- Since the 1960s, the global oceanic oxygen content has declined by more than 2%.
- Oxygen deficiency has been detected in the North Sea and Celtic Sea. Observations are more frequent in the North Sea and have detected more frequent and intense seasonal oxygen depletion but the extent of oxygen depletion elsewhere in UK shelf seas is unknown.
- Ocean warming alongside increased vertical stratification are driving increased oxygen depletion, with seasonally stratified regions more prone to oxygen deficiency than permanently mixed regions.
Low evidence, medium consensus
On a global scale, there is a high level of confidence that the oceans are losing oxygen due to ocean warming. In UK coastal waters, there is a high level of understanding of the seasonal and interannual variability in oxygen dynamics in the permanently mixed and seasonally stratifying waters in the North Sea due to the relatively extensive sampling regime for dissolved oxygen concentrations in this region over the past four to five decades. Repeat sampling at specific sites has provided insight into the occurrence and onset of oxygen deficiency but the spatial extent of oxygen deficiency outside of these specific regions within the North Sea is uncertain due to the paucity of direct observations.
Approximately one third of the historical seasonally focused depletion of oxygen in the North Sea has been attributed to warming but the remaining two thirds are thought to be due to enhanced oxygen consumption. The relative importance of the processes that drive enhanced oxygen consumption (e.g. more organic matter, decreased ventilation of bottom waters), however, remain poorly understood. Finally, whereas the North Sea is well sampled in time and space, the rest of the Northwest European Shelf waters, especially the Celtic and Irish seas, are relatively poorly sampled and offer low levels of confidence in the occurrence or risk of oxygen depletion. Nevertheless, both observations and models agree that the UK coastal and shelf seas are losing oxygen and thus there is a medium level of confidence on the direction of change.
- Average annual mean dissolved oxygen concentration in the global ocean is projected to decline by 1.5 to 4% by 2090 for all RCP scenarios (equivalent to a warming of 1 to 4 ºC). This decline will be most acute below the thermocline away from the surface.
- For UK shelf waters, models project that oxygen depletion will intensify and last longer in the central North Sea and Celtic Sea, regions currently experiencing oxygen depletion. The mean oxygen concentration is projected to decline mostly strongly (3 to 4% by 2100) in the North Sea and Western English Channel. Deeper regions exposed to exchange with the open ocean (the Irish shelf and Shetland shelf) are expected to be less affected, decreasing by ~ 2 %.
- The predicted increase in temperature over this century for UK shelf seas will lead to a decrease in dissolved oxygen through the whole water column because of reduced solubility.
- The risk of oxygen deficiency in summer will increase because of lower oxygen levels experienced during the preceding winter and spring, as well as transport of low oxygen waters from estuaries and potentially adjacent seas.
- Increase in the frequency and intensity of marine heatwaves may exacerbate oxygen loss, driving both thermal stress to marine life alongside low oxygen availability.
- Increased rainfall and runoff would increase the risk of eutrophication and cause oxygen concentrations to locally decrease.
- Continued warming and reduced oxygen availability will affect the metabolism, health, and reproduction of many marine organisms, which could have major impacts on ecosystems and commercial fisheries.
Medium evidence, medium consensus
At the global scale, there is a high level of confidence that an increase in temperature will continue to reduce the solubility of oxygen and enhance stratification and thus lead to the ongoing decline in dissolved oxygen concentrations, especially below the thermocline. On a regional scale appropriate for coastal and shelf seas, there is a consensus that the ocean will lose oxygen. Model simulations can provide estimates of the magnitude and causes of the decline in dissolved oxygen but there is still uncertainty in how well they represent the coupling between physical and biogeochemical processes, biological processes specifically, strong seasonality in nutrient supply in a shallow water column and interaction with the sediment. Therefore, there is a medium level of confidence on the future of dissolved oxygen dynamics on a regional scale relevant to UK marine waters.
- We need to be able to determine the mechanisms driving spatial and temporal trends in dissolved oxygen and confidently identify when and where changes in dissolved oxygen are being driven by human-induced activity such as ocean warming or nutrient enrichment relative to background natural variability.
- Assessing the occurrence, frequency and spatial extent of oxygen deficiency in UK coastal and shelf waters is hampered by the lack of long-term data in regions outside of the North Sea. The poor resolution of dissolved oxygen data also hampers the ability to confidently test coastal and shelf-sea models. An integrated observing system providing high resolution, continuous time-series data using new technologies such as autonomous ocean gliders or instrumented moorings would provide the means to improve detection of oxygen depletion in the future. Recent and current programmes such as the NERC-Defra Shelf Sea Biogeochemistry programme (Kröger et al., 2018) NERC-Defra WWF AlterEco project (https://altereco.ac.uk) and EU H2020 AtlantOS programme (https://atlantos-h2020.eu) are providing emerging insight into best practices on how to operate autonomous ocean gliders to study dissolved oxygen dynamics in UK marine waters.
- There is still uncertainty surrounding the ability of models to simulate the individual processes and coupling between processes that control dissolved oxygen dynamics. To accurately predict dissolved oxygen, models need to simulate each contributing process correctly, in isolation but also coupled to other processes. This is an enormous challenge for ocean models since it is not possible to include all physical, chemical and biological processes in any model. Instead, complex processes must be parameterised to produce net effects that are close to that observed, but that may have differing levels of success dependent on local conditions. We do not yet fully understand all processes contributing to the decline in oxygen in the marine environment and thus representing these processes in models is challenging. The lack of understanding is particularly acute within coastal and shelf sea sediments. The lack of long-term time series data for testing coupled physical-ecosystem models, or the variability in functioning between sites with different conditions is also problematic.