The increased use of chlorofluorocarbons (CFCs) as refrigerants and propellants led to an alarming depletion of stratospheric ozone in the 1980s (Molina and Rowland, 1974) and to a corresponding increase in ultraviolet-B (UVB) radiation reaching the Earth's surface (Madronich et al., 1998). Although the Montreal Protocol led to a dramatic reduction in the emissions of CFCs (UNEP, Environmental Effects Assessment Panel, 2010), the ozone layer will likely not recover to 1980 levels within the coming decades (Weatherhead and Andersen, 2006). Thus, increased UVB radiation will continue to reach the biosphere over the coming years (Weatherhead and Andersen, 2006), although the healing of Antarctic ozone layer is occurring since 2000 (Solomon et al., 2016).

Increased UVB radiation has profound impacts on terrestrial and aquatic organisms (Caldwell et al., 1998; Häder et al., 2007). Bancroft et al. (2007) reported an overall negative effect of UVB on both survival and growth across species differing in life history, trophic position, and habitat. Similarly, a meta-analysis of experimental results showed that the mortality rates of marine biota increased rapidly in response to increased UVB radiation, with protists, corals, crustaceans, fish eggs, and larvae the most sensitive in marine ecosystems to increased UVB radiation (Llabrés et al., 2013).

Some of the mobile organisms may avoid exposure to UVB radiation by seeking shelter at depth and/or in shaded habitats. In contrast, the exposure of photosynthetic organisms to UVB radiation is dependent on light requirement and habitat. For example, obligate upper intertidal and supralittoral species (e.g., green macroalgae, Prasiola crispa) require exposure to air and are consequently exposed to high PAR and UVR (Holzinger et al., 2006). On the other hand, deep-water species (e.g., Eisenia galapagensis, Laminaria brasiliensis) require minimal PAR and consequently do not encounter high PAR and UVR in their habitat (Graham et al., 2007). Depending on their habitat and depth distribution, photosynthetic organisms have developed strong protection mechanisms and/or reparatory systems to avert damage. Because UVB was believed to be absorbed rapidly in aquatic ecosystems, the impacts of UVB radiation were believed to be stronger on terrestrial photosynthetic organisms than on marine photosynthetic organisms (Rozema et al., 2002). However, UVB penetrates to considerable depths in freshwater and marine ecosystems (Llabrés and Agustí, 2006; Tedetti et al., 2007), thereby impacting marine photosynthetic organisms as well, although the penetration of UVB varies with latitude, season and water turbidity (Hanelt et al., 2001; Wiencke et al., 2006).

Indeed, a large number of studies have addressed the effects of UVB on aquatic photosynthetic organisms including micro- and macro-algae both in freshwater and marine ecosystems (e.g., Häder et al., 2007, 2011). Photosynthetic organisms are more resistant than non-photosynthetic organisms to UVB (Llabrés et al., 2013; Agustí et al., 2015), suggesting that photoprotective and reparatory systems designed to counter UVB exposure operate efficiently in photosynthetic organisms. However, UVB radiation causes negative effects on the photosynthetic performance of micro- and macro-algae by reducing their photosystem II efficiency (e.g., Jiang and Qiu, 2005; Jin et al., 2013; Zhang et al., 2013), electron transport rate (e.g., Figueroa et al., 2014; Zhu et al., 2015) as well as photosynthetic carbon fixation rate (e.g., Villafañe et al., 2007; Helbling et al., 2008; Li et al., 2012). Photosynthetic organisms have thus evolved a range of strategies to mitigate UVB damage including increasing non-photochemical quenching (NPQ) (e.g., Jin et al., 2013), UVB-screening pigments (e.g., Klisch et al., 2002), UVB-absorbing/blocking secondary metabolites (e.g., phlorotannin compounds) and DNA repair systems as photolyase activity against UVB-induced DNA damage (e.g., van de Poll et al., 2001; Roleda et al., 2004). Despite these strategies, large variability in responses is found across photosynthetic taxa because the variety of damage induced by UVB is also large, with nucleic acids, proteins and lipids the primary targets of UVB (Buma et al., 2003), all of which can interfere with various cellular processes and metabolisms. For instance, one study reported that nitrogenase activity was completely inactivated under UVB exposure whereas nitrate reductase activity was doubled when compared with the same activities in the control in an N₂-fixing cyanobacterium, suggesting diverse effects of UVB on nitrogen metabolisms (Kumar et al., 1996). At the organismal level, damage to specific proteins of photoreceptor and motor organelles may result in a variety of effects on motile behaviors (Zündorf and Häder, 1991). Moroz et al. (1999) reported that increased UVB decreased the number of gliding cells in the diatom Nitzschia lineariz whereas increased UVB exposure increased the number of immobile and oscillating cells.

The responses of macroalgae to UVB radiation may be dependent on life-history stages. Comparison between different early life-history stages of various kelp species of macroalgae from the Northern Hemisphere showed that spores are more susceptible to UVR damage compared with their corresponding juvenile sporophytic and gametophytic life stages (Dring et al., 1996; Roleda et al., 2007 and references therein). Roleda et al. (2009) also reported that the early life stages, e.g., propagules of a green macroalgae, Urospora penicilliformis, are more susceptible to UVB radiation than the adult life stages. In addition, Swanson and Druehl (2000) reported that kelp species of macroalgae with different meiospore sizes respond differently to UVR. The meiospore size difference was correlated with the depth distribution of adult plants, with the largest meiospores origination from shallow-dwelling adult kelp exposed to high UVR levels. Meiospores from adults growing in high UVR environments displayed greater germination and survival rates than the progeny of adult kelp occupying lower UVR environments. The responses of macroalgae to UVB radiation may be also dependent on thallus thickness and absorption spectra characteristics of the pigments (Roleda et al., 2006). Increasing thallus thickness minimizes UVR damage because the longer pathlength for UVR absorption across the thallus leads to outer cell layers shading inner cells (Franklin and Foster, 1997). Likewise, the vulnerability of microalgae to UVR damage may be dependent on cell size, which determines light absorption properties (Agustí, 1991). Photosynthetic carbon fixation of small phytoplankton species was reported to be much more sensitive to UVB radiation than that of large species in a coastal ecosystem (Li and Gao, 2013), and a recent laboratory study also revealed that the UVB sensitivity of diatoms decreased with increasing cell size (Wu et al., 2015). Small cells are also vulnerable to the accumulation of cyclobutane pyrimidine dimers but photosynthesis is less inhibited by UVB radiation, which may be caused by faster photo-repair in small cells (Buma et al., 2003). However, other factors such as bio-optical characteristics (Figueroa et al., 1997) and intraspecific differences (Halac et al., 2014) may outweigh the effects of cell size. As a consequence, there is currently no agreement on the generality of the relationship between cell size and UVB sensitivity across freshwater and marine microalgae.

The responses of aquatic photosynthetic organisms to UVB also depend on the dose of harmful radiation to which an individual organism is exposed. UVB dose is affected by latitude, altitude and the optical characteristics (i.e., UVB transmittance) of the water body. UVB radiation underwater is primarily absorbed by dissolved organic carbon (DOC) including chromophoric dissolved organic matter (CDOM) (Kirk, 1994) and particles. Since the concentrations of DOC and suspended particles that play key roles in absorbing UVB are much higher in freshwater ecosystems than in marine ecosystems, lower levels of UVB in the water column are expected in freshwater habitats. Thus, the impacts of UVB radiation on photosynthetic organisms are expected be the highest in marine ecosystems and alpine lakes (which have the lowest concentrations of DOC and suspended particles among freshwater habitats; Bancroft et al., 2007).

However, the analysis of the patterns of sensitivity to UVB radiation across aquatic photosynthetic organisms has not yet been published. We hypothesize that differences in responses of photosynthetic organisms to UVB radiation are associated with changes in molecular or metabolic pathways associated to taxa, as well as the differences in cell size of photosynthetic individuals. We also hypothesize that these differences may also be due to major habitat differences between freshwater and marine ecosystems.

Here, we compare the responses of photosynthetic organisms both in freshwater and marine ecosystems to UVB radiation. Specifically, we examine the variability in these responses across phyla and response traits, as well as the variability in responses due to cell size for unicellular algae and form-functional categories for macroalgae, to UVB radiation. We do so through a quantitative meta-analysis of the results of published studies on the responses of aquatic photosynthetic organisms to experimentally increased and reduced UVB. Not all traits that may affect the vulnerability of aquatic photosynthetic organisms to UVB radiation could be considered, because the number of reports providing information on traits such as life-history stages, habit, and thickness of macroalgae was too low to support a quantitative analysis.

The assembled data set included 4,995 observations of responses of aquatic photosynthetic organisms to UVB radiation, including 1980 observations of responses of marine photosynthetic organisms and 3,015 observations of responses of freshwater photosynthetic organisms (Table 1). The experiments on marine photosynthetic organisms reported 1,087 observations on microalgae and 893 observations on macroalgae, whereas the experiments on freshwater photosynthetic organisms reported 2,889 observations on microalgae and 126 observations on macroalgae (Table 1). The phylum Bacillariophyta was the subject of most of the experiments on marine microalgae (56%), whereas cyanobacteria dominated the experiments conducted on freshwater microalgae (79%) (Table S1 in Data Sheet 2). Most experiments on marine macroalgae focused on Rhodophyta (59%); most experiments on freshwater macroalgae targeted Charophyta (59%; Table 2). The overwhelming majority of the experiments showed that increased UVB radiation decreased the organism's performance. Indeed, only 3.06% of the experiments showed improved performance with increased UVB and just 5.78% of the experiments showed reduced performance with reduced UVB. These results demonstrated the expected dominance of the adverse effects of UVB radiation on the performance of aquatic photosynthetic organisms.

Assessment of the Ln effect size in response to a change in UVB radiation indicated that both freshwater and marine photosynthetic organisms improved their performance when UVB was reduced; there were no statistically significant differences between their mean Ln effect size [freshwater: mean (±SE) Ln effect size = −0.537 ± 0.015; marine: mean (±SE) Ln effect size = −0.536 ± 0.012; one-way ANOVA, F = 0.00, p = 0.966; Figure 1]. However, closer inspection of the data revealed that marine microalgae exhibited the greatest improvement in their performance when UVB was reduced (Ln effect size = −0.672 ± 0.021, 22.7% higher than that of freshwater microalgae; one-way ANOVA, F = 25.69, p < 0.0001; Figure 1). Similarly, the improvement in the performance of marine macroalgae when UVB was reduced was 40.2% higher than that of the performance of freshwater macroalgae (one-way ANOVA, F = 4.88, p = 0.027; Figure 1). Both freshwater and marine photosynthetic organisms exhibited damage when UVB was elevated above ambient levels, with the greatest damage observed in marine microalgae (Ln effect size = 0.572 ± 0.037, which was 72.3% higher than that of freshwater microalgae; one-way ANOVA, F = 32.15, p < 0.0001; Figure 1b). In general, marine photosynthetic organisms were more sensitive (by 73.2%) to elevated UVB radiation than were freshwater photosynthetic organisms (one-way ANOVA, F = 55.41, p < 0.0001; Figure 1).

Rhodophyta showed the greatest improvement in their performance when UVB was reduced (Ln effect size = −0.411 ± 0.079; Figure 2a). Charophyta was more resistance than Chlorophyta to increased UVB radiation (Charophyta: Ln effect size = 0.078 ± 0.059; Chlorophyta: Ln effect size = 0.273 ± 0.072; Figure 2a). The responses to increased UVB were generally consistent across marine macroalgae groups (Figure 2b). Diatoms exhibited the largest positive responses to reduced UVB radiation among all aquatic photosynthetic organisms (Ln effect size ~ −0.80), while whole communities exhibited the smallest positive responses (Ln effect size ~ −0.36; Figures 2c,d). Marine green microalgae experienced the greatest impact from increased UVB radiation among all aquatic photosynthetic organisms (Ln effect size ~ 0.79; Figure 2d).

Cellular/molecular traits changed the most and growth rates changed the least when UVB radiation was reduced (Cellular/molecular: Ln effect size −0.599 ± 0.012 for overall dataset, −0.611 ± 0.024 for marine, and −0.594 ± 0.013 for freshwater photosynthetic organisms, Figures 3a–c, 4), although they did not differ to other response traits in some cases. Mortality changed the most in response to increased UVB radiation in marine photosynthetic organisms (0.677 ± 0.135, Figure 3b); no data were available assessing the change in mortality among freshwater photosynthetic freshwater organisms. Chlorophyll a concentrations changed the least in response to either increased or reduced UVB radiation (Figures 5a–c, 6), whereas photosynthetic parameters were more responsive, greatly deteriorating under increased UVB radiation and dramatically improving under reduced UVB radiation (Figures 5a–c, 6).

The response traits were, in general, relative insensitively to changes in UVB radiation [slope (Ln effect size vs. UVB relative) <1; Figures 7, 8], indicating that response traits tend to buffer, rather than to amplify changes in UVB radiation. However, marine photosynthetic organisms exhibited much higher sensitivity to UVB radiation than did freshwater organisms (marine: slope = 0.351 ± 0.021; freshwater: slope = 0.181 ± 0.013; ANOVA, t-test, p < 0.0001; Table 2; Figure 5a), as reflected in a significant interaction between the UVB relative ratio and habitat (marine/freshwater; ANOVA, t-test, p < 0.0001; Table 3, Figure 7a). Specifically, marine microalgae exhibited a similar sensitivity to UVB radiation across phyla (Figure 7a, Table 3). However, freshwater Bacillariophyta (diatoms) exhibited greater sensitivity to increasing UVB radiation (ANOVA, t-test, p = 0.05; Table 3; Figure 7c) than did other freshwater microalgae taxa. Red algae (Rhodophyta) were the most sensitive among the marine macroalgae phyla to UVB radiation with the slope of the Ln effect size in relation to the UVB relative ratio = 0.507 ± 0.023 (Figure 7d), although this result may be influenced by the fewer number of experiments on other taxa in which UVB radiation was increased (Figures 2a,b).

The sensitivity of different response traits to changes in UVB radiation was doubled for marine photosynthetic organisms (slope ~ 0.35) compared with freshwater photosynthetic (slope ~ 0.18) organisms (including microalgae and macroalgae; Figures 8a,b). Furthermore, no significant interactions were found between any response trait and the UVB relative ratio, except for the growth of marine macroalgae, where the Ln effect size was higher under a higher UVB relative ratio than it was for other traits (ANOVA, t-test, p = 0.0148; Table 4; Figure 8e).

The morphology of macroalgae influenced their responses to ultraviolet radiation. When the UVB was reduced, foliose macroalgae (44% of Chlorophyta, 49% of Rhodophyta and 7% of Ochrophyta) exhibited a relatively higher improvement in their performance (foliose: Ln effect size −0.449 ± 0.035 for overall dataset; Figure 9a). Articulated calcareous macroalgae (100% of Rhodophyta) showed a relatively lower response to reduced UVB radiation, especially for marine algae (articulated calcareous: Ln effect size −0.174 ± 0.062; Figure 9b). No differences were observed in responses to reduced UVB radiation between the two freshwater morpho-functional groups (Figure 9c). In addition, no significant differences in the responses to increased UVB radiation were detected among different morpho-functional groups either in marine or freshwater ecosystem. This result is robust against the particularly functional-form classification chosen, as similar results were obtained when the macroalgae were classified according to Steneck and Watling (1982) and Littler and Littler (1980). However, the functional forms of marine macroalgae were not independent, in this data set, of the phylogenetic membership (R² = 0.32, χ²-test, p < 0.0001), as all articulated calcareous algae and most corticated algae were red algae, most leathery algae were brown algae, and most filamentous algae were green algae, so the attribution of differences in response to UVB radiation among marine macroalgae to functional form maybe confounded by phylogenetic differences.

The results confirmed our hypothesis that the response of microalgae to UVB radiation is dependent on cell size (Figure 10). Small-celled microalgae exhibited a much greater sensitivity to changes in UVB radiation than did large-celled microalgae (Figure 10). Indeed, the significant interaction of the log₁₀ of the cell volume and the relative change in UVB radiation suggests that small-celled microalgae are more sensitive to changes in UVB radiation than are large-celled microalgae (ANOVA, p = 0.0074; Table 3; Figure 10).

It has been documented that photosynthetic organisms are, on average, more resistant to UVB radiation than other marine organisms in general (Llabrés et al., 2013). This meta-analysis further found that (1) marine photosynthetic organisms tend to be more sensitive than freshwater photosynthetic organisms to UVB radiation (Figures 1, 7); (2) responses to either decreased or increased UVB radiation vary among taxa (Figure 2); (3) the mortality rate is the most sensitive of the trait responses to elevated UVB radiation, followed by changes in cellular and molecular traits (Figure 3); and (4) as hypothesized, the sensitivity of microalgae to UVB radiation is dependent on size, with small-celled microalgae more sensitive than large-celled microalgae to UVB radiation (Figure 10).

UVB radiation that causes significant mortality among photosynthetic plankton has been reported to penetrate as deep as 60 m in the subtropical Atlantic Ocean (Llabrés and Agustí, 2006). However, the penetration of UVB in freshwater is generally much shallower than in marine waters because the concentrations of DOC and suspended particles, which attenuate UVB penetration in the water column, are typically much higher in freshwater than in marine water (Kirk, 1994). As a consequence, Bancroft et al. (2007) hypothesized that marine photosynthetic organisms experience greater risk of exposure to damaging UVB radiation than do freshwater photosynthetic organisms. Evidence of differential responses of marine and freshwater photosynthetic organisms to UVB radiation has not yet been reported. Indeed, Bancroft et al. (2007) did not find evidence of such differing responses. Although our meta-analysis results show that there were no significant differences between freshwater and marine photosynthetic organisms when UVB was reduced, they do show that marine photosynthetic organisms exhibited stronger sensitivity than freshwater photosynthetic organisms to increased UVB radiation (Figure 1). The discrepancy between Bancroft et al. (2007) results and our results is likely attributable to the vast difference in the scope of the two meta-analyses. We investigated 4995 responses to changes in UVB radiation, whereas the analysis of Bancroft et al. (2007) was based on six observations. Moreover, Bancroft et al. (2007) analyzed primary and secondary producers, whereas we studied micro- and macroalgae. Given our results, we expect that increased UVB radiation due to depletion of the stratospheric ozone layer (SOL) exerts a larger impact on marine primary producers than on freshwater primary producers. The impact on marine primary producers is likely to have been increased by the fact that the oligotrophic ocean, characterized by deep penetration of high UVB doses (Llabrés and Agustí, 2006; Tedetti et al., 2007) encompasses 70% of the surfaces of world's oceans and, thereby, a large fraction of the biosphere. However, ecosystem buffers may include many other factors (such as species-specific resistance, photo-adaptation capacities, nutrient levels, species habit), which were not assessed here, and these may add to the uncertainty in evaluating the impacts of elevated UVB radiation on marine ecosystems.

The Montreal Protocol largely halted further damage to the SOL, with a recovery of the SOL expected in the coming decades. Furthermore, it was reported that the healing of the Antarctic ozone layer is occurring nowadays since 2000 (Solomon et al., 2016). The relationship between the reduction in UVB radiation and the performance of primary producers might, thus, help to forecast the expected response of marine primary producers in a future scenario of ozone recovery (Bais et al., 2011). Although our comprehensive analysis showed that photosynthetic organisms in marine ecosystems are more sensitive than those in freshwater ecosystems to increased UVB, the photosynthetic organisms in alpine lakes, which were characterized by high UVB penetration and transparency, should be at higher risk of exposure to damaging UVB radiation than photosynthetic organisms in lowland freshwater ecosystems or in coastal waters of marine ecosystems.

UVB radiation affects pigment synthesis in aquatic photosynthetic organisms, stimulating the production of UVB-screening pigments (Klisch et al., 2002). However, the concentration of chlorophyll a was less sensitive to changes in UVB radiation than were photosynthetic traits (Figure 6). This finding is in contrast to reports that elevated UVB radiation reduces the chlorophyll a content of photosynthetic organisms through photo-bleaching (Agrawal, 1992, 1996) or that, in contrast, UVB radiation leads to increased chlorophyll a concentration in other organisms (Grobe and Murphy, 1998; Zhang et al., 2013). However, the underlying mechanism for these changes has not been documented (Grobe and Murphy, 1998). Phytoplankton have developed several mechanisms to cope with a wide range of light densities as well as different spectra of irradiance, including massive changes in the cellular content of both light-harvesting and photo-protective pigments, adjustments in their ultrastructures (Berner et al., 1989; Fisher et al., 1998) and regulation of the absorption cross section of photosystem II (σ_PSII) (e.g., Six et al., 2007). Specifically, photosynthetic parameters are sensitive to changes in UVB radiation, suggesting that downstream photosynthesis processes (e.g., photosynthetic electron transport) are more vulnerable than upstream processes to UVB radiation (e.g., light capture by chlorophyll a). Several studies report declines in photosynthetic activity (e.g., a decline in the maximum photosystem II efficiency) with increased UVB radiation (Rai and Mallick, 1998; Sinha et al., 2008; Zhang et al., 2013). However, it has also been documented that photosynthetic activity varies independently from the photosystem II content (Jeans et al., 2014), which is in agreement with our findings.

Cell size has long been proposed as a determinant of sensitivity to ultraviolet radiation (UVR) in phytoplankton. Larger cells are thought to be more resistant to UVR-induced damage due to their smaller surface-area-to-volume ratio, i.e., a. longer path length to the nucleus, which makes them more resistant to accumulation of cyclobutane pyrimidine dimers (CPD) and other forms of damage due to molecular self-shading (Garcia-Pichel, 1994). However, there are also a number of exceptions have been reported (Laurion and Vincent, 1998; Ferreyra et al., 2006). For example, Laurion and Vincent (1998) found that the smallest-celled phytoplankton are relatively resistant to UVR, and this may due to the fact that smaller cells are more vulnerable to DNA damage but more resistant to photosystem damage (Villafañe et al., 2003), or to the fact that smaller cells have faster biosynthetic rates (Peters, 1983). Furthermore, taxon-specific factors for mitigating UVR photodamage have the potential to compensate for size-related vulnerability (Laurion and Vincent, 1998). Our finding that UVB sensitivity in microalgae is dependent on size is consistent with previous reports of the size dependence of the optical properties of microalgae (Agustí, 1991), suggesting higher specific absorption of UVB radiation by small algae, and a recent report that the UV sensitivity of diatoms was negatively correlated with cell size (Wu et al., 2015). Our results also show that the differences in sensitivity between small-celled and large-celled phytoplankton are less important at low UVB irradiances, but that they increased with irradiance as the effect size of UVB radiation increased. Picophytoplankton (cell diameter <2 μm) live both in freshwaters and marine waters, but they dominate the oligotrophic ocean (Agawin et al., 2000). The greater sensitivity of small-celled phytoplankton to UVB radiation is, therefore, consistent with the reported sensitivity of Prochlorococcus and Synechococcus to UV radiation in the oligotrophic ocean, resulting in high cell mortality (Llabrés and Agustí, 2006; Agustí and Llabrés, 2007). Moreover, our results suggest that primary producers may have been heavily impacted by increased UVB radiation due to the erosion of the SOL in the oligotrophic ocean, consistent with evidence of a recent trend toward oligotrophication of the subtropical and tropical ocean (Behrenfeld et al., 2006; Boyce et al., 2010; Henson et al., 2010; Steinacher et al., 2010; Hofmann et al., 2011). Warming of the ocean is predicted to lead to a further expansion of picophytoplankton (Morán et al., 2010; Flombaum et al., 2013; Chen et al., 2014; Nagelkerken and Connell, 2015), suggesting an increase in the sensitivity of marine primary producers to UVB radiation. On the other hand, the predicted recovery of the SOL, with the associated decline in UVB radiation, may lead to improved performance of picophytoplankton in the future, but this improvement may be offset by a deeper penetration of UVB radiation in a warmer, more oligotrophic ocean.

Although the relationships between thickness or spore size and UVB levels in macroalgae were not determined in our analysis due to a limited number of studies reporting these, previous studies showed that these factors also affect their responses to UVB radiation. As translucence or opacity of thallus influences optical properties, such as reflection, attenuation, scattering, absorption, or transmittance of UVB radiation (Caldwell et al., 1983), UVB can, therefore, be attenuated by cell walls of the epidermal tissue effectively reducing UVB effect. For instance, macroalgae with thick thallus offer a longer pathlength to solar radiation resulting in outer cell layers shading inner ones (Roleda et al., 2006). Other protection mechanism described for macroalgae include the development of phlorotannin-containing paraphysis cells towering above the sporangial layer and physodes inside the meiospore-containing sporangium, which protect the soral tissue from UVR exposure (Gruber et al., 2011; Holzinger et al., 2011). Our results showed that foliose macroalgae exhibited the greatest improvement in their performance when UVB is reduced. Gómez and Huovinen (2011) described that foliose forms were able to acclimate rapidly to changing light. Therefore, their performance would be expected to improve rapidly when UVB was reduced, as showed in the present study. Freshwater macroalgae showed limited differentiation in morpho-functional groups, restricted to two groups in our data set, and phyla classification was more informative on their responses to UVB. For marine macroalgae, filamentous algae (e.g., Cladophora), which are typically thinner than the foliose algae (e.g., Porphyra), showed significantly smaller responses to reduced UVB radiation. Leathery algae (e.g., Laminaria), which are much thicker than the filamentous algae, showed comparable responses when UVB was reduced. Our data suggest that morpho-functional grouping account for some of the variation in the responses of marine macroalgae to UVB radiation, but that their responses may be buffered or amplified by many other factors besides functional form. Other factors (e.g., life-history stages, Roleda et al., 2007 and references therein; meiospore size, Swanson and Druehl, 2000) may also affect the responses of macroalgae to UVB radiation, but could not be assessed in present study because of limited number of studies addressing them. Moreover, the covariation between functional forms of marine macroalgae and their phylogenetic membership in our data set implies that the attribution of differences in response to UVB radiation among marine macroalgae to functional form maybe confounded by phylogenetic differences, as functional form classes differ in thallus thickness, which affects light absorption, but phylogenetic differences are associated with relevant differences, such as pigmentary compliments, which may also affect resistance to UVB radiation.

Our analysis provides evidence that sensitivity to UVB radiation is dependent on the phylum in microalgae. The lower sensitivity to UVB radiation of Chlorophyta compared with Chromophyta has been proposed to be due to the larger minimal cross-cell or cross-colony/cenobium distance of Chlorophyta, which can provide shading to the internal structures (Xiong et al., 1996). However, the greater sensitivity of Chlorophyta compared with Bacillariophyta to increased UVB radiation identified here remains unexplained, likely conditioned by the number of experiments (few) and the complex array of UVB damage that targets multiple processes and elicites a large number of protection mechanisms (Häder et al., 2015). Because UVB sensitivity differs among different taxa, it likely plays a crucial role in shaping phytoplankton community composition. Xenopoulos and Frost (2003) found that phytoplankton assemblages incubated under photosynthetic active radiation (PAR) had significantly higher richness, diversity and evenness than those incubated full spectrum irradiance, suggesting possible selection of UVR-tolerant taxa. Consequently, the full phytoplankton community should be more resistant to UVB than a single species or a taxon would be due to the selected effects of UVB. Our results showed that the full community exhibited the smallest responses to UVB (Figure 2). Reports that algal species originally isolated from high-UV radiation locations (Larson et al., 1990; Rautenberger et al., 2013) and marine organisms from the Southern Hemisphere that are subject to higher UVB radiation (Agustí et al., 2015) are particularly resistant to UVB radiation indicate that the history of UV radiation experienced by different taxonomic groups may also account for their diverse responses to UVB radiation. Furthermore, sensitivity to UVB will also imply other aspects than those analyzed in this meta-analysis, as it is expected some responses should be not related to phylum but show large variability among species within the same phylum (e.g., Xiong et al., 1996), therefore, we suggest that more studies on both micro- and macroalgae sensitivity to UVB should be conducted, and on the underlying mechanisms resulting in different responses are worthy of further investigation as well.

Our study confirmed the expected dominance of adverse effects of UVB radiation on the performance of aquatic photosynthetic organisms. However, a small number of observations showed a positive response to enhanced UVB (3.06% of the experiments showed improved performance with increased UVB and 5.78% of the experiments showed reduced performance with reduced UVB). Similarly, Flores-Moya et al. (1999) reported that in the marine alga Dictyota dichotoma, recovery of photoinhibition is delayed if the natural UVB radiation range is removed from the solar-spectrum. This was later confirmed in several macroalgae species (e.g Chara fibrosa), which were collected in shallow waters and have adapted to high UVB radiation environment (Hanelt et al., 2006; Hanelt and Roleda, 2009). A positive effect of UVB radiation has been reported for phytoplankton, where some taxa were most abundant in treatments of intermediate fluxes of UVB radiation (Thomson et al., 2008). Mechanistically, UVB may support repair or photoprotective processes (Sicora et al., 2003) or function as a photoreceptor signal (Hanelt and Roleda, 2009), leading to enhanced recovery of photoinhibition induced by high levels of solar radiation would be better under full solar radiation. Further studies are required to examine the molecular/physiological mechanisms of this photo-physiological function of UVB radiation.

The ocean is undergoing multiple changes including acidification, warming and deoxygenation (Gattuso et al., 2015), which also experiencing elevated UVB levels (Duarte et al., 2009). A vast number of studies have demonstrated that these other stressors could interact antagonistically or synergistically with UVB to influence aquatic biota (e.g., UVB and ocean acidification: Li et al., 2012; Jin et al., 2013; UVB and warming: Banaszak and Lesser, 2009; Xu et al., 2011; UVB and pollutants: Pelletier et al., 2006; UVB and nutrients: Li et al., 2015). Hence, the power of the analysis reported here, focused on the response of aquatic photosynthetic organisms to UVB radiation, is limited by current uncertainties owing to the nature and strength of interactions with other stressors. We, therefore, propose that more comprehensive analysis aiming at the interactive effects of multiple stressors with UVB radiation should be conducted for a better understanding of the consequences stressors for aquatic ecosystem. Our analysis already points to the great vulnerability of marine photosynthetic organisms to UVB radiation, with particularly important consequences for the vast span of oligotrophic ocean dominated by highly UVB-sensitive picophytoplankton.

SA, CD, and PJ conceived and designed the research. PJ, SA, and CD acquired the data and all authors analyzed and interpreted the data. PJ drafted the work and all authors critically revised the work for intellectual content. All authors approved the final version to be published and agreed to accountable for all aspects of the work.

This research was funded by King Abdullah University of Science and Technology (KAUST) through the baseline funding to SA.