Coastal ocean acidification and nitrogen loading facilitate invasions of the non-indigenous red macroalga, Dasysiphonia japonica

Coastal ecosystems are prone to multiple anthropogenic and natural stressors including eutrophication, acidification, and invasive species. While the growth of some macroalgae can be promoted by excessive nutrient loading and/or elevated pCO2, responses differ among species and ecosystems. Native to the western Pacific Ocean, the filamentous, turf-forming rhodophyte, Dasysiphonia japonica, appeared in estuaries of the northeastern Atlantic Ocean during the 1980s and the northwestern Atlantic Ocean during the late 2000s. Here, we report on the southernmost expansion of the D. japonica in North America and the effects of elevated nutrients and elevated pCO2 on the growth of D. japonica over an annual cycle in Long Island, New York, USA. Growth limitation of the macroalga varied seasonally. During winter and spring, when water temperatures were < 15 °C, growth was significantly enhanced by elevated pCO2 (p < 0.05). During summer and fall, when the water temperature was 15–24 °C, growth was significantly higher under elevated nutrient treatments (p < 0.05). When temperatures reached 28 °C, the macroalga grew poorly and was unaffected by nutrients or pCO2. The δ13C content of regional populations of D. japonica was −30‰, indicating the macroalga is an obligate CO2-user. This result, coupled with significantly increased growth under elevated pCO2 when temperatures were < 15 °C, indicates this macroalga is carbon-limited during colder months, when in situ pCO2 was significantly lower in Long Island estuaries compared to warmer months when estuaries are enriched in metabolically derived CO2. The δ15N content of this macroalga (9‰) indicated it utilized wastewater-derived N and its N limitation during warmer months coincided with lower concentrations of dissolved inorganic N in the water column. Given the stimulatory effect of nutrients on this macroalga and that eutrophication can promote seasonally elevated pCO2, this study suggests that eutrophic estuaries subject to peak annual temperatures < 28 °C may be particularly vulnerable to future invasions of D. japonica as ocean acidification intensifies. Conversely, nutrient reductions would serve as a management approach that would make coastal regions more resilient to invasions by this macroalga.


Introduction
The continued anthropogenic delivery of carbon dioxide (CO 2 ) into surface oceans is changing the carbonate chemistry of the ocean, yielding reductions in pH and CO 3 2levels and increases in CO 2 and HCO 3 levels (Meehl et al. 2007). Beyond the combustion of fossil fuels, eutrophication-enhanced microbial respiration can also be a strong regional source of CO 2 and acidification that can result in the seasonal accumulation of respiratory CO 2 to levels that can exceed end of century projections of pCO 2 for the open ocean ([ 1000 latm;Cai et al. 2011;Melzner et al. 2013;Wallace et al. 2014). While ocean acidification can have negative consequences for calcifying organisms (Gazeau et al. 2007;Ries et al. 2009;Talmage and Gobler 2011;Young et al. 2019) elevated pCO 2 can benefit some, but not all, photosynthetic organisms (Hattenrath-Lehmann et al. 2015;Koch et al. 2013;Palacios and Zimmerman 2007), such as macroalgae (Gao et al. 1993;Hepburn et al. 2011;Young and Gobler 2016;Young et al. 2018). However, elevated pCO 2 may also benefit invasive macroalgae and provide them an advantage over their native counterparts (Hall-Spencer et al. 2008).
Invasive macroalgae can be a significant ecosystem threat. Invasive species often tolerate a wide range of abiotic conditions and may be less palatable to grazers than native species (Hayes and Barry 2008;Low et al. 2015;Theoharides and Dukes 2007). The filamentous, turf-forming rhodophyte Dasysiphonia (formerly Heterosiphonia) japonica (henceforth Dasysiphonia) is native to the western Pacific Ocean (Chihara 1970;Choi et al. 2009;Okamura 1936) but was introduced into European waters in 1984 (Sjøtun et al. 2008). Since then, this macroalga has spread throughout the eastern Atlantic Ocean from France north to Norway and Sweden (Husa et al. 2004;Lein 1999), south into the Mediterranean Sea (Sagerman et al. 2015;Sjøtun et al. 2008), and as far west as Scotland (Moore and Harries 2009;Supplementary Fig. S1). Following this introduction, the macroalga spread to the western Atlantic Ocean. In North America, the macroalga has been found as far south as Rhode Island and eastern Long Island Sound (USA) (Newton et al. 2013;Ramsay-Newton et al. 2017;Schneider 2010) and as far north as Maine and Nova Scotia, Canada (Idlebrook 2012; Savoie and Saunders 2013;Supplementary Fig. S1). While the species appears to be a cold-water alga, Bjaerke and Rueness (2004) found that the macroalga's temperature range for growth is 0-30°C and that the optimal temperature for growth is 19-25°C. Given this, it is possible this macroalga will continue to invade the North American East Coast (Bjaerke and Rueness 2004). The role of anthropogenic processes, including ocean acidification and nutrient loading, in facilitating the spread of Dasysiphonia, is unknown.
Excessive nutrient loading can promote the overgrowth of larger indigenous macroalgae (Conley et al. 2009;Valiela et al. 1997) by smaller morphologically simpler species that are capable of rapid growth in the presence of high nutrient concentrations and have a high assimilative capacity for nutrients (Neori et al. 2003;Valiela et al. 1997). Despite the great diversity of macroalgae, eutrophic conditions often lead to dominance by opportunistic chlorophytes and several branching or filamentous genera of rhodophytes such as Gracilaria (Valiela et al. 1997). To date, it has been generally concluded that such macroalgae can dominate estuaries with high nutrient loads and shorter residence times due to their ability to attach to the benthos and form dense beds (MacIntyre et al. 2004;Valiela et al. 1997). The extent to which invasions of Dasysiphonia are related to eutrophication is presently unknown.
Beyond nutrients, elevated pCO 2 has been shown to benefit multiple marine macroalgae, including a variety of non-calcifying chlorophytes (Björk et al. 1993;Holmer et al. 2005;Olischläger et al. 2013;Young and Gobler 2016), phaeophytes (Hepburn et al. 2011), and rhodophytes (Gao et al. 1993;Hofmann et al. 2012;Xu et al. 2010). Of these three divisions, Rhodophyta has the largest number of species that are obligate CO 2 users with a high affinity for CO 2 and thus may respond more positively as CO 2 levels increase compared to species that rely exclusively on HCO 3 or use both forms of inorganic carbon (Badger et al. 1998;Johnson et al. 1992;Koch et al. 2013). The extent to which a macroalga may benefit from elevated pCO 2 is partly dependent on whether inorganic uptake of the macroalga is substrate-saturated at modern day pCO 2 levels (Koch et al. 2013). To date, no study has explored the effects of elevated pCO 2 on the growth of Dasysiphonia.
The objective of this study was to assess how elevated nutrients and pCO 2 levels may individually and collectively affect the growth rate of the invasive rhodophyte Dasysiphonia for the entirety of its growing season (Newton et al. 2013;Ramsay-Newton et al. 2017;Schneider 2010). Surveys of Dasysiphonia were performed across the south shore of Long Island (New York, USA) during which d 13 C and d 15 N content of macroalgal tissue were quantified to evaluate C and N sources. Laboratory experiments were performed monthly over an annual cycle, exposing the macroalga to ambient and elevated pCO 2 levels with and without nutrient enrichment with each experiment performed at ambient water temperatures (4-28°C). Given the invasive nature of Dasysiphonia, further experiments were performed to assess the growth response of co-occurring native macroalgae to elevated pCO 2 and nutrients in both the presence and absence of Dasysiphonia (details below). For those experiments, the foliose rhodophyte Porphyra purpurea (henchforth Porphyra) and cylindrical, branching rhodophyte Agardhiella subulata (henchforth Agardhiella) were examined. Both species are common in estuaries across Long Island (Pedersen et al. 2008;Stewart Van Patten and Yarish 2009;Tang et al. 2015). For all experiments, growth rates, d 13 C signatures, and elemental composition were evaluated and analyzed.

Experimental design and set-up
During spring, summer, and fall 2018, and winter 2019, ten experiments were performed to assess the effects of elevated pCO 2 , nutrients, and temperature on the growth rates of Dasysiphonia. All experiments were performed following previously published methods Gobler 2016, 2017;Young et al. 2018) in 1-L polycarbonate vessels, which were acid-washed (10% HCl) and liberally rinsed with deionized water prior to use. Seven parallel experiments were performed with Porphyra during summer and fall 2018. Vessels were placed in an environmental control chamber set to a constant temperature consistent with ambient water temperatures (4-28°C), light intensity (* 250 lmol photons m -2 s -1 ), and duration (i.e. light: dark cycle) at the macroalgal collection site. The light intensity and temperature measurements were made with a LI-COR LI-1500 light sensor logger and a YSI Pro Series sonde at collections sites (see below) and were obtained via regional monitoring buoys. Vessels were filled with filtered (0.2 lm polysulfone filter capsule, PallÓ) seawater collected from the macroalgal collection sites and were randomly assigned, in quadruplicate, to one of four treatments: a control with ambient pCO 2 (* 400 latm) and ambient nutrients (* 1-5 lM nitrate, * 0.1-1 lM phosphate), a treatment with ambient pCO 2 and elevated nutrients (50 lM nitrate, 3 lM phosphate), a treatment with elevated pCO 2 (* 1800 latm) without nutrient additions, and a treatment with elevated pCO 2 (* 1,800 latm) and elevated nutrients (50 lM nitrate, 3 lM phosphate), resulting in 16 experimental vessels per experiment. The pCO 2 and nutrient levels used during experiments were within the range of levels of US East Coast estuaries (Baumann and Smith 2017;Baumann et al. 2015;Wallace et al. 2014;Wallace and Gobler 2015) and were used during prior experiments that involved macroalgae from Shinnecock Bay, New York, USA Gobler 2016, 2017;Young et al. 2018).
A 3.8 9 1.3 cm air diffuser (Pentair) connected with Tygon tubing to 1-mL polystyrene serological pipettes inserted into the bottom of each vessel was used to deliver dissolved gases to vessels. Vessels were subjected to ambient (400 latm) and elevated (* 2000 latm) pCO 2 via a gas proportionator system (Cole ParmerÒ flowmeter system, multi-tube frame) that mixed ambient air from an air source with 5% CO 2 gas (Talmage and Gobler 2010). Gases were mixed and delivered at a flow rate of 2500 ± 5 mL min -1 , which turned over the volume of the vessels [ 1,000 times daily. Bubbling was initiated 2-3 days prior to the beginning of each experiment to allow pCO 2 levels and pH to reach a state of equilibrium. All experiments lasted approximately one week, a duration consistent with prior studies that yielded significant changes in the growth of macroalgae Gobler 2016, 2017). Measurements of pH within experimental vessels were made daily with a Honeywell Durafet III ion-sensitive field-effect transistor-based (ISFET) solid-state pH sensor (± 0.01 pH unit, pH unit, total scale), which was calibrated with a seawater pH standard (Dickson 1993). Measurements of pH made with the Durafet were compared to measurements made spectrophotometrically using m-cresol purple (Dickson et al. 2007) and were found to be nearly identical and never significantly different.
Discrete seawater samples were collected at the beginning and conclusion of all experiments to directly measure DIC within each experimental vessel in each treatment (n = 4 per treatment). DIC samples were preserved using a saturated mercuric chloride (HgCl 2 ) solution and stored at * 4°C until analysis. DIC samples were analyzed by a VINDTA 3-D (versatile instrument for the determination of total inorganic carbon) delivery system coupled with a UIC Inc. coulometer (model CM5017O) as per Young and Gobler (2018). Final total alkalinity, pCO 2 , and levels of HCO 3 -(Supplementary Table S1) were calculated from measured DIC levels, pH, temperature, and salinity, as well as the first and second dissociation constants of carbonic acid in seawater (Millero 2010) using the program CO2SYS (http://cdiac.ess-dive.lbl. gov/ftp/co2sys/, last access: 23-November-2019). For quality assurance, DIC levels and pH within certified reference material (provided by Andrew Dickson of the University of California, San Diego, Scripps Institution of Oceanography; batches 159 and 173 = 2027.14 and 2042.41 lmol DIC kg seawater -1 , respectively) were measured during analyses of every set of samples. The analysis of samples only continued after complete recovery (99.8 ± 0.2%) of certified reference material was attained. Actual mean pH and pCO 2 were 7.99 and 408 latm, respectively, for ambient conditions, and 7.36 and 1860 latm, respectively, for elevated pCO 2 conditions. These values were within the range found seasonally in some estuarine environments (Baumann and Smith 2017;Baumann et al. 2015;Wallace et al. 2014;Wallace and Gobler 2015).
Assessing the effects of elevated pCO 2 and nutrient levels on Dasysiphonia and Porphyra Dasysiphonia and Porphyra were collected from a shallow-water (depths of 1-2 m) area in Great South Bay, New York, USA ( Fig. 1; 40.73°N, -73.05°W) during low tide. All macroalgae collected were found growing subtidally. 20-L carboys were used to collect water from the collection sites. Large, well-pigmented, robust fronds of Dasysiphonia and Porphyra were collected, placed in coolers of seawater from the collection site, and transported to the Stony Brook Southampton Marine Science Center within 30 min of collection. Identification of macroalgae was based on morphology, microscopy, and known biogeography.
Dasysiphonia is a common, but non-native species of red macroalgae along the North American east coast including around Long Island (Newton et al. 2013;Ramsay-Newton et al. 2017;Schneider 2010). Porphyra is a common species of rhodophyte in estuaries across Long Island (Pedersen et al. 2008;Stewart Van Patten and Yarish 2009;Tang et al. 2015). While Dasysiphonia is present most of the year in NY coastal waters, Porphyra is present seasonally during warmer months. Hence, experiments for Dasysiphonia were performed February through November, whereas Porphyra experiment were performed May through November. Pieces approximately 5 cm in length were cut from larger thalli, rinsed in filtered (0.2 lm) seawater and placed in a salad spinner to remove debris and epiphytes. They were then extensively rinsed with filtered seawater before being spun again to further remove any debris, epiphytes, and excess seawater (Young and Gobler 2016). Additional samples of were cut, cleaned, rinsed, and spun as described, and frozen for later analyses (see below). From the 5-cm pieces of thalli, a sub-sample was removed and weighed on a Scientech ZSA-120 digital microbalance (± 0.0001 g) to obtain initial wet weights which were typically * 1 g. All samples were kept in 100-mL filtered seawater after cleaning and weighing to prevent desiccation prior to introduction to experiments.
Experiments began with the introduction of macroalgae and nutrients into vessels that had been bubbling at the desired pCO 2 levels for several days as described above. Vessels were monitored daily for one week during the experiment as described above. Periodic light measurements were made using a LI-COR LI-1500 light sensor logger and HOBO pendant temperature and light data loggers. At the conclusion of experiments, final pH and temperature measurements were made and a final DIC sample from each bottle was analyzed as described above. Macroalgae samples were removed from their respective vessels and rinsed, spun, re-rinsed, re-spun, and weighed as described above. After being weighed, samples were placed into small freezer bags for further analysis. Weight-based growth rates for all species were determined using the relative growth rate formula (growth d -1 ) = (ln W final -ln W initial )/(Dt), where W final and W initial are the final and initial weights in grams and Dt is the duration of the experiment in days. Normality and equal variance of growth rates were assessed via the use of Shapiro-Wilk and Leven tests; assumptions of equal variance and normality were met for all data. Differences in the growth rates of Dasysiphonia or Porphyra under ambient and elevated pCO 2 and nutrients were assessed via two-way ANOVAs performed using SigmaPlot 11.0 for which the main treatment effects were pCO 2 (ambient or elevated) and nutrients (ambient or elevated). For comparison of growth rates between Dasysiphonia Dasysiphonia japonica and Agardhiella subulata for the October 2019 experiment. All maps were generated using ArcMap (Version 10.6) (Esri) and Porphyra, Tukey Honest Difference (Tukey HSD) tests using R 3.4.0 within 1.0.143 were performed.
Assessing the effects of elevated pCO 2 and nutrients, and competition A final laboratory experiment was performed to assess the effects of competition and elevated pCO 2 and nutrients on the growth of Dasysiphonia and Agardhiella. Identification of macroalgae was based on morphology, microscopy, and known biogeography. Dasysiphonia and Agardhiella were collected during low tide from a shallow-water (depths of 1-2 m) area in Shinnecock Bay, New York, USA ( Fig. 1; 40.85°N, -72.50°W) and prepared in the same manner described above. All macroalgae collected were found growing subtidally. The experiment was performed in 1-L polycarbonate vessels in an environmental control chamber as described above. Containers were randomly assigned, in quadruplicate, to one of six treatments: two controls with ambient pCO 2 (* 400 latm) and ambient nutrients (1 lM nitrate, 0.5 lM phosphate) that received either Agardhiella or Dasysiphonia, a treatment with ambient pCO 2 and nutrients that received Agardhiella and Dasysiphonia, and three treatments with elevated pCO 2 (* 1,800 latm) with daily nutrient additions (20 lM nitrate, 1.2 lM phosphate) that received either Agardhiella only, Dasysiphonia only, or Agardhiella and Dasysiphonia, resulting in 24 experimental units. This level of nutrient addition (20 lM N added daily over six days = 120 lM) is slightly more than twice what was used in the single species experiment (50 lM) to accommodate for the biomass of macroalgae that was twice the level of the single alga experiments (1 v 2 g). Nutrient concentrations were consistent with those observed in waterbodies around Long Island, New York (NY), and with Redfield stoichiometry (Gobler et al. 2006;Wallace and Gobler 2015). As described above, (1) experiments began with the introduction of macroalgae and nutrients into experimental vessels that had been bubbling at the desired pCO 2 levels for several days as described above. Vessels were monitored daily for one week during the experiment; (2) final pH and temperature measurements were made and a final DIC sample from each vessel was analyzed; (3) macroalgae samples were removed from their respective vessels and rinsed, spun, re-rinsed, re-spun, weighed, and placed into small freezer bags for further analysis; (4) Weight-based growth rates for all species were determined using the relative growth rate formula and differences among growth rates were statistically evaluated.

Pre-and post-experimental analyses
For carbon (C), nitrogen (N), d 15 N, and d 13 C analyses, frozen samples of Dasysiphonia from March 2018 through February 2019 experiments were dried at 60°C for 48 h and then homogenized into a fine powder using a mortar and pestle. Tissue C, N, d 15 N, and d 13 C were analyzed using an elemental analyzer interfaced to a Europa 20-20 isotope ratio mass spectrometer at the UC Davis Stable Isotope Facility. For comparison to Dasysiphonia, samples of Fucus vesiculosus, Gracilaria tikvahiae, Porphyra purpurea, Saccharina latissima, and Ulva rigida were collected from Shinnecock Bay, where Dasysiphonia co-occurs during the same period in which the macroalga was collected, and prepared for analysis of d 13 C as described above. These species are common and native to Long Island (Stewart Van Patten and Yarish 2009; Wallace and Gobler 2015). Differences in tissue content of Dasysiphonia among treatments were statistically evaluated as described above. For all experiments involving Dasysiphonia, concentrations of nitrate (NO 3 -), phosphate (PO 4 3-), and ammonium (NH 4 ? ) within experimental vessels, 20 mL of seawater was removed from each vessel and filtered by passing the seawater through pre-combusted (4 h at 450°C) glass fiber filters (GF/F, 0.7-lm pore size) at the beginning and at the end of the experiments. The filtrate was frozen in 20 mL acid-washed scintillation vials for future analysis. A QuikChem 8500 (Lachat Instruments) flow injection analysis system was used to analyze the filtrate colormetrically to measure NO 3 -, PO 4 3-, and NH 3 (Parsons 2013). Initial, ambient concentrations of NO 3 -, PO 4 3-, and NH 3 were 2.65, 0.95, and 0.35 lM, respectively, and final concentrations of NO 3 -, PO 4 3-, and NH 3 were 0.66, 1.18, and 0.60 lM, respectively, with no significant differences in concentrations between ambient or elevated pCO 2 or nutrient treatments.

Field surveys of Dasysiphonia
During October and November 2019, field surveys for Dasysiphonia were conducted along the south shore of Long Island, NY (Fig. 1). A total of 32 land-accessible sites were sampled in Great South Bay, Moriches Bay, Narrow Bay, Quantuck Bay, Shinnecock Bay, and Peconic Bay, NY. Identification of macroalgae was based on morphology and microscopy. The morphology and pigmentation of macroalgal fronds used in this study were fully consistent with prior descriptions of the macroalgae in the region (Newton et al. 2013;Ramsay-Newton et al. 2017;Schneider 2010). Upon arrival at each site, a 100-m transect was performed parallel to the shore. When Dasysiphonia was found, large, well-pigmented fronds were collected, placed in 50-mL centrifuge tubes in a cooler, and transported to the laboratory where they were examined microscopically and subsequently frozen for further analysis. The samples were dried at 60°C for 48 h and then homogenized into a fine powder using a mortar and pestle. Tissue d 13 C and d 15 N was analyzed at the UC Davis Stable Isotope Facility as described above.

Response of Dasysiphonia to elevated nutrients and/or pCO 2
Growth rates of Dasysiphonia were responsive to pCO 2 levels during experiments performed in winter, spring and fall (February, March, April, October, and November), where a higher level of pCO 2 significantly increased growth relative to ambient pCO 2 treatments regardless of nutrient concentrations (Two-way ANOVA; p \ 0.05; Fig. 2; Supplementary  Table S2). For those experiments, growth rates of Dasysiphonia were * 60% higher when it was exposed to elevated pCO 2 compared to ambient conditions (Fig. 2). Across all experiments, growth rates of Dasysiphonia were 26% higher under elevated pCO 2 treatments compared to ambient treatments (Fig. 2). During late spring and summer experiments, growth was not significantly affected by elevated pCO 2 as a single factor (Two-way ANOVA; p [ 0.05; Fig. 2; Supplementary Table S2). Furthermore, for seven of ten experiments, growth rates of Dasysiphonia were significantly higher when exposed to elevated nutrient concentrations, regardless of pCO 2 level (Two-way ANOVA; p \ 0.05; Fig. 2 Table S2). When exposed to elevated nutrient levels, Dasysiphonia growth rates were 40% and * 25% higher in ambient and elevated pCO 2 treatments, respectively (Fig. 2). There were no significant interactions between levels of pCO 2 and nutrients during experiments (Two-way ANOVA; p [ 0.05; Fig. 2; Supplementary Table S2), except for the May experiment when neither pCO 2 nor nutrients alone altered growth rates but the combination of elevated pCO 2 and nutrient levels synergistically increased the growth of Dasysiphonia by 30% relative to the control (Two-way ANOVA; p \ 0.05; Fig. 2; Supplementary Table S2). The August experiment, when water temperatures were * 28°C, was unique in that Dasysiphonia had negative growth rates across all treatments and was not significantly affected by elevated pCO 2 or nutrient levels (Two-way ANOVA; p [ 0.05; Fig. 2; Supplementary Table S2).
The d 13 C content of Dasysiphonia was significantly lower when exposed to elevated pCO 2 (-43.4%) compared to control conditions (-31.7%; Two-way ANOVA; p \ 0.05; Fig. 3; Supplementary Table S3). In contrast, there was no difference in d 13 C content when Dasysiphonia was exposed to elevated nutrients for any experiment (Two-way ANOVA; p [ 0.05; Fig. 3 (Fig. 5). For April through September experiments, average tissue N content of Dasysiphonia under elevated nutrients was 12% and * 16% higher under ambient and elevated pCO 2 , respectively (Fig. 5). Furthermore, C:N ratios of Dasysiphonia were significantly lower when exposed to elevated nutrients for April through July and November to February experiments (Two-way ANOVA; p \ 0.05; Fig. 6; Supplementary Table S5). The average tissue C:N ratio of Dasysiphonia was 10% and * 18% lower under elevated nutrient levels in ambient and elevated pCO 2 treatments, respectively (Fig. 6). Tissue C:N ratios were unaffected by elevated pCO 2 levels (Two-way ANOVA; p [ 0.05; Fig. 6; Supplementary Table S6) except for the May and July experiments when tissue C:N ratios were significantly higher under elevated pCO 2 (Two-way ANOVA; p \ 0.05; Fig. 6; Supplementary Table S6).
Response of Porphyra to elevated nutrients and/ or pCO 2 and comparison to Dasysiphonia Porphyra growth rates was rarely or never responsive to pCO 2 levels and nutrient additions, respectively (Two-way ANOVA; p [ 0.05; Fig. 7; Supplementary  Table S7). Of the seven experiments performed, Porphyra only grew faster in response to elevated pCO 2 during the July and August experiments (Twoway ANOVA; p \ 0.05 for both; Fig. 7; Supplementary Table S7).
When comparing the growth of Dasysiphonia to Porphyra, there were several significant differences among treatments and seasons. For spring experiments, Dasysiphonia growth was 13-fold and significantly higher than that of Porphyra in all treatments, regardless of pCO 2 or nutrient level (Tukey HSD; p \ 0.05 for all; Supplementary Table S8). During the summer experiments, Dasysiphonia growth was 40% higher than that of Porphyra and significantly higher in ambient pCO 2 treatments and treatments that received nutrient additions as well as when all treatments were combined (Tukey HSD; p \ 0.05 for all; Supplementary Table S8). For the fall experiments, Dasysiphonia growth rates were significantly higher than Porphyra in the ambient pCO 2 treatment that received nutrient additions (Tukey HSD; p \ 0.05; Supplementary Table S8) whereas Porphyra growth was significantly higher than that of Dasysiphonia in ambient pCO 2 treatments that did not receive nutrient additions, (Tukey HSD; p \ 0.05 for all; Supplementary Table S8). Overall, during the fall, Porphyra growth was * 70% higher than that of Dasysiphonia (Figs. 2 and 7).

Competition of Agardhiella and Dasysiphonia under ambient and elevated nutrients and/or pCO 2
There were differing growth responses for Agardhiella and Dasysiphonia exposed to combined elevated pCO 2 and nutrients. Agardhiella growth rates were not significantly altered by elevated pCO 2 and nutrient levels or by competition with Dasysiphonia (Two-way ANOVA; p [ 0.05; Fig. 8; Supplementary Table S9). In contrast, Dasysiphonia growth rates were * 120% higher under elevated pCO 2 and nutrients (Two-way ANOVA; p \ 0.05; Fig. 8 Table S9). While growth rates of Dasysiphonia and Agardhiella were similar under ambient levels of nutrients and pCO 2, the growth rates of Dasysiphonia were three times greater than Agardhiella under elevated levels of nutrients and pCO 2 (Two-way ANOVA; p \ 0.05; Fig. 8; Supplementary Table S9).

Field surveys of Dasysiphonia across the south shore of Long Island
The field surveys of the southern shore of Long Island revealed that Dasysiphonia was widespread across the area. The macroalga was present at 24 of 32 sites ( Fig. 9; Supplementary Table S10) including all 15 sites in Great South Bay, two sites in Narrow Bay, and two sites in western Moriches Bay. It was absent from eastern Moriches Bay, Quantuck Bay, and western Shinnecock Bay but was present at all four sites across eastern Shinnecock Bay and the single site in the southern section of Great Peconic Bay ( Fig. 9; Supplementary Table S10). In Great South Bay, Dasysiphonia d 15 N signatures varied by site, with no noticeable trends across the bay. Signatures of d 15 N in Great South Bay ranged from 6.4 to 11.7% (sites 1-15; Fig. 10 Table S10). The overall minimum, maximum, and average d 15 N content of Dasysiphonia was 6.4, 11.7, and 9.1%, respectively ( Fig. 10; Supplementary Table S10).

Discussion
Given that the growth of Dasysiphonia generally increased when exposed to elevated pCO 2 and nutrient levels in contrast to indigenous macroalgae, this study provides insight regarding the factors driving invasions of this macroalga. During this study, elevated Columns represent means ± standard deviation. Two-way ANOVA performed with significant differences (p \ 0.05) in main treatment effects, (CO 2 and Nutrients, abbreviated as N ? P), appearing at the top right of each panel pCO 2 was found to significantly enhance the growth rates of the invasive red macroalga Dasysiphonia during experiments performed when water temperatures were at or below 15°C. In contrast, elevated nutrient concentrations significantly increased growth rates during experiments across a wider range of temperatures (4-23°C). Tissue N and C:N content were significantly increased and decreased, respectively, when Dasysiphonia was exposed to elevated nutrient concentrations. The growth rates of Dasysiphonia were generally greater than those of two other indigenous macroalgae, especially when levels of nutrients and pCO 2 were elevated. The d 13 C and d 15 N Two-way ANOVA performed with significant differences (p \ 0.05) in main treatment effects, (CO 2 and Nutrients, abbreviated as N ? P), appearing at the top right of each panel of this macroalga indicated that it relied on CO 2 and wastewater-derived N for in situ growth. The response of macroalgae to pCO 2 levels may depend on their mode of C acquisition as well as the extent to which inorganic carbon uptake of the organism is substrate-saturated at current pCO 2 levels (Badger 2003;Koch et al. 2013). When exposed to elevated pCO 2 , macroalgae may downregulate their carbon concentrating mechanisms (CCMs) that convert HCO 3 to CO 2 , allowing more energy to be available for processes such as vegetative growth (Björk et al. 1993;Cornwall et al. 2012;Koch et al. 2013). Values of d 13 C are often used to assess the types of inorganic carbon used by macroalgae for Two-way ANOVA performed with significant differences (p \ 0.05) in main treatment effects, (CO 2 and Nutrients, abbreviated as N ? P), appearing at the top right of each panel photosynthesis, with a d 13 C signature of -10% being indictive of HCO 3 use and signatures of -30% or lower characteristic of diffusive uptake of CO 2 (Bricelj et al. 2001;Hepburn et al. 2011;Maberly et al. 1992;Raven et al. 2002). In the present study, Dasysiphonia had significantly lower d 13 C signatures when exposed to experimentally elevated pCO 2 , evidencing uptake of the d 13 C-enriched experimental CO 2 source (Young and Gobler 2016). Furthermore, under ambient conditions, the d 13 C signatures of Dasysiphonia were consistently at -30% or lower, a finding consistent with prior studies of this macroalga (Kang et al. 2008;Leclerc et al. 2013) and indicating that Dasysiphonia predominately uses CO 2 for Columns represent means ± standard deviation. Two-way ANOVA performed with significant differences (p \ 0.05) in main treatment effects, (CO 2 and Nutrients, abbreviated as N ? P), appearing at the top right of each panel Fig. 7 Growth rates of Porphyra purpurea exposed to ambient and elevated pCO 2 with and without nutrient additions for May through November 2018 experiments. Columns represent means ± standard deviation. Temperatures represent means ± standard deviation for all treatments. Two-way ANOVA performed with significant differences (p \ 0.05) in main treatment effects, (CO 2 and Nutrients, abbreviated as N ? P), appearing at the top right of each panel photosynthesis (Bricelj et al. 2001;Hepburn et al. 2011;Maberly et al. 1992;Raven et al. 2002). Rhodophytes are known to have a high affinity for CO 2 and RuBisCO that minimizes CO 2 loss from photorespiration despite not using HCO 3 or having a CCM (Badger et al. 1998;Johnson et al. 1992;Koch et al. 2013). While not all rhodophytes may benefit from increased pCO 2 , (Britton et al. 2019;Koch et al. 2013), the results of the present study demonstrate that Dasysiphonia is likely not substrate-saturated at Fig. 8 Growth rates of Agardhiella subulata and Dasysiphonia japonica exposed to ambient and elevated CO 2 and nutrient conditions, with and without competition from the other macroalga for the October 2019 experiment. Columns represent means ± standard deviation. Two-way ANOVA performed with significant differences (p \ 0.05) in main treatment effects, (CO 2 and Nutrients, abbreviated as N ? P), for each species appearing at the top left of the panel Fig. 9 Dasysiphonia japonica field surveys conducted during October and November 2019. Site numbers correlate with the locations, bays, and coordinates in Supplementary Table S10 ambient pCO 2 levels (* 400 latm) given the seasonally increased growth rates in response to elevated pCO 2 . Dasysiphonia differs from many other macroalgal species that comprise the assemblages in the waters around New York, including Fucus vesiculosus, Gracilaria tikvahiae, Porphyra purpurea, Saccharina latissima, and Ulva rigida, which have d 13 C signatures ranging from -12 to -18% Gobler 2016, 2017;Young et al. 2018;Fig. 11). This finding suggests that among all of these co-occurring macroalgae, Dasysiphonia is the most reliant on CO 2 as a carbon source (Bricelj et al. 2001;Hepburn et al. 2011;Maberly et al. 1992;Raven et al. 2002) and may, therefore, be the most likely to benefit from higher pCO 2 in coastal waters. In addition to the rapid rise in atmospheric pCO 2 during the past half century (Meehl et al. 2007), it has become well recognized that eutrophication-enhanced organic-matter loading to coastal waters also causes seasonally elevated pCO 2 (Cai et al. 2011;Melzner et al. 2013;Wallace et al. 2014). It seems plausible that high pCO 2 levels seasonally present today in the coastal waters may be favoring the proliferation of Dasysiphonia over other macroalgae that are substrate-saturated at present pCO 2 levels or are unable to utilize CO 2 for photosynthesis under acidified conditions (Cornwall et al. 2012;Koch et al. 2013).
Elevated nutrient concentrations significantly increased and decreased tissue N and C:N content, respectively, of Dasysiphonia, indicating that while both N and P were added during experiments, N was a critical element. Beyond these changes in N content relative to the control treatment, the absolute N content of Dasysiphonia usually increased to C 0.03 g N per g tissue following nutrient additions, a level known to represent N saturation (Lyngby et al. 1999;Wallace and Gobler 2015). These trends align with growth rates in most experiments of the present study, where Dasysiphonia responded positively to elevated nutrient levels during most experiments. Dasysiphonia has been shown to have a high nitrate uptake efficiency consistent with rapid growth compared to native macroalgal species (Low et al. 2015). In the present study, while the presence of Dasysiphonia did not significantly affect the growth rates of a common native rhodophyte (Agardhiella) in ambient or elevated pCO 2 and nutrient conditions, the former  Supplementary Table S10. 'N/A' denotes sites where Dasysiphonia japonica was not collected during the field surveys outgrew the latter by nearly two-fold when high levels of nutrients and pCO 2 were present. Similarly, the growth rates of Dasysiphonia exceeded those of Porphyra during most seasons when nutrients and pCO 2 levels were high. The filamentous morphology of Dasysiphonia affords it a surface-area-to-volume ratio greater than bladed and branching macroalgae and allows for a higher nutrient uptake capacity (Low et al. 2015;Rosenberg and Ramus 1984). These traits may allow Dasysiphonia to undergo rapid growth in eutrophic settings (Valiela et al. 1997;Wallace and Gobler 2015). Eutrophication can promote coastal ocean acidification due to the accumulation of respiratory CO 2 due to microbial degradation of excessive organic matter, resulting in seasonal pCO 2 levels ([ 1000 latm) not predicted to occur in open ocean regions for more than a century (Wallace et al. 2014). As such, eutrophication may promote seasonallyelevated pCO 2 levels that may yield nutrient-and CO 2 -mediated enhanced growth of Dasysiphonia, a scenario that would make eutrophied estuaries more susceptible to invasion by the macroalga. Furthermore, given that climate change is expected to increase incidences of eutrophication in many estuaries due to increased precipitation-driven N loading (Sinha et al. 2017), these systems may become more susceptible to invasion by macroalgae such Dasysiphonia.

Seasonal response to nutrients and CO 2
The response of Dasysiphonia varied seasonally with CO 2 responses being limited to cool water conditions, and nutrient effects present during warmer periods of the year with these changing responses being reflective of seasonal cycles in nutrients and acidification. For example, continuous measurements of pH in Great South Bay during this study, the collection site for Dasysiphonia, revealed pH values decreased from [ 8.1 during early April to \ 7.9 from May through September, and values [ 8 during the fall (Fig. 12a).
Given that values of pH and pCO 2 are inversely related in oceans (Caldeira and Berner 1999;Marsh 2008;Martz et al. 2010) and Long Island estuaries (Baumann et al. 2015;Gobler and Baumann 2016;Wallace et al. 2014), these patterns in pH suggest in situ pCO 2 levels were low during the colder months and higher during the warmer months of the study, a trend consistent with other studies in the region (Baumann et al. 2015;Gobler and Baumann 2016;Wallace et al. 2014). In Fig. 11 The d 13 C content of various species of macroalgae (Dasysiphonia japonica, Fucus vesiculosus, Gracilaria tikvahiae, Porphyra purpurea, Saccharina latissima, and Ulva rigida) under ambient pCO 2 levels without nutrient additions. Columns represent means ± standard deviation. Red dashed lines represent divisions in macroalgae that solely rely on HCO 3 -(0 to -10%), can use both HCO 3 and CO 2 (-10 to -30%), or solely rely on CO 2 (-30% or less) for photosynthesis Bricelj et al. 2001;Raven et al. 2002;Hepburn et al. 2011) Fig. 12b). This trend is consistent with prior coastal studies showing nutrients drawn down by enhanced rates of productivity during warmer months (Carpenter and Dunham 1985;Glibert et al. 2007;Woodwell et al. 1979). Collectively, these seasonal patterns in nitrogenous nutrients and pH/CO 2 align well with the experimental outcomes of this study. During the colder months (\ 15°C; February, March, April, October, and November), pH levels were higher and pCO 2 levels were likely lower, and CO 2 was likely a limiting resource for the growth of the obligate CO 2user, Dasysiphonia. During the warmer months, when pCO 2 levels were higher but DIN concentrations levels were low, growth of Dasysiphonia was limited by N but not pCO 2 levels. The only experiment that revealed co-limitation of nutrients and CO 2 was during the transition month of May, when nutrients and CO 2 combined to synergistically enhance the growth of Dasysiphonia. Lastly, the d 15 N content of macroalgae is \ 3% when utilizing N from fertilizer or atmospheric deposition, and [ 3% when utilizing N from wastewater Kendall and McDonnell 2012;Lapointe et al. 2004). Based on samples collected during field surveys of the present study, the d 15 N content of Dasysiphonia across Long Island was up to * 12% and averaged * 9%, indicating that wastewater was the primary source of N for the macroalga, an outcome consistent with the finding that wastewater supplies * 70% of the N to Long Island estuaries (Kinney and Valiela 2011). Given that N is often a limiting resource for macroalgae in estuaries (Valiela et al. 1997) and given the significantly higher growth rates of Dasysiphonia in the majority of experiments following N enrichment, it seems likely that while pCO 2 levels can limit Dasysiphonia growth, N may be the more important resource.
During this study, Dasysiphonia populations were observed in Great South Bay throughout the year, during which the macroalga was exposed to temperatures ranging from 0°C to 28°C (Fig. 12c). Dasysiphonia is an invasive species, in part, due to its ability to tolerate a wide temperature range (0°C to 30°C), although reduced growth and cellular damage can occur at temperatures exceeding 28°C (Bjaerke and Rueness 2004;Husa et al. 2004). This upper temperature limit is consistent with the present study, where negative growth rates occurred during the August experiment, when temperatures were * 28°C. This is also consistent with continuous measurements of temperature in Great South Bay, with the highest daily temperatures (27-28°C) occurring throughout August and the beginning of September (Fig. 12c). While Bjaerke and Rueness (2004) found that the optimal temperature for growth of Dasysiphonia (as Heterosiphonia japonica) was 19-24°C, in this study growth rates were similar across a wide range of seasonal experiments (4-24°C), suggesting that the optimal temperature range for growth may be wider than previously thought. This difference may be, in part, because the current study sampled macroalgae from the field over a complete annual cycle, allowing for natural acclimation at the population or clonal level to in situ temperatures. Regardless, as water temperatures increase this century and some northern estuaries more regularly enter the 19-24°C range, they may become vulnerable to invasion by Dasysiphonia. The upper temperature limit of Dasysiphonia in the present study and others (Bjaerke and Rueness 2004) suggests that the timing of future invasions in areas south of Long Island may be limited to some, but not all, months of the year. Isotherms in the North Atlantic Ocean are expected to shift as much as 600 km northwards by the end of the twenty-first century (Hansen et al. 2006), with annual mean water temperatures increasing by 4°C on rocky shores in the North Atlantic (Müller et al. 2009). This increase in temperature is expected to cause poleward shifts in pelagic and benthic communities, including macroalgal assemblages (Jueterbock et al. 2013;Southward et al. 1995;Wernberg et al. 2011). Considering the enhanced growth of Dasysiphonia under elevated pCO 2 and nutrient levels (current study), the thermal tolerance of Dasysiphonia (Bjaerke and Rueness 2004; Husa et al. 2004), and the higher levels of pCO 2 at higher latitudes across North America (Gledhill et al. 2015), eutrophic estuaries of the northeast US and eastern Canada may be the most vulnerable to future invasion by Dasysiphonia. Given the ability of this macroalgae to persist at lower levels during the warmest month, invasions to the south of its southern-most site in the North Atlantic (Long Island, NY; this study) may also be possible.

Past, present, and future invasions of Dasysiphonia
Dasysiphonia is native to East Asian waters, where it occurs sporadically throughout the year and generally comprises a small fraction of macroalgal communities (Kang et al. 2008;Kim et al. 2008). Since the 1980s, the macroalga has spread throughout the eastern Atlantic Ocean (Moore and Harries 2009;Sagerman et al. 2015;Sjøtun et al. 2008; Supplementary Fig. S1), the US East Coast, (Idlebrook 2012) and Canada (Savoie and Saunders 2013). Schneider (2010) documented the macroalga's presence in the waters around Rhode Island in 2007 but did not find the macroalga in the waters around Long Island. However, Newton et al. (2013) reported the presence of the macroalga in Long Island Sound in 2012 ( Supplementary Fig. S1). As evident by the present study, the macroalga has continued its western North Atlantic invasion of more than two dozen sites across the south shore of Long Island since its first appearance in Long Island Sound, representing the southernmost observations of Dasysiphonia in North America.
Non-indigenous macroalgae can outcompete native species for space and with respect to growth and nutrient acquisition (Hayes and Barry 2008;Low et al. 2015;Theoharides and Dukes 2007). High levels of biomass of Dasysiphonia have been shown to exceed that of native species by an order of magnitude (Sagerman et al. 2014) and can lower native macroalgal species richness, evenness, and diversity (Low et al. 2015;Sagerman et al. 2014). In the present study, spring, and summer growth rates of Dasysiphonia exceeded those of Porphyra by as much as 13-fold and * 25%, respectively. Additionally, Dasysiphonia outgrew Agardhiella by * 70% under high nutrient and high pCO 2 conditions. We note, however, that this experiment assessed interspecific competition but did not establish a control with double the levels of each seaweed individually to assess intraspecific competition (Underwood 1986). Ramsay-Newton et al. (2017) found that, following invasion of Dasysiphonia in Nahant, MA in 2011, there was a significant decline in macroalgal species richness. Introduced macroalgae such as Dasysiphonia can form dense mats that limit light availability and inhibit the recruitment and establishment of other macroalgae species, such as kelp (Ambrose and Nelson 1982;Britton-Simmons 2004). Dasysiphonia can occupy up to 80% available substrate and can grow epiphytically on other macroalgae (Moore and Harries 2009;Newton et al. 2013; Supplementary Fig. S2). The literature regarding management of Dasysiphonia is currently limited, but it appears that the most effective means of limiting invasions by the macroalga, and others, is prevention of introduction through means such as transoceanic shipping, aquaculture and aquarium trade, fishing/boating activities, and the opening of canals (Doelle et al. 2007;Mathieson et al. 2008;Mineur et al. 2014;Petrocelli et al. 2013;Sant et al. 1996;Vranken et al. 2018). Given the results of this study, it would seem efforts to reduce N loading to coastal waters may also reduce metabolic CO 2 production, and may slow or reverse invasions of Dasysiphonia, thereby mitigating the negative effects of this macroalga on marine ecosystems. Reductions in N inputs to coastal ecosystems have been shown to reduce the frequency and intensity of harmful algal blooms (Heisler et al. 2008) caused by microalgae (Gobler et al. 2005;Nuzzi and Waters 2004), as well as macroalgae (Harlin 1995).
There are numerous ecosystem implications associated with the ability of nutrient loading and ocean acidification to promote invasions of Dasysiphonia. In the North Atlantic Ocean, the replacement of kelp beds with invasive species, such as Dasysiphonia, may alter the ecosystem services provided by kelp, such as a refuge from predation (Watanabe 1984). The ability of macroalgae to serve as refuge from predation can partly depend on the complexity of the macroalga's thallus. The thalli of Dasysiphonia are complex and filamentous (O'Brien et al. 2018), which, in higher densities, can form a biogenically complex habitat capable of hosting some meso-invertebrates due to smaller interstitial volume and area compared to macroalgae with branched or blade morphology (Dijkstra et al. 2017;Holmlund et al. 1990;Veiga et al. 2014;Ware et al. 2019;Warfe et al. 2008). Dijkstra et al. (2017) found that invasive filamentous red macroalgae, such as Dasysiphonia, supported two to three times more meso-invertebrate individuals and species that form the base of food webs, which may benefit higher trophic levels. However, several studies have found that while cunner prefer to refuge in the tall, simple blades provided by kelp (Saccharina latissima), they will use turf macroalgae (i.e. Dasysiphonia) if kelp becomes scarce, although this is not preferred, as turf macroalgae do not provide as much protection from predation as kelp O'Brien et al. 2018). While Dasysiphonia may provide a suitable habitat for some organisms, its value as a food source could alter herbivore assemblages. Some herbivores avoid grazing on the invasive species (Low et al. 2015) and Sagerman et al. (2015) reported that Dasysiphonia (as Heterosiphonia japonica) is not a preferred food source for several common invertebrates, an outcome that may allow for continued proliferation of the macroalga. Given Dasysiphonia can grow epiphytically on other macroalgae (Moore and Harries 2009;Supplementary Fig. S2) and seagrass (García-Redondo et al. 2019), the further invasion, proliferation, and overgrowth of this macroalga in response to acidification and eutrophication could disrupt seaweed assemblages and/or seagrass beds due to decreased light availability (Valiela et al. 1997) and/or direct competition for nutrients (Duarte 1995;Young and Gobler 2017;Young et al. 2018). Macroalgal overgrowth can also smother benthic habitats and promote diel hypoxia/anoxia (Liu et al. 2009;Valiela and Cole 2002;Valiela et al. 1997). Lastly, the overgrowth of Dasysiphonia may be problematic for shellfisheries, as it can overgrow oysters (Haydar and Wolff 2011). A study by Glenn et al. (2020) found that within the Great Bay Estuary in New Hampshire, USA, 81% of all biomass on oyster farm gear consisted of non-native macroalgae, including Dasysiphonia, which may provide additional substrate for the macroalga to propagate.
In conclusion, the invasive red macroalga Dasysiphonia experienced enhanced growth compared to other red macroalgae when exposed to elevated pCO 2 and nutrient levels. Based on the d 13 C signatures of Dasysiphonia, the macroalga appears to predominately use CO 2 as a source of carbon for photosynthesis, which may give it a competitive advantage over other macroalgae in high CO 2 ecosystems. Its d 15 N signatures suggest its growth is being promoted by excessive wastewater-derived N loading. Dasysiphonia has spread across estuaries of the northwestern and northeastern Atlantic Oceans since the 1980s. Given predicted trends in N loading, acidification, and warming of coastal ecosystems in the coming decades (Doney et al. 2009;Sinha et al. 2017), it may continue to spread into eutrophied estuaries further north and south due to its ability to rapidly grow under elevated pCO 2 and nutrient conditions and its wide thermal tolerance.