Monitoring changes in regional sea surface temperature is essential for evaluating climate variability and its regional impacts. Over more than the past quarter of a century, a major fraction of the surface ocean in the northeastern Atlantic Ocean and adjacent seas has been warming at rates that are equal to or higher than the global rate (WMO, 2023).

Over the past 4 decades (1982–2023), the Mediterranean Sea surface temperature warmed at a rate of 0.41 ± 0.01 °C per decade, corresponding to a mean surface temperature warming of about 1.7 °C. The spatial pattern of the Mediterranean Sea surface temperature trend shows a general uneven warming tendency following an eastward increased tendency. The strongest surface warming is reported in the eastern Mediterranean Sea, such as in the Adriatic, Aegean, Levantine, and Ionian seas (MedECC, 2020b). Between 2022 and 2023, the Mediterranean Sea experienced an exceptionally long-lasting and intense marine heatwave that hit the whole basin. This marine heatwave started in May 2022, when the mean anomaly increased abruptly from -0.1 ± 0.2 °C (April) to 0.8 ± 0.1 °C (May), reaching the highest values during June (1.7 ± 0.1 °C) and July (1.45 ± 0.07 °C) and persisting until spring 2023 (Marullo et al., 2023).

The Baltic Sea sea surface temperature is rising at a rate of 0.38 ± 0.04 °C per decade (over the period 1982–2023), which represents an average warming of about 1.60 °C. Sea surface warming in the Baltic Sea is characterized by a distinct north–south gradient, and the largest increase in sea surface temperature has been reported for the northern part of the Gulf of Bothnia, the Gulf of Finland, the Gulf of Riga, and the northern part of the Baltic Proper (Meier et al., 2022; von Storch et al., 2015). A long-term increase in surface ocean temperature has also been detected in the North Sea at rates exceeding the global mean values (Huthnance et al., 2016). The rise in sea surface temperature in areas of so-called Large Marine Ecosystems has even been classified within a “superfast warming category”, such as for those located in the Norwegian Sea, North Sea, and along the Celtic-Biscay Shelf (Kessler et al., 2022).

The Black Sea is experiencing the most intense sea surface temperature trend among the European seas, estimated at 0.65 ± 0.02 °C per decade (period: 1982–2023), which corresponds to an average increase of about 2.7 °C during the last 42 years. Together with the Baltic and the Mediterranean seas, the Black Sea represents a hot spot of global warming, where the surface warming trend largely exceeds both the northeastern Atlantic Ocean and adjacent seas and the global mean sea surface temperature trend (0.25 ± 0.03 and 0.13 ± 0.01 °C per decade, respectively).

The Iberia–Biscay–Irish (IBI) and the European North West Shelf (NWS) seas are characterized by milder warming rates estimated at 0.24 ± 0.02 and 0.22 ± 0.02 °C per decade, respectively. These trends are strongly affected by decadal variability superposed on the long-term trend (Fig. 2).

Variations from interannual to decadal scales play a critical role in shaping the evolution of regional sea surface temperature. Major climate modes affect northeastern Atlantic Ocean and adjacent seas sea surface temperature, such as the Atlantic Multidecadal Oscillation (AMO) or the North Atlantic Oscillation (NAO) (Pisano et al., 2020; Yan and Tang, 2021; Valcheva and Slabakova, 2020; Kniebusch et al., 2019), and can, together with ocean processes and cryosphere–ocean interactions, challenge the detection of long-term trends from global warming. For example, in the Baltic Sea, changes in the interplay of atmospheric circulation and ocean processes (e.g. changes in the frequency of coastal upwelling) have resulted in regional patterns of lower-than-average surface warming, which is manifested northeast of Bornholm up to and along the Swedish coast (von Storch et al., 2015). In the North Sea, changes in sea surface temperature are largely affected by decadal variations driven by water mass exchange from the adjacent northeastern Atlantic, superposing any long-term surface warming trend (Huthnance et al., 2016). In the North Atlantic, the so-called “warming hole” – a large pattern of unusual cold sea surface temperature – is discussed as being driven by ocean processes (e.g. circulation change), air–sea interactions, cryosphere–ocean interactions, and changes in anthropogenic forcing (Li et al., 2022; Allan and Allan, 2019; Caesar et al., 2018; Hu and Fedorov, 2020; Qasmi, 2023; Gutiérrez et al., 2021).

Good knowledge of sea surface temperature spatial patterns and trends is also critical to correctly assess European climate trends and projections, as they determine variations in regional weather and climate at a variety of scales that have societal implications (O'Carroll et al., 2019). For example, knowledge of the temporal evolution of sea surface temperature can provide important handles to increase the predictability of certain events, such as wet and dry extremes in Europe (Ionita et al., 2012, 2015; Volosciuk et al., 2016).

The rise in sea surface temperature is not homogeneous, and the rate of regional surface ocean warming is largest in the eastern part of the European zone and decreases further towards the western part (Fig. 3). The latter is also affected by the extension of the so-called North Atlantic warming hole. However, nearly the entire northeastern Atlantic Ocean and adjacent seas area is affected by an increase in sea surface temperature, and about 43 % of the area is warming at a rate 2 times faster than the global mean warming rate, i.e. in the Mediterranean Sea, North Sea, Baltic Sea, and Black Sea areas (Fig. 4, Table 1).

Regional ocean warming is not homogeneous, and the North Atlantic area between 20–50° N has been identified as one of the regions that is affected by unusual strong warming rates as compared to the rest of the global ocean (Cheng et al., 2022, Fig. 5). However, subsurface water temperature changes at various timescales in the subpolar North Atlantic as triggered by several factors, such as anthropogenic warming (Cheng et al., 2022), heat redistribution (Messias and Mercier, 2022), climate variability (e.g. the North Atlantic Oscillation; Hurrell and Deser, 2010), and dynamical processes linked to gyre dynamics and large-scale ocean circulation (Yeager, 2020). In the central subpolar North Atlantic, a decadal-scale cooling of the surface ocean has been reported, also known as the so-called “North Atlantic cold blob” or North Atlantic warming hole (Hansen et al., 2010; Drijfhout et al., 2012; Li et al., 2022; Fan et al., 2023), which impacts North Atlantic storm track activities and weather patterns in Europe (Gervais et al., 2019, 2020; Woollings et al., 2018). The cold blob has been also associated with variations in ocean heat content down to about 1000 m depth (Dubois et al., 2018). Several drivers of the cold blob have been hypothesized, such as evidence for an Atlantic Meridional Overturning Circulation (AMOC) slowdown (Drijfhout et al., 2012; Menary and Wood, 2018; Rahmstorf et al., 2015; Caesar et al., 2018; Chemke et al., 2020), variations in the subpolar gyre circulation (Hu and Fedorov, 2020; Keil et al., 2020; Ma et al., 2020), and changing process at the air–sea interface (Fan et al., 2023; Sanders et al., 2022).

The strong interplay of natural and anthropogenic warming is reflected during the regionalization approach for the ocean warming indicator. Over the period 1960–2023, the area-averaged results for the northeastern Atlantic Ocean and adjacent seas indicate ocean warming at a rate of 0.3 ± 0.1 W m-2 (Fig. 6, ensemble 1). However, temporal changes in ocean warming are strongly affected by natural variations (Fig. 5), and consequently warming rates quantified since the year 2005 are not statistically significant in the northeastern Atlantic and adjacent seas (-0.1 ± 0.4 W m-2, Fig. 6, ensemble 1). The dominant role of large natural variations is also again reflected in the high sensitivity of the ocean warming trend evaluation in this region when considering different time periods (de Pascual-Collar et al., 2023) (Fig. 6).

In the adjacent seas, ocean warming emerges more clearly in the time series. The sea surface of the Black Sea is experiencing the most intense warming since about the 1980s, leading to an average increase of 2.7 °C over the past 42 years (see results on the indicator sea surface temperature in this issue), and basin-wide ocean warming of the upper 300 m shows the highest warming level in the study area over the period 2005–2023 (Fig. 6). Whether this high level is the result of both the ongoing warming trend and natural variability, which is known to play a central role (Lima et al., 2020, 2021), remains a matter of research (Cheng et al., 2024), with a marked increase over the past few decades (Pinardi et al., 2015; Simoncelli et al., 2018; von Schuckmann et al., 2016; Fig. 6).

At a regional scale, the sea level trends show spatial patterns, with regions rising at a faster or slower rate than the global mean rates or even dropping in some areas. The regional sea level trends depart from the global mean sea level rise as a result of the superposition of geographical trend patterns caused by different processes (Cazenave and Moreira, 2022). For example, the sea level trends in the Mediterranean Sea differ from global mean trends as a result of the semi-enclosed conditions of the Mediterranean Sea (Pinardi et al., 2014). The process contributing to the regional variability in sea level trends are regional changes in ocean circulation, temperature, and salinity (stereodynamic changes); redistribution of ocean water mass by the ocean circulation (manometric changes); atmospheric loading; and changes in Earth gravity, Earth rotation, and viscoelastic solid Earth deformation (GRD) due to ongoing changes in the solid Earth caused by past changes in land ice (GIA) and contemporary changes in the mass of water stored on land as ice sheets, glaciers, and land water storage (GRD sea level fingerprints) (Gregory et al., 2019).

Depending on the measurement technique used, different information on sea level change can be obtained. For example, relative sea level as obtained from tide gauges is referred to the height of the sea surface relative to the sea floor, and thus to land, at a given location. Sea level from the satellite is linked to the so-called absolute sea level, which is the height of the sea surface at a given location relative to the reference ellipsoid. At a regional scale, these are important concepts allowing for further understanding on the causes of sea level change. For example, in the Baltic Sea the rise in absolute sea level has been assessed to a rate of 1.3 to 1.8 mm yr-1 over the period 1993–2010 (Hünicke et al., 2015; The BACC II Author Team, 2015), whereas relative sea level in the Baltic Sea shows the lowest increase as compared to other ocean areas in Europe due to land uplift (Ranasinghe et al., 2021; Vousdoukas et al., 2017). The Baltic Sea is strongly affected by vertical land motion, especially due to the GIA caused by the melting of past ice sheets (Ludwigsen et al., 2020), and relative sea level trends have been shown to be strongly negative (Passaro et al., 2021). Accordingly, the IPCC (2021a) projected that relative sea level will rise in all European areas except for the Baltic Sea. However, absolute sea level trends show statistically significant positive trends (Fig. 7) (Passaro et al., 2021).

A recent assessment report for the Mediterranean climate has reported a mean sea level rise of 6 cm over the past 20 years, and this trend is likely to accelerate (with regional differences) by the global rate of 43 to 84 cm until 2100 (MedECC, 2020c). In the North Sea, an absolute mean sea level trend of 2.61 ± 0.95 mm yr-1 has been reported over the period 1995–2019, which varies regionally between 1.5 and 3.5 mm yr-1 over the region, with the highest trends in the German Bight and around Denmark and lower trends around the southern part of Great Britain (Dettmering et al., 2021). Sea level change in the northeastern Atlantic up to the European shelf area is known to be largely affected by variations driven by ocean dynamics and processes at the air–sea interface, which have induced large sea level variations at decadal and smaller timescales, superposing the long-term trend in this area (Chafik et al., 2019). In the Black Sea, a recent study has reported a mean rate of total sea level rise of 2.5 ± 0.5 mm yr-1 over the entire basin (Avşar and Kutoglu, 2020), with large interannual- to decadal-scale variations driven by climate variability, freshwater fluxes from land, and sea level fluctuations in the Mediterranean Sea (Volkov and Wilson, 2019).

The Copernicus Marine Service sea level data for the European regional seas have been adjusted for the TOPEX-A instrumental drift based on Ablain et al. (2017) and adjusted for the GIA using latitude and coastal weighting on the map of GIA trend from Spada and Melini (2019). Therefore, the European weighted-mean GIA corrections amount to -0.49 mm yr-1. At basin level, the correction amounts to 0.21 mm yr-1 in the Baltic Sea, whereas in the Mediterranean Sea it is on the order of -0.21 mm yr-1. On the other hand, in the Black Sea the correction is estimated to 0.08 mm yr-1, and it ranges between -0.5 and -0.69 mm yr-1 in the North West shelf and Iberia–Biscay–Irish regions, respectively.

Most recent Ocean State Report estimates of absolute sea level recorded by altimeters onboard satellites show sea level rise over all the European seas ranging from 1.04 to 4.11 mm yr-1 with uncertainties on the order of 0.8 mm yr-1. However, space-based radar altimetry does not provide reliable sea level data within 20 km of the coast, and the amount of valid data strongly decreases due to land contamination within the radar echo in the vicinity of the coast (Cazenave et al., 2022; The Climate Change Initiative Coastal Sea Level Team, 2020; Vignudelli et al., 2019). The sea level rise presents spatial variations with differences across basins, with the Baltic Sea absolute sea level rising at a faster rate and the Black Sea at a comparably slower pace (Fig. 7).

The northeastern Atlantic Ocean and adjacent seas area-averaged sea level time series (Fig. 7a) presents dominant interannual variability that in this region is dominated by the North Atlantic Oscillation (NAO) (Calafat et al., 2022; Chafik et al., 2019; Meli et al., 2023; Passaro et al., 2021; Volkov and Landerer, 2015). Karimi et al. (2022) showed that the NAO dominates the interannual variations of barystatic sea level in the Baltic Sea, whereas steric sea level plays a key role in decadal variations. Additionally, Chafik et al. (2019) proposed that the decadal sea level changes are driven by steric variability in the subpolar North Atlantic. Masina et al. (2022) presented an anticorrelation between the Mediterranean mean sea level and the upper branch of the Atlantic Meridional Overturning Circulation at 26.5° N.

Over the January 1993–June 2023 period, the regional rates of rise are positive across the northeastern Atlantic and adjacent seas region, except for some eddy-scale areas in the eastern Mediterranean Sea (Fig. 8a), where negative trends are observed. The rate of sea level rise in the Mediterranean Sea shows spatial variations that reflect changes at basin-scale circulation. Based on data over the altimetry era, Meli et al. (2023) showed that the mass component is the dominant driver in this area, except for in the Levantine and Aegean subbasins, where the steric component is the major contributor to the sea level trends.

In the Black Sea, the major drivers of change have been attributed to steric variability, water mass changes related to the Mediterranean inflow, and freshwater fluxes (precipitation, evaporation, and river discharge) (Tsimplis et al., 2004; Volkov and Landerer, 2015). Most of the Mediterranean Sea and Black Sea are rising a slower rate than the global mean sea level (Fig. 8b). On the other hand, the entire Baltic Sea is rising faster than the global mean sea level trend, estimated to be 3.4 mm yr-1 over 1993–2023. Large areas of the Iberia–Biscay–Irish regional seas and the Gulf Stream region show sea level trends larger than the 4.3 mm yr-1 observed at global level for the last 10 years (June 2013–June 2023; Fig. 8b).

In addition to the polar seas, there are several marine regions around the world where sea ice occurs, such as the Baltic Sea. Sea ice occurs in a wide range of types and forms and significantly and directly affects marine transport and navigation. Specifically, the presence of sea ice cover sets special requirements for navigation, both for the construction of the ships and their behaviour in ice, as in many cases merchant ships need icebreaker assistance (JCOMM Expert Team on Sea Ice, 2017). Variations in sea ice conditions at different timescales in the Baltic Sea can also have implications for regional weather and climate as sea ice influences the way the ocean and atmosphere interact via fluxes of heat, water, or carbon. There is a tight link between surface air temperature and sea ice loss in the Baltic Sea that makes sea ice monitoring a critical climate indicator for this region (Granskog et al., 2006). Changes in sea ice are also known to affect the ecosystem in the Baltic Sea (Pärn et al., 2022; Granskog et al., 2006; Climate Change in the Baltic Sea, 2021; Meier et al., 2022; Eilola et al., 2013). For example, extreme events at the land–sea ice interface that adversely affect infrastructure at the coast happened several times between 2010 and 2019 (Girjatowicz and Łabuz, 2020). They are usually linked to onshore riding up of sea ice from shore to land, i.e. ice ridges, through a combination of specific sea ice conditions, storms, currents, and sea level variations (Leppäranta, 2013).

The importance of sea ice is reflected in the fact that sea ice variations have been monitored regularly in the Baltic Sea since the late 19th century (Meier et al., 2022). Basin-wide assessments have revealed that sea ice in the Baltic Sea undergoes large seasonal and interannual variations, and a long-term trend in sea ice loss has also been observed for this part of the world ocean (Meier et al., 2022​​​​​​​; The BACC II Author Team, 2015). In the past, the entire Baltic Sea was covered in ice, as it was in the 1940s (Vihma and Haapala, 2009), whereas in most recent periods only the northern areas of the Baltic Sea are mostly covered in ice (Raudsepp et al., 2020).

Sea ice coverage in the Baltic Sea is strongly seasonal, and sea ice starts to form in October and may last until June of the following year. The ice season 2022/23 had the moderate maximum ice extent in the Baltic Sea and reached a maximum area of about 65 000 km2, which amounts to 15 % of total Baltic Sea ice coverage. Usually, the maximum sea ice extent is reached at the end of February, but in 2023 maximum ice cover was recorded a couple of weeks later (Fig. 9). Although results indicate a decrease in sea ice extent over the period 1993–2023, the linear trend is not statistically significant.

The Mediterranean Sea is a semi-enclosed basin characterized by peculiar thermohaline circulations consisting of a basin-scale zonal overturning connected to the Atlantic through the Strait of Gibraltar and two internal meridional cells, one for the west and another for the eastern basin, driven by deep-water formation in the Gulf of Lion and the Adriatic Sea, respectively (see Tsimplis et al., 2006; Schroeder et al., 2012; Pinardi et al., 2023 for general information). The latter experienced a major shift in the early 1990s with deep-water formation temporarily occurring in the southern Aegean Sea (the so-called Eastern Mediterranean transient; Roether et al., 2014). A recent investigation of the dynamics of intermediate and deep-water formation can be found in Waldman et al. (2018).

Masina et al. (2022) presented changes in the Gibraltar inflow transport from the Atlantic Ocean into the Mediterranean Sea that trigger basin-mean sea surface height variability, which is anti-correlated with variations in the Atlantic Meridional Overturning circulation (AMOC). They show that during years of weaker (stronger) AMOC and higher (lower) sea surface height in the Mediterranean Sea, a stronger (weaker) Azores Current results in stronger (weaker) Gibraltar inflow transport.

The zonal overturning cell enables the connection between the inflow transport at the Gibraltar Strait and remote areas of the eastern Mediterranean basin. The wide and shallow Strait of Sicily sill divides the western Mediterranean basin from the eastern Mediterranean basin (Pinardi et al., 2019). The western sub-basin clockwise meridional overturning circulation is associated with the deep-water formation area of the Gulf of Lion, while the eastern clockwise meridional overturning circulation is composed of multiple cells associated with different intermediate- and deep-water sources in the Levantine, Aegean, and Adriatic seas (Lyubartsev et al., 2020).

Based on these characteristics, the Mediterranean meridional overturning circulation can be monitored by the western Mediterranean overturning index (WMOI) and eastern Mediterranean overturning index (EMOI) (Lyubartsev et al., 2020) (Fig. 12). These indices are useful for monitoring the health of the Mediterranean Sea in terms of stratification and deep-water formation, as these factors mediate the exchange of oxygen and other tracers between the surface and the deep ocean, fostering the establishment of early warming indices for the Mediterranean Sea to support the Sustainable Development Goal (SDG) 13 Target 13.3 (Lyubartsev et al., 2020).

Both the EMOI and the WMOI show no long-term trend, and both records are characterized by large interannual- to decadal-scale change (Fig. 12). Overall, the overturning circulation in the eastern Mediterranean Sea is stronger as compared to the western basin. This is demonstrated by the generally higher values of EMOI as compared to WMOI (Fig. 12). The difference in both indices is most prominent during the major climatic event in the circulation and water mass properties of the Mediterranean Sea in the last century, i.e. the so-called Eastern Mediterranean Transient (EMT; Roether et al., 2014; Incarbona et al., 2016). The pronounced peak in EMOI in the year 1992 is driven by the EMT. In 1999, the difference between EMOI and WMOI started to decrease because EMT water masses reached the Strait of Sicily flowing into the western Mediterranean Sea (Lyubartsev et al., 2020).

During the past 2 decades, WMOI reached another maximum in the year 2006, potentially linked to anomalous deep-water formation during the Western Mediterranean Transition (Lyubartsev et al., 2020). In 2019, the EMOI reached another high value, which is discussed as being triggered by subduction of the Levantine Intermediate Water along the Cretan Sea and the Rhodes Cyclonic Gyre (Taillandier et al., 2022). In 2022, the EMOI peaked again, possibly linked to deep-water formation events that took place in the northern and central Aegean Sea during the winter of 2021–2022 (Potiris et al., 2024).

The circulation of the water masses and distribution of heat and salinity between the North Atlantic and the Arctic and in the Arctic Ocean are governed by the northward flow of relatively warm and saline Atlantic Water through the Nordic Seas across the Iceland–Faroe–Scotland Ridge and into the Arctic, balanced by outflow of cold Arctic Water masses through the western Nordic Seas (Lien and Raj, 2018; Mauritzen et al., 2011; Rudels, 2012). Moreover, the exchanges between the North Atlantic and the Arctic have broad effects on sea ice, ecology, biogeochemistry, and atmospheric climate (Smedsrud et al., 2013; Polyakov et al., 2023; Gerland et al., 2023; Lundesgaard et al., 2022). Specifically, the Atlantic Water flow plays an integral part in defining both the physical and biological border between the boreal and Arctic realms. Variability of Atlantic Water flow to the Barents Sea has been found to move the position of the ice edge and influences the sea ice cover in the Barents Sea (Onarheim et al., 2015; Lien et al., 2017) and habitats of various species in the Barents Sea ecosystem (Fossheim et al., 2015; Jørgensen et al., 2022).

The northward flow of Atlantic water into the Arctic through the Faroe–Shetland Channel amounts to 2.5 Sv (1 Sv = 106 m3 s-1) over the period 1993–2022 (Fig. 13), which is in within the uncertainty range of observation-based estimates of 2.7 ± 0.5 Sv (1993–2015) (Berx et al., 2013). In addition, a small but significant negative trend of -0.17 Sv per decade is reported for the period 1993–2022 (Fig. 13). This is in comparison to a reported non-significant trend of -0.006 Sv yr-1 based on observations during the period 1993–2015 (Østerhus et al., 2019). The modelled transport in the most recent year did not differ from the historical values. The hydrographic properties of the inflowing Atlantic Water affects the hydrographic properties in the downstream Norwegian and Barents seas, and these areas are currently recovering from record low salinity levels advected from the North Atlantic (Holliday et al., 2020). Moreover, such changes may also be a precursor for different water mass composition and hence nutrients, phytoplankton, and zooplankton communities being advected into the Nordic Seas (Hátún et al., 2017). However, the strength of the inflow itself has also been shown to impact the hydrographic properties downstream in the Norwegian and Barents seas, where stronger inflow causes increased salinity and temperature (Sundby and Drinkwater, 2007) in addition to regional ocean–atmosphere interaction (Segtnan et al., 2011; Mork et al., 2019).

In the Barents Sea Opening, the model indicates a long-term average net Atlantic water inflow of 2.2 Sv compared to an observation-based estimate of 1.8 Sv (Smedsrud et al., 2013). Note the different time window used for averaging. The model results indicate a small, non-significant negative trend of -0.04 Sv per decade, which is in agreement with observations indicating no trend over the 1998–2020 period (ICES, 2022). According to Østerhus et al. (2019), a positive, significant trend of 0.016 Sv yr-1 was found for the period 1998–2013. Note, however, the different time window used for the calculation. The modelled transport in the most recent year did not differ from historical values. In the Fram Strait, the model data indicate a positive trend in the Atlantic Water transport to the Arctic of 0.45 Sv per decade for the period 1993–2022. This trend is partly explained by increased temperature in the West Spitsbergen Current during the period 2005–2010 (e.g. Walczowski et al., 2012), which caused a larger fraction of the water mass to be characterized as Atlantic Water (T > 2 °C). Moreover, the strong recirculation in the Fram Strait area further complicates the interpretation of the results.

In the Baltic Sea, the salinity and its changes are key factors in determining the overall stratification conditions, which strongly influence different ecosystem processes and play an important role in the energy and water cycles in this area (Lehmann et al., 2022). The long-term change in water salinity of the Baltic Sea is determined by the inflow of saline water from the North Sea and its encounter with freshwater originating from numerous rivers across the Baltic coast and from the net precipitation (Lehmann et al., 2022). A major Baltic inflow (MBI) is an event that transports large amounts of saline water into the Baltic Sea through the Danish Straits (Mohrholz, 2018), shaping the stratification and oxygen conditions and hence influencing the marine ecology of the Baltic Sea. MBIs occur sporadically, usually many years apart, and are usually of barotropic origin, initiated by a special sequence of large-scale meteorological events in winter and spring. Since 1996, summer inflows of baroclinic origin have been observed. These summer inflows inject more highly saline water with higher temperatures and low oxygen content into the halocline of the Baltic Sea (Lehmann et al., 2022).

MBI can be detected and monitored in the bottom-layer salinity of the Arkona Basin and the Bornholm Basin and through the vertical distribution of temperature, salinity, and dissolved oxygen concentration in the Gotland Basin. Temperature, salinity, and dissolved oxygen profiles in the Gotland Basin enable the estimation of the amount of the major Baltic inflow water that has reached the central Baltic, the depth interval that has been the most affected, and how much the oxygen conditions have been improved. Major Baltic inflows were identified in 1993, 2002, and 2014, showing salinity peaks in the three basins (Figs. 14, 15) and displaying a very clear signal in the Gotland Basin with water salinity, temperature, and dissolved oxygen conditions up to 100 m depth (Raudsepp et al., 2018). The bottom salinity has decreased since 2017 when the effect of the last MBI in 2014 has receded (Fig. 15). There was no MBI in winter 2022/23 and in 2023 salinity continued to decrease and oxygen was completely consumed below the depth of 75 m in the Gotland Basin.

Marine heatwaves can have dramatic impacts on marine ecosystems, such as reduction of primary production, migration or reduction of endemic species, emergence of species coming from other regions, and mass mortality of organisms (Smith et al., 2023; Oliver et al., 2019; Garrabou et al., 2022), which in turn can have adverse impacts on human systems like fisheries and aquaculture (Cheung and Frölicher, 2020; IPCC, 2022; Wakelin et al., 2021). Like in the global ocean, the frequency, intensity, and duration of MHW events have increased in recent years in the northeastern Atlantic Ocean and adjacent seas and are particularly increased in the semi-enclosed seas like the Mediterranean and Black seas (Peal et al., 2023; Yao et al., 2022; Oliver et al., 2018; IPCC, 2021b; Darmaraki et al., 2019; Dayan et al., 2023; Juza et al., 2022). These events are expected to continue to increase in the future (Yao et al., 2022; Oliver et al., 2019).

From 1982 to 2023, the fraction of the ocean surface that did not experience any MHW event over the year in the northeastern Atlantic Ocean and adjacent seas (black polygon in Fig. 11) dramatically dropped, from more than 80 % in 1982 to less than 10 % in 2023 (Fig. 11​​​​​​​). The major drop occurred between 1982 and the early 2000s. Since then, the yearly fraction of ocean surface without any MHW event has remained relatively stable at around 20 % ± 10 % (i.e. 80 % ± 10 % of the region is affected by MHW events every year). The fraction of ocean surface that was affected by moderate MHW events increased from about 10 % to about 45 % between 1982 and 2001. It remained relatively stable between 2000 and 2010 and then started to drop in 2010 down to less than 20 % in 2023. The fraction of the ocean surface experiencing strong MHW events has been continuously increasing over the whole period, from less than 5 % in 1982 to more than 45 % in 2023. Severe and extreme MHW events also show a positive trend over the last decade. However, the trend of extreme category events is strongly dampened when excluding sea-ice-covered regions (Fig. 11), as is the case for the global estimates. In addition, similar to the global estimates, the impact is minor for the other categories (not shown). Since 2013, the cumulative fraction of ocean surface affected by strong, severe, and extreme MHW events has been almost systematically larger than that affected by moderate MHW events. As we consider a fixed baseline for the climatology, these results are consistent with the fact that this part of the ocean is one of those that have been warming at the fastest rates due to global warming.

Regarding the marine heatwave durations (Fig. 16), the yearly averaged maximum continuously increased from 1982 (less than 10 d) to 2023 (about 40 d). Until the mid-2010s, most of the longest MHW events lasted less than 15 d. Since then, the ocean surface fraction affected by short events at most has dropped, down to about 12 % in 2023. The ocean surface fraction affected by events lasting between 2 weeks and 1 month has increased from 1982 (less than 5 %) to the early 2000s, when it stabilized around 25 %. The ocean surface fraction affected by events lasting between 1 and 2 months has continuously increased from almost no occurrence in 1982 to more than 30 % in 2023, and these events now prevail on other MHW durations in the region. Similarly, the ocean surface fraction experiencing MHW events lasting between two and four months started to rise in the 1990s and reached more than 20 % in 2023. Events lasting more than four months remain at a low level but have also been more observed since 2015. Similar results are obtained when excluding sea-ice covered regions from the computation (not shown).

In 2022 and 2023, 91 % and 93 %, respectively, of the ocean surface in the northeastern Atlantic Ocean and adjacent seas were affected by at least one moderate marine heatwave (Table 2). These numbers rise to about 95 % when considering only the European shelf seas, i.e. regions at depths between 0 and 200 m. In 2022, about 60 % of the northeastern Atlantic Ocean and adjacent seas were affected by strong (49 %) or even severe (7 %) and extreme (5 %, mostly in the Arctic region) marine heatwaves. Some marine heatwave events prevailed for a long period of about 4 to 5 months, such as in the western Mediterranean Sea, the English Channel, and the southern Arctic. Other areas affected by strong to extreme marine heatwaves and events lasting about 1 to 2 months include the central Mediterranean Sea, the North Sea, and large parts of the Iberia–Biscay–Irish areas (Fig. 11​​​​​​​).

In 2023, almost 80 % of the ocean surface in the northeastern Atlantic Ocean and adjacent seas experienced MHW of at least strong (45 %), severe (24 %), or extreme (8 %) categories (Table 2). Compared to 2022, the Bay of Biscay and the northeastern Atlantic region off the Iberian coast were particularly affected by marine heatwaves in 2023 (Fig. 17), with events lasting more than 5 months (Fig. 17). This corresponds to the anomalously warm surface waters that have been observed in the tropical North Atlantic Ocean since March 2023. Other regions hit by MHW events lasting 2 to 3 months in 2023 include the English Channel and the Irish Sea, the Alboran Sea, the Adriatic Sea, the Levantine basin, and the western part of the Black Sea (Fig. 17). All of these regions also experienced strong to severe MHW events over the year 2023 (Fig. 17).

The Mediterranean Sea is an area where also a multidecadal increase in the frequency and intensity of marine heatwaves has been reported (Juza et al., 2022; Dayan et al., 2022). This basin experienced an exceptionally long-lasting and intense series of marine heatwave events between 2022 and 2023 (Fig. 17), as MHWs have been continuously detected since June 2022 until the end of 2023, particularly in the western basin (Marullo et al., 2023). In 2022, the heatwave intensity was highest in the northern parts of the western Mediterranean, decreasing to the south and to the east. In 2023, almost the entire Mediterranean Sea experienced at least one marine heatwave of strong category, and the strongest events happened in the Ionian Sea basin (severe category) (Fig. 17).

Thermal stress on marine organisms can lead to a decrease in production (Islam et al., 2022; Smith et al., 2023), and in the worst cases unusual warming of the waters can lead to lower seawater dissolved oxygen quantities, which can cause asphyxia (Keeling et al., 2009; Roman et al., 2019) and/or foster the development of toxic algae or parasites that can prevent the consumption of farmed shellfish and finfish (Cavole et al., 2016; Boudouresque et al., 2024). Some regions of high production of shellfish and/or finfish (Fig. 12) were affected by strong to severe marine heatwaves in 2022 and 2023 (Baltic Sea, Irish Sea, English Channel, northwestern Mediterranean Sea, Adriatic Sea) or by marine heatwave events that sometimes lasted more than 2 or 3 months in a row (English Channel, Irish Sea, Balearic Sea, Adriatic Sea) (Fig. 17).

In 2022, more than 66 % of the shelf ocean surface (regions with depth between 0 and 200 m) in the northeastern Atlantic Ocean and adjacent seas experienced strong (58 %) and severe or extreme (8 %) marine heatwave events. Here we exclude sea-ice-covered regions as they represent 40 % of the considered shelf area and the marine heatwave detections may be less accurate there; see Table 2 and Fig. 17. In 2023, 88 % of these shelf regions were affected by strong (62 %), severe (24 %), or extreme (2 %) marine heatwave events. There is a clear ascending trend in the ocean surface fraction affected by marine heatwave events in these shelf regions over the last 40 years, with about 20 % of the area experiencing MHW events in the 1980s and more than 90 % in 2023 (Fig. 18). The areas affected by strong and severe events have increased over the whole period, while the extreme events are relatively stable and rare. The annual mean maximum duration of the marine heatwave events has also increased, from about 10 d in 1982 to about 40 d in the most recent years, with larger fractions of the ocean surface affected by longer events, in particular those lasting between 2 weeks and 4 months. In general, more interannual variability can be noted in these shelf estimates than when considering the whole northeastern Atlantic Ocean and adjacent seas region (Fig. 16), as the shelf regions are more sensitive to local oceanic and atmospheric dynamics and drivers.

Monitoring extreme ocean surface winds, including their evolution and change over time, is of particular importance to ensure accurate representation of all linked ocean parameters (e.g. waves, storm surges, sea surface temperature, currents; Alvarez-Fanjul et al., 2019, 2022; de Alfonso et al., 2020; Staneva et al., 2020; Berta et al., 2020; Raudsepp et al., 2021; Mourre et al., 2023) in order to better understand the associated disaster risks (e.g. coastal flooding and impacts on offshore and coastal infrastructure, activities, and ecosystems), design marine infrastructures, and feed early warning systems and risk prevention and adaptation plans (de Alfonso et al., 2020; Drago et al., 2021), in line with the priorities identified in the United Nations Sendai Framework for Disaster Risk Reduction 2015–2030 (UNDRR, 2024).

When considering the northeastern Atlantic Ocean and adjacent seas, the regions that show the highest extreme ocean surface winds over the 2007–2022 period are the central and northern parts of the North Atlantic Ocean (more than 22 m s-1), especially between Greenland and Iceland and along the southern coast of Greenland (more than 26 m s-1, which corresponds to storms in the Beaufort wind scale; see Table 3) (Fig. 19a). The tropical North Atlantic Ocean (0 and 30° N) is characterized by lower extreme winds in the 99th percentile climatology, ranging between 12 and 14 m s-1.

Another area with strong extreme winds (about 20 m s-1) is the Gulf of Lion in the northwestern Mediterranean Sea, which is a region of strong cyclogenesis known to experience the most severe winds in the entire Mediterranean Sea (e.g. Zecchetto and De Biasio, 2007) and the highest extreme wind waves (Barbariol et al., 2021). In the Mediterranean Sea, a sub-group of cyclonic storms is called Medicanes (Mediterranean hurricanes). Although there is no official physical definition for the meteorological term Medicane, it is generally associated with systems that, despite being smaller in size and intensity (wind speed), share similarities with tropical cyclones, such as a circular shape with a warm core and low wind at the centre, surrounded by strong winds (MedECC, 2020a; Flaounas et al., 2022). As their definition may vary depending on the studies, there is still uncertainty on the statistics of the number or intensity of Medicanes, but they seem to generally occur one to three times per year (Cavicchia et al., 2014; Gaertner et al., 2018; Zhang et al., 2021).

In early September 2023, Storm Daniel was identified as the deadliest Medicane in recent history, hitting the coast of Libya with massive rainfall that led to the destruction of two dams, flooding the coastal city of Darna and causing more than 12 350 deaths in the area (EM-DAT, 2024). In comparison, for the whole year 2022, 43 deaths were reported by this dataset due to storm events in Europe. According to IPCC AR6, the number of Medicane events is likely to decrease, but their intensity is likely to increase under global warming (Seneviratne et al., 2021) as atmospheric convection and moisture are expected to be enhanced by higher sea surface temperatures (Flaounas et al., 2022).

At the regional scale, the detection of long-term changes also remains a challenge due to high variability and the limited length of the time series. Over the past 16 years (2007–2023), increasing extreme wind speeds are observed in several areas over the northeastern Atlantic Ocean and adjacent seas (Fig. 19). These include the central North Atlantic Ocean Gulf Stream region (30–35° N, 30–50° W), the sub-polar Atlantic region between Greenland and south of Iceland, the Norwegian Sea, the Baltic Sea, and the eastern Mediterranean Sea (Ionian Sea and Levantine Sea). Such changes in extreme wind intensity in the North Atlantic Ocean may be associated with the poleward migration of tropical and extra-tropical cyclone tracks that is observed as the tropics expand with global warming (Seneviratne et al., 2021). However, further analysis, specific detection, and attribution studies, as well as longer time series, are needed in the future for the identification of long-term trends in observed extreme wind speeds (Gentile et al., 2023).

Another approach to spatially diagnose extreme ocean surface winds is to evaluate the storm track density (Fig. 19) (Hodges, 1999, 1995; Hoskins and Hodges, 2002). During the spring to summer season (i.e. April to September), the average storm density exceeds 140 south of Greenland and around 40° N at the Canadian coast (Fig. 19c). Areas where the storm density exceeds 100 include the North Sea, the Baltic Sea, the Barents Sea, the Kara Sea, and the northern Mediterranean Sea (Fig. 19c). During the autumn to winter season (i.e. October to March), these patterns intensify and increase in size, spanning the entire Atlantic Ocean across 40° N from the Canadian coast up to the coast of Europe, and the average storm density increases to more than 120 in the adjacent seas (Fig. 19d).