Surface conditions in the ocean are largely dictated by its interaction with the atmosphere. Such interaction is determined by a number of factors, a very important one being the horizontal patterns of mean sea level pressure (mslp) as it determines barotropic currents in the ocean and is the main driver of the wind field (both zonal, U10 and meridional, V10 components). Hence, daily mslp and wind components were downloaded from cds.climate.copernicus.eu over the ROI (30°N-56°N and 19°W-10°E, Figure 1 ) from the ECMWF, ERA-5 reanalysis (Hersbach et al., 2020) at 1/12° spatial resolution and covering the period 01/1980 – 12/2020 (Supplementary Figure 1 ).

In addition, the monthly values of the North Atlantic Oscillation index (NAO) were obtained from the NOAA Climate Prediction Center website ( https://www.cpc.ncep.noaa.gov/products/precip/CWlink/pna/nao.shtml ) from 01/1980 to 12/2020 (Supplementary Figure 1 ).

Surface ocean currents (u and v velocities) and significant wave height (Hs) are two key oceanographic variables that can influence P. physalis dispersion and that are currently available for our ROI ( Figure 1 ) from COPERNICUS marine data store (https://resources.marine.copernicus.eu/) for the whole available period (01/1993 – 12/2020). Copernicus oceanographic models represents an open-access source of high-quality ocean reanalysis data including multi-layered data assimilation schemes and rigorous quality control protocols. Details of the used models are provided in Supplementary Figure 1 , Supplementary Table 2 and here below.

Surface currents were obtained from the Copernicus Marine Environmental Monitoring Service (CMEMS) product IBI_MULTIYEAR_PHY_005_002 (https://doi.org/10.48670/moi-00029). The downloaded dataset consists of monthly values of surface u and v velocities over the ROI ( Figure 1 ) for the period 01/1993 to 12/2020 ( Supplementary Figure 1 ) at a spatial resolution of 1/12°. This product is based on the Nucleus for European Modelling of the Ocean (NEMO) v3.6 ocean general circulation model and assimilates altimeter data, in situ temperature and salinity vertical profiles and satellite sea surface temperature.

Wave height data come from the product IBI_MULTIYEAR_WAV_005_006 ( https://doi.org/10.48670/moi-00030 ). We subsampled the whole dataset over the GoC region (35.5°N-38°N, 5°W-10°W) downloading monthly data for the period 01/1993 to 12/2020 ( Supplementary Figure 1 ) at 1/20° spatial resolution. This model configuration is based on the Météo-France WAve Model (MFWAM) which assimilates significant wave height (SWH), altimeter data, and wave spectral data.

Also, in order to provide a general description of the overall conditions over the North Atlantic, surface current data (u and v) for the region 15°-65°N, 0°-80°W were downloaded from the GLOBAL-MULTIYEAR-PHY-001-030-MONTHLY (https://doi.org/10.48670/moi-00021). Data for spring months (April and May) were downloaded for the period 01/1993 to 12/2020 ( Supplementary Figure 1 ) at 1/12° resolution. This dataset is produced by the same physical model (NEMO) driven at the surface by ERA-interim/ERA-5 (ECMWF) reanalysis but covering the whole earth. It assimilated observations using a reduced-order Kalman filter including along-track altimeter data (sea level anomaly), satellite sea surface temperature, sea ice concentration and in situ temperature and salinity vertical profiles.

A comprehensive dataset of P. physalis sightings was compiled from various sources, including published scientific articles, media reports, national and regional agencies, and personal communication (Prieto, 2021). This dataset encompasses 14 complete years (2009-2022) and covers the geographical area of the North East Atlantic Ocean and the Mediterranean Sea.

A subset of this dataset focuses on the exceptional summer 2019 event of swarms in the Gulf of Cadiz (GoC). During this period, the coastline was intensively monitored by technicians from the Andalusian regional government and lifeguards from the province of Cádiz, who systematically recorded and measured stranded colonies. Additionally, P. physalis sightings from 2019 were retrieved from three online databases: the Jellywatch Program (http://jellywatch.org), the PERSEUS Jellyfish Spotting website (http://www.perseus-net.eu/en/jellyfish_map/index.html), and the Medusapp dataset (https://www.medusapp.net/).

The stranding data from the summer 2019 event in the GoC is considered comprehensive, as the coastline was continuously monitored. This means that the absence of recorded stranding (i.e., zero values) can be taken as a reliable indication of no colonies arriving on shore. The complete dataset is publicly available at the following link (Álvarez-Trasobares and Prieto, 2024 upon acceptance of this manuscript).

In a very similar analysis to the one performed for winds, climatological (1993 – 2020) June-July surface current intensity and for the specific year 2019 are presented in Supplementary Figure 5 . For 2019, the southward current along the Portuguese coast and the eastward coastal current in the GoC are enhanced in summer.

More details can be analyzed in the surface currents anomalies maps presented in Figure 5 . Starting with the overall map of the ROI ( Figures 5A, D ), no clear patterns could be identified in the open Atlantic Ocean. However, a particular dynamic could be easily spotted, the high intensity of the AJ and the associated Western Alboran Gyre in the Strait of Gibraltar – Alboran Sea region. As commented in section 3.1.1, the strong increase of the eastward velocity of the AJ (+0.15/0.2 m/s) is most likely connected to the low mslp over the Western Mediterranean Sea via the ‘inverse barometer effect’ (Dastis et al., 2018; Bolado-Penagos et al., 2021).

For both months, also the eastward flowing current along the northern coast of the GoC (García-Lafuente et al., 2006) is reinforced (+0.05/0.1 m/s, Figures 5B, E ). A closer look at this northern region of the GoC reveals that also over the continental shelf the direction of the anomaly is towards the east, i.e. coastward ( Figures 5C, F ). It is worth mentioning that this current anomaly is concurrent in direction with the wind anomalies described above ( Figures 4C, F ).

There is also a significant statistical relationship between the zonal (eastward) velocities of the AJ (surface values at the eastern exit of the Strait of Gibraltar) and of the surface current in the northern GoC (<100m depth) ( Figure 6A ). For both months (June and July) the correlation between both velocities is moderate (r²>0.6) but significant (p<0.01) and indicates an acceleration of the eastward currents in the northern coast of the GoC when the AJ velocity increases.

Furthermore, both months of summer 2019 show the higher mean u velocity on the northern coasts of the GoC during the 1993–2020 period ( Figures 6A, B ). These values of the coastal current are, indeed, clearly identified as outliers in the statistical analysis ( Figure 6C ). The zonal velocity of the AJ is not an outlier in June ( Figure 6D ) although it is outside the 75% percentile range, while zonal AJ velocity for July is, indeed, identified as an outlier in the analysis ( Figure 6D ).

Finally, and on a more local scale, the actual arrival of floating P. physalis colonies to the Cádiz beaches could also be impacted and determined by the waves’ action (e.g., through Stokes drift). Figure 7 shows the anomalies maps of significant wave height (Hs in m) for June and July 2019 in the GoC. In both months, Hs was higher (10–20 cm) than their climatology along the coasts of the Cádiz province, precisely in the region where P. physalis beachings were recorded (green circles in Figure 7 ).

The contemporaneous presence of potentially dangerous marine organisms and beachgoers is a source of health and economic problems that is not usual along the southern Iberian coasts. Therefore, when in summer 2019 the presence of P. physalis forced the closure of several beaches, local authorities and stakeholders were concerned. The social alarm triggered numerous questions about why and how this event could happen and raised much interest in understanding if similar cases could likely occur in the future.

Fortunately, and thanks to the continuous investments in public research, nowadays we have access to datasets with enough quality to scientifically elucidate the causes of this rare event. Indeed, the analysis of atmospheric and oceanographic conditions in the region shown above depicts a clear picture of the circumstances that lead to the unusual arrival of P. Physalis colonies to the GoC shores in June and July 2019.

We hypothesize that a major driving of the full incident seems to be the very peculiar (for the summer period) atmospheric pressure distribution in the ROI, with persistent low mslp in those two months of 2019 ( Figure 2 ) associated to a south-west displacement of the Azores high from its usual summer position ( Supplementary Figure 2 ). Actually, the summer 2019 atmospheric situation resembles that of the typical winter situation in the ROI, with low mslp anomalies on the Atlantic side of the ROI indicative of stormy conditions passing through the region. Such low pressure atmospheric systems typically move from west to east and are associated with more frequent and intense westerly winds at the surface connected to the Ferrel cell circulation. These westerlies were particularly conspicuous in the northern area of the GoC ( Figure 4 ) where wind anomalies for the two studied months were clearly directed towards the coasts of Cádiz.

However, this is not the only effect mslp had on the oceanographic conditions of the GoC – Alboran Sea region. The persistent low mslp over the Western Mediterranean Sea ( Figure 3 ) may be responsible for the strong acceleration of the AJ entering the Mediterranean through the Strait of Gibraltar ( Figure 5 ). This is a barotrophic phenomenon already known and described in the literature (e.g., Candela et al., 1989; Garcia-Lafuente et al., 2002a; Macias et al., 2016; Bolado-Penagos et al., 2021). What is less known, however, is the impact it has on the currents along the northern shores of the GoC.

As shown by the scatter plot on Figure 6A , there is a moderate but significant correlation between the zonal velocity of the AJ and the zonal intensity of the current on the northern shores of the GoC. This relationship has also been described by a very recent study (Sirviente et al., 2023) carried out using high frequency radar (HFR) measurements. The authors combined HFR analyses with high resolution modelling experiments in order to identify the Empirical Orthogonal Functions (EOFs) of the currents in the GoC to discretely study the subinertial signal (hence the orthogonal modes) and their forcing agents. They found significant correlations between the mslp in Liguria and the wind components in the GoC and Strait of Gibraltar.

This relationship between the currents in the northern coast of the GoC and the AJ velocity could be understood if the general circulation pattern in the GoC is considered. There, the water flows around the Cape San Vincent in southern Portugal and mainly follows the isobaths as an eastward flowing current towards the entrance of the Strait itself (García Lafuente and Ruiz, 2007, see also sketch in Figure 1B ). If the water flowing into the Mediterranean is accelerated within the Strait of Gibraltar by the barotrophic effect of the low mslp described above, this will imply a general increase of the eastward water movement in the GoC, particularly on its northern shores, where the water already moves eastwards (e.g., Sirviente et al., 2023).

It is also worth considering that wind anomalies were directed eastward in this northern region of the GoC ( Figure 4 ), aiding the acceleration of the water along the coastal fringe. The concurrence of both elements (hydrological and atmospheric) surely contributed to the fact that eastward velocities in the northern GoC for the months of June and July 2019 were identified as outliers in the statistical 27 year record analysis ( Figure 6C ).

Concomitant with the increased westerly winds and eastward surface current during summer 2019, higher than normal waves (+10/20 cm Hs) are simulated for the northern coasts of the GoC ( Figure 7 ). It is, therefore, understandable that P. physalis colonies floating in the surface waters of the region were dragged to the coasts of the northern GoC (marked as green circles in Figure 7 ) where their arrival created alarm among beachgoers and forced beach closures.

One question remains open, however: why P. physalis colonies were close to the European continental margin and if the particular atmospheric and oceanographic conditions of the North Atlantic during summer 2019 had any influence on that circumstance. To evaluate this hypothesis, surface current data for the spring months (April and May) where downloaded from the global product available at CMEMS as described in section 2.2.

The mean currents for the spring period (April-May) in the North Atlantic are presented in Figure 8A . The strong Gulf Current flowing northwards along the American coasts, the eastward North Atlantic Drift and the Azores Currents are clearly identifiable. On the eastern boundary of our ROI ( Figures 8C, D ) the presence of eastward currents and typical westerly winds could be seen in the area between 40° and 50° north (Headlam et al., 2020).

During the months of April and May 2019, the general circulation patterns described above are reinforced in the overall North Atlantic ( Figure 8B ) with stronger eastward currents and westerly winds in the eastern boundary of our ROI (between 40-50° north) ( Figures 8E, D ). Furthermore, a report from an opportunistic vessel (fishing boat) signaled the presence of P. physalis on the position of the white cross in Figure 8B at the beginning of June (Prieto, pers. comm.).

We do not have confirming data, but it is likely that the region marked with the white cross in Figure 8B also contained P. physalis colonies during the spring months, as it corresponds to the typical open-ocean environment these organisms like to occupy (Mapstone, 2014). From there, and given the stronger east-flowing currents and the enhanced westerly winds (associated to the more regular presence of storms as indicated by the mslp anomalies), the colonies were likely transported towards the vicinity of the western European coasts, where they followed the main currents and winds towards the GoC and the Cádiz beaches as described by Prieto et al. (2015) for a winter event in 2010.

The area highlighted in Figure 8F (and Figure 1A ) has been identified as a potential source region for P. physalis colonies ending up on Basque shores (Ferrer and Pastor, 2017) or arriving to Irish beaches (Headlam et al., 2020). Other reports point to the North Atlantic Drift as the main carrier for sea turtle juveniles (e.g., Hays and Marsh, 1997) arriving from the eastern coast of the United States (Monzón-Argüello et al., 2012), while others indicate that the Gulf Stream and North Atlantic Drift could carry tropical seeds from the Sargasso Sea to the European coasts (Quigley and Gainey, 2018). All of the aforementioned reports support our hypothesis that P. physalis colonies found on the Cádiz shores in June/July 2019 came from the tropical central Atlantic carried by the stronger than average surface currents and westerly winds prevalent during the previous spring (April/May).

A notable aspect of the 2019 beaching events concerns the absence of P. physalis colonies in the Mediterranean waters of the Alboran Sea. Notably, previous research (Macias et al., 2021) has demonstrated that when large-scale aggregations of Physalia are detected in the Gulf of Cadiz (such in winters of 2010, 2013 or 2018), a significant proportion can be transported through the Strait of Gibraltar into the Mediterranean Sea. The lack of stranded colonies in the Alboran Sea during the summer 2019 event may be attributed to the relatively low number of individuals involved in this particular incident.

When the summer event of 2019 occurred, there was a debate regarding whether this was a result of global change causing P. physalis to inhabit areas closer to the EU coasts and whether such events could become more frequent in the future. Our study indicates that there are currently no signs of P. physalis altering its habitat distribution and that the particular event in summer 2019 was the result of particular atmospheric and oceanographic conditions in the North Atlantic. However, there are indications that due to global change, similar situations may occur more frequently in the future.

Analysis of historical reanalysis (ERA5) wind fields in the area of interest indicates that over the last 43 years (1980 – 2022), wind intensity in the ‘source area’ (identified in Figure 8F ) has, indeed, increased ( Table 1 ). In addition, the zonal (u10) component of the wind field shows a trend that is one order of magnitude larger than the trend calculated for overall wind velocity ( Table 1 ). It is also noteworthy that the trend analysis conducted only for the summer months (JJA), indicates increasing trends that are 50% (for total intensity) to 166% (for zonal intensity) larger than those observed for the annual values ( Table 1 ).

Although the computed trends are not particularly large and there is substantial inter-annual variability, with only the summer u10 trend being significant at 95% confidence, this analysis suggests that stronger westerlies in the source area are becoming more prevalent, particularly during the summer months. This could create more favorable conditions for the arrival of P. physalis colonies to the EU coasts during summer.

In addition to the likely increase in zonal wind speed, there have been reports of an increased risk of coastal flooding (Vousdoukas et al., 2018) as a consequence of increased storminess in the ROI. Moreover, there has been an increase of tropical cyclone exposure in the region during the last few decades (Jing et al., 2024). The increase in these types of atmospheric phenomena could provide another mechanism for the enhanced arrival of P. physalis colonies to European coasts and beaches.

Forecasting future wind intensity and direction is extremely challenging due to inherent limitations and shortcomings of global circulation models (Lowe and Gregory, 2005; Marcos et al., 2011). However, for our ROI, there are certain patterns and atmospheric links that can be used to provide insights into plausible future conditions. One such condition is the correlation between wind intensity and the NAO index in the region (see Supplementary figure 6). Our results align with previous studies (e.g., Marcos et al., 2011; Mentaschi et al., 2017) and essentially indicate that with more positive NAO values, there is an increase in wind (and westerlies in particular) velocities in the source area for P. physalis (identified in Figure 8F ). Indeed, our studied year 2019 is in the middle of a sustained positive NAO phase covering the period 2015 – 2021.

There are indications that the NAO will tend to become more positive in the future (Mentaschi et al., 2017; Shimura et al., 2013; Woodwort et al., 2007), and that warmer surface Atlantic waters will accelerate this trend (Rodwell et al., 1999). More positive NAO values in the future could be associated with increased westerly intensity according to the past trends (Supplementary Figure 6 ).

Despite the uncertainty in the future projections mentioned above, there are already some suggestions that wind intensity (McInnes et al., 2011), and the associated significant wave heights (Hemer et al., 2013) will likely increase in the coming decades in the western European coasts. In fact, McInnes et al. (2011) even indicate that 90% of the investigated models show a substantial increase in wind intensity in our ROI, particularly during the summer months.

If the projected changes of the NAO and associated wind and wave fields into the future materialize, the conditions encountered in summer 2019 could become more frequent. Consequently, the probability and risks of human interaction with drifting P. physalis colonies in the summer months will also increase. This dictates the need for better monitoring and warning systems to be put in place in order to minimize economic impacts and health threats for the coastal communities of Western Europe.