Competitive interactions between corals and turf algae depend on coral colony form

Site description

This research was carried out in Ninh Van Bay (12.356°N; 109.277°E), part of the South Chinese Sea (Fig. 1A) and located nine kilometers northeast of Nha Trang, the seventh largest city of Vietnam. All surveys were conducted between March and April 2013. The reef at our study location extended over approximately 500 meter parallel to the wave-sheltered eastern side of Ninh Van Bay between depths of zero and five meters (Fig. 1B). Fishing is prohibited at this site, but occurs in other sections of Ninh Van Bay and adjacent waters (Ngoc, Flaaten & Anh, 2009).

Benthic cover

Photoquadrats were used to quantify the composition of the benthic community (Preskitt, Vroom & Smith, 2004). Four transects of 50 m were deployed with at least 50 m in between at a depth between two and five meters. Along each transect, 25 quadrats (0.9 × 0.6 m) were laid down at 2 m intervals and subsequently photographed using a digital underwater camera (Nikon AW100 Coolpix). Benthic cover and composition of all major functional groups were analyzed underneath 100 randomly placed points overlain on each picture using Coral Point Count with Excel extensions (Kohler & Gill, 2006). Corals and macroalgae were identified to genus-level whereas CCA and turf algae were classified as individual functional groups. Other benthic organisms (e.g., sponges, soft corals) were rare at our study site and not detected in our surveys. Non-biological substrates (e.g., sand, rubble, dead coral) were specified as such.

Surveys of coral algae interactions

To study coral–algal interactions, we used a line intercept approach described by Barott et al. (2009) and Barott et al. (2012). Along a depth range between two and five meter forty-two transects (25 m) were haphazardly positioned across the reef in various directions so that individual transects never bisected others. Each coral colony on each transect was photographed against a 30 cm ruler for scale. A top view photo was taken along with photos from various directions and distances to capture the entire coral–algal boundary. The proportion of coral border involved in each type of coral–algal interaction was later measured in ImageJ 1.48 (Abramoff, Magalhaes & Ram, 2004) using the top view photo, whereby side photos and close-ups were used to confirm the initial assessment if necessary. The putative outcome of each interaction was estimated by eye from the same pictures (see below). Algae were classified to genus for macro algae and to a single functional group for turf algae and CCA. The only exception to this method was made for branching Acropora corals since colony bases were generally hidden under an entanglement of branches. These Acropora colonies could exist of more than 100 individual branches. For colonies with more than 40 branches we photographed 40 individual primary branches and used the percentage of branches with algal growth as a proxy for the percentage of the coral border involved in competitive interactions. For smaller colonies, we analyzed all individual primary branches for algal growth. Branches that were completely overgrown from the primary branch upwards were considered dead and not included in the calculations.

For all colonies interacting with neighboring algae, the putative outcome of each competitive interaction was noted following the classifications of Barott et al. (2012). In short, competitive outcomes were classified as: (I) coral outcompeted neighboring algae, (II) algae outcompeted neighboring coral, or (III) if there was no obvious “winner” the interaction was classified as neutral. When healthy coral tissue was growing over the algal turfs, corals were assumed to outcompete neighboring algae (see also Fig. 2 in Barott et al. (2012)). Algae were regarded as winning the competitive interaction when the coral colony’s edge suffered from bleaching, discoloration, tissue necrosis or when algal turfs overgrew the coral surface. The first two categories are considered ‘directional competitive interactions,’ whereby one organism overgrew or killed its neighbor and eventually would take its place. Note that a single coral colony could be involved in multiple competitive interactions with multiple algal genera or functional groups and that each interaction could have multiple competitive outcomes. Each interaction and their outcomes were included proportionally in our analyses (Supplemental Information). We stress that we focused on describing the short-term dynamics of interaction zones and not on the long-term fate of individual colonies. Lastly we measured the maximum coral diameter to investigate a potential relationship between coral colony size and competitive outcome.

Definitions of colony growth forms and size classes

Every coral colony was classified into one of six growth forms; branching (B), encrusting (E), massive (M), plating (P), solitary (S) and upright (U) (Fig. 2). These classifications are based on commonly used typologies of coral growth forms (e.g., McCook, Jompa & Diaz-Pulido, 2001; Muko et al., 2013), that were binned to represent the morphological variation at our study site. Corals showing clear digitate-, corymbose- or branched patterns were grouped as branching (Fig. 2A), whereas all vertical orientated corals lacking these patterns (e.g., columnar- and foliating corals) were classified as upright (Fig. 2B). Massive corals were hemispherically shaped (Fig. 2C). Both plating and encrusting corals grew horizontally, whereby the former were elevated above the benthos and the latter grew over the benthos (Figs. 2D and 2E). Lastly, solitary corals were non-attached, free-living coral colonies moving over the top of the substratum (Fig. 2F) and in this study only included the genera Fungia and Ctenactis. All colonies were categorized into one of six size classes (‘0–5 cm,’ ‘5–10 cm,’ ‘10–20 cm,’ ‘20–40 cm,’ ‘40–80 cm,’ ‘80+ cm’) following Barott et al. (2012). For branching Acropora colonies it was not possible to identify individual colonies in dense Acropora thickets so that this group was excluded from the size class comparisons.

Statistical analyses

We compared the general abundance of algal groups to their relative abundance in coral–algal interactions using a two-tailed binomial test to test whether certain algal groups were over- or underrepresented in coral–algal interactions. Non-parametric, multiple comparisons Kruskal–Wallis tests with Bonferroni corrections (to compensate for multiple comparisons) were used to test for differences in the average proportion of coral edge interacting with turf algae for each of the six growth forms and the six size classes. Non-parametric multiple comparisons Kruskal–Wallis tests were also used to test which different growth forms and different size-classes were more successful in winning competitive interactions with turf algae. For these last tests we used the fraction of corals winning the interaction divided by the total amount of directional competitive interactions (average % of corals winning / (average % of corals winning + average % of algae winning)). Aforementioned tests were not only performed on growth morphology whereby different taxa were binned into one morphological category, but also within morphologically plastic coral genera that harbored multiple growth forms. We only compared growth forms within a single genus, if the growth forms were represented by at least 20 individuals each. Based on this comparison we could detect variation in the performance amongst growth forms without the possibility of falsely accrediting the differences to genus-specific reactions to turf algae. All statistical analyses were conducted in R (R Development Core Team 2010).

Benthic cover and coral–algal interactions

The reef community of Ninh Van Bay was comprised of macro- and turf algae (46.8%), stony corals (37.5%) and the remaining 15.7% consisted of non-biological substrates. The algal community was dominated by turf algae (74.6% of the total algal cover) accounting for 34.9% of the total reef community. After turf algae, green- (6.5%; Dictyosphaeria spp.; Halimeda spp. Valonia spp.), brown-(5.0%; Dictyota spp., Padina spp., Sargassum spp., Turbinaria spp.) and red algae (0.4%; Amphiroa sp., crustose coralline algae) were the most dominant algal taxa respectively (Table 1).

The circumferences of the 1,046 measurable coral colonies (i.e., all coral colonies except Acropora with a branching growth form) added up to 50,538 cm, of which 24,049 cm (47.6%) was in direct contact with algae. Again, turf algae were the dominant algal group, accounting for 87.6% of the total algal community along these coral borders. In the 230 branching Acropora coral colonies an estimated 47.0% of all branches were interacting with algae, of which 91.2% were algal turfs.

Coral growth form and coral–algal interactions

Of the six growth forms, encrusting corals were most commonly engaged in competitive interactions with turf algae (along 79.3% of their edges; Fig. 3A). In terms of the occurrence of competitive interactions, encrusting corals were followed by massive-(59.1%), upright-(49.7%) and branching corals (27.3%). Plating-(12.0%) and solitary corals (12.4%) experienced the least amount of turf algal interaction of the six growth forms (Fig. 3). Encrusting corals always had more turf algae growing along their borders than all other growth forms, i.e., from 1.34 times more than massive corals up to 6.61 times more than plating corals. Massive and upright coral colonies have more competitive interactions with turf algae than branching, plating and solitary growth forms (Table 2). Only in a few occasions did different morphologies experience a similar amount of coral–algal competitive interaction, i.e., branching and solitary, plating and solitary and upright and massive colonies (Table 2).

The effect of coral growth form on competitive outcomes

Similar to above, a colony’s success in outcompeting neighboring turf algae depended on its growth form. Encrusting corals were the most successful competitors against turf algae and won 35% of the interactions (Fig. 3B), followed by plating-(17%), massive-(13%), upright-(5%) and branching corals (2%). Solitary corals were only observed in neutral interactions or being outcompeted by turf algae (Fig. 3B). Turf algae were least successful in winning competitive interactions with encrusting corals (35%) and were increasingly better at outcompeting neighboring corals for solitary-(36%), massive-(38%), upright-(47%), plating-(66%), and branching growth forms (82%) (Fig. 3B).

Overall, turf algae won competitive interactions with corals more often than vice versa (Fig. 4), but growth forms that experienced more competitive interactions with turf algae were also more likely to be successful during such interactions. To illustrate, encrusting corals experienced the highest average amount of turf algal interaction along their perimeter, but they were the most successful growth form competing against turf algae, winning 49% of all competitive interactions (Fig. 4). The opposite pattern was observed for branching corals that had a relatively low number of their branches partly covered by turf algae, but could only be qualified as winning in 3% of all competitive interactions.

Taxon or growth form as the main driver of competitive success?

For six coral genera (Acropora, Favia, Favites, Montastrea, Montipora and Porites) at least twenty individuals of more than one growth form could be found. Comparing coral–turf algal interactions among growth forms within individual genera again showed differences in competitive outcomes among growth forms within the same genus (Fig. 5). Acropora was the most abundant coral genus on the reef (79% of total coral cover) and branching corals in this genus experienced three times more competitive interactions with turf algae along there edges than plating Acropora’s (Fig. 5A). Corals of the genus Favia were found as massive- and encrusting growth forms that experienced similar amounts of interaction with algal turfs. However, encrusting Favia colonies successfully outcompeted algal turfs more than twice as often as massive colonies (19.8% vs. 9.6% respectively) (Fig. 5B). Massive coral colonies within the genera Favites and Montastrea experienced 42.9% and 33.3% more interactions with turf algae respectively than encrusting colonies in the same genus (Figs. 5C and 5D). Within the genus Montipora, encrusting colonies had on average 2.45 times more interactions with turf algae along their edges than branching colonies. However, the former outcompeted turf algae more often than the latter, winning 70.2% and losing 16.8% of the competitive interactions whereas the branching colonies were never observed winning an interaction and losing 86.8% of them (Fig. 5E). Porites was the most variable genus in terms of growth forms. Of the four growth forms we observed, encrusting colonies again experienced more competitive interaction along their borders (76.1%) than the other three growth forms and branching corals experienced the least (21.4%) relative to upright-(39.2%) and massive colonies (47.5%) (Fig. 5F). Encrusting corals again performed better in these interactions, winning 4.8 times more often than massive colonies and up to 55 times more often compared to branching colonies (Fig. 5F).

Quantification and outcome of coral–algal interactions per coral colony size class

The relative percentage of a colony’s border involved in competitive interactions with neighboring turf algae decreased as colonies increased in size (Table 3; Fig. 6A). Colonies larger than 80 cm experienced 7.5 times less competitive interactions with turf algae along their borders compared to the two smallest size classes (Fig. 6A). However, such differences could not be statistically supported, suggesting that growth form is foremost important in determining the outcome of competitive interactions with neighboring turf algae at our study site (Figs. 6B and 6C; Table 3).