceanographic proxy records from across the North Atlantic, and will evaluate their utility and robustness through a collaborative proxy-modelling investigation with colleagues at the University of Reading.

Last updated April 2026 | Written by Jack Wharton, Postdoctoral Research Fellow in Paleoceanography at University College London (UCL).