Microhabitat change alters abundances of competing species and decreases species richness under ocean acidification

Science of the Total Environment 645 (2018) 615–622

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Microhabitat change alters abundances of competing species and decreases species richness under ocean acidification Ivan Nagelkerken ⁎, Silvan U. Goldenberg, Ericka O.C. Coni, Sean D. Connell Southern Seas Ecology Laboratories, School of Biological Sciences, and The Environment Institute, DX 650 418, The University of Adelaide, Adelaide, SA 5005, Australia

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Niche segregation allows species to coexist and maintain diversity. • Ocean acidification could modify niche availability and niche segregation. • Natural CO2 vents showed altered microhabitat availability and fish abundances. • Competitively dominant fishes increased in density but others decreased. • Fish species diversity decreases due to niche alteration under elevated CO2.

a r t i c l e

i n f o

Article history: Received 10 May 2018 Received in revised form 13 July 2018 Accepted 13 July 2018 Available online xxxx Editor: D. Barcelo Keywords: Species interactions Niche overlap Biodiversity CO2 vents Habitat use Diet

a b s t r a c t Niche segregation allows competing species to capture resources in contrasting ways so they can co-exist and maintain diversity, yet global change is simplifying ecosystems and associated niche diversity. Whether climate perturbations alter niche occupancy among co-occurring species and affect species diversity is a key, but unanswered question. Using CO2 vents as natural analogues of ocean acidification, we show that competing fish species with overlapping diets are partially segregated across microhabitat niches and differently-orientated substrata under ambient CO2 conditions. Under elevated CO2, benthic microhabitats experienced a significant increase in non-calcifying turf and fleshy algae but a sharp reduction in calcareous algae. The increased availability of turf and fleshy algae supported increased densities of a competitively dominant species, whilst the reduction in calcifying algal microhabitats decreased densities of several subordinate species. The change in microhabitat availability also drove an increased overlap in microhabitat use among competing fishes at the vents, associated with a reduced fish species richness on horizontal substrates. We conclude that loss of preferred microhabitat niches, exacerbated by population proliferation of competitively dominant species, can drive population losses of less common and subordinate species, and reduce local species richness. The indirect effects of ocean acidification on microhabitat availability can therefore impair maintenance of species populations, and drive changes in local community and biodiversity patterns. © 2018 Elsevier B.V. All rights reserved.

1. Introduction

⁎ Corresponding author. E-mail address: [email protected] (I. Nagelkerken).

https://doi.org/10.1016/j.scitotenv.2018.07.168 0048-9697/© 2018 Elsevier B.V. All rights reserved.

Ocean acidification can affect species through direct effects on their physiology (Pörtner and Farrell, 2008) and behaviour (Briffa et al., 2012; Clements and Hunt, 2015; Nagelkerken and Munday, 2016) as

616

I. Nagelkerken et al. / Science of the Total Environment 645 (2018) 615–622

well as indirect effects through resources (bottom-up), predators (topdown), and competitors (Doropoulos et al., 2012; McCormick et al., 2013; Nagelkerken and Connell, 2015). Elevated CO2 is known to drive community homogenisation, often because of increased abundance of dominant species, to the detriment of rarer species (Kroeker et al., 2011; Inoue et al., 2013; Enochs et al., 2015). For example, as a result of ocean acidification temperate rocky reefs consisting of mosaics of kelp forests, urchin barrens and macroalgae, undergo phase shifts towards turf algal dominated systems (Connell and Russell, 2010), and altered fish species communities (Nagelkerken et al., 2016b, 2017). Such habitat simplification is likely to reduce niche diversity and hence affect species richness and abundance of associated animals. Yet, we still know little about how habitat modification resulting from ocean acidification alters niche diversity and associated species communities. Such insights are important for an understanding of how species-level impacts might scale up to community-level responses (Riebesell and Gattuso, 2015), as ecological responses of individual species can be modified in nature when inter-species and species-habitat interactions are considered (Gilman et al., 2010; Goldenberg et al., 2018; Ullah et al., 2018). Niche diversity has important implications for species diversity and abundances. According to the Volterra model of interspecific competition, n species cannot co-exist on fewer than n resources (Stewart and Levin, 1973; Armstrong and McGehee, 1976). However, niche segregation allows for a higher number of species to coexist than the number of resources available. This is because resources can be partitioned along multiple axes (e.g. space, time, resources, and natural enemies) to reduce inter-specific competition (Ross, 1986; Chesson, 2000). Whilst such partitioning allows for co-existence, ocean acidification is reducing the diversity of habitat niches (Kroeker et al., 2011; Connell et al., 2018). Hence, species diversity may be sensitive to environmental change through altered resource availability. How changes in abiotic and biotic conditions driven by ocean acidification affect niche segregation, local species diversity, and species abundances is not well understood, in particular for ocean acidification caused by increasing human atmospheric CO2 emissions. Closely related species can co-exist through a diversity of mechanisms (Munday et al., 2001), but habitat (32%) and diet (57%) alone account for 89% of niche segregation in fishes (Ross, 1986). Ocean acidification has the capacity to enhance both habitat and food availability (Nagelkerken et al., 2016b; Connell et al., 2017; Goldenberg et al., 2017) and thus affect resource partitioning. Due to CO2 enrichment, food and habitat resources could become less limiting and we would expect that niche segregation between competing consumer species might be relaxed. Alternatively, because niche overlap is inversely correlated to the degree of species interactions (Pianka, 1974), and because species competition remains unaltered by ocean acidification (Nagelkerken et al., 2017), niche segregation might not change under elevated CO2. Here we test whether niche segregation (in diet, habitat surface orientation, and microhabitat type) among temperate benthic fish species and niche availability is altered by ocean acidification and how this affects fish species abundances and richness. To incorporate ecological complexity and intragenerational acclimation, we observed the three most abundant territorial fish species that had been living at volcanic CO2 vents their entire juvenile and adult lives among their natural prey, predators, competitors, and habitats: common triplefin Forsterygion lapillum, Yaldwin's triplefin Notoclinops yaldwyni, and blue-eyed triplefin Notoclinops segmentatus. The CO2 elevation at the vents was close to end-of-century projections under a high-emission scenario (RCP 8.5) with an average pCO2 of 977 μatm. We show that changes in microhabitat availability under ocean acidification is directly associated with altered abundances of fish species, increased overlap in habitat niches among species, and a concurrent decrease in species richness.

2. Materials and methods 2.1. Study system White Island is a volcanic island located in the Bay of Plenty of the North Island of New Zealand. Two independent vent and two control sites were identified along the north-eastern coast of the Island (see Fig. S1 in Connell et al., 2018). The CO2 plumes at vent sites were ~24 × 20 m in dimension and located at 6–8 m depth. The control sites were located adjacent to the vents (~25 m away) where pH levels represented ambient oceanic conditions. Under the representative concentration pathway (RCP) 8.5 emission scenario (high emission scenario), human CO2 emissions into the atmosphere will increase from the current levels of ~400 ppm (parts per million) to ~936 ppm by the end of the century (IPCC, 2014). This will lead to a decrease in ocean-surface pH of ~0.33 ± 0.003 units by 2100 compared to the 1990s (Bopp et al., 2013). pH levels at the two vent sites were close to RCP 8.5 projections and showed a pH reduction of 0.28 ± 0.06 units (mean ± SD) across years, and were not confounded by elevated temperatures (Table S1). Measurements from different time periods showed similar pH values, suggesting the vents are relatively stable over time. The study area represents a hard-substratum rocky reef ecosystem, and the substratum at control sites was characterised by a mosaic of kelp (Ecklonia radiata), turf-forming macroalgae (b10 cm in height), and hard-substratum sea urchin barrens devoid of vegetation. Our three focal species (common triplefin Forsterygion lapillum, Yaldwin's triplefin Notoclinops yaldwyni, blue-eyed triplefin Notoclinops segmentatus) accounted for 93% and 95% of the total benthic fish density at control and vent sites, respectively. These species co-exist on temperate shallow rocky reefs and show high overlap in niche use (i.e., water depth, wave exposure, and substratum type) as well as micro-positioning on different substrate types (Wellenreuther et al., 2007). 2.2. Carbonate chemistry measurement and analysis The CO2 concentration in the water was calculated using the values of temperature, salinity, pHNBS and Total Alkalinity (TA) measured from samples collected in the field (Table S1). The software CO2SYS was used to estimate seawater pCO2 (partial pressure of CO2) with constants K1 and K2 from Mehrbach et al. (1973) and refit by Dickson and Millero (1987). Alkalinity was measured by dynamic endpoint titration using a Titrando (Metrohm) titrator. During the study, values for standards were maintained within 1% accuracy from certified reference materials from Dr A. Dickson (Scripps Institution of Oceanography). Water samples were collected during May 2013, November 2013, February 2015, and March 2016. TA water samples were collected during years 2013 and 2015 (this study), and were fixed with mercuric chloride and preserved in Duran glass bottles (Schott) until analysis, according to standard operating procedures (Dickson et al., 2007). Because alkalinity was not measured in 2016, values from previous years (2013, 2015) were used instead to calculate seawater pCO2. Ocean alkalinity is known to be relatively stable across years (Pearson and Palmer, 2000) and any potential differences in alkalinity between 2016 and previous years would only slightly affect the absolute values of pCO2 in 2016, but not the relative difference between vent and control treatments. Salinity was measured with a SR6 refractometer (Vital Sine). 2.3. Diet analysis The three focal fish species were collected for diet analysis in 2015 and in 2016 (for additional replicates) at the same control and CO2 vent sites where niche availability and niche use was quantified (see Sections 2.4 and 2.5). Total number of fish collected was 82 common triplefins, 57 Yaldwin's triplefins, and 24 blue-eyed triplefins (Table 1). Gut contents were assessed under a stereo microscope by

I. Nagelkerken et al. / Science of the Total Environment 645 (2018) 615–622

617

counting each prey organism and assigning it to one of several prey categories based on taxonomy and size (see Table S3). Across fish species and CO2 treatments, a subsample of well-preserved prey individuals was photographed for each prey category, and their length and width measured using ImageJ (n = 745 across prey categories). Average prey mass was then determined for each prey category using the formula for the most similar geometric form for the respective prey type (e.g. Alcaraz et al., 2015). Prey mass was then multiplied by prey numbers to attain total prey biomass in each fish gut. Finally, the relative contribution to the fishes' diet was calculated for each of the 10 principle prey taxa for use in the analysis (Acaridae, Amphipoda, Bivalvia, Copepoda, Gastropoda, Isopoda, Ostracoda, Polychaeta, Tanaidacea, and other prey).

space, and hence the sampling of boulders accounted for much of the hard-bottom benthic habitat that was available to fishes. To determine how fishes use differently orientated hard-substratum surfaces, abundances of all benthic territorial species were visually quantified using SCUBA (self-contained underwater breathing apparatus) during February 2015 (same time period as when the majority of fish were collected for diet analysis) in replicate 1 × 1 m quadrats at the two vent and two control sites. At each site, 25 quadrats were sampled on flat surfaces located on top of large boulders (‘horizontal’: ~0–10° orientation), on the sides of large boulders (‘vertical’: ~80–90° orientation), and on angled surfaces of large boulders (‘angle’: ~25–45° orientation, relative to horizontal), for a total of 300 quadrats (Table 1). A short follow-up study in February 2018 further quantified the use of microhabitat niches on the three differently-oriented surfaces for the same three common fish species as used for the diet analyses. Because we were only interested in potential differences in microhabitat use and because the abundances of common triplefin were 5–10 times higher than that of the other species, we focussed on a predetermined number of individuals per species rather than using fixed 1 × 1 m transects. Using free-roaming visual surveys on SCUBA we searched for individuals of the three species (until target numbers were reached), and for each individual observed we noted the orientation of the substratum (horizontal, angle, vertical surfaces, respectively) and the microhabitat type they were occupying. The a priori target number for each of the two CO2 treatments was ~60 common triplefins, 60 Yaldwin's triplefins, and 30 blue-eyed triplefins (because these were less common) with the final number of individuals shown in Table 1. The approach of a targeted number of individuals across species enabled us to attain sufficient replicate individuals of less common species and allow us to calculate their relative microhabitat use. The seven microhabitat niches identified were: bare rocky substratum, cobbles (~0.5–2 cm in diameter), crustose coralline algae (CCA), encrusting fleshy green algae (Codium sp.), erect calcareous algae (largely Amphiroa sp.), sand, and turf algae b10 cm in height (largely Gigartina sp.). Together this resulted in 17 different habitat–surface orientation combinations, with no values for sand and cobble at angled and vertical surfaces as these two only occurred as horizontal surfaces.

2.4. Niche use by fishes

2.5. Niche availability

A limitation of using natural CO2 vents to represent future acidified ecosystems is that many animals move in and out of the vent areas and are therefore not continuously exposed to high CO 2. To work within this limitation we focussed on site-attached species that occupy a territory directly after settlement, show little movement, and have small home ranges. For both vent locations this was an acceptable approach as the benthic fish communities were dominated by blennies (Blenniidae) and triplefins (Tripterygiidae). All of these species maintain territories of a few square metres and are highly site-attached (Syms and Jones, 1999). Whilst the fish densities in this study were only quantified during one year, we showed earlier that fish density of the common triplefin (p = 0.1920), Yaldwin's triplefin (p = 0.4393), blue-eyed triplefin (p = 0.1466), and all fish species combined (p = 0.3225) did not differ across the years 2013, 2015, and 2016 (Table S1 in Nagelkerken et al., 2017). Furthermore, by comparing controls and CO2 vents we substitute space for time to reveal potential changes in species population sizes and diversity resulting from future ocean acidification effects. Nevertheless, space-for-time substitutions remain a powerful tool and have high accuracy compared to time-for-time predictions (Blois et al., 2013). The rocky reefs at the study sites were dominated by mediumsized boulders (~0.5–1 m in diameter). Because of the large numbers of boulders on the substratum there was little between-boulder

To separate the direct effects of elevated CO2 on fish behaviour (e.g. altered niche preference) vs. the indirect effects on altered niche availability (i.e. different abundance of substrate surfaces or different cover of microhabitats) we quantified niche availability at control and vent sites. For surface orientation (year 2015), we used 10 replicate 1 × 1 m quadrats at each of the four sites (2 vent and 2 control), for a total of 40 quadrats (Table 1). The quadrat was randomly dropped on the substratum and for each quadrat the relative % horizontal vs. angled vs. vertical surfaces was estimated (up to 100%). For microhabitat availability (February 2018), we used 20 replicate 1 × 1 m quadrats for each treatment, for a total of 40 quadrats (Table 1). Each quadrat was photographed at a distance of ~1 m above the substratum using a Canon G1X camera. The relative cover of the seven microhabitat niches was estimated from the photos using the Coral Point Count with Excel Extensions (CPCe) Software, using 80 randomly distributed points per image. Because turf cover more than doubled under elevated CO2 (Nagelkerken et al., 2016b), we also measured turf algal height in 2015 in 10 replicate 1 × 1 m quadrats for each of three surface orientations at each of the four sites (2 vent and 2 control), for a total of 120 quadrats (Table 1). In each quadrat, turf height was measured at five random positions in situ using a ruler. Average turf height per 1 m2 quadrat was used for the analyses.

Table 1 Sample size for the different experiments performed at control and vent sites. Control

Diet analysis (# individuals) Forsterygion lapillum Notoclinops yaldwyni Notoclinops segmentatus

CO2 vent

Feb. 2015

Feb. 2016

35 11 11

6 9 1

Feb. 2018

Surface orientation use (# quadrats) Horizontal surfaces 25 Angled surfaces 25 Vertical surfaces 25

Competitive interactions (# videos) Baited videos

Feb. 2016

34 15 12

7 22 0

Feb. 2018

25 25 25

Microhabitat use (# individuals) Forsterygion lapillum Notoclinops yaldwyni Notoclinops segmentatus Niche availability (# quadrats) Surface orientation availability Turf height horizontal surfaces Turf height angled surfaces Turf height vertical surfaces Microhabitat availability

Feb. 2015

64 57 30

20 20 20 20

73 54 18

20 20 20 20 20

12

20

12

I. Nagelkerken et al. / Science of the Total Environment 645 (2018) 615–622

A baited remote underwater video (BRUV) device was constructed (see Nagelkerken et al., 2017) to quantify the total number of competitive interactions in 2016 between the three most common fish species (common, blue-eyed, and Yaldwin's triplefins), and quantify which species won or lost the competitive battles. A transparent vial (diameter 4.25 cm, height 5.5 cm) with ~20 small holes (~2 mm), was attached on top of a square fibre cement plate (20 × 20 cm). For stabilisation, the plate was connected with a 70 cm long metal strip to a cinder block. A GoPro camera (Hero 4) was mounted on top of the cinder block with the lens directed towards the food vial. Six identical BRUVs were constructed. They were placed randomly on the substratum at vent and control sites. The BRUVs were always placed concurrently at control and vent sites, and replicated across 2 days. Each deployment of a BRUV was considered a separate replicate, and a total of 12 vent and 12 control replicates were performed (Table 1). At the start of each BRUV deployment, a fresh piece of fish (~2 × 4 cm) was placed in the food vial to attract fish from the directly surrounding environment and concentrate them within the field of view of the camera. Filming was performed under the wide angle setting with a resolution of 1080p at a speed of 25 frames/s and lasted 10 min. During this period we quantified the total number of competitive interactions among the three species. Furthermore, we determined for each observed interaction which species lost (being chased away) vs. won (chaser) the battle. Because only one interaction was observed between the blueeyed and Yaldwin's triplefins we only report the results for the common triplefin vs. the blue-eyed and Yaldwin's triplefins, respectively. The results of this study differ from those by Nagelkerken et al. (2017). Whilst the present study specifically quantified species-specific competitive losses towards just the behaviourally dominant common triplefin, Nagelkerken et al. (2017) focussed on lost interactions independent of the aggressor's identity and did not quantify the total number of species interactions. All experiments were performed under animal ethics approval numbers S-2015-222 and S-2015-019, and according to the University's animal ethics guidelines. 2.7. Statistical analyses Differences in diet composition among the three most abundant species were tested with a 4-way MANOVA (multivariate analysis of variance), with ‘CO2 treatment’ as a fixed factor, and ‘year’, ‘site’ nested within treatment, and ‘fish species’ as random factors. Differences in surface orientation use were tested with a 3-way MANOVA, with ‘CO2 treatment’ as a fixed factor, and ‘fish species’ and ‘site’ nested within treatment both as random factors. Species was designated as a random factor because we selected these species from a ranked list of species abundances in 2015 rather than an a priori selection of three particular species. Differences in use of the 7 microhabitats and 3 substrate orientations (with a total of 17 habitat–surface orientation combinations) were tested with a 2-way MANOVA, with ‘CO2 treatment’ as a fixed factor, and ‘fish species’ as a random factor. The above three analyses were all performed using MANOVA, as they incorporated multiple dependent variables (i.e. different diet components, surface orientations, and microhabitats) within a single analysis. Differences in relative availability of differently oriented habitat surfaces were tested with a 3-way ANOVA (analysis of variance) with ‘CO2 treatment’ and ‘surface orientation’ as fixed factors, and ‘site’ nested within treatment as a random factor. Differences in relative availability of the seven microhabitats were tested with a 2-way ANOVA with ‘CO2 treatment’ and ‘microhabitat’ as fixed factors. Differences in turf algal height on the differently oriented substrata at control and vent were tested with a 3-way ANOVA, with ‘CO2 treatment’ and ‘surface orientation’ as fixed factors, and ‘site’ nested within treatment as a random factor.

For differences in absolute density of each of the three triplefin species and species richness between control and vents, 2-way ANOVAs were used (fixed factors: ‘CO2 treatment’ and ‘surface orientation’). To test for differences in total number of competitive interactions and % interactions lost between control and vents, 2-way ANOVAs were used with ‘CO2 treatment’ as a fixed factor and ‘experimental BRUV’ nested within treatment as a random factor. Post-hoc pooling (Winer et al., 1991) of lower-order interaction terms (if p N 0.25) was performed to enable a more powerful test of the main effects and interactions. All main tests were followed by pair-wise comparisons of the means. PRIMER version 7 was used to perform all statistical analyses. 3. Results The three most common benthic fish species had highly overlapping trophic niches as shown by lack of a significant difference in their diet composition (MANOVA, p = 0.339; Tables S2, S3), irrespective of year (p = 0.248) or CO2 treatment (p = 0.335). In contrast, species differed significantly in their use of differentlyorientated habitat surfaces (Fig. 1; MANOVA, species: p = 0.0003, Table S4a), irrespective of CO2 treatment (p = 0.274). The common triplefin differed in its surface use from the blue-eyed triplefin (posthoc test: p = 0.0194) and the Yaldwin's triplefin (p = 0.0001). The common triplefin was most abundant on horizontal and angled surfaces

Control

a) 70

Relative density (% )

2.6. Competitive interactions

60

common blue-eyed Yaldwin's

50 40 30 20 10 0 Horizontal

Angle

Vertical

CO2 vent

b) 70 60

Relative density (% )

618

50 40 30 20 10 0 Horizontal

Angle

Vertical

Fig. 1. Relative densities (mean ± SE) of fishes on differently-oriented hard substratum surfaces at controls (a) and vents (b) in 2015. The relative densities across surfaces sum up to 100% for each fish species. Details of statistical analyses are shown in Table S4a.

I. Nagelkerken et al. / Science of the Total Environment 645 (2018) 615–622

Relative abundance (%)

a)

50

Control

45 40

619

common Yaldwin's blue-eyed

35 30 25 20 15 10 5 0

Relative abundance (%)

b)

50

CO2 vent

45 40 35 30 25 20 15 10 5 0

microhabitat-orientation combination Fig. 2. Relative abundances (mean ± SE) of fishes on different microhabitats at different habitat surface angles for controls (a) and vents (b) in 2018. Total number of microhabitat– orientation combinations is 17, and the relative abundances across microhabitats sum up to 100% for each fish species. Details of statistical analyses are shown in Table S4c. Microhabitat abbreviations: T = turf algae b10 cm in height (largely Gigartina sp.), Ca = erect calcareous algae (largely Amphiroa sp.), Cca = crustose coralline algae, B = bare rocky substratum, Fa = encrusting fleshy green algae (Codium sp.), S = sand, Co = cobbles (~0.5–2 cm in diameter). Habitat surface orientation abbreviations: hor = horizontal, ang = angle, ver = vertical.

(Fig. 1), the blue-eyed triplefin showed a similar distribution across the three surfaces, and the Yaldwin's triplefin predominated on the angled surfaces. The availability of habitat substrates with different orientations did not differ between controls and vents (Fig. S1a; ANOVA: p = 0.4199, Table S4b). Comparable to habitat surface orientation, the three species differed in their use of various microhabitats (Fig. 2; MANOVA, species × treatment interaction: p = 0.0013, Table S4c). All three species differed from one another at control (post-hoc tests: p b 0.0011) as well as vent (p b 0.0233) sites, with microhabitat use also differing between controls and vents for the common triplefin (p = 0.0018) and the Yaldwin's triplefin (p = 0.0001) but not for the blue-eyed triplefin (p = 0.3708). At the control sites, the common triplefin was mainly found on turf algae (especially on horizontal and angled surfaces) and to a lesser extent on calcifying algae (erect calcareous and crustose coralline algae), the Yaldwin's triplefin mainly on calcifying algae (especially on vertical surfaces), and the blue-eyed triplefin mainly on crustose coralline algae (horizontal and vertical surfaces). At vents, the common triplefin showed a sharp increase in relative abundance on vertical surfaces covered by turf and on all surfaces with fleshy algae compared to controls, whilst their relative abundance was lower on surfaces with calcifying

algae. Likewise, the Yaldwin's triplefin showed higher relative abundance on turf and fleshy algae on vertical surfaces and lower abundance on erect calcareous algae on vertical surfaces. The blue-eyed triplefin also showed higher relative abundance on turf algae (on horizontal surfaces) and fleshy algae, and lower abundance on calcifying algae, at vents vs. controls. The change in relative microhabitat use between controls and vents by the triplefins matched that of altered microhabitat availability at vents. Microhabitat availability differed between control and vents (ANOVA, microhabitat × treatment interaction: p = 0.0001, Table S4d), with cover of turf and fleshy green algae more than doubling, and that of erect calcareous algae and crustose coralline algae decreasing by more than half at vents (Fig. 3a). The height of turf algal microhabitat did not differ across the surface orientations under ambient conditions, but increased at least threefold at CO2 vents on the horizontal and angled surfaces (Fig. 3b; ANOVA, treatment × surface orientation interaction: p = 0.0002, Table S4e). The density of the common triplefin at each of the three surface orientations was higher at vents than controls (Fig. S2a; ANOVA: p = 0.0006, Table S5a). In contrast, density of the less common blue-eyed (Fig. S2b; ANOVA: p = 0.0493, Table S5b) and Yaldwin's (Fig. S2c;

620

I. Nagelkerken et al. / Science of the Total Environment 645 (2018) 615–622

Microhabitat availability

a) 60

*

40 30

* 20

*

10

* 0 Turf algae

Calc algae

CCA algae

Fleshy Bare algae substr

Turf algal elevation

b) 2.5

Turf height (cm)

b

Sand Cobbles Other

Control CO2 vent

b

2.0

1.5

a

a

a b b

b

1.0

0.5

0.0 Horizontal

Angle

Vertical

a

a a

a

0.5

0.0 Horizontal

2.0

Control CO2 vent

Fig. 4. Fish species richness (mean ± SE) on differently-orientated habitat surfaces at controls and CO2 vents in 2015. Letters above bars indicate significant differences (p b 0.05) if letters are not shared, based on post-hoc analyses; details of statistical analyses are shown in Table S5d.

1.5

1.0

2.5

Control CO2 vent

*

No. of species (m-1)

Cover (% )

50

Fish species richness

Angle

Vertical

Fig. 3. Relative cover (mean ± SE) of benthic microhabitats in 2018 at controls and vents (a). The relative coverages across microhabitats sum up to 100% for vents and controls, respectively. Details of statistical analyses are shown in Table S4d. Microhabitat abbreviations: Turf algae = turf algae b10 cm in height (largely Gigartina sp.), Calc algae = erect calcareous algae (largely Amphiroa sp.), CCA algae = crustose coralline algae, Fleshy algae = encrusting fleshy green algae (Codium sp.), Bare substr = bare rocky substratum. (b) Turf algal height (mean ± SE) on differently-orientated habitat surfaces in 2015. Letters above bars indicate significant differences (p b 0.05) if letters are not shared, based on post-hoc analyses; details of statistical analyses are shown in Table S4e.

ANOVA: p = 0.0281, Table S5c) triplefins was lower at the vents across all three surface orientations. Fish species richness of all benthic fishes was 1.5 times higher at controls than vents on horizontal surfaces (Fig. 4; ANOVA, treatment × surface orientation interaction: p = 0.0336, Table S5d). The common triplefin was the winning species in, on average, 81–100% of the competitive interactions with the blue-eyed and Yaldwin's triplefins, irrespective of CO2 treatment (Fig. S3a, Table S6). The total number of competitive interactions between species did not differ between controls and vents (Fig. S3b, Table S6). 4. Discussion 4.1. Niche use as a function of niche availability Niche segregation allows species to co-exist (Schoener, 1974; Amarasekare, 2003), but we show that although significant habitat niche segregation (in microhabitat as well as habitat surface orientation) was evident among the three fish species under ambient and elevated CO2 conditions, species diversity and densities of some species still experienced a decrease at the CO2 vents. This reduction might be explained by at least three non-exclusive mechanisms: (1) alterations in preferred microhabitat availability, (2) increased abundances of behaviourally dominant competitor species, and (3) changes in microhabitat preference under elevated CO2.

Firstly, fish abundances changed according to microhabitat changes at vents compared to controls. All three fish species experienced reduced densities at vents on calcifying algal microhabitat (erect calcareous and crustose coralline algae) whilst showing increased densities on non-calcifying algae (turf and fleshy algae). Likewise, availability of calcifying algal habitat decreased and that of non-calcifying algae increased at the vents. Such changes in microhabitat cover are expected as elevated CO2 reduces the ability of species to build calcareous skeletons (Doropoulos et al., 2012), whilst it acts as a nutrient for fleshy algae (Connell and Russell, 2010). Hence, changes in relative abundance of calcifying vs. non-calcifying microhabitats are probably the primary driver of altered benthic fish densities at the CO2 vents. Secondly, the three fish species had only partially segregated niches (microhabitats and surfaces of different orientation) under contemporary conditions, and the behaviourally dominant fish species increased in density across all habitat surfaces under elevated CO2. At the same time, relative densities of the two subordinate species increased at the preferred habitat of the dominant species at the vents (i.e. turf and fleshy algae). Together, these changes increased the degree of overlap in habitat use (orientation and microhabitat) between dominant and subordinate fish species. Such small increases in niche overlap can be sufficient to enable species exclusion when species markedly differ in their competitive ability (Bulleri et al., 2016). The competitive dominance and increased abundance of the common triplefin may thus have exacerbated the loss of preferred calcareous microhabitat of subordinate species. The fact that densities of the dominant species increased considerably on vertical surfaces where its preferred microhabitat (turf algae) did not increase in cover, suggests that changes in microhabitat availability are not the only driver of fish abundances at vents. Nagelkerken et al. (2017) showed that reduced predator densities at vents contribute to the population expansion of the common triplefin. With increased niche overlap and loss of microhabitat, alternative refuge niches may have become limiting to maintain population sizes of subordinate species at local scales, with reduced species abundances and richness as a consequence. Thirdly, elevated CO2 is known to alter a wide range of behaviours, including habitat choice (Nagelkerken and Munday, 2016; Goldenberg et al., 2018). Studies have shown a reduced selection of preferred habitat by fishes under elevated CO2 (Devine and Munday, 2013; Nagelkerken et al., 2016a), but this likely played a minor role in the current study if at all. In an earlier study, we showed that in the absence of

I. Nagelkerken et al. / Science of the Total Environment 645 (2018) 615–622

other fish species, the common and Yaldwin's triplefins preferred turf habitat over bare substratum and this preference did not change under elevated CO2 (Nagelkerken et al., 2017). Likewise, the common triplefin did not alter its preference for turf algae that prospered at the vents and the blue-eyed and Yaldwin's triplefin still occupied the little remaining calcifying algal microhabitats at the vents. Ecological complexity is known to buffer the negative impacts of ocean acidification on fish behaviour (Goldenberg et al., 2018), which might explain the absence of a clearly altered habitat preference in our species under elevated CO2 in nature. 4.2. Competitive interactions In contrast to many other species and ecosystems (Tylianakis et al., 2008; Gilman et al., 2010; McCormick et al., 2013), the competitive ranking of species remained unaltered (Nagelkerken et al., 2017) under elevated CO2, suggesting that the way in which these species use their resources is intrinsic to the species itself (Amarasekare, 2003). The common, expanding species won by far the majority of competitive interactions (in the presence of food) with other less-common species, and all species showed overlap in diet, which could explain the partial segregation in habitat use among species under ambient CO2 conditions. Such an unaltered competitive ranking that is linked to resource acquisition can help explain why the dominant species was able to expand (rather than maintain) its populations at the vents to the detriment of behaviourally subordinate species. At first, when the dominant species is experiencing an increase in population size, an increase in total number of competitive interactions with other species would be expected (Milazzo et al., 2013). However, we observed the fish community at an evolved rather than initial stage, where population sizes of subordinates had already decreased resulting in fewer subordinates for the dominant to interact with. The total number of species interactions therefore remained the same at vents and controls. Whilst species competitive strength was not affected at vents, we show that under elevated CO2 species behaviour (i.e. competitive dominance) can exacerbate the negative effects of microhabitat loss, together acting as key pathways towards altered species populations and local diversity. 5. Conclusions In conclusion, the inherent ecological complexity of natural communities can propagate or buffer CO2 enrichment effects on individual species, and their senses and behavioural responses. Whilst some behaviours are altered by elevated CO2 (Nagelkerken and Munday, 2016) we found little support of this effect on relative niche use and competitive interactions. More importantly, relative habitat use largely changed according to microhabitat availability, with a reduced use of diminishing calcifying algae and increased use of expanding fleshy, non-calcifying algae. The expansion of a competitively dominant species in its preferred and less-preferred habitat niches under elevated CO2 likely exacerbated the reduction in densities of numerically and competitively inferior species, with loss of local species diversity on horizontal substrates. We show that ocean acidification can act as a powerful driver of species population and community changes by altering microhabitat niches and increasing niche overlap among competing species, with potential local species reductions as the ultimate outcome. Acknowledgements Financial support was provided by an Australian Research Council Future fellowship to I.N. (grant no. FT120100183) and a grant from the Environment Institute. S.D.C. was supported by an ARC Future Fellowship (grant no. FT0991953). We thank Camilo Ferreira for logistic support in the field.

621

Appendix A. Supplementary results Supplementary results to this article can be found online at https:// doi.org/10.1016/j.scitotenv.2018.07.168. References Alcaraz, C., Gholami, Z., Esmaeili, H.R., García-Berthou, E., 2015. Herbivory and seasonal changes in diet of a highly endemic cyprinodontid fish (Aphanius farsicus). Environ. Biol. Fish 98, 1541–1554. Amarasekare, P., 2003. Competitive coexistence in spatially structured environments: a synthesis. Ecol. Lett. 6, 1109–1122. Armstrong, R.A., McGehee, R., 1976. Coexistence of species competing for shared resources. Theor. Popul. Biol. 9, 317–328. Blois, J.L., Williams, J.W., Fitzpatrick, M.C., Jackson, S.T., Ferrier, S., 2013. Space can substitute for time in predicting climate-change effects on biodiversity. Proc. Natl. Acad. Sci. U. S. A. 110, 9374–9379. Bopp, L., Resplandy, L., Orr, J.C., Doney, S.C., Dunne, J.P., Gehlen, M., Halloran, P., Heinze, C., Ilyina, T., Séférian, R., Tjiputra, J., Vichi, M., 2013. Multiple stressors of ocean ecosystems in the 21st century: projections with CMIP5 models. Biogeosciences 10, 6225–6245. Briffa, M., de la Haye, K., Munday, P.L., 2012. High CO2 and marine animal behaviour: potential mechanisms and ecological consequences. Mar. Pollut. Bull. 64, 1519–1528. Bulleri, F., Bruno, J.F., Silliman, B.R., Stachowicz, J.J., 2016. Facilitation and the niche: implications for coexistence, range shifts and ecosystem functioning. Funct. Ecol. 30, 70–78. Chesson, P., 2000. Mechanisms of maintenance of species diversity. Annu. Rev. Ecol. Syst. 31, 343–366. Clements, J.C., Hunt, H.L., 2015. Marine animal behaviour in a high CO2 ocean. Mar. Ecol. Prog. Ser. 536, 259–279. Connell, S.D., Russell, B.D., 2010. The direct effects of increasing CO2 and temperature on non-calcifying organisms: increasing the potential for phase shifts in kelp forests. Proc. R. Soc. B 277, 1409–1415. Connell, S.D., Doubleday, Z.A., Hamlyn, S.B., Foster, N.R., Harley, C.D.G., Helmuth, B., Kelaher, B.P., Nagelkerken, I., Sarà, G., Russell, B.D., 2017. How ocean acidification can benefit calcifiers. Curr. Biol. 27, R95–R96. Connell, S.D., Doubleday, Z.A., Foster, N.R., Hamlyn, S.B., Harley, C.D.G., Helmuth, B., Kelaher, B.P., Nagelkerken, I., Rodgers, K.L., Sarà, G., Russell, B.D., 2018. The duality of ocean acidification as a resource and a stressor. Ecology 88, 1005–1010. Devine, B.M., Munday, P.L., 2013. Habitat preferences of coral-associated fishes are altered by short-term exposure to elevated CO2. Mar. Biol. 160, 1955–1962. Dickson, A.G., Millero, F.J., 1987. A comparison of the equilibrium-constants for the dissociation of carbonic-acid in seawater media. Deep Sea Res. Part A 34, 1733–1743. Dickson, A.G., Sabine, C.L., Christian, J.R., 2007. Guide to best practices for ocean CO2 measurements. PICES Special Publication vol. 3. Doropoulos, C., Ward, S., Diaz-Pulido, G., Hoegh-Guldberg, O., Mumby, P.J., 2012. Ocean acidification reduces coral recruitment by disrupting intimate larval-algal settlement interactions. Ecol. Lett. 15, 338–346. Enochs, I.C., Manzello, D.P., Donham, E.M., Kolodziej, G., Okano, R., Johnston, L., Young, C., Iguel, J., Edwards, C.B., Fox, M.D., Valentino, L., Johnson, S., Benavente, D., Clark, S.J., Carlton, R., Burton, T., Eynaud, Y., Price, N.N., 2015. Shift from coral to macroalgae dominance on a volcanically acidified reef. Nat. Clim. Chang. 5, 1083–1088. Gilman, S.E., Urban, M.C., Tewksbury, J., Gilchrist, G.W., Holt, R.D., 2010. A framework for community interactions under climate change. Trends Ecol. Evol. 25, 325–331. Goldenberg, S.U., Nagelkerken, I., Ferreira, C.M., Ullah, H., Connell, S.D., 2017. Boosted food web productivity through ocean acidification collapses under warming. Glob. Chang. Biol. 23, 4177–4184. Goldenberg, S.U., Nagelkerken, I., Marangon, E., Bonnet, A., Ferreira, C.M., Connell, S.D., 2018. Ecological complexity buffers the impacts of future climate on marine animals. Nat. Clim. Chang. 8, 229–233. Inoue, S., Kayanne, H., Yamamoto, S., Kurihara, H., 2013. Spatial community shift from hard to soft corals in acidified water. Nat. Clim. Chang. 3, 683–687. IPCC, 2014. In: Barros, V.R., Field, C.B., Dokken, D.J., Mastrandrea, M.D., Mach, K.J., Bilir, T.E., Chatterjee, M., Ebi, K.L., Estrada, Y.O., Genova, R.C., Girma, B., Kissel, E.S., Levy, A.N., MacCracken, S., Mastrandrea, P.R., White, L.L. (Eds.), Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Kroeker, K.J., Micheli, F., Gambi, M.C., Martz, T.R., 2011. Divergent ecosystem responses within a benthic marine community to ocean acidification. Proc. Natl. Acad. Sci. U. S. A. 108, 14515–14520. McCormick, M.I., Watson, S.-A., Munday, P.L., 2013. Ocean acidification reverses competition for space as habitats degrade. Sci. Rep. 3, 3280. Mehrbach, C., Culberson, C.H., Hawley, J.E., Pytkowicx, R.M., 1973. Measurement of apparent dissociation – constants of carbonic-acid in seawater at atmospheric-pressure. Limnol. Oceanogr. 18, 897–907. Milazzo, M., Mirto, S., Domenici, P., Gristina, M., 2013. Climate change exacerbates interspecific interactions in sympatric coastal fishes. J. Anim. Ecol. 82, 468–477. Munday, P.L., Jones, G.P., Caley, M.J., 2001. Interspecific competition and coexistence in a guild of coral-dwelling fishes. Ecology 82, 2177–2189. Nagelkerken, I., Connell, S.D., 2015. Global alteration of ocean ecosystem functioning due to increasing human CO2 emissions. Proc. Natl. Acad. Sci. U. S. A. 112, 13272–13277.

622

I. Nagelkerken et al. / Science of the Total Environment 645 (2018) 615–622

Nagelkerken, I., Munday, P.L., 2016. Animal behaviour shapes the ecological effects of ocean acidification and warming: moving from individual to community-level responses. Glob. Chang. Biol. 22, 974–989. Nagelkerken, I., Pitt, K.A., Rutte, M.D., Geertsma, R.C., 2016a. Ocean acidification alters fish–jellyfish symbiosis. Proc. R. Soc. B 283, 20161146. Nagelkerken, I., Russell, B.D., Gillanders, B.M., Connell, S.D., 2016b. Ocean acidification alters fish populations indirectly through habitat modification. Nat. Clim. Chang. 6, 89–93. Nagelkerken, I., Goldenberg, S.U., Ferreira, C.M., Russell, B.D., Connell, S.D., 2017. Species interactions drive fish biodiversity loss in a high-CO2 world. Curr. Biol. 27, 2177–2184. Pearson, P.N., Palmer, M.R., 2000. Atmospheric carbon dioxide concentrations over the past 60 million years. Nature 406, 695–699. Pianka, E.R., 1974. Niche overlap and diffuse competition. Proc. Natl. Acad. Sci. U. S. A. 71, 2141–2145. Pörtner, H.O., Farrell, A.P., 2008. Physiology and climate change. Science 322, 690–692. Riebesell, U., Gattuso, J.P., 2015. Lessons learned from ocean acidification research. Nat. Clim. Chang. 5, 12–14.

Ross, S.T., 1986. Resource partitioning in fish assemblages - a review of field studies. Copeia 1986, 352–388. Schoener, T.W., 1974. Resource partitioning in ecological communities. Science 185, 27–39. Stewart, F.M., Levin, B.R., 1973. Partitioning of resources and outcome of interspecific competition - model and some general considerations. Am. Nat. 107, 171–198. Syms, C., Jones, G.P., 1999. Scale of disturbance and the structure of a temperate fish guild. Ecology 80, 921–940. Tylianakis, J.M., Didham, R.K., Bascompte, J., Wardle, D.A., 2008. Global change and species interactions in terrestrial ecosystems. Ecol. Lett. 11, 1351–1363. Ullah, H., Nagelkerken, I., Goldenberg, S.U., Fordham, D.A., 2018. Climate change could drive marine food web collapse through altered trophic flows and cyanobacterial proliferation. PLoS Biol. 16, e2003446. Wellenreuther, M., Barrett, P.T., Clements, K.D., 2007. Ecological diversification in habitat use by subtidal triplefin fishes, Tripterygiidae. Mar. Ecol. Prog. Ser. 330, 235–246. Winer, B.J., Brown, D.R., Michels, K.M., 1991. Statistical Principles in Experimental Design. third ed. McGraw-Hill, New York.