Every summer, a vast stretch of the northeast Atlantic quietly holds its breath. Across the Celtic Sea, the wide temperate shelf off the coasts of Britain and Ireland, the water column splits into two worlds: a warm, sunlit surface layer and a cold, dark pool of water hugging the seabed. Between them sits the seasonal thermocline, a thin veil of density that all but severs the deep water from the atmosphere above. Cut off from the air and from the oxygen-producing surface ocean, the bottom waters begin a slow, months-long decline in dissolved oxygen that lasts from spring until the autumn storms finally stir the sea back together. A new study, published in the journal Biogeosciences, has now produced the most complete accounting yet of where that oxygen goes — and the results carry a stark warning about what a warming ocean will do to the seafloor.
The research, led by Xin Meng of the University of Liverpool together with Claire Mahaffey, Juliane Wihsgott and Jonathan Sharples, draws on an unusually rich observational record. Between March 2014 and August 2015, as part of the UK Shelf Sea Biogeochemistry programme, the team repeatedly crossed the Celtic Sea aboard research vessels, lowering conductivity-temperature-depth instruments through the water column at dozens of stations. Oxygen sensors mounted on the instruments were painstakingly calibrated against hundreds of bottle samples analysed by Winkler titration, the classical chemical method for measuring dissolved oxygen, so that the final profiles were accurate to within a few micromoles per litre. One station in the central Celtic Sea was visited on every voyage, giving the researchers a time series detailed enough to build a full oxygen budget for the bottom water — a ledger of every process adding or removing oxygen over the stratified season.
The seasonal cycle they documented is strikingly regular. In winter, the water column is fully mixed and oxygen sits at saturation, roughly 282 millimoles per cubic metre at around 10 degrees Celsius. When stratification sets in during April, the decline begins almost immediately. By late summer the bottom water has lost tens of millimoles per cubic metre, with the steepest losses in the shallower northern shelf, where a thinner pool of deep water offers less oxygen to be consumed. The lowest concentrations, below 220 millimoles per cubic metre, appear in the north, while deeper southern waters near the shelf edge hold on to more oxygen, partly because vigorous internal tides there keep stirring oxygen downward. Only when autumn winds deepen the surface layer and the sea remixes in December do concentrations recover above 280 millimoles per cubic metre.
To understand what drives the decline, the team constructed a budget with four main terms: oxygen consumed by microbes respiring organic matter in the water column, oxygen taken up by seafloor sediments, oxygen carried in or out by horizontal currents, and oxygen supplied by turbulent mixing across the thermocline from the subsurface chlorophyll maximum. This last term is the surprise hero of the story. The subsurface chlorophyll maximum is a persistent layer of phytoplankton living right at the base of the thermocline, where nutrients diffuse up from below and light is still sufficient for photosynthesis. The oxygen these microscopic algae produce accumulates in the pycnocline because weak mixing traps it there, creating a hidden reservoir of oxygen directly above the depleted bottom water.
At the central Celtic Sea site, the bottom water lost oxygen at a net rate of 44 plus or minus 4 millimoles per square metre per day during the 2015 stratified period. Respiration and remineralisation of organic matter dominated the ledger, consuming 54 plus or minus 19 millimoles per square metre per day. Turbulent fluxes from the subsurface chlorophyll maximum supplied roughly 30 plus or minus 18 millimoles per square metre per day, offsetting more than half of the biological demand. Benthic oxygen demand, estimated at about 7 millimoles per square metre per day from sediment incubations, and horizontal transports, contributing a further 6 to 7 millimoles per square metre per day each, played only minor roles, together accounting for perhaps 10 to 15 percent of the total decline.
The budget also revealed how dramatically episodic weather can tip the balance. When winds exceed about 12 metres per second for at least 12 hours, inertial motions of the surface layer generate shear across the thermocline, multiplying turbulent diffusivity by a factor of roughly 20. Drawing on wind records from the ERA5 reanalysis, the researchers found that such events occurred on about eight days in each of the two study summers, and that including them nearly doubled the average downward oxygen flux. Seafloor topography matters too: banks and slopes generate breaking internal waves that locally boost mixing, and the team identified a persistent patch of elevated bottom-water oxygen near their central station that they attribute to such topographically driven mixing. Over the longer view, an analysis of wind events from 1960 to 2020 showed that interannual variability alone could swing late-summer bottom-water oxygen concentrations by roughly plus 41 to minus 27 millimoles per cubic metre.
The most consequential part of the study, however, looks forward. Using a 2 degree Celsius warming scenario for the region, consistent with business-as-usual projections for the northwest European shelf by 2100, the researchers calculated how each budget term might change. Warmer water simply holds less oxygen: a 2 degree rise in the spring surface temperature that sets the starting concentration for the whole season would lower bottom-water oxygen by about 12 millimoles per cubic metre by October. An earlier onset of stratification, projected at about five days, adds only a further 2 millimoles per cubic metre. But the respiration term is where the numbers become alarming. Microbial metabolism is strongly temperature dependent, and applying standard Q10 relationships — a factor of about 2 for bacterial respiration and up to 5 for mixed natural communities — suggests that a 2 degree warming could accelerate oxygen consumption by 15 to 38 percent, potentially driving an additional decline of 21 to 51 millimoles per cubic metre, provided enough labile organic matter is available to fuel it.
That caveat matters. Faster metabolism needs fuel, and a further 51 millimoles per cubic metre of oxygen loss would require roughly 35 millimoles per cubic metre of organic carbon in the bottom water. Measurements from the region suggest that much of the dissolved organic carbon present in late autumn may be refractory, meaning microbes cannot easily digest it. Whether future primary production will supply more or less organic matter to the deep water is itself uncertain, because warming, changing nutrient supplies and shifting wind patterns all interact in complicated ways: stronger winds could fertilise the subsurface chlorophyll maximum with nutrients, boosting both oxygen production and the organic rain that feeds deep-water respiration, while simultaneously deepening the thermocline and dimming the light that phytoplankton there depend on.
Combining the solubility and respiration effects, the authors estimate that in a 2 degree warmer ocean, autumn bottom-water oxygen in the central Celtic Sea could fall to around 174 millimoles per cubic metre — below the oxygen deficiency threshold of 192 millimoles per cubic metre used for UK and Irish waters. In the shallower northern shelf, where the same dynamics play out in a thinner bottom layer, concentrations would likely drop even further before autumn mixing arrives. Even modest oxygen loss below about 190 micromolar can stress fish and other high-demand organisms, and sustained depletion degrades habitats, reduces biodiversity and, in extreme cases, produces hypoxic dead zones. The findings also sharpen a scientific debate: while existing model projections for the shelf attribute most future oxygen decline to reduced solubility, this budget suggests that the temperature sensitivity of microbial respiration could be the dominant driver — a process that remains poorly constrained in many Earth system models.
What emerges is a picture of a shelf sea living on a knife edge each summer, its deep-water oxygen sustained only by the delicate interplay of microbial appetite, hidden photosynthesis and the occasional storm. The study underscores two priorities for future research: better measurements of how respiration rates respond to warming in real shelf-sea communities, and a clearer understanding of how storm tracks and wind-driven mixing over the northwest European shelf will change in a warming climate. For the creatures that live on and above the Celtic Sea floor, and for the fisheries that depend on them, those two questions may determine how breathable their summers remain.
Subject of Research: Summer dissolved oxygen dynamics and climate-driven depletion in the bottom waters of the Celtic Sea
Article Title: Summer oxygen dynamics in the bottom waters of a wide, temperate shelf sea
Article References: Meng, X., Mahaffey, C., Wihsgott, J., & Sharples, J. (2026). Summer oxygen dynamics in the bottom waters of a wide, temperate shelf sea. Biogeosciences, 23(19), 6915-6929. https://doi.org/10.5194/bg-23-6915-2026
Image Credits: AI Generated
Keywords: Celtic Sea, dissolved oxygen, shelf sea, thermal stratification, microbial respiration, subsurface chlorophyll maximum, diapycnal mixing, ocean deoxygenation, hypoxia, climate warming, oxygen solubility, biogeochemistry
News Source: Violet Maxwell. (October 8, 2026). Microbes, not just warmth, may drain oxygen from a warming Celtic Sea. Scienmag.
