The physical–chemical parameters were sampled across the south of La Palma, located in the west of the Canary Islands (north-eastern Atlantic Ocean) (Fig. 1a, File S1 in the Supplement). The sampling took places between a 0 and 2 m depth, at three different times (March 2018, December 2018 and June 2019) and during low and high tide when it was necessary to assess the continuity of the natural emissions (Fig. 1b, Appendix A). Following the previous studies in the area (Hernández et al., 2016; Pérez, 2017; González-Delgado et al., 2018a, b; Hernández et al., 2018; Viotti et al., 2019), a sampling network was created for the first time. It is formed by the following sites: Playa del Faro, Los Porretos and surroundings (that together with the Las Cabras site are known as the Punta de Fuencaliente system or PFS), Playa Echentive and the two Echentive lagoons (Fig. 1c).

Scuba diving was used for sampling all bottles except in the reference station off Playa del Faro (Fig. 1cA), where a CTD rosette was used. For the scuba sampling, the bottle was previously rinsed three times at the sampling location and then the bottle was immersed with the mouth down and turned at 1 m depth for sampling. Samples were poisoned with 100 µL of saturated HgCl2 solution, sealed, kept in darkness and analysed in the lab. In March 2018, this was performed on the same day, while in December it was performed 2 d later. For pH, 100 mL borosilicate glass bottles were filled with seawater.

In March and December 2018, the total dissolved inorganic carbon concentration (CT), total alkalinity (AT), pH, salinity and temperature were measured, whilst in June 2019 only the pH and temperature were measured. Total alkalinity and CT were determined by potentiometric and coulometric methods, respectively, using a VINDTA 3C system (Mintrop et al., 2000). The calibrations were made using certified reference material batch no. 163 (González-Dávila et al., 2007). The pH was measured at a constant temperature of 25 ∘C within 1 h of sampling, using an Orion pH meter with a combined Orion glass electrode (pHT,is). The calibration was performed on the total seawater scale using a Tris artificial seawater buffer (salinity 35) according to the Guide to Best Practices for Ocean CO2 Measurements (Dickson et al., 2007, SOP 6a).

Salinity and temperature were measured in situ using a handheld conductivity meter (Hanna Instruments HI98192). Furthermore, 200 mL salinity bottles were measured in the laboratory within 2 d and using a high-precision Portasal salinometer, accurate to ±0.001. The pH under in situ conditions, the partial pressure of carbon dioxide (pCO2) and the saturation states of calcium carbonate forms (Ω aragonite and Ω calcite) were determined from AT and CT data using the CO2SYS program (Pierrot et al., 2006).

Atmospheric CO2 concentrations used for flux calculations were those measured at the Izaña station on the island of Tenerife (IZO site and available in the World Data Centre for Greenhouse Gases).

We used the linear interpolation method to represent the AT, CT, pHT,is, Ω aragonite and calcite parameter measurement when anomalies were found.

Temperature and salinity on Playa del Faro and Los Porretos do not present major changes between the different time points (File S2). During March 2018, Playa del Faro had an average temperature of 19.00±0.20 ∘C with colder values of 18.70 ∘C near the shore; Los Porretos was not measured at this time. In December 2018, both Playa del Faro and Los Porretos presented an average temperature of 21.50±0.02 ∘C. However, salinity values present a minor diminution from 37.05 to 36.51 on Playa del Faro and from 37.05 to 36.07 on Los Porretos (File S2). Both sites presented colder and slightly less saline water near the coast. Regarding the Echentive lagoons, only the biggest lagoon was measured, where the salinity varied from 31.00 to 32.00 units (File S2). The same lagoons presented warmer temperatures than the coastal waters during June 2019, 26.40±0.70 and 22.00±0.10 ∘C, respectively.

In both studied shore areas of the PFS (Playa del Faro and Los Porretos) the parameters of the carbon dioxide system, pHT,is (Fig. 2a, b), AT, CT, and Ω aragonite and calcite (Figs. 3, 4B), were strongly affected by the entrance of the SGD with less salinity.

The CO2 flux was calculated for Playa del Faro. We assumed two endmembers, the open ocean endmember and the SGD endmember. Soler-Liceras (2007) discovered an aquifer near this area with brackish water (salinity of 30). Considering the bathymetry, the volume occupied by seawater was 19 700 m3. We also assumed that groundwater discharge only occurred at low tide. The average salinity changed from 36.93 (equivalent to 745.8 t of sea salt) at low tide to 37.02 at high tide (747.5 t of sea salt). The decrease in salinity at low tide could be accounted for by the emission of 57 m3 of brackish groundwater.

The brackish groundwater was also responsible for the AT and CT changes (Fig. 3a, b). Alkalinity increased by 219 µmol kg-1 from high tide (2465 µmol kg-1) to low tide (2684 µmol kg-1). Considering 57 m3 of brackish water, 4.40 kmol of alkalinity was required; therefore, the brackish groundwater had an AT concentration of 76 mmol kg-1. Similarly, the CT on the beach increased by 333 µmol kg-1, from high tide (2190 µmol kg-1) to low tide (2523 µmol kg-1). The brackish water caused the increase of 6.7 kmol of inorganic carbon on the beach and, therefore, had an endmember concentration of 116 mmol kg-1.

Considering the in situ temperature (20.67 ∘C), the pHT,is decreased by 0.25 from 8.01 at high tide to 7.76 at low tide. This meant that the acidity increased by 80 %. This pH reduction meant that the water discharged on the beach had a pH of 5.57. The medium partial pressure of carbon dioxide for the area increased from 459 µatm at high tide to a value of 988 µatm at low tide. Considering an average wind speed at the beach of 7 m s-1 (https://datosclima.es/Aemethistorico/Vientostad.php, last access: 5 March 2020), Playa del Faro acts as a strong source of CO2, emitting 5.70 mmol CO2 m-2 d-1 at high tide and increasing by an order of magnitude at low tide (57 mmol CO2 m-2 d-1; Wanninkhof, 2014). Consequently, Playa del Faro with its small area of only 0.01 km2 is responsible for an atmospheric CO2 emission flux varying between 2.80 and 28 kg CO2 d-1.

Although CO2 emissions on the Fuencaliente coast had already been detected (e.g. Hernández et al., 2016; Viotti et al., 2019), this is the first time that this naturally acidified system has been described chemically and physically. Previous works have focused on specific questions; Hernández et al. (2016) published for the first time the presence of CO2 SGD in Fuencaliente, specifically on Las Cabras beach. Later, in the thesis by Pérez (2017), as well as in the conference papers by González-Delgado et al. (2018a, b) and in the article by Viotti et al. (2019), new points of acidification were discovered on Playa del Faro and Los Porretos. However, in none of them was a chemical characterization of the whole area made as it was here. Our results reveal the continuous influence of brackish water discharge in the acidification process of the Punta de Fuencaliente system (PFS), which had been missed before (Fig. 5). Similarly to aerial remnant volcanic activity on La Palma that generates high CO2 diffusive atmospheric concentration (Padrón et al., 2015), submarine remnant volcanic activity causes the acidification process found here, as indicated by the chemical composition of the groundwater analysed, which is less than 200 m from the coast (Soler-Liceras, 2007). The activity of this SGD is comparable with other CO2 vent and seep systems worldwide (references within González-Delgado and Hernández, 2018). Moreover, the presence of the acidic water flow of La Palma also has a slight resemblance to the acidification phenomenon found in Mexico, originating from a karstic groundwater discharge (Crook et al., 2012). Furthermore, the highly alkalized and bicarbonate waters found in Echentive lagoons are an artefact of water discharge from the hydrothermally affected aquifers of the area (Soler-Liceras, 2007), as found in Las Cañadas del Teide, in Tenerife (another island of the same archipelago) (Marrero et al., 2008).

In the PFS there is a decrease in salinity due to brackish water discharges. Hence, there is a constant filtration of brackish acidified waters through highly permeable volcanic rocks (Carracedo et al., 2001; Marrero et al., 2008), with chemical features due to underground volcanic activity, such as a 5.57 pH and a concentration of 76 mmol kg-1 of AT and 116 mmol kg-1 of CT. However, the effect on the surrounding seawaters depends upon tidal pressure and, more likely, upon other oceanic forces such as wind and waves (Moore, 2010; Mulligan et al., 2019).

In the case of the PFS, the water with lower salinity (36.79–36.45) and high concentrations of CT and AT affect the surroundings, decreasing the seawater pH by up to 0.8 and reducing the carbonates' saturation state by up to 1.1 for calcite and 0.7 for aragonite. This situation generates a carbon imbalance affecting carbonated organisms, especially those that precipitate aragonite on their calcareous structures (Kroeker et al., 2010). When the saturation values are below 1, the formation of carbonates is not thermodynamically possible, although certain species require much higher saturation levels (Kroeker et al., 2010). The calcifying organisms that could live in these acidified areas may present weaker shells, skeletons and/or others solid structures, as we have recently observed in the mollusc Phorcus sauciatus (Viotti et al., 2019), as well as in other calcifying organisms (Pérez, 2017). This excess of CO2 has also modified the community composition and trophic structure, causing a loss of ecological and functional diversity in the benthic marine ecosystem (González-Delgado et al., 2021).

In the case of Echentive lagoons, the anomaly is amplified due to a lower tidal influence and insulation. These acidified lagoons, which are at around a 200 m distance from the coast (Fig. 1), have a salinity of 32 and CT and AT concentrations 5 times higher than normal ocean values. The CT and AT concentrations are so high that they compensate for the decrease in pH with the content of carbonates in the water. These singular characteristics create a unique marine ecosystem. The environment is dominated by a biofilm of microorganisms, predominantly microalgae, cyanobacteria and diatoms (Sangil et al., 2008) and probably other bacteria and fungi. Nonetheless, some marine invertebrates persist, such as the common errant polychaete Eurythoe complanata and the anemone Actinia sp. (Sangil et al., 2008). A more in-depth physiological study of these species could help us to better understand their adaptation process to these conditions and to give insights into what we might expect in future ocean acidification conditions, especially in the PFS area.

The natural CO2 gradients south of La Palma have been characterized from shore to offshore, varying for CT from 2120.10 to 3794.00 µmol kg-1, for pH from 7.12 to 8.07, for Ω aragonite from 0.71 to 3.28 and for Ω calcite from 1.09 to 5.02. This high local variability is in line with other acidified natural systems. For example, the CO2 vent of Ischia (Italy) has pH levels from 6.07 to 8.17, Ω aragonite from 0.07 to 4.28 and Ω calcite from 0.11 to 6.40 (Hall-Spencer et al., 2008). The one from the island of Vulcano (Italy) has pH values between 6.80 and 8.20, Ω aragonite from 1.49 to 4.65, and Ω calcite from 2.28 to 7.00 (Boatta et al., 2013). Meanwhile the CO2 seeps from Papua New Guinea have pH levels between 7.29–7.98, Ω aragonite between 1.2–3.4 and Ω calcite between 1.36–5.12 (Fabricius et al., 2011). Those from Shikine-jima (Japan) have pH values between 6.80 and 8.10, Ω aragonite from 0.20 to 2.22, and Ω calcite from 0.30 to 3.45 (Agostini et al., 2015). Although these systems are far from being perfect predictors of the ocean future due to their chemical variability and physical limitations, they have proven to be important tools for the study of ocean acidification (Foo et al., 2018; González-Delgado and Hernández, 2018; Aiuppa et al., 2021). These naturally acidified systems, such as the Punta de Fuencaliente system (PFS), can be used as natural analogues of climate change scenarios predicted by the IPCC (2014) (Fig. 6). Therefore the PFS can be considered a very useful spot for large-scale and long-term adaptation experiments, as seen in other CO2 systems (e.g. Ricevuto et al., 2014; Uthicke et al., 2019). Moreover, the acidified system of La Palma is highlighted by the absence of bubbling, since the volcanic degasification takes place in the aquifers and not directly on the coast as in other acidified systems of volcanic origin (e.g. Hall-Spencer et al., 2008; Fabricius et al., 2011) (Fig. 5). This could give us new insights into the effect of acidification in situ avoiding the effects of bubbling (González-Delgado and Hernández, 2018). Nevertheless, several caveats for future prediction experiments should be considered, here as well as in other naturally acidified systems, especially those related to increased alkalinity values in the submarine discharge.

First, there is a clear tidal influence; this is an important force that controls the acidified brackish water discharges. Although a fluctuation in the emission is observed, normal ocean conditions can occur for a short time, about 2–4 h d-1, during high tide and depending on the oceanic conditions (Viotti et al., 2019). The pHT is severely affected by the location, reaching down to ∼7.2 in the emission points, so a careful selection of the study sites is recommended, depending on the study objectives (Fig. 6). This tidal phenomenon has also been reported in other acidified natural systems such as Puerto Morelos in Mexico (Crook et al., 2012) and Ischia (Kerrison et al., 2011). However, the pH time fluctuation can be used to our advantage, as a daily and seasonal fluctuation in the pH is normal in coastal habitat environments (Hofmann et al., 2011). So, it could be considered very useful to incorporate pH variability in ocean acidification studies as environmental fluctuations that can have a large impact on marine organisms (Hofmann et al., 2011).

Second, one of the most common concerns with CO2 seeps and SGD areas is the presence of other gases or elements associated with volcanic emissions, such as nitrogen (N2), mercury (Hg) or methane (CH4) (e.g. Fabricius et al., 2011; Boatta et al., 2013; Aiuppa et al., 2021). Although there are no traces of the presence of volcanic elements such as methane or sulfates that are harmful to marine organisms in the seawater of the PFS (Hernández et al., 2016), there is an extra supply of different elements such as Mg that comes from groundwater (Soler-Liceras, 2007). Groundwater has 10 times more magnesium than normal, but when mixed with seawater, the supply is considerably lower compared to CO2. Nevertheless, Mg plays an important role in the calcification of marine organisms that have magnesite–calcite, such as echinoderms (Weber, 1969) and some Bryozoa species (Smith et al., 2006). Similarly, Hernández et al. (2016) found an increase in silicates in the nearby area of Las Cabras. In these cases, Si could participate in the calcification of diatoms (Paasche, 1973) as well as of many sponges (Smith et al., 2013). The increase in these essential elements for certain calcifying species can allow their survival and growth in the PFS while buffering the effects of acidification (Smith et al., 2016; Ma et al., 2009). Therefore, measurements of other metals in seawater should be considered in the following studies.

The high concentration of bicarbonate in the brackish waters also implies an extra contribution of alkalinity and carbonate that can buffer the effect of acidification in the area, so it is necessary to take this into account when making predictions of the future. These values together with calcium content are especially important factors in the case of the saturation state for both calcite and aragonite, which shows high values for seawater with low pH values. Hence, despite the fact that we are dealing with a subtropical ecosystem, the values obtained in both saturation states are more similar to the predictions for a tropical ecosystem, such as the values found in Papua New Guinea seeps (Fabricius et al., 2011; IPCC, 2014).

Finally, the area is not very large and only one type of rocky benthic habitat, the most typical community of the Canary Islands, is present at the PFS (Sangil et al., 2018). Therefore, all conclusions derived from this acidified system should be interpreted with caution and acknowledging local effects. Hence, it is crucial to establish a collaborative network of researchers who are working in other naturally acidified systems worldwide to have a more realistic interpretation of future ocean scenarios.

The Echentive lagoons are an oversaturated carbonate system. Like hydrothermal alkalinity vents (Martin et al., 2008), they could help us to understand early life on Earth from the Precambrian, 4000 million years ago, when the atmosphere was rich in CO2 (Kasting, 1993; Nakamura and Kato, 2004) (Fig. 6). These studies could allow us to disentangle the adaptation and evolution of marine life to the changing carbonate conditions over time (Gattuso et al., 1998).

Additionally, to our knowledge, this is the first time that a brackish water discharge altered by volcanic activity has been studied. Each studied beach with a contribution of 150 t CO2 yr-1 becomes an important source of carbon into the sea. Correspondingly, Playa del Faro is emitting 28 kg CO2 d-1 in each tidal flow to the atmosphere. This may seem very scarce compared to volcanic eruptions such as the most recent in the Canaries that occurred on the neighbouring island, El Hierro, in 2010, which was emitting 6.0×105±1.1×105 kg d-1, and now the emissions of the PFS are unappreciated (Santana-Casiano et al., 2016). However, the flux of CO2 from La Palma seems to have started before the islands were conquered in 1493 (Soler-Liceras, 2007), being in a more advanced degassing phase than El Hierro with fewer emissions, and continued over time. Therefore, if we consider its timescale, La Palma becomes a significant CO2 source. For all these reasons, the PFS and the lagoons are an interesting area for future hydrological and oceanographic research, helping in new studies focusing on groundwater fluxes, the oceanic water cycle and oceanic carbon fluctuation (Moore, 2010; Santana-Casiano et al., 2016; Mulligan et al., 2019).