Oxygen concentrations in Europe's coastal and marine waters

The occurrence of reduced oxygen levels in near-seafloor waters is increasing, owing mainly to a combination of natural causes and human-induced pressures, including excess nutrient inputs and climate change. Around 18% of assessed areas reveal reduced concentrations (<6mg/l), below the level needed to support marine life with minimal stress. The Baltic and Black seas are most affected, with over 33% of assessed areas falling below this level. Oxygen depletion can severely impact marine life and disrupt ecosystems, leading to significant environmental and socio-economic consequences.

Figure 1. Occurrence of reduced oxygen concentrations in Europe's coastal and marine waters (average for the years 2012-2023)

Please select a resource that has a preview image available.

Excessive nutrient inputs into marine ecosystems from agriculture runoff, wastewater and industrial discharges can lead to harmful algal blooms. As these blooms die and decompose, they consume oxygen. This can be further aggravated by ocean warming, which:

decreases the solubility/content of oxygen in seawater;

increases the metabolic oxygen demand of most marine organisms;

intensifies stratification of the water column, reducing oxygen exchange with deeper waters.

Ocean deoxygenation is considered one of the three major stressors, along with warming and ocean acidification, and are collectively referred to as the ‘deadly trio'. These stressors can severely impact marine life and disrupt ecosystems, leading to significant environmental and socio-economic consequences. This includes biodiversity loss and changes in the food-web, and changes in the abundance and distribution of fisheries resources.

Reduced oxygen concentrations indicate the indirect effects of nutrient enrichment and, consequently, eutrophication. It is key for assessing progress towards improved water quality in line with EU policy objectives, such as the Water Framework and the Marine Strategy Framework directives.

These directives aim to achieve ‘good ecological status’ and ‘good environmental status’ of Europe’s waters, respectively. The European Green Deal supports these by introducing ambitious targets for reducing nutrient use in agriculture and losses to the environment, outlined in key policies including the EU Biodiversity 2030 and the Farm-to-Fork strategies, and Zero Pollution Action Plan.

Oxygen concentrations monitored from July to October were considered as this period is most likely to experience low oxygen levels due to higher water temperatures. Results show that large parts of the Baltic and Black seas suffer from severe oxygen depletion. During 2012-2023, 14% of the assessed spatial area in the Baltic Sea and 24% in the Black Sea recorded near critical concentration conditions, below 4mg/l (Figure 1). These areas typically occur in deeper, denser water layers where oxygen inflow or its downward movement is irregular or limited.

Reduced oxygen levels in the Mediterranean Sea are more localised, mainly occurring in coastal areas. Around 24% of the assessed area in this region experiences reduced concentrations (<6mg/l). While the North-East Atlantic region experiences some localised and short periods of oxygen deficiency, most assessed areas show concentrations above the 6mg/l threshold value.

Figure 2. Trends in oxygen (O₂) concentrations in the near-bottom layer and number of trends, by concentration group (2000-2023)

Please select a resource that has a preview image available.

Analysing trends is crucial for understanding whether the state of Europe's seas is improving or declining. Figure 2 shows stations with average oxygen concentrations in three classes: (1) below 4mg/l (includes <2mg/l class); (2) between 4 and 6mg/l; and (3) above 6mg/l. For all marine regions combined, over the period 2000-2023, no significant change is found for 91% of the stations assessed. However, 7% show a decreasing trend.

Stations in the Baltic Sea and some Danish fjords facing severe oxygen depletion (<4mg/l) see conditions worsening in 4% of the cases and some improvement in 3%. Most stations with concentrations between 4-6mg/l were also found in these areas, with 4% showing improving trends and 4% showing decreases.

North-East Atlantic Ocean stations with low oxygen concentrations (<6mg/l) see conditions improving in two locations, while the rest show no significant change. Baltic Sea stations with levels above 6mg/l show 6% worsening and 2% improving, while in the North-East Atlantic Ocean, 12% of such are deteriorating and 2% improving.

Concerted action against nutrient pollution is needed to combat oxygen deficiency, alongside efforts to mitigate ocean warming. Improved data collection and monitoring across all regional seas are required to better understand the impacts of these measures and the broader effects of climate change.

Supporting information

Definition

This indicator displays the geographical distribution and trends in summer-autumn oxygen concentrations, measured in milligrams per litre (mg/l), in near-bottom waters of Europe’s regional seas.

Threshold values (TVs) for dissolved oxygen concentrations in coastal waters are set under the Water Framework Directive (WFD). The Marine Strategy Framework Directive (MSFD) aligns its TVs for coastal waters with those set under the WFD and extends these beyond coastal waters to ensure consistency. Member States establish TVs through (sub)regional cooperation.

Methodology

Data on oxygen concentrations during the summer-autumn months (July-October) are used as this period has the highest probability of oxygen depletion due to higher water temperatures.

For each monitoring site, the mean of the 5-percentile of observations for the years 2012-2023 was used. Results are aggregated at the level of 100x100km grid cells. For each marine region, the 5-percentile of oxygen concentrations in a grid cell is used to classify the grid cells by the oxygen concentration classes: O₂ < 2mg/l, 2 ≤ O₂ < 4mg/l, 4 ≥ O₂ ≤ 6mg/l, and > 6mg/l.

The main sources of data include:

the International Council for the Exploration of the Sea (ICES);
the European Marine Observation and Data Network (EMODnet) data sets; and
WISE SoE – Water Quality (WISE-6).

Data maintained by ICES are collected through the Eionet Central Data Repository (Eionet CDR) from the marine conventions and represent a sub-set of national data compiled to provide comparable indicators of the condition and impacts on transitional, coastal and marine waters (TCM data) across Europe. When data from both ICES and EMODnet are available for the same station (defined by position and time), ICES data are used.

The procedures of data extraction, data selection and aggregation, trend analysis and the plotting of results are carried out using the R programming language.

Geographical classification: sea region, coastal or offshore and station

Stations are geographically defined by their longitude and latitude in decimal degrees. All geographical positions in the data are assigned to marine (sub)regions based on their coordinates.

The primary aggregation process involves:

identifying (clusters of) stations and assigning them to marine (sub)regions;
creating statistical estimates for each combination of station and year.

Statistical aggregation per station and year includes:

selecting the season (months July-October);
selecting the sample depth (0-20m above the sea floor for depths less than 100m; 0-50m above the sea floor for depths greater than 100m);
selecting data for each station and year that fall within the lower quartile (≤ 25th percentile);
calculating the mean of the data selected per station and year.

Trend analysis

Trend analysis is conducted for each station in regions for which there were at least data for five or more years in the period 2000-2023. Trends are detected using the non-parametric Mann-Kendall trend test. The tests reported here are two-sided (testing for both positive and negative trends). Data series with p-values of <0.05 are reported as significantly positive or negative. The test analyses only the direction and significance of the change, not the magnitude of the change. It assesses whether there is a monotonic upward or downward trend over time without assuming any specific distribution for the data. This makes it useful for environmental data analysis, such as detecting trends in oxygen concentrations.

Methodology for gap filling

Not applicable

Policy/environmental relevance

Analysing oxygen concentrations and their changes over time is key to assessing progress towards improved marine and coastal water quality in line with EU policy objectives, such as those under the Marine Strategy Framework Directive (MSFD) , and the Water Framework Directive (WFD). The WFD mandates the achievement of good ecological status or the good ecological potential of transitional and coastal waters across the EU, and specifically identifies dissolved oxygen concentrations as one of the physio-chemical parameters for assessing ecological status. The MSFD requires the achievement or maintenance of good environmental status in EU sea basins and designates dissolved oxygen concentration in the bottom of the water column as one of the primary criteria (D5C5) for Descriptor 5 human-induced eutrophication.

Other EU directives are also related to the control of eutrophication by aiming to reduce the loads and impacts of the nutrients, and include: the Urban Waste Water Treatment Directive aimed at reducing pollution from sewage treatment works and certain industries; the Nitrates Directive aimed at reducing nitrate pollution from agricultural sources and the Integrated Pollution Prevention and Control Directive aimed at controlling and preventing the pollution of water from industry. In addition, the EU Biodiversity Strategy 2030, Farm to Fork Strategy and Zero Pollution Action Plan are main policies under the EU Green Deal setting ambitious targets for reducing nutrients from agriculture.

EU policies and legislation also support the implementation of the Regional Seas Conventions and Action Plans (RSCAPs) — the Oslo Paris Convention (OSPAR), the Helsinki Convention (HELCOM), the Barcelona Convention (UNEP-MAP) and the Bucharest Convention, which also outline measures that aim to reduce the loads and impacts of nutrients.

Excessive nutrient flow into the sea, primarily from agricultural fertilizers, can trigger large phytoplankton blooms, increasing primary production—a process known as eutrophication. When these organisms die and sink to the seafloor, oxygen is consumed during decomposition. If the water column cannot mix adequately to replenish the supply of oxygen at the seafloor, this can lead to significantly reduced oxygen levels that limit biological activity (hypoxia) and may even result in complete oxygen depletion (anoxia).

Oxygen-depleted areas demonstrate how one type of anthropogenic pressure (eutrophication) is intensified by climate change effects, such as rising water temperature. Increased water temperatures affect various biological and chemical processes in the marine environment. For example, warmer water decreases oxygen solubility, reducing oxygen concentrations, while simultaneously increasing organisms' metabolic demand for oxygen. Most marine organisms rely on oxygen for metabolism, so lower oxygen levels can adversely affect their physiology, species composition and abundance. Insufficient oxygen supply can lead to broader ecological and economic impacts, affecting productivity, species interactions and community composition at the ecosystem level .

Reduced oxygen concentrations serve as an indicator of the indirect effects of nutrient enrichment and, consequently, eutrophication. It is part of a set of indicators assessing the state and pressures on Europe's seas. It is closely linked to the other two EEA marine indicators on eutrophication: 'Nutrients in transitional, coastal and marine waters’ and 'Chlorophyll in transitional, coastal and marine waters’. Oxygen depletion can also interact with other anthropogenic stressors, such as climate change, overfishing and the introduction of invasive alien species, further impacting marine ecosystems and fisheries. Thus, it is also associated with indicators of: Ocean acidification, Marine non-indigenous species in Europe's seas, Status of marine fish and shellfish stocks in Europe's seas and Changes in fish distribution in Europe's seas.

Accuracy and uncertainties

Methodology uncertainty

Recent discussions in OSPAR have suggested that the 25th percentile, used previously in this assessment, is not precautionary enough and the 5th percentile should be used, and is now used in this update. In HELCOM, minimum concentrations are assessed.

Additionally, OSPAR and HELCOM use a sample depth of 10m above the seafloor. However, due to the uncertainty in sample depth data, 20m was used in this assessment. These uncertainties will be further explored in future updates of the indicator.

Geographical comparability

Data for this assessment are still limited considering the large spatial and temporal variations inherent in transitional, coastal and marine waters surrounding Europe. This lack of data means that long stretches of coastal and marine waters remain uncovered by the analysis. Most of the available time series data are concentrated in the Greater North and Baltic seas, particularly in the Kattegat, and the Dutch, German and Danish parts of the North Sea, as well as the central and western parts of the Baltic Sea. In the other regions, longer time series data are limited.

For the analysis, only data for the lower 20m or 50m of the water column are considered. However, not all available data have reliable attributes on sampling depth and bathymetry. In shallower waters, selecting data from the lower 20m of the water column may not be optimal as vertical mixing tends to reduce oxygen depletion.

Stations are defined geographically based on longitude and latitude in decimal degrees; however, the datasets do not always contain reliable and consistent station identifiers. Coordinates for supposed identical stations may vary between visits due to actual positioning versus target positioning, which can fragment time series. To improve data aggregation into time series, data are grouped into squares of approximately 1.375km for stations within 20km of the coastline and 5.5km for open-water stations beyond 20km. Although this method reduces errors in data aggregation and fragmentation of time series due to minor positional shifts, assessing broader assessment areas might be more effective than the current station-based approach.

Comparability over time

The Mann-Kendall test is a robust and accepted approach, however, due to the multiple trend analyses, with a probability level of p <0.05, approximately 5% of the tests conducted will appear significant (identify a trend) even if in fact there is no trend.

To better understand trends in dissolved oxygen in relation to climate change, more data are required than currently available. This would require expanding the monitoring network to all regions to ensure long-term changes are captured.

Data sources and providers

Metadata

DPSIRTopicsTagsTemporal coverageGeographic coverage

TypologyUN SDGsUnit of measureFrequency of dissemination

References and footnotes

<div class="csl-bib-body" style="line-height: 1.35; "> <div class="csl-entry">EU, 2008, Directive 2008/56/EC of the European Parliament and of the Council of 17 June 2008 establishing a framework for community action in the field of marine environmental policy (Marine Strategy Framework Directive), OJ L 164, 25.6.2008, p. 19-40.</div> </div>
^↵
<div class="csl-bib-body" style="line-height: 1.35; "> <div class="csl-entry">EU, 2017, Commission Decision (EU) 2017/848 of 17 May 2017 laying down criteria and methodological standards on good environmental status of marine waters and specifications and standardised methods for monitoring and assessment, and repealing Decision 2010/477/EU, OJ L 125, 18.5.2017, p. 43-74.</div> </div>
^↵
<div class="csl-bib-body" style="line-height: 1.35; "> <div class="csl-entry">Díaz, R., 2001, 'Overview of hypoxia around the world.', <i>Journal of environmental quality</i>.</div> </div>
^↵
<div class="csl-bib-body" style="line-height: 1.35; "> <div class="csl-entry">Diaz, R. J. and Rosenberg, R., 2008, 'Spreading Dead Zones and Consequences for Marine Ecosystems', ( <a key=link-1 href=https://www.science.org/doi/10.1126/science.1156401?cookieSet=1 rel="noopener"> https://www.science.org/doi/10.1126/science.1156401?cookieSet=1 </a>) accessed July 20, 2022.</div> </div>
^↵
<div class="csl-bib-body" style="line-height: 1.35; "> <div class="csl-entry">Vaquer-Sunyer, R. and Duarte, C. M., 2008, 'Thresholds of hypoxia for marine biodiversity', <i>The Proceedings of the National Academy of Sciences (PNAS)</i> 105(40) ( <a key=link-1 href=https://www.pnas.org/doi/full/10.1073/pnas.0803833105 rel="noopener"> https://www.pnas.org/doi/full/10.1073/pnas.0803833105 </a>).</div> </div>
^↵