All biological materials were collected from Phuket, Thailand from May to June 2019. The location map of the sampling sites is indicated in Figure 1A. Marine species belonging to three groups of photosynthetic organisms, including corals, seagrasses, and macroalgae, were selected based on their dominance in the shallow areas of the sampling sites, which makes them more susceptible to warming and consequently relevant candidates for this study. The recorded annual sea surface temperature (SST) of Phuket, Thailand from 2017 to 2020 ranges from 27.45°C ± 0.02°C to 32.00°C ± 0.02°C (Figure 1B).

The corals P. acuta and P. lutea and the lightly-calcified macroalga P. boryana were collected from Panwa, Phuket (7°48′6.26′′N 98°24′23.75′′E). Healthy colonies of P. acuta and P. lutea were collected from 3 to 5 m depth using a stainless bone cutter, chisel, and hammer. Thalli of P. boryana with holdfast intact were hand-picked haphazardly at the sandflat areas at low tide and placed in plastic bags containing the seawater collected on site. The seagrasses H. ovalis and T. hemprichii and the macroalga U. intestinalis were collected from Pakhlok subdistrict, Phuket (8°01′18.3′′N 98°24′47.2′′E). Monospecific patches of seagrass located at least 5 m apart with similar density were chosen. Intact plants (leaves and root/rhizome complex) with their sediment were collected using a hand shovel, placed in plastic boxes, and covered with wet paper towels. Thalli of U. intestinalis, free-floating green tides-forming macroalgae, that are found in the seagrass meadow were hand-picked at low tide and kept in plastic bags containing the seawater collected on the site. All biological materials were transferred with natural seawater to the aquarium facility of the Coastal Oceanography and Climate Change Research Center (COCC), Prince of Songkla University within 12 h of collection. Epiphytes were washed off the collected samples with sterile seawater (salinity of 32 PSU), followed by acclimation, as described below.

All coral colonies were initially acclimated in 600 L holding tank for a week in flowing seawater pumped directly from the seawater stock, prepared from natural seawater filtered with Nomex Filter Bag, and treated with 50 ppm chlorine (50 mg chlorine L^–1). Water flow within the tank was generated with recirculating pumps (WP-300M, SOBO, China and AT-107, atman, China). Temperature, salinity, and pH were controlled at levels recorded at the sampling site (29°C, 32 PSU, and pH 8.2). Temperature was regulated using a Heater chiller (CS-160CIRV1, atman, China). Irradiance of 150 μmol photons m^–2 s^–1 (measured with Light Sensor Logger, LI-1500, LI-COR, United States) was provided by COB light (TS-A600, Aquarium lamp, China), with 12/12 of light/dark cycle. This irradiance level was non-photoinhibitory (pre-determined in the pilot study) and corresponds to average daily irradiance at the respective sampling periods (155.83 ± 8.81 μmol photons m^–2 s^–1 in May and 156.20 ± 7.31 μmol photons m^–2 s^–1 in June, Supplementary Table 1). After 1 week, coral colonies were cut into nubbins of 3–5 cm and distributed into experimental tanks (36 pieces/species/tank, 60 L/tank). The setting of the experiment tank condition was similar to the holding tank, except for the models of temperature chillers (JMC-02, JBA, China) and water pumps (AP-1200, SONIC, China). The coral nubbins were allowed to acclimate for an additional week. One-fourth of the seawater volume was renewed at the end of the week and all physical and chemical parameters were monitored every 2–3 days throughout the acclimation period.

Three plastic boxes containing sods of each seagrass species (six shoots of T. hemprichii or approximately 30 leaf pairs of H. ovalis in each box) and sediment were combined into an aquarium (40 L, six boxes in total in each aquarium). All aquaria were filled with sterile seawater (salinity of 32 PSU, pH 8.2). Seawater in the aquaria was regularly mixed and filtered using a submersible pump coupled with a microfiber filter (Atman HF-0600, China). Non-photoinhibitory irradiance (pre-determined in the pilot study) was provided by aquarium LEDs (A601, Chihiros, China) set at approximately 200 μmol photons m^–2 s^–1 and 12/12 light/dark cycle. This level corresponds to average daily irradiance at the respective sampling period (199.80 ± 43.59 μmol photons m^–2 s^–1, Supplementary Table 1) and is comparable to minimal saturating irradiances [E_k], pre-determined in situ from the RLCs (218.39 ± 19.72 μmol photons m^–2 s^–1 in T. hemprichii and 225.09 ± 19.79 μmol photons m^–2 s^–1 in H. ovalis, Diving-PAM, Walz, Effeltrich, Germany). A constant temperature at 30°C throughout the acclimation period of the aquaria was maintained using submersible heating rods (EHEIM 3617, EHEIM, Germany). The acclimation period lasted 5 days.

Approximately 20 g of each macroalgal species were placed in each glass container and anchored to the bottom of the containers with plastic mesh. Other settings followed the seagrass setups.

As this study aims to evaluate vulnerability to warming in shallow-water marine organisms, we focused on their stress responses to an increase in temperature within a range recorded in their natural settings encompassing the future warming scenarios. The testing temperatures were chosen based on past SST (27.45 ± 0.02°C to 32.00 ± 0.02°C, Figure 1B), records of temperature profiles of the sampling sites (28.84–30.25°C in shallow water coral reefs and 26.96–36.70°C in tidal flat areas at the respective sampling periods, Supplementary Table 1), our previous investigations (Rasmusson et al., 2020, 2021; unpublished data) and other studies from the tropical shallow waters (Sutthacheep et al., 2013a; Collier and Waycott, 2014; Pedersen et al., 2016; George et al., 2018; Kong et al., 2019). Extreme temperatures, exceeding 40°C, have been reported in tropical seagrass meadows (Collier and Waycott, 2014; Pedersen et al., 2016; George et al., 2018) whereas the highest temperature reported in shallow water coral reefs in Thailand was 32.7°C (Sutthacheep et al., 2013a). As corals are exposed to smaller thermal variations in situ compared to seagrasses and macroalgae (Supplementary Table 1), narrower range of temperature with smaller increment was thus adopted for coral experimental design.

The experiment was continued in the same setups for acclimatization and divided into two phases: 5 days of stress and 5 days of recovery. The stress period consisted of four treatments (1) 27°C (2) 29.5°C (2) 32°C, and (4) 34.5°C for corals and (1) 27°C (2) 32°C (2) 37°C, and (4) 42°C for seagrasses and macroalgae. During the stress period, the maximum quantum yield of photosystem II (F_v/F_m) and photosystem II potential activity (F_v/F₀) were assessed on days 0, 3, and 5 before the light was turned on, whereas the effective quantum yield (ϕPSII) and RLCs were assessed in the middle of the photoperiod on the same day. At the end of the stress period, the biological materials were randomly collected from all treatments for further analysis of photosynthetic pigments, while zooxanthellae density was assessed in the coral samples. Growth rates and biomass loss rates (when applicable) were determined at the end of the stress period (as detailed below). A recovery period was initiated by adjusting the temperature of all aquaria to 27°C for corals and 32°C for seagrasses and macroalgae (pre-determined in the preliminary study as the optimal temperature). All species were subjected to the new temperature regime for an additional 5 days and the same measurements were done as described for the stress period. The seawater was renewed before starting the stress period and before starting the recovery period. Salinity was measured every day using a refractometer and, when necessary, adjusted to 32 PSU by adding distilled water. The experimental design and timeline for the measurements are summarized in Figure 1C.

Integrated Biomarker Response (IBR) approach was used to quantify the overall effects of temperature on each species through an integrated response of F_v/F_m, chlorophyll content, and growth rate. The biomarkers used in this analysis were chosen as the representatives for the responses at physiological, biochemical, and whole-organism levels. The IBR indices were analyzed according to the method of Beliaeff and Burgeot (2002) as described by Madeira et al. (2018). For each biomarker, the general mean (m) and standard deviation (s) of all treatments combined were calculated and standardized (Y) as Y = (X − m)/s; where X is the mean value for the biomarker at a given treatment. Z was then calculated as Z = −Y or Z = Y in the case of biological effects corresponding to inhibition or activation, respectively. The score (S) was computed as S = Z + | Min|; where S ≥ 0 and | Min| is the absolute value for the minimum of all Y calculated in a given biomarker. The score results (S) were presented as star plots and used to calculate the IBR as:

Where

Where Si and S_{i + 1} are two consecutive clockwise scores (radius coordinates) of a given star plot; A_i corresponds to the area connecting two scores; n is the number of biomarkers used for the calculations; and α = 2π/n (Beliaeff and Burgeot, 2002; Madeira et al., 2018).

All statistical analyses were performed separately for each group of organisms (corals, macroalgae, or seagrasses). The photosynthetic parameters derived by PAM measurements (F_v/F_m, F_v/F₀, ϕPSII, ETR_max, E_k, and α) and growth rates were analyzed by repeated-measures analysis of variance (ANOVA) using time of measurements as the within-group factor and thermal treatments and species as the categorical factors. The statistical significance of differences in total chlorophyll content and coral zooxanthellae density was determined by factorial ANOVA using temperatures, species, and treatment period (stress and recovery) as categorical factors. Fisher’s least significant difference (LSD) test was used for post hoc comparisons. Cochran’s test was used to test the assumption of homogeneity of variances before performing the ANOVAs. All analyses were performed with Statistica Academic Version 13.0.

Total chlorophylls in both coral species (Figure 4) varied according to temperature (factorial ANOVA, p < 0.05), time of sampling (p < 0.001), and interactions between the two factors (p < 0.001). At the end of the stress period, total chlorophyll of P. acuta exposed to 27, 29.5, and 32°C remained similar and was higher than total chlorophyll of individuals exposed to 34.5°C (Fisher’s LSD test). A further decline in the total chlorophyll after the recovery period was detected in individuals pre-exposed to 34.5°C (Fisher’s LSD test). On the other hand, chlorophyll content of P. lutea decreased after exposure to 32 and 34.5°C, showing a significant difference between the two treatments (Fisher’s LSD test). Nevertheless, no further change in the total chlorophyll was detected at the end of the recovery period. A reduction in chlorophyll content in coral species was in accordance with a reduction in zooxanthellae density and bleaching (Supplementary Figure 1). Total chlorophylls in seagrasses exhibited differences according to temperature (factorial ANOVA, p < 0.001) and species (p < 0.001). At the end of the stress period, a decrease in total chlorophyll in H. ovalis was found at 37 and 42°C (Fisher’s LSD test). A recovery was observed in the plants exposed to 37°C. At the end of the stress period, a decrease in total chlorophyll in T. hemprichii was found only at 42°C, which showed a further decline at the end of the recovery period (Fisher’s LSD test). Total chlorophylls in macroalgae exhibited differences according to temperature (factorial ANOVA, p < 0.001) and species (p < 0.001). Both macroalgae showed a significant decrease in total chlorophyll only at 42°C treatment (Fisher’s LSD test). While P. boryana showed no significant change in total chlorophyll at the end of the recovery period, a further decline in the total chlorophyll was detected in U. intestinalis pre-exposed to 37°C (Fisher’s LSD test).

Growth rates of all species were affected by increasing temperature (Figure 5). Growth rates of the corals (Figure 5A) were influenced by species (repeated ANOVA, p < 0.05), temperature (p < 0.01), time of measurement (p < 0.001), and interactions between these factors (p < 0.001). At the end of the stress period, the net biomass increments of P. acuta exposed to 27, 29.5, and 32°C remained similar, whereas a significant decline was observed in individuals exposed to 34.5°C (Fisher’s LSD test). At this temperature, P. acuta exhibited a net loss in biomass. No significant change was detected after the recovery period. On the other hand, the growth rates of P. lutea showed a stepwise decline at 27–34.5°C (Fisher’s LSD test), reaching negative values at 34.5°C indicating a net loss in biomass. Complete recovery was seen in individuals pre-exposed to 29.5 and 32°C, whereas growth rates of the individuals pre-exposed to 34.5°C did not improve.

The growth rates in seagrasses (Figure 5B) showed differences according to species (repeated ANOVA, p < 0.01), temperature (p < 0.001), time of measurement (p < 0.001), and interactions between these factors (p < 0.001). The biomass increments of H. ovalis measured at the end of the stress period decreased linearly with increasing temperature from 27 to 42°C (Fisher’s LSD test). Further reductions in growth rates were detected in plants pre-exposed to 27°C at the end of the recovery period (Fisher’s LSD test). On the contrary, negative effects of temperature on biomass increments of T. hemprichii were detected only at 42°C (Fisher’s LSD test). Further reductions in biomass increments were detected in individuals pre-exposed to 37 and 42°C at the end of the recovery period (Fisher’s LSD test). In addition, a negative effect of moderate warming (37°C) was also seen in senescence rates in T. hemprichii, which was alleviated after the recovery period (Figure 5C). The highest senescence rates were found in H. ovalis at 42°C, which was likely due to mortality.

The growth rates in macroalgae (Figure 5D) showed differences according to temperature (repeated ANOVA, p < 0.001), species (p < 0.01), and interaction of the two factors (p < 0.001). At the end of the stress period, the net biomass increments of P. boryana exposed to 27 and 32°C were comparable, whereas a net loss in biomass was observed at 37 and 42°C (Fisher’s LSD test). No significant change was detected after the recovery period. On the other hand, a negative effect of increasing temperature on growth rates of U. intestinalis was detected only at 42°C (Fisher’s LSD test), where there was a net loss in biomass (Figure 5D). In the recovery period, U. intestinalis pre-exposed to 37°C suffered a further decline in growth, resulting in a net loss in biomass (Fisher’s LSD test).

For all categories, the most sensitive species to increases in temperature were P. lutea, H. ovalis, and P. boryana, as indicated by higher IBR index at 32 and 34.5°C for coral and at 37 and 42°C for seagrass and macroalga. There was recovery at 37 and 32°C only in H. ovalis and P. lutea, while all species showed no recovery capacity after exposure to the highest temperatures (Figure 6A). The most responsive biomarkers during the stress period were growth in P. lutea, H. ovalis, and P. boryana and chlorophyll content in P. acuta, T. hemprichii, and U. intestinalis. The most responsive biomarkers during the recovery period were growth in P. acuta, H. ovalis, and P. boryana and chlorophyll content in P. lutea, T. hemprichii, and U. intestinalis. On the other hand, F_v/F_m was a less responsive biomarker during the stress and recovery period in all species (Figure 6B).

In line with previous studies (Robison and Warner, 2006; Sinutok et al., 2012; Collier and Waycott, 2014; Rasmusson et al., 2020; Danaraj et al., 2021), our results suggest that short duration of elevated temperature poses significant adverse effects on the tropical shallow-water corals, seagrasses, and macroalgae. The heat stress effects, though complex, were manifested as lower efficiency of photosynthesis and photoprotection, chlorophyll degradation, slower growth, and mortality. Our results suggested that heat stress threshold lies between 37 and 42°C in seagrasses and macroalgae and between 32 and 34.5°C in corals, where the photosynthetic performance and growth were significantly affected. Our results are comparable to the thermal optima and threshold in tropical shallow-water organisms reported in previous studies (Mayfield et al., 2013; Collier et al., 2014; Buerger et al., 2015; Kong et al., 2019; George et al., 2020; Rasmusson et al., 2020, 2021; Keng et al., 2021).

It is established that photosynthesis is a major target of heat stress, since its associated components are heat sensitive (Allakhverdiev et al., 2008). Based on F_v/F_m and F_v/F_0, corals showed partial photoinhibition at 32°C, with subsequent recovery and irreversible photodamage at 34.5°C. On the contrary, the effect of increasing temperature on seagrass and macroalgae went from no effect or a slightly positive effect from 27 to 37°C to an acute and lethal effect at 42°C. These were accompanied by a decrease in RLC-derived parameters (ETR_max, E_k, and α) and chlorophylls contents, indicating an overall decline in photosynthetic activity. As an increase in temperature from 27 to 42°C can lead to a significant reduction in HCO₃^– in seawater (Pierrot et al., 2006), such HCO₃^– limitation may also contribute to a decrease in photosynthetic activities and growth rates observed in this study, particularly in HCO₃^– users (Beer and Rehnberg, 1997; Sinutok et al., 2011, 2012). The impacts on photosynthetic functions were worsened with the experimental duration, indicating accumulative stress effects. It is worth noting that F_v/F₀, although showed higher variability, provided a more sensitive indication of heat stress, which is in agreement with the recent study by Rasmusson et al. (2020), in which F_v/F₀ instead of F_v/F_m was adopted to trace the effects of warming in temperate and tropical seagrasses. In addition, a decrease in chlorophyll content implies that the ability of the photosynthetic tissue to absorb light also decreased. By using constant absorption factor (AF), the ETR values in RLC fittings were likely to be over-estimated. Direct comparison of ETR_max across treatments (Table 1) should be interpreted with care as heat stress may have more pronounced effects on ETR than reported in our results. This points to limitation of the use of RLCs in repeated measurements in experiment which may induce variation in AF, in line with previous discussions by Enríquez and Borowitzka (2010) and González-Guerrero et al. (2021).

Negative effects were also observed in growth rates of all the tested organisms, which appeared to be more heat sensitive than photosynthesis. This may be linked to the negative carbon balance caused by higher respiratory carbon consumption relative to photosynthetic carbon assimilation (Scheufen et al., 2017; Hammer et al., 2018; Rasmusson et al., 2020; Costa et al., 2021), while photorespiratory activity, a competing process with carbon fixation, is known to be enhanced by increased temperature (Zhang et al., 2018; Hu et al., 2020). In addition, metabolic depression and energy allocation to certain defensive mechanisms (Bernardet et al., 2019; Marín-Guirao et al., 2019; Innis et al., 2021) may have long-term consequences on energetic balance and growth. Previous studies highlight that maintaining metabolic balance in coral holobionts (host, symbiotic algae, and associated microbes) is crucial under stress conditions (Levas et al., 2013). Undoubtedly, energy deficiency could be exacerbated during bleaching in corals (Innis et al., 2021), which was visible at 34.5°C. As coral calcification is known to be tightly coupled with photosynthetic and carbon translocation rates (Tremblay et al., 2016), a decrease in integrated photosynthetic capacity (photodamage, chlorophyll content, and zooxanthellae density) is likely to be among the key drivers for a decline in growth rates in the two corals at this temperature.

A more prominent heat stress effects on growth rates than on photosynthesis was also reported in other tropical species (Sinutok et al., 2012; Terada et al., 2016), including P. acuta, and T. hemprichii (George et al., 2018; Poquita-Du et al., 2020; Ho et al., 2021). This implies that the general notion that photosynthesis is one of the most sensitive physiological processes (Bhagooli et al., 2021) and that its response may precede the response at growth level may not entirely hold true for heat stress or could not be supported by the chlorophyll fluorescence parameters measured in the present study. Tropical shallow-water organisms may be able to maintain an intact photosynthetic machinery during warming events, albeit slower biomass production, which may persist in some species after the warming period is over (e.g., H. ovalis; P. boryana, and U. intestinalis at 37°C). In seagrasses, senescence was found to be among the significant effects of warming; not only seen as higher number of dead leaves in the two species, but higher number of shed leaves in T. hemprichii at 37°C. Anoxia in intercalary meristematic tissue has been identified as among the mechanisms contributing to detachment of shoots in seagrasses subjected to heat stress (Hammer et al., 2018). From our results, we proposed that whole-organism-related parameters, such as growth rates and senescence, may serve as good measurements for heat stress responses in relatively fast-growing species. These parameters are not only quantifiable but also ecologically relevant since they give an integrated indication of the organism’s performance and health status (Madeira et al., 2018; Dias et al., 2020).

Different sensitivities to warming in conjunction with biotic interactions are likely to be the key predictors of climate change impacts on shallow-water marine ecosystems. Among the six species from three different groups of organisms, the seagrass T. hemprichii may be the least affected by future warming, whereas the coral P. lutea appears to be the most sensitive to increments in temperature.

Coral bleaching and mortality generally occur when the ambient temperature exceeds the thermal thresholds (Skirving et al., 2019). Mass bleaching of shallow-water corals in Thai waters corresponded to the thermal anomalies, with an average SST of 30.1°C in 1991 and 1995, 30.4°C in 1998, and 31.1°C in 2010 and 2016 (Brown et al., 1996; Phongsuwan and Chansang, 2012; Sutthacheep et al., 2013b; Putchim et al., 2017). In addition, a long-term comparative study conducted in Phuket in 1984–1986 and 2003–2005 revealed that a gradual increase in temperature resulted in a significant reduction in skeletal growth of the coral P. lutea (Tanzil et al., 2009). The average SSTs showed an increasing trend during the past three decades (Brown et al., 1996; Tanzil et al., 2009; Phongsuwan and Chansang, 2012; Sutthacheep et al., 2013b; Putchim et al., 2017) and is expected to reach 32°C in the near future (Hoegh-Guldberg et al., 2018). Our field records of variability in temperatures in the natural setting indicate that shallow-water corals are already exposed to an occasional rise in ambient temperature to 32°C (Unpublished data by authors). The future scenarios may drive radical change of local reef structure as recorded in the Great Barrier Reef (Hughes et al., 2018). Such higher thermal sensitivity observed in corals compared to seagrasses and macroalgae might be because they often grow in deeper habitats, thus experiencing smaller fluctuations in temperature and are rarely exposed during low tide (Smale et al., 2019; Roth et al., 2021). Although seagrasses and macroalgae can withstand a wide range of increasing temperature, a rapid deleterious effect was observed when the testing temperature reached 42°C, implying that the irregular warming events, in which temperature rises above 40°C, though uncommon and brief, will put immense pressure on intertidal seagrasses and macroalgae, thereby affecting productivity and leading to mortality. Although the seagrass and macroalgal die-off as a result of warming event was never reported in Thailand, it has been recorded in other locations in the tropical bioregions (Carlson et al., 2018; Buckee et al., 2021).

It is therefore expected that, as a result of moderate warming, seagrasses and macroalgae could outcompete shallow-water corals as the temperature increases (Anton et al., 2020; Roth et al., 2021). Although our results suggest that T. hemprichii might be the most persistent species in warming scenarios because of their relatively high thermotolerance, successful expansion, and colonization in the subtidal areas requires acclimation to different arrays of environmental factors (Beca-Carretero et al., 2020). One of the key limiting factors is light availability, which has been identified as the fundamental driver for growth and distribution of many marine primary producers (Falkowski and Knoll, 2007). Due to their relatively higher light requirement, slower growth rate, and preference for soft substrates (Collier et al., 2012; Badalamenti et al., 2015), seagrasses are less likely to expand to deeper coral reef areas, implying that direct interaction between corals and seagrasses is improbable. On the contrary, there is growing evidence of enhanced algal occupation of coral reefs following warming events (Hughes et al., 2019; Anton et al., 2020; Wasim et al., 2021). This interaction with macroalgae has been shown to worsen coral health status and physiological performance and is often associated with coral reef degradation (Fong and Todd, 2021; Roth et al., 2021). Although they appear less tolerant to extreme warming, opportunistic characteristics, such as rapid nutrient uptake, effective carbon utilization, and high reproduction capacity of Ulva species and their free-floating trait (Buapet et al., 2008, 2013; Zhang et al., 2019; Kambey et al., 2020), may allow them to proliferate in the coral reef habitat under moderate warming scenarios, particularly where eutrophication is not properly regulated. As previously stated, light is one of the key controlling factors (Falkowski and Knoll, 2007) and studies have shown that interaction with light availability can determine heat stress responses in marine species (Dove et al., 2006; Moreno-Marín et al., 2018; Costa et al., 2021). The extrapolation of our results obtained from controlled experiment using a single non-photoinhibitory level of irradiance to natural conditions, where organisms are exposed to different light environments, must be taken with caution.

When comparing the two species of corals, P. acuta maintained their photosynthesis and growth rates relatively well when facing moderate heat stress (29.5–32°C), although more rapid bleaching (Supplementary Figure 1) and earlier mortality were observed in 34.5°C compared to P. lutea. Other biological attributes, such as photosynthetic efficiency and growth rates, suggested that P. lutea was more sensitive to warming, particularly at moderate levels. It has been suggested that Pocillopora species are more susceptible to thermal stress than the massive Porites species (Hill et al., 2012). This is evident in the mass coral bleaching events from 1998 to 2010 during which severe bleaching and mortality of pocilloporids in Thai waters were recorded (Yeemin et al., 2009; Chavanich et al., 2012; Sutthacheep et al., 2013b). Nevertheless, their susceptibility seems to reduce as they were repeatedly exposed to thermal stress (Pratchett et al., 2013; McClanahan, 2017; Putchim et al., 2017). Association with Clade D types of zooxanthellae, a more heat-resistant clade, which have been increasingly reported in Thai coastal reefs (Chankong et al., 2018, 2020), may contribute to their enhanced thermotolerance (Cunning et al., 2015; Wham et al., 2017). On the contrary, field records showed that P. lutea often suffered partial bleaching and mortality (Phongsuwan and Chansang, 2012; Brown et al., 2014) and that this phenomenon has gradually increased during the past decade (Putchim et al., 2017). Significant impacts of warming on Porites species were manifested as a decline in calcification measured as linear extension, density, and hiatus (Cantin and Lough, 2014; Xu et al., 2018), which is in line with our results. It is worth noting that growth rates of P. lutea pre-exposed to 32°C displayed complete recovery, thus corroborating previous investigations reporting the subsequent recovery of P. lutea after thermal stress in their natural settings (Phongsuwan and Chansang, 2012; Brown et al., 2014).

The different life traits of the two seagrasses may contribute to their difference in thermotolerance. Halophila are considered the colonizing species, while Thalassia are considered the persistent ones (Kilminster et al., 2015; O’Brien et al., 2018). It has been suggested that the former often exhibit lower physiological resistance to environmental disturbances compared to the latter (Kilminster et al., 2015; O’Brien et al., 2018). This notion agrees with the findings in the present study and is in line with our previous investigations in which H. ovalis showed lower physiological tolerance to stressors, such as desiccation and high irradiance, than T. hemprichii (Wuthirak et al., 2016; Phandee and Buapet, 2018). Nevertheless, colonizing species are expected to recover from stress more rapidly than persistent species, owing to their fast growth rates and usually large seed banks (Kilminster et al., 2015; O’Brien et al., 2018; Ong et al., 2020). Even at a short duration of recovery, physiological improvement was already seen in H. ovalis exposed to 37°C. In this study, while growth rates were not yet restored, their photosynthesis and chlorophyll contents showed almost complete recovery after 5 days at optimal temperature.

Padina species are lightly-calcified brown macroalgae (Wichachucherd et al., 2010), whereas Ulva species are green macroalgae known for their opportunistic traits and capacity to generate “green tides” (Zhang et al., 2019; Kambey et al., 2020). Higher thermotolerance in U. intestinalis compared to P. boryana may be a result of their local acclimation. In their respective sampling sites, U. intestinalis were mostly found free-floating or attached to seagrass leaf blades or mangrove roots and directly exposed to sunlight and heat at low tide, while P. boryana remained attached to the substrate, forming assemblage structure, which may buffer drastic changes in environmental factors influenced by tidal cycle. Underlying mechanisms of heat tolerance in Ulva species have been extensively studied and vast arrays of molecular regulations and metabolic adjustment pathways have been identified (Fan et al., 2017; He et al., 2018). Previous studies also provide evidence that calcified macroalgae are more negatively affected by warming than fleshy macroalgae (Kram et al., 2016; Dove et al., 2020), partly due to the sensitivity of the calcification process to heat (Sinutok et al., 2011, 2012). Although P. boryana and U. intestinalis did not show recovery potential after 5 days at optimal temperature, both species have high reproductive outputs and high recruitment potential (Wichachucherd et al., 2010; Mayakun, 2019), which may support their long-term recolonization once the environmental factors become favorable.

Devastating loss of ecosystem services are expected if the corals, seagrasses, and macroalgae undergo degradation. As Porites and Pocillopora species serve as dominant reef-building corals in Thai waters (Yeemin et al., 2009; Chavanich et al., 2012; Sutthacheep et al., 2013b; Putchim et al., 2017), negative effects on their productivity and growth may be translated to an overall decrease in the health of coral habitats, thereby affecting a range of services, such as nursery ground (Lin et al., 2021), fisheries (Edgar et al., 2014), and coastal protection (Zhao et al., 2019). In addition to these services, seagrass meadows and some macroalgal habitats contribute as effective sinks for CO₂ (Krause-Jensen and Duarte, 2016), and their decline would eventually reduce the overall capacity for carbon sequestration and promote a release of greenhouse gasses (Arias-Ortiz et al., 2018). Thus, the effects of warming may not only limit to coastal habitat degradation but may also accelerate the changing climate. While recovery by recolonization is possible, reports suggest that ecosystem services provided by fast-growing pioneer species may differ from those provided by the climax species (Kilminster et al., 2015; O’Brien et al., 2018). Therefore, long-term shifts of structure, functions, and processes of coastal ecosystems may be expected following ECEs.

The IBR index was adopted to quantify and visualize an overall integrated stress effect of warming on all shallow-water marine organisms. This tool has been successfully applied in assessing the health of corals, reef calcifiers, and marine invertebrates exposed to climate change related stressors, including warming (Madeira et al., 2018; Dias et al., 2020; Marques et al., 2020; Zhang et al., 2021); however, its application in seagrass and macroalgal research is yet to be explored. It is established that various biomarkers at different biological organization levels, ranging from cellular to organismal levels, are affected by increasing temperatures (George et al., 2018; Madeira et al., 2018; Savva et al., 2018; Dias et al., 2020). Early investigations on corals and other marine invertebrates also suggested that biomarkers both associated with cellular mechanisms and whole-organism performance should be combined in IBR indices to better elucidate the effects of heat stress (Madeira et al., 2018; Dias et al., 2020). Therefore, we tested the IBR approach using F_v/F_m, chlorophyll content, and growth rate as biomarkers encompassing physiological, biochemical, and organismal responses, respectively.

In general, the IBR index increased with increasing temperature in all tested organisms (Figure 6A). The highest IBR index was found at 34.5°C in corals and at 42°C in seagrasses and macroalgae, which indicates the highest stress level at this temperature. These integrative analyses, which corroborate the previously discussed results, indicate that an increase of 2.5°C from the maximum SST (32°C) will produce lethal effects in corals, while an increase of 10°C from the maximum SST will potentially cause seagrasses and macroalgae to die-off. In addition, a slightly higher IBR value was observed at 32°C in the coral P. lutea and at 37°C in the seagrass H. ovalis and the macroalga P. boryana, thus suggesting a higher sensitivity of these species under moderate heat stress. Therefore, poorer health of these organisms may be already seen within a range of the current maximum SST to an increase in the maximum SST by 5°C. Recovery from five consecutive days of moderate heat stress, however, was observed in P. lutea and H. ovalis, implying that the overall health status can be restored once the warming period is over. This emphasizes that the ability of the organisms to recover from stress significantly contributes to their resilience to moderate warming (Brown et al., 2014; Xu et al., 2020; Nguyen et al., 2021) and that recovery should be considered when evaluating the potential impacts of ECEs.

The IBR analysis showed an inhibition of photosynthetic activity, degradation of chlorophyll, and slower growth in all the tested species (Figure 6B). Deleterious effects at 34.5°C for corals and 42°C for seagrasses and macroalgae were manifested as a marked decline of all three biomarkers, while sublethal effects (32°C in the corals and 37°C in the seagrasses and macroalgae) were largely driven by growth rate and chlorophyll content for most of the tested species. These results, which support the suggestions made by previous studies (Madeira et al., 2018; Dias et al., 2020; Marques et al., 2020; Zhang et al., 2021), highlight the importance of using multiple biomarkers across the biological organizations and inclusion of the biomarkers at the organismic level.

Our results demonstrate that the IBR approach provides quantifiable indices for monitoring the health status of shallow-water marine organisms under warming scenarios and is therefore proven a useful tool for both research and management. Nevertheless, improvement of its efficiency is warranted, and careful consideration is necessary when interpreting the results within complex environments of a natural setting. While the biomarker suite adopted in the present study seems sufficient to elucidate the overall stress effects in the tested species, since it displayed a temperature-dependent response, all biomarkers are not specific to heat stress. Incorporations of biomarkers, such as heat shock proteins, energy metabolisms-related makers, and certain regulatory molecules (van Oppen and Oakeshott, 2020; Danaraj et al., 2021; Innis et al., 2021), particularly those at the molecular level, will enhance the precision and resolution of the IBR index. However, the selection of suitable biomarkers relies on advanced knowledge of the cause-effect relationship between temperature and complex biological responses of given species and populations, which remains inadequate, particularly in the tropical Southeast Asian region (Guan et al., 2020; Nguyen et al., 2021). An in-depth understanding of the underlying physiological mechanisms associated with warming will therefore not only be crucial for predicting how future warming scenarios will affect the coastal ecosystems, but also for developing a more effective tool to assess and mitigate the long-term impact of warming events on shallow-water marine organisms.