Indicator Assessment

Oxygen concentrations in European coastal and marine waters

Indicator Assessment
Prod-ID: IND-476-en
  Also known as: MAR 012
Published 15 Nov 2019 Last modified 18 Nov 2021
21 min read
This page was archived on 18 Nov 2021 with reason: No more updates will be done
  • Widespread oxygen depletion, partly due to natural conditions (stratification), occurs in the Baltic Sea and the Black Sea.
  • In the Baltic Sea, oxygen concentrations in the water layer near the sea floor decreased during the period 1990-2017 at 11 % of stations, mainly in the Bothnian Bay, the Bothnian Sea, the Gulf of Finland and the Baltic Proper, and in some parts of the south-western Baltic Sea.
  • In the Greater North Sea area, decreases in oxygen concentrations during the period 1990-2017 were observed at 9 % of stations, mainly in fjords in Denmark and along the Norwegian and Swedish Coasts and at some stations in the German Bight.
  • Limited data were available for the Celtic Seas and the Adriatic Sea.
  • Reduced oxygen concentrations were observed at some stations in the coastal waters of the Black Sea, but there were no significant trends in oxygen concentrations during the period 1990-2017.
  • No significant trends in concentrations were observed for the majority of stations in all regions during the period 1990-2017.
  • Data coverage is not sufficient in all regional seas; it is sufficient for the Baltic and the North Seas, while data for only coastal waters are available for the Adriatic and Black Seas.  

Summer/autumn oxygen concentrations in water column near the seafloor, 2013-2017

Note: The map shows dissolved oxygen concentrations in the water column near the seafloor, observed in summer/autumn of the years 2013-2017. Purple: <2 mg/l; Light purple: 2-4 mg/l; Light green: 4-6 mg/l; Green: >6 mg/l

Data source:

Percentage of stations in each marine (sub)region with significant negative, positive or no trends in summer/autumn oxygen concentrations in the water column near the sea floor, 1990-2017

Data sources:
Data sources:

Baltic Sea

Because of the large input of fresh water and the limited inflow of saline water from the North Sea, which is controlled by the North Atlantic Oscillation, the Baltic Sea is permanently stratified. These conditions favour the development of hypoxia. Hypoxia in the Baltic Sea is a perpetual phenomenon, in particular in the deeper basins as a result of the permanent stratification and very long renewal times. However, its occurrence has increased since 1950 as a result of enhanced nutrient inputs (Conley et al., 2009Helcom, 2018), and since the 1990s an increase in spatial extent has also been observed (Conley et al., 2009). Carstensen et al. (2014) have shown that the increase in hypoxia in the Baltic Sea is strongly linked to increased nutrient inputs. HELCOM developed ‘oxygen debth’ as an indicator for the indirect effects of eutrophication on oxygen levels in the deeper waters of the Baltic Sea (Helcom, 2018). In addition to eutrophication, hypoxia has been increasing because of higher respiration rates due to increased temperatures. Life in deep water habitats is highly dependent on the aeration provided by inflows of marine water from the North Sea. Since 2013, a series of inflow events has improved oxygen conditions in the deep waters of the Baltic Sea (Helcom, 2018; Naumann et al., 2018).

The results in Fig. 1 show that there were large areas with low oxygen concentrations (< 2 mg/l) in the water layer near the seafloor of the Baltic Sea in the period 2013-2017. Hypoxic conditions are predominantly found in the Gulf of Finland and the Baltic Proper.

The analysis shows negative trends in oxygen concentrations during the period 1990-2017 in parts of the Baltic Sea, such as the Bothnian Bay and the Bothnian Sea, the Gulf of Finland and the Baltic Proper, and Kiel Bay/Great Belt. Of a total of 718 stations with sufficient data for trend analysis, 11 % showed a negative trend towards lower oxygen concentrations during the period 1990-2017. About two thirds of these stations have time series of more than 10 years. At 1 % of the stations, a positive trend towards higher concentrations was observed.

Greater North Sea

In the Greater North Sea, hypoxia generally does not occur, but several areas have a higher risk of hypoxia due to stratification, in particular the Oyster Grounds and Skagerrak (Topcu and Brockmann, 2015). OSPAR developed a common indicator for oxygen that also showed localised areas of oxygen deficiency, particularly in the eastern part of the southern North Sea (OSPAR, 2017).

The data for 2013-2017 show that hypoxic (< 2 mg/l) or reduced (< 4 mg/l) oxygen concentrations were observed in the water layer near the sea floor at a few stations in fjords along the Norwegian and Swedish coasts and in fjords in Denmark. In addition, there were a few observations of reduced oxygen concentrations (< 4 mg/l) in the North Sea near the Oyster Grounds.

Trends towards lowering oxygen concentrations during the period 1990-2017 were observed at several stations. Of a total of 301 stations, 9 % showed a decrease in oxygen concentrations, and nearly half of those cases were based on a time series longer then 10-years. Only 3 % of the stations showed an increase on oxygen concentrations. Decreasing concentrations were mainly observed in Danish fjords and at some stations in the German Bight. In Skagerrak/Kattegat, both decreases and increases were observed, which is probably related to variability in the exchange of water between the North Sea and the Baltic Sea.

Atlantic waters: Celtic Seas, Bay of Biscay and the Iberian coast

In the other parts of the North-East Atlantic, data were available for 2013-2017 for only the Celtic Seas, with only a few observations of reduced (< 4 mg/l) oxygen concentrations in the water column near the sea floor along the Irish coast.

Only shorter time series (of 5 years) were available for three stations, which showed no significant trends.

Mediterranean Sea

In the Mediterranean Sea, data were available for 2013-2017 for only the Adriatic Sea. In a few cases, oxygen concentrations of < 4 mg/l were observed.

Out of 23 stations with available time series, one showed an increase and one a decrease in oxygen concentrations. Both of these stations were found along the Croatian coast.

Black Sea

The Black Sea has the largest volume worldwide of naturally permanent anoxic waters (Friedrich et al., 2014). Seawater exchange is limited through exchange with the Mediterranean though the Bosporus strait, and freshwater inputs from the Danube, Dnieper and Dniester rivers result in strong thermohaline stratification which limits vertical transport and the input of oxygen into deep waters. Persistent anoxia below the pycnocline occurs in the central part of the Black Sea Basin as a result of those (natural) hydrodynamic characteristics. The north-western shelf experiences seasonal bottom-water hypoxia as a consequence of both strong thermohaline stratification in the summer and eutrophication due to nutrient inputs from the Danube river (Capet et al., 2013).

The data for 2013-2017 do not include data from the central part of the Black Sea. Only some data for coastal waters were available, showing stations with reduced oxygen concentrations and in some cases hypoxia (< 2 mg/l) at several points, with the highest concentration near the Bosporus strait.

Time series were available for stations along the north-west shelf only, and did not show significant trends in oxygen concentrations for the period 1990-2017.

References in key assessment text

  • Capet, A., et al., 2013, 'Drivers, mechanisms and long-term variability of seasonal hypoxia on the Black Sea northwestern shelf — Is there any recovery after eutrophication?', Biogeosciences 10, pp. 3943-3962,
  • Carstensen, J., et al., 2014, 'Hypoxia in the Baltic Sea: Biogeochemical cycles, benthic fauna, and management', Ambio 43, pp. 26-36.
  • Conley, D. J., et al., 2009, 'Hypoxia-related processes in the Baltic Sea', Environmental Science & Technology 43, pp. 3412-3420,
  • Friedrich, J., et al., 2014, 'Investigating hypoxia in aquatic environments: diverse approaches to addressing a complex phenomenon', Biogeosciences 11, pp. 1215-1259,
  • Helcom, 2018, 'State of the Baltic Sea — Second Helcom holistic assessment 2011-2016', Baltic Sea Environment Proceedings 155, 155 pp,
  • Naumann, M. et al., 2018. ‘Water exchange between the Baltic Sea and the North Sea, and conditions in the deep basins’,  HELCOM Baltic Sea Environment Fact Sheets, .
  • OSPAR, 2017,  ‘Intermediate assessment. Concentrations of Dissolved Oxygen Near the Seafloor’,
  • Topcu, H. D. and Brockmann, U. H., 2015, 'Seasonal oxygen depletion in the North Sea, a review', Marine Pollution Bulletin 99, pp. 5-27,

Supporting information

Indicator definition

The indicator illustrates the geographical distribution and trends in summer/autumn concentrations of oxygen in the near-bottom waters of the regional seas of Europe. It uses oxygen concentrations in the near-bottom layer during the period July-October.

The following marine regions and subregions are covered, in line with the Marine Strategy Framework Directive (MSFD) (sub)regions: Baltic Sea, North-East Atlantic Ocean, Mediterranean Sea and Black Sea (see 'Regional seas surrounding Europe' map: Regional Seas).


The concentration of oxygen is expressed as mg/l in the bottom 10 m of the water column during summer/autumn.


Policy context and targets

Context description

With respect to eutrophication, there are a number of EU directives specifically aimed at reducing the loads and impacts of nutrients, such as the Nitrates Directive (91/676/EEC), the Urban Waste Water Treatment Directive (91/271/EEC) and the Integrated Pollution Prevention and Control Directive (96/61/EEC).

The Water Framework Directive (WFD; 2000/60/EC) requires the achievement of good ecological status or the good ecological potential of transitional and coastal waters across the EU and the MSDF (2008/56/EC) requires the achievement or maintenance of good environmental status in European sea basins by the year 2020. The WFD mentions dissolved oxygen concentrations as one of the physico-chemical parameters for measuring ecological status. In the MSFD, dissolved oxygen concentration in the bottom of the water column is one of the primary criteria (D5C5) for Descriptor 5 human-induced eutrophication: 'The concentration of dissolved oxygen is not reduced, due to nutrient enrichment, to levels that indicate adverse effects on benthic habitats (including on associated biota and mobile species) or other eutrophication effects.'

With respect to climate change, the European Commission has developed the EU Adaptation Strategy with the overall aim of contributing to a more climate-resilient Europe. One of the objectives is better-informed decision-making, e.g. by sharing knowledge on (1) observed and projected climate change and its impacts on environmental and social systems and on human health, (2) relevant research, (3) EU, transnational, national and subnational adaptation strategies and plans, and (4) adaptation case studies.


Under the WFD, targets for dissolved oxygen concentrations in coastal waters have been set by Member States. Under the MSFD, threshold values for coastal waters are set in accordance with the WFD, and beyond coastal waters threshold values must be consistent with those under the WFD. Member States establish those values through (sub)regional cooperation.

Related policy documents



Methodology for indicator calculation

The two main sources of data for this indicator are the International Council for the Exploration of the Sea (ICES) and the European Marine Observation and Data Network (EMODnet) data sets.

Data kept by the ICES are collected through the Eionet Central Data Repository (Eionet CDR) from the marine conventions and are therefore sub-samples of national data assembled for the purpose of providing comparable indicators of the state of and impacts on transitional, coastal and marine waters (TCM data) on a Europe-wide scale. In addition, data supplied by EMODnet are combined with ICES data. The latest EMODnet data set submission (by 22 June 2018) was used.

In cases in which both ICES and EMODnet data were available for the same station (defined by position and time), ICES data were used.

Concentrations are presented on maps showing means of the 25th percentile per station and year for the most recent 5-year period (2013-2017).

Consistent time series of the concentrations (mean of the 25-quartile) per station and year are used as the basis for the assessment of developments over time. The trend analyses are based on time series from 1990 onwards. Stations with data from at least 5 or more years in the period since 1990 were selected for trend calculations.

The primary aggregation involves:

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

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


Geographical classification: sea region, coastal or offshore and station

All geographical positions defined in the data are assigned to marine (sub)region by coordinates.

Stations are defined geographically by position, given as longitude and latitude in decimal degrees; however, the data reported do not contain reliable and consistent station identifiers. The coordinates reported for what is intended to be the same station may vary between visits, because the exact position achieved is recorded, not the target position. Identifying stations strictly by position may fragment time series too much, as the position of the same station may vary slightly over time.

To improve aggregation into time series, data are aggregated into squares with sides of approximately 1.375 km for coastal stations within 20 km from the coastline and approximately 5.5 km for open-water stations more than 20 km from the coastline. This procedure however does not completely prevent the erroneous aggregation of data belonging to stations close to each other or the erroneous breakup of time series into fragments due to small shifts in position, but does reduce the problem considerably.

The following marine regions and subregions are covered, in line with the MSFD (sub)regions:

  • Baltic Sea (none);
  • North-East Atlantic Ocean (Greater North Sea, Celtic Seas, Bay of Biscay and the Iberian coast, Macaronesian biogeographic region);
  • Mediterranean Sea (Western Mediterranean Sea, Adriatic Sea, Ionian Sea and Central Mediterranean, Aegean-Levantine Sea);
  • Black Sea (none).


Statistical aggregation per station and per year

The aggregation includes:

  • selecting the season (months July-October);
  • selecting the sample depth (0-20 m above the sea floor when the depth of the sea floor is less than 100 m; 0-50 m above the sea floor when the depth of the sea floor is more than 100 m);
  • selecting data per station and year within the lower quartile (≤ 25th percentile);
  • calculating the mean of the data selected per station and year.


Trend analysis

Trend analysis was carried out for each station in regions for which there were at least some data from the past 6-year period (2013 or later) and data for 5 or more years from the period since 1990. Trends were detected in each time series using the non-parametric Mann-Kendall trend test.

The Mann-Kendall method is a non-parametric test suggested by Mann (1945) and has been extensively used to analyse environmental time series (Helsel and Hirsch, 2002; Hipel and McLeod, 2005). The Mann-Kendall method tests for monotonic trends in a time series, y(x), which in this analysis is oxygen concentration (y) as a function of year (x). The test is based on Kendall's rank correlation, which measures the strength of monotonic association between the vectors x and y. In the case of no ties in the x and y variables, Kendall's rank correlation coefficient, tau, may be expressed as tau = S/D, where S = sum {i<j} (sign(x[j]-x[i]) * sign(y[j]-y[i])) and D = n(n - 1)/2, S being the score and D being the denominator; D is the maximum possible value of S. The p-value of tau is computed by an algorithm given by Best and Gipps (1974). 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 size of the change.

The Mann-Kendall test is a robust and accepted approach. Because of 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.

Methodology for gap filling

Not applicable

Methodology references



Methodology uncertainty

Only data for the lower 20 m or 50 m of the water column are selected. Not all available data are associated with reliable attributes on sampling depth and bathymetry.

In shallow water, the selection of the lower 20 m of the water column may not be optimal. However, low oxygen concentrations generally occur in deeper waters, as vertical mixing will diminish the occurrence of oxygen depletion in shallow waters.

Data sets uncertainty

Data for this assessment are still scarce considering the large spatial and temporal variations inherent in European transitional, coastal and marine waters. Long stretches of European coastal and marine waters are not covered by the analysis because of this lack of data. Trend analyses are consistent only for parts of the Greater North Sea, the Baltic Sea, the Celtic Seas, the Adriatic Sea and the north-west shelf of the Black Sea.

Some of the stations are also in transitional water bodies

Rationale uncertainty

Not applicable

Data sources

Other info

DPSIR: Impact
Typology: Efficiency indicator (Type C - Are we improving?)
Indicator codes
  • MAR 012
Frequency of updates
Updates are scheduled every 3 years
EEA Contact Info


Geographic coverage

Temporal coverage



Document Actions