The Effects of Climate Change on Macroalgal Growth, Trophic Interactions and Community Structure

Global climate change is threatening the structure, function, and health of ecosystems. While factors of climate change have been studied extensively over the past few decades, most research has focused on the response of single organisms or populations; as our ecosystems are comprised of complex interactions and relationships, it is of critical importance to understand how entire communities are going to be impacted by climate change. Ocean acidification (a by-product of increased atmospheric carbon dioxide, CO2), and nutrient loading are two major forces of global change that are projected to have detrimental impacts on coastal marine species and ecosystems. Most work on ocean acidification has focused on the response of calcifying organisms, where the changes in ocean chemistry associated with acidification enhance shell dissolution and impair growth. However, while calcifying species are expected to exhibit negative responses to acidification, primary producers, like macroalgae, are expected to flourish. Both ocean acidification and nutrient loading can stimulate the growth and productivity of opportunistic, fast-growing, ephemeral macroalgae at the expense of foundational species such as corals, seagrasses, and long-lived, perennial macroalgae (i.e. kelps). As a result, these ecosystems will likely undergo major shifts in structure, function, and diversity. Few studies have investigated the interactive effects of ocean acidification and nutrient loading, particularly in terms of community response and trophic interactions. Despite increasing the growth rates of macroalgae, the presence and diversity of herbivores within an ecosystem has the potential to control this

expected algal growth. The research described in this dissertation aims to: 1) quantify the combined effects of ocean acidification and nutrient loading on the growth, tissue quality, and competition of two abundant macroalgal species with different life histories; 2) test whether or not an abundant grazer can enhance consumption of macroalgae under future conditions of acidification and nutrients, promoting community resilience; 3) describe the impact of ocean acidification on the growth and diversity of reef-associated turf algal communities.
Using a laboratory mesocosm design, the response of Ulva (an ephemeral, opportunistic green alga) and Fucus (a long-lived, perennial brown alga) to the interaction of two levels of ocean acidification and two levels of nutrients was tested.
Individual, field-collected algal thalli were placed in flow-through seawater systems with one of four experimental conditions: high pCO 2 (~1100 µatm) or background pCO 2 (~390 µatm) and high nutrients (200 µM TN) or low nutrients (10 µM TN), in a fully factorial design. Three experiments were run: the first two investigated the response of Ulva and Fucus in monocultures; the third tested the response of both species cultured together (biculture). Growth rates and tissue quality (via carbon to nitrogen ratios, C:N) were measured after 21 days of exposure to treatments. Ocean acidification and nutrient loading significantly increased the growth in Ulva, where growth rates under high pCO 2 and high nutrients were about 3X greater than those grown under ambient conditions, with the environmental factors appearing to have an additive impact on Ulva growth. Growth rates of Fucus were unaffected by environmental conditions. Both species exhibited an increase in tissue quality as a result of decreased C:N when exposed to high nutrients. Response variables were compared between monoculture and biculture experiments for both species. Growth rates of Ulva and Fucus were unaffected by the treatment culture, but tissue C:N of Fucus was significantly higher when grown with Ulva, indicating potential resource competition, where Ulva outcompetes Fucus.
The enhanced growth exhibited by Ulva supports previous work indicating the enhanced growth of opportunistic algal species under future climate conditions. While this could be troubling for species inhabiting coastal ecosystems (such as seagrasses, non-bloom forming macroalgae, and fish), grazers may hold the key to mitigating algal growth and keeping ecosystems in balance. Consumption rates and feeding preferences of a common marine snail (Littorina littorea) were tested using the same experimental design and environmental parameters detailed above. Snails were placed in treatment mesocosms for seven days and were given a choice of Ulva and Fucus.
High pCO 2 levels significantly reduced macroalgal consumption by about 50% and snails switched from a mixed algal diet to feeding exclusively on Ulva. Respiration rates for L. littorea were measured, and under high pCO 2 respiration was significantly reduced. Artificial food trials were run to help explain the diet preference change. No difference was found comparing consumption of the artificial food, pointing to the preference shift being driven by algal tissue toughness.
Reef ecosystems have been well studied with respect to ocean acidification. This research shows that corals are overtaken by fleshy macroalgae and fast-growing turf algae. Here, we tested the response of turf algal communities to three levels of ocean acidification. Natural turf communities growing on coral rubble from the Great Barrier Reef, Australia were collected and exposed to ambient, medium, and high pCO 2 treatments in mesocosms for 41 days. Communities were assessed for biomass and genus diversity. Biomass of turf communities was significantly higher under high pCO 2 . Turf community evenness and diversity significantly increased under high pCO 2 . This change in community structure is likely due to the decline in abundance of Polysiphonia (a filamentous, branched red algae). The results indicate that enhanced turf growth under conditions of acidification will aid in the growth and expansion of macroalgae at the expense of corals in reef ecosystems. Changes in turf diversity should inform how larger macroalgal communities may be structured in the future. This research highlights the success that opportunistic macroalgae and turf algae will have under future climate conditions. The success of macroalgae, however, comes at the expense of other critical and foundational species within a community. In addition, macroalgal communities are likely to undergo assemblage shifts as well, due to species-specific responses to environmental change, where species gain a competitive advantage. Top-down forces, such as grazing, may protect against changes in community structure and macroalgal assemblages, but only if grazers are not negatively impacted directly by environmental change. If grazers exhibit decreased consumption and are unable to keep up with anticipated algal growth, this will ultimately enhance macroalgal abundance in coastal ecosystems.  Abstract:

LIST OF TABLES
Coastal ecosystems are subjected to global and local environmental stressors, including increased atmospheric carbon dioxide (CO 2 ) (and subsequent ocean acidification) and nutrient loading. Here, we tested how two common macroalgal species in the Northwest Atlantic (Ulva spp. and Fucus vesiculosus Linneaus) respond to the combination of ocean acidification and nutrient loading. We utilized two levels of pCO 2 with two levels of nutrients in a full factorial design, testing the growth rates and nutritional quality of Ulva and Fucus grown for 21 days in monoculture and biculture. We found that the opportunistic, fast-growing Ulva exhibited increased growth rates under high pCO 2 and high nutrients, with growth rates increasing threefold above Ulva grown in ambient pCO 2 and ambient nutrients. By contrast, Fucus growth rates were not impacted by either factor. Both species exhibited a decline in carbon to nitrogen ratios (C:N) with elevated nutrients, but CO 2 concentration did not alter nutritional quality in either species. Species grown in biculture exhibited similar growth rates to those in monoculture conditions, but the Fucus C:N increased significantly when grown with Ulva, indicating potential resource competition. Our results suggest that the combination of ocean acidification and nutrients will enhance abundance of opportunistic algal species in coastal systems and will likely drive macroalgal community shifts, based on species-specific responses to future conditions.

Introduction:
Increasing amounts of carbon dioxide (CO 2 ) in the earth's atmosphere are the driving force behind global climate change (Pachauri et al., 2014). Ocean acidification, a decrease in pH brought about by increased atmospheric CO 2 , has garnered attention due to the overwhelmingly negative effects predicted for calcifying organisms (Comeau et al., 2014;Diaz-Pulido et al., 2011;Hoegh-Guldberg et al., 2007;Waldbusser et al., 2015). Changes in ocean chemistry associated with ocean acidification, such as lowered saturation states, are causing reductions in growth, increased shell dissolution, and declines in fitness and performance of many marine calcifying species Waldbusser et al., 2015).
Conversely, less attention has been paid to non-calcifying autotrophic organisms. These species may benefit from ocean acidification and the subsequent change in ocean chemistry, as increased concentrations of both aqueous CO 2 and bicarbonate (HCO 3 ) may enhance photosynthesis and growth in primary producers.
However, the response of primary producers to ocean acidification is highly species specific, ultimately dependent on carbon limitation and carbon acquisition ability as well as developmental stage (Gaitán-Espitia et al., 2014;Olischläger et al., 2012). As such, negative and neutral responses to ocean acidification have also been observed (Gutow et al., 2014;Rautenberger et al., 2015). Divergent responses of fleshy macroalgae to acidification are correlated with the presence and efficiency of the carbon concentrating mechanism (CCM) Raven and Beardall, 2003). Due to highly abundant bicarbonate ions, most macroalgae rely on CCMs to convert HCO 3 to CO 2 for use in photosynthesis . In addition, many species also have the ability to passively diffuse CO 2 and may gain an advantage under future conditions due to reduced reliance and down-regulation of CCMs. While most marine macroalgae have CCMs, a few species within the Rhodophyta rely on passive diffusion of CO 2 for photosynthesis Raven and Beardall, 2003). These species, among others, should experience enhanced growth and photosynthesis due to the increased concentration of CO 2 associated with ocean acidification.
While ocean acidification is projected to impact all marine systems, the effects will likely vary across ecosystems (Hofmann et al. 2011). The signal of ocean acidification is easy to determine in the open ocean; unlike in the open ocean, coastal pH is highly variable due to daily and seasonal shifts in photosynthesis and respiration, and coastal acidification may be driven more by eutrophication than increases in atmospheric CO 2 (Feely et al., 2010;Cai et al., 2011;Wallace et al. 2014).
Nutrient loading (and potential eutrophication events) also impacts coastal bays and estuaries with low flow and low turnover (Lee and Olsen, 1985). Nutrients can enter these waterways via agricultural and urban runoff and sewage treatment discharge, pumping excess nitrogen and phosphorous into the water column (Nixon, 1995).
While these nutrients are critical to algal growth, excess concentrations can facilitate harmful algal blooms (toxic or non-toxic), either composed of micro and/or macroalgae (Anderson et al., 2002). Decomposition or respiration from large macroalgal blooms can lower oxygen levels in the water column potentially leading to hypoxic events, with detrimental impacts on coinhabitants (Granger et al., 2000;Thomsen et al., 2006;Valiela et al., 1997). Macroalgal blooms can also act as a deterrent to coastal recreation (Valiela et al., 1997;Worm and Lotze, 2006).
Our understanding of climate effects on coastal zones is critical, as these ecosystems hold high value in biodiversity as well as economic and societal importance. Increased CO 2 from acidification, combined with increased concentrations of limiting nutrients, could act in conjunction to stimulate and enhance growth in primary producers. While acidification studies are beginning to incorporate additional environmental stressors such as light intensity and warming Rautenberger et al., 2015;Roleda et al., 2012;Sarker et al., 2013), the combined effects of acidification and nutrients on primary producers are less well understood (but see Campbell and Fourqurean, 2014;Falkenberg et al., 2013;).
In coastal zones, the green macroalga Ulva and the brown macroalga Fucus have different life history and ecological traits. Ulva is a fast-growing, opportunistic, ephemeral genus that thrives in a wide range of environments. Fucus is a long-lived, slow growing, perennial genus that creates complex, three-dimensional habitat for other organisms. These genera, among others, form the base of coastal marine food webs in the northwest Atlantic and are commonly grazed by herbivores and omnivores (Bracken et al., 2014;Lubchenco, 1983;Watson and Norton, 1985). Both genera use CCMs (Koch et al., 2013), but exhibit divergent responses, with increased growth rates for Ulva lactuca and decreased growth rates for Fucus vesiculosus under high CO 2 conditions (Gutow et al., 2014;. Similarly, Ulva lactuca has increased growth rates under high nutrients (Steffensen, 1976). Fucus vesiculosus experiences a reduction in growth and cover due to the indirect effects of added nutrients, such as increased turbidity and increased growth of epiphytic algae (Berger et al., 2004).
The objective of our research was to quantify the impacts of the interaction of ocean acidification and nutrients on Ulva spp. and F. vesiculosus, by assessing growth rates, tissue quality (tissue C:N ratio), carbon and nitrogen content of algal tissues, and potential competitive impacts. While U. lactuca was chosen for this study, recent invasions of the cryptic U. australis have nullified our initial identification (Guidone et al., 2013;Hofmann et al., 2010). It is likely that the tested specimens are a mix of two species: U. lactuca and U. australis. We will hereafter refer to our test organisms as Ulva and Fucus. We predicted that the growth rate and tissue quality of Ulva would increase with increased pCO 2 and nutrients, as the combination of the two environmental factors would result in a synergistic effect on its growth rate (Neori et al., 1991;. By contrast, we predicted that increased pCO 2 will decrease the growth rate of Fucus, which is likely to occur due to potential pH sensitivity of its carbon concentrating mechanism CCM (Axelsson et al., 2000;Gutow et al., 2014), but increase tissue quality of Fucus (as seen in Gutow et al., 2014). We expect Fucus growth to be unaffected by nutrient loading as this species is adapted to low-nutrient environments (Savage and Elmgren, 2004), and nutrients may indirectly reduce growth rates by promoting the growth of competitors (Hemmi et al., 2005;Pedersen and Borum, 1996;Worm and Lotze, 2006). Growth rates and tissue quality of both Ulva and Fucus will be tested in a biculture experiment, where we expect opportunistic, fast-growing Ulva to outcompete Fucus resulting in lower growth rates and tissue quality of Fucus (Connell and Russell, 2010;Falkenberg et al., 2013;Worm and Lotze, 2006). We interpret our results in the context of shifting macroalgal assemblages and ecosystem structure.

Algal Collection and Experimental Design
We conducted three temporally distinct experiments using the flow-through  Gutow et al., 2014). Fucus tips and Ulva thalli were cleaned of any epiphytes, and transferred into separate 20L glass aquaria with flow-through seawater and aeration. Algal individuals were acclimated to lab conditions for five days prior to the start of each experiment.
To set up each experiment, we first spun algae 20x in a salad spinner (Thornber et al., 2008), removed a small piece (~10% of starting mass) of the thallus which was then dried at 38°C for 24 hours and then placed in a desiccator for C:N analysis (see below), and then recorded the initial algal wet mass of each remaining piece. We placed individuals into 20L aquaria with one individual per tank (for monocultures) or one individual of each species (for the biculture experiment).
Starting wet mass was 0.50g and 0.75g for Ulva and Fucus, respectively.
Narragansett Bay has high variation in pCO 2 and DIN on both spatial and temporal scales. Annual average pCO 2 concentration is around 400 µatm but ranges from 150 -1000 µatm (Turner, 2015). DIN in Narragansett Bay runs along a north south gradient, where water has an annual average of 70 µM DIN in the north and annual average of 4-10 µM DIN in the south, where certain parts of the bay can exceed 180 µM DIN on occasion (Krumholz, 2012 (Table 1).
Each experiment ran for a total of 21 days, and tanks were supplemented with artificial light (Sylvania Full Spectrum) at 128.7 + 2.7 µmol photons m -2 sec -2 with a light/dark rhythm of 14:10h (L:D). Tanks were scrubbed and cleaned every two days and any epiphytes growing on the algae were removed. Every seven days, algae were briefly removed, weighed, and a small piece (<10% of total wet mass) was removed for future C:N analysis. Mass of removed tissue was included in the calculation of total growth; however, this likely resulted in an underestimation of total algal growth.
Algal tissue quality was determined by drying tissue samples for 24 hours at 38°C. All dried samples were preserved in glass vials and placed in a desiccation chamber.
Samples were ground into a powder and placed in tin capsules. All samples were analyzed for carbon and nitrogen concentrations by Dr. Brad S. Moran's laboratory at the University of Rhode Island Graduate School of Oceanography.

Carbonate Chemistry
Water samples were collected and preserved, using mercuric chloride (HgCl 2 ), over the course of each experiment to measure dissolved inorganic carbon (DIC) and total alkalinity (TA) following the Best Practices Guide ( this data ensured that we were able to capture the natural variability of coastal pH in our treatments (data not presented).

Statistical Analysis:
Relative growth rate (RGR, % day -1 ) was calculated for each alga based on change in wet mass between initial and final masses for each experiment (Lüning et al., 1990). Mean RGR, final carbon and nitrogen concentrations, and final C:N ratios for each species were assessed using three-way analyses of variance (ANOVA) with pCO 2 (ambient or high), nutrient level (ambient or high), and culture (or community; monoculture or biculture) experiment as fixed factors. Initial C:N ratios of either species did not differ among environmental treatments (two-way ANOVA, p>0.05).
By including culture experiment as a factor in our analysis, we were able to determine potential resource competition between the two species. All statistical analyses were performed using JMP v 11 (www.jmp.com).

Seawater parameters
We were able to obtain elevated and relevant environmental parameters for each of our treatments.

Algal Tissue Content (C:N, C, N)
The C:N tissue content of both Ulva and Fucus was significantly lower in high nutrient treatments (p < 0.0001 and p < 0.0001, respectively;  Table 3).
We found evidence of resource competition between the two species for Fucus tissue C:N; Fucus grown in biculture had tissues with significantly higher C:N (p = 0.0021; Table 3). Conversely, Ulva tissue C:N was not impacted by community (p = 0.21;  Fig. 3A and 3B). Nitrogen content of Ulva and Fucus tissues were unaffected by community, pCO 2 , or any interactions of factors (Table 3). There was a significant interaction between community and nutrient level influencing the carbon content in Ulva (p = 0.03; Table 3) as well as a significant three-way interaction between pCO 2 , nutrients, and community (p = 0.05; Table 3).

Algal Growth
Non-calcifying primary producers are predicted to benefit from changes in seawater chemistry due to ocean acidification (see Kroeker et al., 2013), as we found for Ulva, with growth rates doubled under 1200 µatm pCO 2 conditions. This matches the response of Ulva to ocean acidification in other systems Xu and Gao, 2012) as well as other macroalgal species (Campbell and Fourqurean, 2014;Swanson and Fox, 2007).  observed a doubled growth rate of U. lactuca when exposed to 700 µatm pCO 2 , and Xu and Gao (2012) found U. prolifera exhibited increased growth rates of about 40% when exposed to 1000 µatm pCO 2 . These studies used concentrations of pCO 2 expected in the next 50-100 years, whereas our study focused on the more extreme pCO 2 projection for the year 2100. However, the response of macroalgae is still highly species-specific, and neutral or negative impacts of ocean acidification on growth rate have been observed in non-calcifying macroalgae (Cornwall et al., 2012;Gutow et al., 2014;Israel and Hophy, 2002;Mercado and Gordillo, 2011). Divergent responses of macroalgae to acidification are likely due to the differences in CCM effectiveness, potentially giving certain species more independence from the environment, or CCMs are optimized for higher pH conditions and their activity is sensitive to pH (Axelsson et al., 2000;Moulin et al., 2011).
How Ulva takes advantage of ocean acidification conditions may be due to changes in physiological processes, such as down-regulation of CCM activity and reallocation of energy, increased nitrogen assimilation, and/or slight increases in photosynthetic activity Xu and Gao, 2012  . Ulva, however, has a highly efficient CCM and does not appear carbon limited (Axelsson et al., 1999(Axelsson et al., , 1995. Photosynthetic rates of Ulva, as determined by oxygen production, increased by a factor of 1.2 under 700 µatm pCO 2 , but this increase was not statistically significant   (Kawamitsu and Boyer, 1999). It is possible that long-lived species such as Fucus may respond more slowly (e.g. months vs. weeks) to an altered environment.
Pedersen and Borum (1996) found that fast-growing species, like Ulva, are nutrient limited and increase growth rates when nutrient concentrations are high, unlike Fucus, which is adapted to low-nutrient environments. Our Ulva growth rates were also significantly enhanced by the addition of nutrients, similar to prior experimental studies (Steffensen, 1976;Teichberg et al., 2010) and field observations (Díaz et al., 2005). As a fast-growing opportunistic species, Ulva can absorb of excess nutrients in the water column and form blooms that are detrimental to community coinhabitants (Teichberg et al., 2010;Valiela et al., 1997) and can lead to eelgrass declines (Hauxwell et al., 2001;McGlathery, 2001). By contrast, we found no effect of nutrient treatment on the growth rate of Fucus. Fucus can take up excess nutrients, but at a marginal rate (~3%) that may not enhance growth rates (Savage and Elmgren,

2004).
Of several studies on the combined impacts of ocean acidification and nutrients (Campbell and Fourqurean, 2014;Falkenberg et al., 2013;, only  found a synergism between these two factors, with turf algal percent cover multiplied when both of these factors are increased. Campbell and Fourquean (2014) and Falkenberg et al. (2013) found that at least one of the factors increased growth rates, but with no significant interaction. Similarly, our results do not point to a synergism, but the mean growth rates for each treatment show an additive affect of ocean acidification and nutrients for Ulva ( Fig. 1A and 1C).

Algal Tissue Content
Overall, we found that increased pCO 2 did not affect the C:N ratio of either Ulva spp. or Fucus vesiculosus. Although adding carbon is counterintuitive to decreasing the C:N ratio, Gordillo et al. (2001) found that increased pCO 2 resulted in increased uptake of nitrate in U. rigida, thus lowering the tissue C:N ratio. Our results indicate that nutrient level was the primary driver of tissue C:N ratio in both Ulva and Fucus. High nutrient treatments resulted in decreased C:N ratios, which increases tissue quality. Falkenberg et al. (2013) found that increasing each factor resulted in decreased C:N ratios for turf algae, but only nutrients effectively lowered the C:N in the kelp, Ecklonia radiata. We did not find a significant interaction between acidification and nutrients loading on tissue C:N, similar to Falkenberg et al. (2013).
Our results contradict those from Gutow et al. (2014), where high pCO 2 resulted in a decreased C:N ratio for F. vesiculosus.
Our findings indicate that both acidification and nutrients have the ability to alter tissue composition with respect to concentrations of carbon and nitrogen. The addition of pCO 2 significantly increased the percent carbon found in Ulva and Fucus tissues, and increased nutrient concentrations raised tissue nitrogen in both. While we observed higher concentrations of N within high pCO 2 treated individuals, the difference was not significant and our results do not match Gordillo et al. (2001) and Xu and Gao (2012), who found that high pCO 2 facilitates nitrate uptake in U. rigida and U. prolifera, respectively.
We found evidence of resource competition in C:N ratios, which were significantly higher in Fucus tissues when grown with Ulva. This is likely due to the inability of Fucus to acquire nutrients in the presence of fast-growing, nutrient limited species like Ulva (Duarte, 1995). Intraspecific competition and interactions for space and resources drive algal community composition and function (Olson and Lubchenco, 1990;Stachowicz, 2001), and algal communities may undergo assemblage shifts under climate change (Connell and Russell, 2010). However, we did not find any evidence of competition affecting the growth rates of either species..

Impacts to Coastal Ecosystems
In coastal systems where anthropogenic nutrient loading is prevalent, algal blooms likely occur. The role of nutrients in facilitating macroalgal blooms is well established (Lapointe, 1997;Lapointe and Bedford, 2010). Both nutrients and CO 2 are both critical resources for primary producers, but few studies have assessed the role of pCO 2 in contributing to blooms. As our study shows, the interaction of nutrients and pCO 2 results in an additive growth effect on Ulva. As such, ocean acidification may ultimately end up contributing to the timing, frequency, and duration of macroalgal blooms. Ocean acidification has been shown to not only enhance growth rates in some algal species but also may enhance the uptake of nutrients (Gordillo et al., 2001), increasing growth rates indirectly. In addition, acidification can alleviate the cold-driven temperature stress in a species of red algae, resulting in higher growth rates in colder waters . These factors may ultimately contribute to macroalgal bloom events in coastal ecosystems.
Seaweed takeover of critical ecosystems such as coral reefs  and the turf algal dominance in kelp forest ecosystems (Connell and Russell, 2010) highlight the prediction that many macroalgal species are expected to flourish under future climate scenarios. Our results support this, with major increases in the growth rates of an opportunistic species when pCO 2 and nutrients are high. Field studies of algal diversity at naturally low pH vent sites indicate shifting assemblages as water become more acidic (Porzio et al., 2011). Based on our results, we expect a similar shift where small, fast-growing species are able to take advantage of changing conditions, outcompeting larger, slow-growing species (Connell and Russell, 2010;Falkenberg et al., 2013). In this system, long-lived species like Fucus may not be directly impacted by environmental change, but they may be indirectly affected by the overgrowth of epiphytes (Berger et al., 2004) or competing species. Fucus dominates space in the intertidal and shallow subtidal and helps create complex, 3-D structure that is critically important as a source of refuge and habitat for other organisms in the community. Ocean acidification and nutrient loading is likely to have an indirect effect on long-lived, complex species like Fucus, where the growth and spread of algal epiphytes and fast-growing turf species may outcompete, overgrow, and shade out other species (Falkenberg et al., 2013;McSkimming et al., 2015). Our results do not describe a direct, negative effect of acidification or nutrients on Fucus, but these indirect effects may ultimately result in Fucus decline, altering algal community assemblages and ecosystem services. However, to fully understand how macroalgal communities will respond to change more work needs to be done investigating how larger, more diverse communities respond. It is necessary to quantify algal community response over long periods of time, as the seasonality of algal species in temperate coastal ecosystems will likely play a role in determining community dynamics.

Acknowledgements:
This research has been supported by a grant from the U.S. Environmental Protection Agency's Science to Achieve Results (STAR) program. This material is based upon work supported in part by the National Science Foundation EPSCoR Cooperative Agreement #EPS-1004057. We thank the US EPA Atlantic Ecology Division for support and aide in experimental set up and design, and A Pimenta for laboratory assistance.     We observed a significant, positive effect of nutrient addition and increased pCO 2 on the RGR of Ulva (* represents p < 0.001, ** represents p < 0.0001).  N = 7). We observed a significant, negative effect of nutrient addition on the C:N of both Ulva and Fucus (** represents p < 0.0001). In addition, we observed a significant difference between the C:N of Fucus tissue by culture treatment where C:N was lower in monoculture (C < D, p = 0.002).  1. Ocean acidification (increased pCO 2 ) and eutrophication (nutrient loading) have direct, positive effects on the growth rates of many marine macroalgae, potentially causing shifts in ecosystem structure and function. This enhanced growth of macroalgae, however, may be controlled by the presence of grazers, where indirect effects of acidification and eutrophication result in increased consumption.

References
2. We tested whether a common marine herbivorous snail, Littorina littorea, could increase consumption rates of macroalgae under ocean acidification and eutrophication conditions. Choice experiments were run giving snails both Ulva (an opportunisitic, ephemeral green alga) and Fucus (a slow-growing, perennial brown alga). We measured total consumption rates, diet preference (live algal tissue and artificial, reconstituted algae), and respiration rates in snails.
3. High pCO 2 (acidification) resulted in a 50% reduction in the total consumption rate. Under high pCO 2 conditions, snails switched from a mixed diet to feeding almost exclusively on Ulva. Despite increasing the quality of both Ulva and Fucus, high nutrients did not affect the consumption rates or feeding preference of snails. Snails consumed similar amounts of artificial Ulva and Fucus across environmental treatments, indicating that physical characteristics of algal tissues were driving diet shifts. Snails also showed signs of stress as respiration rates were significantly reduced under high pCO 2 conditions. 4. Synthesis. This study shows an inability of L. littorea to increase consumption to match expected growth rates of macroalgae. In this system, decreased consumption, coupled with increased growth of macroalgae, will ultimately enhance algal growth and spread. Here, the direct, physiological effects on snails had a greater effect than the indirect effects of increased food quality. Grazer diversity will be important in determining the structure and function of coastal communities in the future, and although L. littorea may not be able to increase consumption, other grazers may be able to fill the void. The direct effects of ocean acidification and nutrient loading on macroalgal growth may be counteracted by an enhancement of top-down forces, such as grazing, resulting from the indirect effects of ocean acidification and nutrient loading on grazers and herbivores Falkenberg et al. 2014). Grazers have the ability to control growth and expansion of macroalgae and epiphytes (Hughes et al. 2003). Grazer presence and grazer diversity act to quell increased algal growth under future climate conditions (Falkenberg et al. 2014;Baggini et al. 2015). have investigated this interaction on grazing (but see Falkenberg et al. 2014). It is critical to quantify how acidification and nutrient loading, two factors that can enhance macroalgal growth, impact grazing organisms and their ability to impose top-down forces; these interactions will help in our understanding of community and ecosystem response to climate change.
Intertidal and shallow subtidal communities in the northwest Atlantic are supported by a diverse assemblage of macroalgae which provide both food and habitat (Watt & Scrosati 2013 (Lubchenco 1978(Lubchenco , 1983Watson & Norton 1985;Bracken, Dolecal & Long 2014). Littorina littorea can manipulate algal communities and competitive interactions by preferential grazing and clearing space within the intertidal and shallow subtidal (Lubchenco 1983). However, Littorina has exhibited reduced growth and reduced respiration when exposed to ocean acidification conditions (Bibby et al. 2007;Melatunan et al. 2011;Landes & Zimmer 2012), which may ultimately impact their grazing ability. Therefore, while the quality of algal food may increase under future climate conditions, the direct physiological effects on herbivores may disrupt their ability compensate for excess algal growth.
Here, we investigated the interactive effects of ocean acidification and nutrients on the consumption rates, feeding preference and physiology of a common snail (Littorina littorea) when feeding on algae (Ulva and Fucus). We tested the hypotheses that under novel environmental conditions: (1) grazing organisms can control algal growth under these climate scenarios by consuming at higher rates (Falkenberg et al. 2014) and (2) altered tissue quality under high nutrients will drive diet preferences from a diet of mostly Ulva to a mixed diet, resulting in a partial release from grazing pressure, which may have implications in terms of growth and abundance in this species. We also tested the direct physiological effects of ocean acidification and nutrient loading on grazers by measuring respiration rates.
Identifying these aspects of consumption by an abundant grazer will help determine the influence of top-down forces in structuring coastal communities in the future.

SNAIL CONSUMPTION (LIVE ALGAL THALLI)
To test the impacts of ocean acidification and nutrients on snail consumption rates and feeding preferences, we conducted a series of consumption experiments at a flow-through seawater facility at the US Environmental Protection Agency Atlantic Ecology Division in Narragansett, RI. Ulva, Fucus, and L. littorea were collected from the shallow subtidal zone at the University of Rhode Island's Narragansett Bay Campus beach (41°29'26"N, -71°25'11"4W) in August 2015. Non-reproductive tips of Fucus (~3-5cm in length) were cut from adult thalli (Gutow et al. 2014). Fucus tips and Ulva thalli were cleaned of any epiphytes, and experimental organisms were transferred into separate 20L glass aquaria with flow-through seawater and aeration.
Individuals were acclimated to lab conditions (ambient pCO 2 and nutrient levels, artificial lighting) for five days prior to the start of each experiment.
After five days of acclimation, algae and snails were exposed to one of four environmental treatments, based on the factorial combination of two pCO 2 levels (ambient ~ 400 µatm pCO 2 , pH 8.10, and representative concentration pathway 8.5  (Krumholz, 2012 photons m -2 sec -2 with a light/dark rhythm of 14:10h (L:D). Tanks were scrubbed and cleaned every two days and any epiphytes growing on the algae were removed. After one week, algae were removed, blotted dry, spun, and weighed for final mass.
Two trials were run for this experiment, with new individuals collected and acclimated for the second trial. In this trial, consumer tanks and non-consumer control tanks were switched, resulting in seven replicates per consumer treatment per environmental treatment. There was no significant difference in consumption between trials, therefore, we pooled our data. Total consumption, based on the total algal biomass consumed by snails (both Ulva and Fucus), was calculated as (S i x C f x C i -1 ) -S f , where S i and S f were the initial and final mass of the algae exposed to consumption and C i and C f were the initial and final mass of the non-consumer control algae (Stachowicz & Hay 1999;Jones & Thornber 2010). These calculations were repeated for Ulva and Fucus consumption separately.

SNAIL CONSUMPTION (ARTIFICIAL FOOD)
To test the mechanisms behind potential shifts in snail feeding preference, we ran a series of consumption experiments with artificial (or reconstituted) food. By using artificial food, we were able to remove any physical characteristics of the algal tissue but retain the chemical composition of tissue (i.e. nutritional quality and secondary compounds consumer tanks and non-consumer control tanks, in order to quantify any potential loss of artificial food to treatment conditions. This experiment was run twice in order to obtain seven replicates per consumer treatment per environmental treatment. We observed no differences in consumption between trials and therefore pooled our data.
There was no loss of artificial foods squares in non-consumer control tanks; therefore, we determined artificial food consumption for each species by calculating percent change between initial squares count and final square count.

SNAIL RESPIRATION
To determine the individual physiological effects of environmental treatments on snails, we measured respiration rates in L. littorea from field-collected individuals in November 2015. Using the same experimental design as above, snails were lab acclimated for five days at 18°C before being exposed to environmental treatments.
Snails were kept in one of the four environmental treatments for two weeks and were fed Ulva and Fucus throughout. To measure respiration, small oxygen optodes (PreSens Fibox 3) were placed in 50mL glass vials. Vials were filled with treated water from one of the four environmental treatments. One snail was then placed in the vial of its respective treatment and the vial was sealed. An instantaneous reading of initial oxygen percent within the vial was measured (PreSens Software). Vials were then placed in a water bath kept at a constant 18°C. After 30 minutes, a final instantaneous measurement of oxygen percent within the vial was measured. Overall, measurements of respiration were taken for five snails under each environmental treatment, with an equal number of "blank" vials run (vials without snails) in order to corrected measurements of oxygen consumption. All snails were weighed. Respiration rates were calculated by determining the initial and final amounts of oxygen within the vials (umol L -1 ) and were standardized by snail weight and time for a final measurement of oxygen consumed (umol g -1 L -1 hr -1 ).

STATISTICAL METHODS
Mean total consumption rate (mg snail -1 day -1 ) was compared across environmental treatments using two-way fixed factor analysis of variance (ANOVA) with pCO 2 and nutrient levels as categorical, fixed factors. To determine diet preference, mean consumption of Ulva and mean consumption of Fucus were compared across environmental treatments using a factorial MANOVA to account for non-independence of algal consumption (Roa 1992 By adding nutrients to our aquaria we effectively increased concentrations of DIN above ambient levels. DIN in high nutrient treatments averaged 202.4 + 6.9 µM and 196.6 + 11 µM, respectively; whereas ambient nutrient treatments averaged 6.92 + 0.78 µM and 6.96 + 0.5 µM DIN, respectively (Table 1).
We found a significant effect of pCO 2 on the diet preference of L. littorea (p < 0.0001; Table 3). While there was no significant effect of environment on snail consumption rates of Ulva ( Fig. 2; Table 3), consumption rates of Fucus were significantly decreased (p < 0.0001; Table 3) and almost nonexistent under conditions of acidification, as rates dropped from 53.6 mg snail -1 day -1 under ambient pCO 2 to 8.0 mg snail -1 day -1 under high pCO 2 ( Fig. 2; Table 3). Snail consumption rates of Fucus were not affected by nutrient level or the interaction of pCO 2 and nutrients (Table 3).
The proportion of artificial Ulva consumed ranged between 0.45 and 0.70 while the proportion of artificial Fucus consumed ranged between 0.58 and 0.73 (Fig. 3). There was no significant change in algal diet across environmental treatments based on the chemical composition of tissue (i.e. nutritional quality and secondary compounds).

DISCUSSION:
In marine ecosystems, ocean acidification and nutrient loading are predicted to enhance the growth and spread of ephemeral and turf algae at the expense of foundational species like corals, seagrasses, and perennial macroalgae (i.e. kelps), as well as reduce overall community diversity (Worm & Lotze 2006;Connell & Russell 2010;Diaz-Pulido et al. 2011;Hale et al. 2011;Koch et al. 2013). Increasing evidence points to the indirect effects of climate change on herbivores to exert top-down control on excessive algal growth, promoting community resilience (Russell & Connell 2005;Falkenberg et al. 2014;McSkimming et al. 2015). In this system, Ulva grows three times faster under high pCO 2 and nutrients compared to growth under ambient conditions, and Fucus maintains constant growth or exhibits a reduced growth rate (Ober & Thornber in review). We found an inability of L. littorea to increases consumption and compensate for increased growth of macroalgae as its total algal consumption rate decreased by 50% under high pCO 2 , with no effect of nutrient level. Russell et al. (2013) found a similar decrease in consumption of biofilms by L. littorea under high CO 2 but simultaneously observed consumption differences based on the exposure time of snails, where a "shock" of low pH resulted in lower consumption than snails that were acclimated to treatments for five months.
In rocky shore communities, nutrient loading can stimulate the growth of bloom-forming Ulva (Enteromorpha) spp., but grazer consumption does not control or slow blooms, leading to decreases in diversity within these communities (Worm & Lotze 2006). When exposed to high nutrients, both Ulva and Fucus tissues decreased . Marine grazers, including L. littorea, exhibit a preference for higher quality food (Watson & Norton 1985) and early successional, ephemeral species (Lubchenco 1978). Littorina littorea consume both Ulva and Fucus, with a preference for Ulva over Fucus, but also preferentially feed on algal epiphytes and microalgae (Watson & Norton 1985;Russell et al. 2013;Bracken et al. 2014). Our study shows that, under high pCO 2 , Ulva is selected over Fucus as a result of the physical characteristics of the algal tissue, with Fucus having tougher tissue (Watson & Norton 2009). Despite a lack of preference for Fucus, our study remains ecologically relevant as L. littorea can appear in such high densities along the coastline of the northwest Atlantic (greater than 200 individuals m -2 ) that even minimal consumption plays a large role in structuring the algal communities (Perez et al. 2009;Poore et al. 2012). While our study showed changes in algal preference (or choice) under acidification conditions, we did not test aspects of algal use or selection. The amount of time L. littorina spends using either Ulva or Fucus could provide insight as to how these two algal species are used by snails. In addition, we did not manipulate the abundance of the algal species.
In terrestrial snails, availability (abundance) is related to the proportion consumed, but preference does not rely on abundance (Speiser & Rowell-Rahier 1991).
Detrimental effects of acidification on the physiology of many marine species have been reported Stumpp et al. 2011;Barton et al. 2012). We observed a significant decrease in the respiration rate of L. littorea when exposed to acidification, similar to Leung et al. (2015) and Melatunan et al. (2011). stages, as well as increasing feeding rates (Bezemer & Jones 1998;Stiling & Cornelissen 2007). Increased CO 2 can alter tissue composition of both terrestrial and marine primary producers, which can indirectly influence feeding rates of herbivores (Bezemer & Jones 1998;Falkenberg et al. 2014). Our study highlights the overwhelming direct effects of climate change on L. littorea, which lead to decreased consumption of macroalgae, despite positive indirect effects (increased food quality).
Coupling decreased consumption with increased growth rates of opportunistic species under future CO 2 conditions will ultimately enhance algal growth in this ecosystem and could lead to shifts in macroalgal assemblages, favoring ephemeral species at the expense of more long-lived species that are critical in creating suitable habitat for other organisms (Worm & Lotze 2006).
As herbivores play a major role in structuring communities (Poore et al. 2012), quantifying the direct physiological effects and the indirect effects (altered food quality) will be critical in determining the response of communities and ecosystems . Ultimately, community resilience will depend on a diversity of herbivores (Baggini et al. 2015), as climate change is likely going to affect organisms differently ). While certain species might not have an ability to increase consumption under high CO 2 , other herbivores in this ecosystem may be able to compensate. As such, our understanding of ecosystem response to climate change will be dictated by the species involved and the diversity of the community.

ACKNOWLEDGMENTS:
This research has been supported by a grant from the U.S. Environmental Protection Agency's Science to Achieve Results (STAR) program (awarded to G Ober). This material is based upon work supported in part by the National Science Foundation EPSCoR Cooperative Agreement #EPS-1004057. We thank the US EPA Atlantic Ecology Division for support and aide in experimental set up and design, and A Pimenta for laboratory assistance. We also thank M Birk and E McLean for their help in obtaining physiological measurements.     will aid the overall expansion and growth of fleshy macroalgae in coral reef ecosystems, as opportunistic algae may have an advantage over other reef-associated species. Changes in turf community diversity will help provide insight into how macroalgal communities may be structured in the future, highlighting genera primed to take advantage of the changes in ocean chemistry associated with ocean acidification.

REFERENCES
Key Words: ocean acidification, climate change, turf algae, diversity, coral reef

INTRODUCTION:
As atmospheric carbon dioxide (CO 2 ) continues to rise, ocean pH continues to decrease (Pachauri et al. 2014). As a result, ocean acidification (OA) is happening at unprecedented rates Hofmann et al. 2011). OA and associated changes in ocean biogeochemistry, specifically changes in aragonite and calcite saturation states, have been targeted as problematic and potentially lethal for many calcifying marine organisms. Under future projected OA levels, impaired calcification and shell dissolution are likely to occur (Fabry et al. 2008;Waldbusser et al. 2015).
Most OA research has focused on the physiological responses of calcifying organisms (Orr et al. 2005;Hoegh-Guldberg et al. 2007;). Concurrently, noncalcifying primary producers, like fleshy macroalgae, seem primed to take advantage of chemical changes associated with OA (Gao et al. 1991;Kroeker et al. 2013;. The response of fleshy macroalgae to OA is highly species specific (Raven 1997;, but many species have exhibited positive, or at least neutral, impacts on growth rates (Gao et al. 1991;Johnson et al. 2014). OA shifts the proportion of carbon species in the water column, increasing the relative amount of bicarbonate (HCO 3 ) and CO 2 (Zeebe and Wolf-Gladrow 2001;Sabine et al. 2004). Overwhelmingly, macroalgae use HCO 3 to obtain carbon for photosynthesis, facilitated by the use of carbon concentrating mechanisms (CCM) . However, many primary producers are able to utilize the excess CO 2 in the water column by passive diffusion and can reallocate CCM energy to growth, reproduction, or defense (Johnston 1991;Magnusson et al. 1996;Hurd et al. 2009). Some species, however, are not able to take advantage of the excess CO 2 , or the change in pH disrupts the activity of the CCM, and as a result macroalgae exhibit reduced growth (Swanson and Fox 2007;Gutow et al. 2014).
Ultimately, however, it is expected that future ocean environments will favor the growth and spread of fleshy macroalgae (Harley et al. 2006;Diaz-Pulido et al. 2011).
The divergent responses of calcifying and non-calcifying organisms may lead to major shifts in communities and assemblages (Hall-Spencer et al. 2008;Connell and Russell 2010;Anthony et al. 2011;Porzio et al. 2011). One such community is the coral reef, where foundation species such as corals and crustose coralline algae are susceptible to OA and are predicted to be replaced by fleshy macroalgae (Hoegh-Guldberg et al. 2007;Anthony et al. 2008;Kuffner et al. 2008;Diaz-Pulido et al. 2011). Shifting from a coral dominated system to an algal dominated system will likely yield changes in ecosystem productivity and diversity (Fabry et al. 2008).
Filamentous turf algae are an important component in many coastal ecosystems (Airoldi et al. 1995). Despite morphological diversity, turf algae are lowlying (0.5-10cm), can grow rapidly and cover a wide area, and are typically good at trapping sediment ). Turf algae act as a key food source for grazers (Carpenter 1986;Morrison 1988) and since they have a shorter life span and due to their rapid turnover, turf algae contribute significantly to nutrient cycling in reef systems (Klumpp et al. 1987;Connell et al. 2014). Turf algae are one of the most abundant benthic components in coral reefs (Klumpp et al. 1992;Vermeij et al. 2010) and are able to quickly colonize space, taking over bare rock or dead coral (Diaz-Pulido and McCook 2002;Carilli et al. 2009). Algal turfs can directly or indirectly affect coral settlement and growth (Birrell et al. 2005;Vermeij et al. 2010;Venera-Ponton et al. 2011). Turf algal communities are expected to thrive under future OA scenarios in many ecosystems (Connell and Russell 2010), potentially blocking the settlement and growth of habitat forming kelps (Connell and Russell 2010;Falkenberg et al. 2014) and corals.
Despite a diverse species composition, turf algal communities are typically treated as one entity (Westphalen and Cheshire 1997;Connell et al. 2014;Fricke et al. 2014). Research focusing on turf algae and OA has focused on the response of the community (Connell and Russell 2010;Kroeker et al. 2013) or one dominant species (Falkenberg et al. 2014). Few studies, however, have investigated turf assemblages at different pH ranges (but see Porzio et al. 2011;Porzio et al. 2013;Bender et al. 2014).
Due to the diversity of turf communities, species within the community are likely The overall goal of this research was to investigate how turf algal communities respond to OA. Using three levels of CO 2 (ambient, medium, and high), representing three OA scenarios, we tested the response of reef-associated turf communities in the Great Barrier Reef, Australia in terms of species diversity, richness, total biomass, and organic content. We hypothesized that under medium and high CO 2 scenarios, turf algal communities would exhibit increased growth rates and a higher organic content (an indicator of algal health; Taylor et al. 2002) than in ambient CO 2 treatments, due to the excess carbon and potential reallocation of resources. We also hypothesized that turf species diversity would be lower under higher CO 2 treatments (Porzio et al. 2011) and that turf communities exposed to higher CO 2 would become red algal dominated (Raven and Beardall 2003). We interpret our results in the context of the resilience of turf algal communities.

Turf algae collection
Turf algal communities were collected in the reef slope of Coral Gardens, Heron Island, Great Barrier Reef, Australia (23 o 26.698' S, 151 o 54.533' E) at a depth of 6-8m. We used turf algal communities growing on the tips of dead Acropora spp.
coral branches. Dead coral branches were collected by hand or using cutters and were placed in plastic bags underwater. We used two criteria for selecting the turf algal communities: 1) turfs exhibited comparable densities of algal growth (Fig. 1) and 2) dead coral branches were of around 7-10cm long. Turfs were then immediately transported to the outdoor flow-through aquaria facilities of Heron Island Research Station (HIRS) and acclimatized in tanks with running seawater during one week prior to the experiment. Dead coral fragments covered in turf algae were then placed in the bottom of 10L plastic tanks and randomly assigned to CO 2 treatments.

Biomass and organic content of turf communities
Pieces of coral branches covered in turf algae were thawed and first measured for total length and circumference to standardize turf community biomass results. One half of the turf-covered coral then was scraped with a razor until all turf algae were removed. The non-scraped half of the coral was saved for assessing turf community assemblage (see Turf community composition). As the purpose of this study was to investigate non-calcifying turf species, scraped turf algae were decalicified with a 1M HCL solution for ten minutes in order to dissolve any existing calcifying species and coral (Bender et al. 2014). After dissolution, the HCL solution was rinsed with DI water and turf algae were blotted dry and placed in pre-weighed aluminum tins. Wet mass of each replicate was taken and then standardized to mg/cm 2 based on surface area measurements from coral rubble.
Tins containing the turf algal community replicates were then placed in a 60°C drying oven for approximately 48 hours. After 48 hours, tins were removed and weighed to determine the dry mass of the turf community. Tins were placed in a muffle furnace at 550°C for 2.5 hours and turf ash mass for each replicate was recorded. Organic content of turf communities was determined by calculating the difference between total dry mass and ash mass (inorganic tissue) (Taylor et al. 2002).

Turf community composition
Using the same pieces of coral rubble, we thawed turf algal replicates and removed a 1cm X 1cm square of turf algae from the covered half of the rubble via scraping. The removed turf algae was rinsed for 10 minutes in a 1M HCL solution to dissolve and remove any unwanted calcified algae and coral (Bender et al. 2014).
Turfs were then rinsed and blotted dry. Turf algae was then placed on a glass microscope slide and spread into one thin layer, with one slide per replicate (as per, Diaz-Pulido and McCook 2002). Turf algae were stained aniline blue dye, allowing algal features to be determined, ultimately aiding in the identification of turf algal genera and cyanobacteria. Once the dye was set, slides were examined at a magnification of 40X and turf algae were identified down to the genus.
We determined the relative abundance of present genera by selecting five fields of view on each slide, recording the genera present, and calculating percent cover within the field of view. Since empty spaces on each slide were a byproduct of mounting the turf community, percent cover of empty space in each field of view was also calculated in order to determine the relative abundance of each genus. Relative abundance of each genus was averaged over the five fields of view to establish an overall breakdown of turf community for each replicate. Genus was used as the taxonomic unit of classification as many turf algal species are cryptic and unidentifiable without molecular analysis.

Statistical analysis: Biomass and organic content
Turf community biomass was standardized to mg/cm 2 (dry mass) and organic content (%C of tissue) was calculated. Due to the nature of the experimental set-up and potential pseudoreplication, we ran our data with a nested analysis of variation (ANOVA), nesting tank within CO 2 treatment to test whether the treatment tank was a significant factor. Our nested ANOVA showed tank was not a significant factor (F 3,25 = 1.752, p = 0.182) and we were able to proceed with our statistical analysis with each piece of coral rubble as a replicate. For each response variable, a one-way ANOVA model was performed using JMP v 11 (www.jmp.com). If the statistical model indicated a significant effect, Tukey's HSD was run post-hoc.

Statistical analysis: Diversity metrics and community structure
We calculated mean genus richness for each CO 2 treatment from the observed present genera. Genus evenness for each CO 2 treatment was determined by the observations and abundances within the five fields of view. The Shannon-Weiner Index (H') was used to assess the diversity of turf genera. H' was calculated for each replicate based on average relative abundances of each present genera in the five fields of view. We assessed the difference among treatments for mean richness, evenness and diversity (H') with one-way ANOVAs. Where significant differences occurred, we used a Tukey's HSD test as a means of post-hoc analysis.
Turf algal community structure was analyzed within and across treatments based on the relative abundance of present genera using Primer v6 (http://www.primer-e.com; accessed January 2013). A Bray-Curtis similarity index was used to create an MDS plot to visualize similarities among turf communities treated under the three levels of CO 2 (Clarke 1993). To determine whether there were significant differences in community structure between CO 2 treatments we ran a oneway analysis of similarity (ANOSIM) using the vegan package in R (version 2.14, R Development Core Team, 2016, R-project.org). For genera of interest, we ran one-way ANOVA on percent cover among CO 2 treatments.

Seawater parameters
Elevated CO 2 and decreased pH were realized for each one of our treatments.

Turf community biomass and organic content
We observed a significant, positive affect of OA on turf community biomass (F 2,26 = 3.24, p = 0.05 Fig. 2), where turf communities grown under the most extreme acidification scenario (the highest CO 2 treatment) had the greatest biomass. The high CO 2 treatment had a mean community biomass that was 50% and 20% greater than in the medium and control (ambient) CO 2 treatments, respectively. Post-hoc analysis revealed a significant increase in turf biomass between medium and high CO 2 treatments (p = 0.04), but no significant difference between ambient CO 2 and high CO2 treatments.
While ocean acidification is expected to have detrimental effects on calcifying organisms, the effect on non-calcifying, fleshy macroalgae appears to be more species specific (see Johnson et al. 2014). Depending on their ability to acquire carbon from the water column, some algal species have shown increased growth rates under elevated CO 2 levels , while other species appear stressed and their growth rates decline (Gutow et al. 2014). Under future conditions of ocean acidification, turf algal communities flourish via enhanced growth and primary productivity Connell and Russell 2010;Falkenberg et al. 2014).
Our observed response of a greater mean turf biomass under high CO 2 scenarios is similar to results from Falkenberg et al. (2014) and Connell and Russell (2010). Our analysis indicates that greater community biomass only appears under the high CO 2 treatment. We interpret this as a potential threshold for turf algal communities, wherein modest increases in CO 2 are not enough to influence community growth, but as CO 2 levels continue to increase, turf communities may flourish.
An indicator of algal health and nutritional quality, organic content would likely be higher when more resources are available .
Community composition has the potential to play a large role in determining collective organic content; turf algal species have different levels of baseline organic content and respond differently to ocean acidification . For example, Falkenberg et al. (2014) found the C:N within the brown filamentous alga Feldmannia spp. was reduced under high CO 2 and nutrient addition. Alternatively, Gutow et al. (2014) found that nutritional quality (as determined by C:N:P) was unaffected by CO 2 .
These results point to resilience of turf species, where under acidification conditions, turfs are able to maintain their tissue quality and health.

Diversity and community structure
Overall, OA had no effect on genus richness (F 2,26 = 0.08, p = 0.92). Turf communities grown under all three OA scenarios had means of ~12 genera (Table 2).
Despite genus richness consistency across treatments, CO 2 had a significant effect on genus evenness (F 2,26 = 3.818, p = 0.036), with increased evenness found with increased CO 2 (Table 2). Turf genus diversity (Shannon H') was also significantly impacted by CO 2 level (F 2,26 = 3.4285, p = 0.0495) with medium and high CO 2 levels resulting in significantly larger H' than control communities (Table 2). Thus, despite having similar numbers of genera present, the abundance of these genera was influenced by OA.
We constructed an MDS plot (Fig. 3) to assess turf community similarities.
Based on present genera and relative abundance, we expected communities of the same CO 2 treatment to group together. Despite significant differences in H' and evenness, the resulting plot yielded no strong patterns grouping communities by treatment, likely due to the overall variability within treatments. Results from our ANOSIM support the lack of difference found between treatment communities (p = 0.86). These results agree with Bender et al. (2014), showing no pattern of community similarity under different CO 2 treatments.
We had initially hypothesized that our turf communities would become less diverse with increased CO 2 as has been observed in other macroalgal communities (Hall-Spencer et al. 2008;Porzio et al. 2011). However, we found that increasing CO 2 results in increased turf community H ' . Although CO 2 and ocean acidification act as a major source of stress for many marine organisms, fleshy macroalgae, like those comprising turf communities, are expected to either flourish or persist. Most macroalgae use CCMs to convert the abundant bicarbonate molecules for use in photosynthesis; in addition, many algal species also have the ability to passively diffuse CO 2 .. Species that can passively diffuse CO 2 are the ones expected to thrive under ocean acidification conditions, where CO 2 is more readily available and existing CCMs can be down-regulated (Cornwall et al. 2012). The turf community analysis in this study suggests that the dominant present genera represented algae that have CCMs, but more work needs to be done to determine the contribution of passive diffusion and/or quantifying CCM down-regulation. By investigating genus evenness of turf communities, we were able to determine that increased levels of CO 2 had a positive effect on evenness. Despite observing diverse communities across all treatments, those communities grown under ambient CO 2 were dominated by only a few of the present genera, particularly Polysiphonia spp. Abundances within medium CO 2 and high CO 2 communities were more evenly distributed. Similar to genus evenness, H' for medium and high CO 2 treatments was than H' in ambient CO 2 treatments. However, this may be due entirely to the abundance of Polysiphonia spp., the genus accounting for 30% of the community under ambient CO 2 treatments. Under medium and high CO 2 , Polysiphonia abundance is decreased. Statistical analysis of Polysiphonia cover over CO 2 treatment reveals a trend where abundance decreases as CO 2 increases (F 2,26 = 2.810, p = 0.077). Species within the genus Polysiphonia have CCMs (as inferred from carbon stable isotope (δ13C) signatures, Raven et al. 2002), and other studies have indicated either a lack of Polysiphonia in low pH environments (Porzio et al. 2013) or reduced cover under high CO 2 (Bender et al. 2015). We can thus infer that the increased evenness and diversity of our treated turf communities in low pH environments is correlated to the decrease in cover of Polysiphonia.
We hypothesized that red algae would dominate turf communities under future projections of OA due to the larger relative portion of species within this phylum that have the ability to diffuse CO 2 (Raven 1997;Raven and Beardall 2003). We separated genus relative abundance by phyla, including cyanobacteria (Fig.   4). Our study showed a significant decline in the relative abundance of red turf algae from the ambient CO 2 treatments to the medium and high CO 2 treatments (Fig 3, F 2,26 = 3.428, p = 0.048). The relative abundance of green turfs and brown turfs were unaffected by CO 2 treatment (F 2,26 = 1.021, p = 0.37, F 2,26 = 1.115, p = 0.34, respectively). We observed significantly greater abundances of cyanobacteria with increasing CO 2 (F 2,26 = 3.328 , p = 0.05), supporting evidence of Bender et al. (2014).
Here, cyanobacteria abundance was significantly higher in medium than ambient CO 2 treatments (p = 0.02). Several studies have investigated the relative abundance of different phyla at different pH levels (Hall-Spencer et al. 2008;Porzio et al. 2011;Bender et al. 2014). Porzio et al. (2011) found higher abundance of brown algae at extreme low pH (6.7) from field surveys but little change in community structure between normal pH zones (8.1) and slightly more acidic zones (7.8). Hall-Spencer et al. (2008) found similar success for brown algae and found some resilient green algae. Bender et al. (2014) showed a trend shifting from higher abundances of brown turf species to higher counts of red turf algae and cyanobacteria as pH decreased. Our analysis of turf algal communities supports Bender et al. (2014) findings of increased abundance of cyanobacteria with increased pH, but we found an opposite trend in red algal abundance. This discrepancy points to the need for determining species-specific responses, in addition to continued studies of entire communities for a better understanding of algal community dynamics to climate change.
Overall, we found there to be slight, but significant changes to turf communities when exposed to different levels of CO 2 . These findings support the positive effect of elevated CO 2 on growth of turfs Connell and Russell 2010;Falkenberg et al. 2014). Under future climate conditions, where coral cover will likely decline as a result of bleaching caused by global warming (Hoegh-Guldberg and Bruno 2010), resilient and opportunistic turf algal communities may play a large role in the phase shift expected in this ecosystem ). The shift in dominance is also likely to occur in kelp forest ecosystems (Connell and Russell 2010;Falkenberg et al. 2014), where long-lived kelps are expected to be replaced by opportunistic turf communities. Turf algal communities can increase genus diversity and evenness under future ocean acidification by reducing the abundance of dominant species, opening up valuable space and resources. While the mechanism behind this shift is unclear, whether it is a result of decline of a   Figure 1. Turf algal community replicates for ambient CO 2 treatment (left), medium CO 2 treatment (middle), and high CO 2 treatment (right) at the beginning of experimentation.

Figure 2.
Mean biomass of turf algal communities (mg cm -1 +/-1 SE) after 41 days of experimentation under three levels of CO 2 . Letters represent significant differences between CO 2 treatments (post-hoc).  . Mean relative abundance (proportion of present genera) of turf algal phyla and cyanobacteria (+/-1 SE) by CO 2 treatment. A significant effect of CO 2 was observed in the relative abundance of Rhodophyta where relative abundance under medium and high CO 2 are less than that of the ambient treatment (p = 0.048).