The studied rhodolith bed, dominated almost exclusively by Phymatolithon lusitanicum (Carro et al., 2014; Pardo et al., 2014; Peña et al., 2015), has an estimated area of 3 km² and extends from 13 to 25 m depth. It is located at 4.7 nautical miles offshore of Armação de Pêra, in the Algarve, Southern Portugal (37°01′N, 8°19′W; Figure 1A). The region is influenced by coastal counter currents due to opposite alongshore winds that cause sharp current inversions in short time scales (<2–15 days), resulting in changes from upwelling- to downwelling-favorable conditions, especially during May-Oct (Sánchez et al., 2006; Garel et al., 2016). As shown by in situ data, these quick changes in coastal currents cause strong daily and weekly fluctuations in temperature and light availability at the location of the rhodolith bed (Figures 1B–D), but also strong temperature fluctuations along the day, with recorded variations of up to 3–4°C (Supplementary Figures 1A–C). The temperature variation is accompanied by a large variability of maximum in situ daily light intensities that range between no detectable light to maximum recorded values of 240 μmol quanta m^–2 s^–1 (Supplementary Figure 1D).

Climatological data from 1982 to 2020, based on satellite data of sea surface temperature (SST) retrieved from the Marine Heatwave Tracker (Schlegel, 2018) show that the southern Portuguese region experiences the influence of marine heatwaves (MHWs). Our analysis of the MHWs detected since 1982 indicate that not only the frequency, but also the severity of these events exhibits an increasing trend (Supplementary Figure 2A). In the 2010’s decade, 31 MHW events were recorded, 7 of which are considered strong events, with maximum temperature rises above the climatological threshold ranging between 1.1 and 4.3°C. The occurrence of MHWs shows a seasonal variability, with the majority of the events occurring during summer and autumn (Supplementary Figure 2B). Most of the MHWs are of short duration (<2 weeks; Supplementary Figure 2C), though there is a large variability regarding the time passed between events. Often 1–4 weeks pass between events, while longer time intervals (>11 weeks) are also quite frequent (Supplementary Figure 2D). These occurrences of thermally anomalous events, which are identified based on SST data, are consistent with strong in situ temperature increases (Figures 1B–D).

Individuals of P. lusitanicum were collected by SCUBA at 22 m depth in October 2020 (recorded in situ temperature: 17°C) and transported in coolers to the Center for Marine Sciences marine station Ramalhete (∼50 km from Armação de Pêra). Here, the rhodolith were maintained in seawater flow-through indoor tanks at 17°C and 50 μmol quanta m^–2 s^–1 (11 h light: 13 h dark; LEDVANCE Flood LED IP65, 50 W, 6500 K, Ledvance, Germany) during 2 weeks for pre-acclimation to the tank conditions. The chosen light intensity represents an intermediate maximum light level, based on previous records of natural light conditions at the location of the rhodolith bed (Supplementary Figure 1D).

To test for potential impacts of MHWs (sensu Hobday et al., 2016) on rhodolith physiology and the influence of recurrent thermal stress on those impacts, three different treatments were designed, (1) Control- kept at 17°C during the whole experiment (Figure 2A), (2) HT- simulation of a high-temperature event for 12 days, with a gradual temperature increase/decrease of 1.5°C day^–1 and T_max = 24.5°C for 4 days (Figure 2B). With 3.5°C above the climatological threshold of the sampling location and month (Oct- 21.2°C), this treatment simulated a MHW lasting 8 days. And (3) 2HT- simulation of two high-temperature events: the 1st event represented a moderately naturally occurring temperature increase during 12 days (max. 21°C for 6 days, gradual temperature increase/decrease: 1°C day^–1), while the 2nd event simulated the same high-temperature event as in HT (Figure 2C).

During the experiment, the rhodoliths were kept in a set-up consisting of 150–L treatment tanks (n = 5 per temperature treatment), connected to a seawater flow-through system. Before entering the experimental tanks, the seawater was pumped into a header tank (2000 L), in which the temperature was kept at 17°C by a heat pump (i-Komfort, Kripsol, Hayward), after passing through a preliminary mechanical filtration (sand filters), two in-line cartridge filters (10–20 μm and 5 μm) and two UV filters (16 and 8 W). From there, the water entered the experimental tanks, each equipped with a submersible air pump to ensure water circulation and an aquarium heater that was used to maintain and manipulate the temperature in the different treatments, which was monitored by a temperature data logger (HOBO, Onset, United States) positioned in each tank (Supplementary Figure 3).

To determine the treatment effects on rhodolith physiology, the responses in photosynthetic, respiratory and calcification rates of the same rhodoliths (n = 5 per treatment, n = 1 per tank) were monitored over the entire experimental period. For this purpose, short-term incubations of the individuals were performed at nine different time points, including measurements at the beginning of the experiment, during and in-between the different high-temperature event simulations and after a recovery period of 7 days (Figure 2).

Individual rhodoliths were incubated (n = 5 per treatment and time point) at their respective treatment temperature with filtered seawater (0.45 μm) in sealed custom-made chambers (V = 70 mL) with internal circulation, provided by a magnetic stirrer. Firstly, the rhodoliths were exposed for 1 h to 50 μmol quanta m^–2 s^–1 (Lexman LED, 20 W, 6500 K) to determine photosynthetic and light calcification rates and subsequently, they were incubated in darkness for another hour to determine respiratory and dark calcification rates. Previous tests assured that 1 h-incubation time provides a high enough signal to noise ratio to allow detection of changes in both oxygen concentration and alkalinity, while ensuring that pH changes during incubations were <0.1 units. At the beginning and end of the incubation, the oxygen concentration was measured with an oxygen meter (Fibox 4, PreSens, Germany). Control incubations (without rhodolith samples) were also performed, showing that oxygen changes in the incubations water, e.g., due to microbial activity, were neglectable. Seawater samples were taken at the beginning and end of every incubation, poisoned with HgCl₂ and stored in borosilicate tubes (two tubes per incubation chamber, V = 25 mL each) for later estimation of calcification rates. At the end of the experiment the rhodoliths were oven-dried for 48 h at 60°C and the estimated dry weight was used to normalize metabolic and calcification rates.

The calcification rates of the rhodoliths were determined from alkalinity measurements of seawater samples before and after the different incubations, using the alkalinity anomaly principle based on the ratio of two equivalents of total alkalinity per each mole of CaCO₃ precipitation (Smith and Kinsey, 1978). For alkalinity (TA) measurements, duplicate analyses of each sample were performed, using the Gran titration method (Hansson and Jagner, 1973; Bradshaw et al., 1981). The samples were titrated with HCl 0.1 M, using an automated titration system (Titroline 7000, SI Analytics, Mainz, Germany), coupled to an autosampler (TW alpha plus, SI Analytics, Mainz, Germany). Data were captured and processed with a computer, using Titrisoft 3.2 software (SI Analytics, Mainz, Germany). For quality control, a certified reference material of known total alkalinity was used to calibrate the method (CRMs, Batch No. 160; supplied by the Marine Physical Laboratory, Scripps Institution of Oceanography, United States).

During the experiment, the physiological responses of the same rhodolith individuals were tracked over time and thus, the data of every treatment were analyzed using one-way repeated measures ANOVA, with time as the within-subjects factor. In cases where significant differences were found, a Tukey’s HSD post hoc test was conducted to determine significant groupings. A priori, data were tested for normality and heteroscedasticity, using the Shapiro-Wilk and Levene’s tests, respectively. All statistical analyses were performed using STATISTICA 7.

Most studies on the physiological responses to temperature increase in coralline algae have been performed either by comparing their seasonal performance (i.e., winter vs. summer) or by assessing their responses to experimentally-induced temperature increases above a mean control or ambient temperature (Martin and Hall-Spencer, 2017; Cornwall et al., 2019). In contrast, information on coralline algal tolerance/susceptibility to MHW events is almost non-existent (but see Rendina et al., 2019; Ragazzola et al., 2021). Regardless of the approach, the available evidence suggests that coralline algae express a large variability in their responses to thermal stress, which may be related to species-specific differences, but also, as pointed out by Cornwall et al. (2019), to temperature variability within experimental set-ups and/or the past thermal history of the organisms.

In the present study, the acute thermal stress treatment, simulating a realistic MHW event (HT-treatment), showed little and only short-term negative effects on rhodolith net primary production, but a strong decrease in net calcification. The former was related to an initial, though not significant, increase in respiration, a response often found in coralline algae during exposure to higher temperatures (Steller et al., 2007; Vásquez-Elizondo and Enríquez, 2016). However, within a few days at the high temperature P. lusitanicum was able to adjust its metabolic demand, allowing the species to revert to the prior daily net primary production rates (Figure 5C). On the other hand, rhodolith calcification was strongly impacted by the MHW simulation, due to a decline in both light and dark calcification and consequently, daily net calcification (Figure 5D). The stronger impact on calcification, compared to the metabolic processes, has also been reported in other coralline algae (Vásquez-Elizondo and Enríquez, 2016; Schubert et al., 2019) and indicates an uncoupling of physiological processes, such as photosynthesis and light calcification under thermal stress conditions. On the other hand, the ∼40% recovery of daily net calcification at the time when temperature was back to the pre-MHW level suggests that recovery started already during the gradual temperature decrease at the end of the MHW. This, together with the 90% recovery 1 week after the event, indicates a high recovery capacity of P. lusitanicum and hence, species’ resilience in the face of MHW-associated heat stress, most likely related to an adaptive tolerance to the strong and fast temperature fluctuations of its natural habitat.

The influence of an organism’s thermal history on its response to acute thermal stress has recently been studied in coralline algae through prior pre-acclimation to higher or varying temperatures (naturally or experimentally) before exposure to severe heat stress. The experimental evidence indicates contradictory responses, either demonstrating reduced physiological performance (Rendina et al., 2019; Page et al., 2021) or no substantial effects on the studied species in response to acute heat stress (McCoy and Widdicombe, 2019), probably related to differences in the experimental design or the thermal history of the species. Experiments involving long-term acclimation of tropical and temperate species to an increase in mean temperature prior to acute thermal stress exposure reported negative effects (Rendina et al., 2019; Page et al., 2021), while an acute thermal stress did not exert any negative effects on individuals of the temperate species Ellisolandia elongata, collected from a strong intertidal gradient (McCoy and Widdicombe, 2019).

Direct comparison among studies regarding the responses of coralline algae to MHWs is very difficult due to (1) the currently extremely scarce number of studies, (2) the often large differences in the experimental design, e.g., MHW intensity and duration, and (3) the differences in the specific time points during the MHW (and after a recovery period), when physiological assessments were performed. To date, there are only two studies on temperate articulated coralline algae that have assessed the physiological responses to realistic scenarios of MHWs, based on climatological data and thresholds for the respective study location (Rendina et al., 2019; Ragazzola et al., 2021). In general terms, the lack of significant responses of P. lusitanicum’s metabolic rates to the MHW-associated temperature increase found here agrees with those observed in the temperate articulated coralline alga Ellisolandia elongata (Ragazzola et al., 2021), while a significant response has been reported in the articulated Corallina officinalis (Rendina et al., 2019). The latter species exhibited a decrease in gross photosynthesis, as well as respiratory demand immediately after a simulated MHW event, without signs of recovery after 2 weeks. Also, the calcification response of these two articulated coralline species differed greatly from that of the here studied rhodolith, as the former did not show any responses in either, light or dark calcification in response to a simulated MHW event or responded with a significant decline in dark calcification only (Rendina et al., 2019; Ragazzola et al., 2021). These contradictory findings are not surprising, considering the large variability in magnitude and direction of thermal stress responses in this group that in large part may be related to inherent species-specific differences (Martin and Hall-Spencer, 2017; Cornwall et al., 2019), also supported by a recent study showing that even different rhodolith species from the same location express differences in their responses to seasonal temperature variations (Qui-Minet et al., 2021).

The comparison of the rhodolith responses upon exposure to a realistic MHW scenario, with and without prior moderate thermal stress, clearly showed that (1) a temperature increase consistent with naturally occurring rises in temperature boosted the rhodolith’s net primary production, and (2) this prior exposure mitigated MHW impacts on rhodolith calcification, DNP and DNC, and allowed for a quicker recovery of calcification and derived DNC, indicating the species’ capacity to acquire thermal stress memory that lasted for at least 1 week.

Regarding the first point, our results were mostly consistent with P. lusitanicum’s physiological responses to temperature levels within the species’ natural range (16–22°C) after a month-long acclimation, which showed that similarly to our study gross photosynthesis and light calcification were significantly increased at 22°C (Sordo et al., 2019). These findings are also supported by evidence from previous field and experimental studies, showing that the response of coralline algae to rising temperature is dependent on the magnitude of increase. A temperature rise within the range experienced in natural habitats, can increase growth, photosynthesis and calcification in both temperate and tropical species (Adey, 1970; Adey and McKibbin, 1970; Martin et al., 2006; Steller et al., 2007; Martin and Gattuso, 2009), while further increasing temperature results in a decline (Adey, 1970; Adey and McKibbin, 1970) and can have severe detrimental effects (Martin and Hall-Spencer, 2017).

The lower MHW impact on rhodolith physiology due to prior exposure to moderate temperature increase was consistent with the well-studied concept in terrestrial plants that appropriate responses to recurrent environmental fluctuations rely on the ability of an organism to keep “a memory” of prior exposure to certain conditions for a certain length of time (Bruce et al., 2007; Trewavas, 2009). Prior exposure to moderate heat stress led to acquired thermotolerance (or thermal priming) that enhanced the tolerance of P. lusitanicum to the subsequent challenge of a more severe heat stress event. The study of this phenomenon in marine science is a very recently emerging field, though ample ecological evidence exists from field observations that suggest weaker negative heat stress impacts on marine organisms exposed to recurrent thermal stress events (Oliver and Palumbi, 2011; Hughes et al., 2019) or high-frequency temperature variability, such as daily temperature range (Safaie et al., 2018). So far only a few studies on benthic primary producers have experimentally assessed the effect of thermal priming on acute (or MHW) stress responses (Kishimoto et al., 2019; Nguyen et al., 2020; Yu et al., 2020). The red alga Bangia fuscopurpurea acquired short-term thermal stress memory, which enhanced the species’ thermotolerance against subsequent lethal high-temperature stress (Kishimoto et al., 2019). In two seagrass species, Posidonia australis and Zostera muelleri, a similar experimental design as used in our study showed that non-preheated plants suffered a significant reduction in photosynthetic capacity, leaf growth and chlorophyll a content, while plants exposed to a prior mild thermal stress were able to cope better with the recurrent stressful event (Nguyen et al., 2020). Similar findings of increased thermal tolerance upon pre-exposure to moderate stress were also reported for the coral Acropora pruinosa (Yu et al., 2020). This increased thermotolerance through thermal stress memory has been associated to molecular modifications via activating, enhancing or speeding up responses to coping with environmental stressors (Crisp et al., 2016), e.g., higher expression of heat stress transcription factors regulating the expression of heat-shock proteins and antioxidant genes (Kotak et al., 2007; Song et al., 2012; Wang et al., 2014). Indeed, in the aforementioned studies regarding thermal priming in marine organisms, Nguyen et al. (2020) demonstrated that methylation-related genes were significantly regulated in seagrasses in response to thermal stress, while in B. fuscopurpurea modification of membrane fluidity via changes in membrane fatty acid composition was seemingly involved in the establishment and maintenance of heat-stress memory (Kishimoto et al., 2019).

Phymatolithon lusitanicum exhibits a wide latitudinal distribution, spanning from Scotland and Ireland to the Iberian Peninsula and the Western Mediterranean Sea, but also a large vertical distribution from the intertidal zone down to 64 m (Peña et al., 2015; Bunker et al., 2018). This suggests that the species has a wide environmental tolerance, which is supported by our study that shows that it is well adapted to the highly fluctuating environment at the southern coast of Portugal. In fact, it seemingly benefits from the often strong temperature increases, due to inversions from upwelling to downwelling conditions, as these conditions not only boost the species productivity, but also allow it to cope better with more intense temperature rises, such as MHW events, ensuring lower impacts on net primary production and calcification, as well as a quicker recovery. The capacity of P. lusitanicum to remember past thermal incidents and store this knowledge to adapt to new challenges represents a great advantage for this species in its natural habitat, where environmental conditions, i.e., temperature and light availability, fluctuate fast and repeatedly. Thus, the P. lusitanicum population in the Algarve can be considered as resilient (sensu Hodgson et al., 2015), combining both, resistance to immediate physiological impacts and recovery after the stress exposure. The species’ resilience is also evidenced by the maintenance of relatively constant seasonal growth rates, despite the large environmental fluctuations and MHWs (Sordo et al., 2020). However, future research is needed to assess if other populations of P. lusitanicum along the species’ latitudinal distribution show a similar thermotolerance and resilience or if the naturally variable environment of this specific population is the major driver for its resistance to thermal stress.