Introduction

Coral reefs are under increasing threat from many stressors, including more frequent heatwaves (Hughes et al. 2018), crown of thorns starfish outbreaks (Mellin et al. 2019), stronger cyclones (Cheal et al. 2017), and lethal diseases (Muller et al. 2020). The effects from these stressors can be devastating and widespread (De’ath et al. 2012) and reefs can take decades for reefs to recover to their pre-disturbance state (Robinson et al. 2019). There is a growing need to support the diversity and function of tropical reefs by developing interventions that assist coral recruitment and increase coral cover as the climate continues to change (Kleypas et al. 2021; Bay et al. 2023). While efforts to restore reefs have been successful at small scales (e.g., individual reef patches), the scale of reef restoration has failed to match the widespread and precipitous decline of coral reefs (Randall et al. 2020; Suggett and van Oppen 2022).

Reef restoration projects often transplant corals or split colonies into fragments to establish ‘new’ colonies on degraded reefs (Bayraktarov et al. 2019; Boström-Einarsson et al. 2020). Small coral fragments grow quickly and have relatively high survival (Edwards et al. 2015); however, the fragmentation method is labor intensive and sacrifices healthy corals for propagation (Randall et al. 2020). Seeding reefs with coral larvae or early recruits (spat) can be a more efficient approach, as a few coral colonies can produce hundreds of thousands of individuals with comparable genetic diversity to the natural populations (Zayasu and Suzuki 2019). However, spat have a low chance of survival due to grazing, sediment smothering, and competition with benthic algae. It has been estimated that approximately fifty percent of spat settled on the reef die within days of settlement (Martinez and Abelson 2013; Cooper et al. 2014); therefore, methods to improve spat survival will increase the utility of sexually produced spat for reef restoration (Randall et al. 2020; Banaszak et al. 2023).

A number of environmental factors, including temperature, water flow, and light intensity, affect the performance of corals across their life history, with optimum conditions restricted to a narrow range (Leal et al. 2016; Barton et al. 2017). Light is particularly important due to the intracellular symbiosis between corals and Symbiodiniaceae, which provides corals with a large proportion of their energy budget through photosynthesis (Roth 2014). While higher light intensity can translate to greater photosynthesis and faster coral growth (Wijgerde 2013), excessive light also brings the additional costs of photosystem repair, bleaching, and potential damage to coral tissue (Gómez-Campo et al. 2022). The responses of corals to light can be influenced by species-specific and environmental factors, including morphology, its symbiont type, light acclimation history, the exposure intensity and duration, as well as the quantity and quality of the light spectrum (Roth 2014). Therefore, coral responses to different light exposures are variable, with some studies showing positive associations between light intensity and coral growth (Izumi et al. 2023), others showing a negative association (Wijgerde 2013), and some show no clear relationship (Forsman et al. 2012; Schutter et al. 2012; Wijgerde et al. 2012; Kuanui et al. 2020). For example, in a direct comparison, only three of 10 coral species had positive responses to increasing light intensities (Wijgerde 2013).

For spat, the response to light intensity may be even more complex than for adult corals. It can take 2‒3 weeks for spat to establish symbioses (Koch et al. 2022), and during this period, spat rely on energetic reserves and heterotrophic feeding. Spat may be effectively asymbiotic, having only established a small population of symbionts, or may have partnered with a suboptimal symbiont; therefore, the light intensity required for growth and survival may change over the course of spat development (McMahon 2018; Kreh 2019; Brunner et al. 2022). As seen with adult corals, the response of spat to light intensity is mixed, with some evidence of faster growth and/or higher survival under low light intensity (McMahon 2018; e.g., 10 µmol m−2 s−1; Kreh 2019; Rahnke et al. 2022) and other studies showing better performance under high light (Hancock et al. 2021; Koch et al. 2022). These studies and the range of light intensities that spat experience, from 5 µmol m−2 s−1 for spat in a reef crevice (Doropoulos et al. 2016) to beyond 1000 µmol m−2 s−1 for adult corals on a reef flat (Carpenter 1985; Cooper et al. 2008), suggest that there is scope to optimize light conditions in aquaria to improve survival and growth of coral spat for restoration.

In addition to light intensity, the spectral properties of light can also affect coral growth and survival. Sunlight is relatively evenly distributed across the visible spectrum from 400 to 700 nm; however, the available light spectrum in marine ecosystems changes with depth and with the absorption characteristics of the water column (i.e., algal or particulate content; Jones et al. 2021). The spectra of LED light sources commonly used to grow corals often emphasize blue wavelengths (400‒500 nm), which have been associated with more rapid growth in adult corals compared to individuals grown under natural (full) spectrum profiles (Kinzie et al. 1984; Rocha et al. 2013; Mendes et al. 2017). However, blue light does not accelerate coral growth at all light intensities or in all species (Wijgerde et al. 2012; Izumi et al. 2023). For instance, Galaxea fascicularis grew faster in a blue-dominated light but only at intensities ≥ 125 µmol m−2 s−1 (Wijgerde et al. 2012). Coral aquaculture, for the production of seeding units for reef restoration, can utilize either natural or artificial light sources and both can be manipulated in intensity and spectral quality, offering opportunities to optimize the growth and survival of spat prior to deployment on the reef.

Light can also indirectly affect corals by promoting the growth of other photosynthetic benthic competitors, including macroalgae (Chadwick and Morrow 2011). For example, crustose coralline algae (CCA) are commonly used to induce coral larval settlement (Heyward and Negri 1999; Abdul Wahab et al. 2023), but can subsequently compete with spat for space on settlement substratum (Harrington et al. 2004; Jorissen et al. 2020). Other macroalgae can exclude corals for space on the reef and algal removal leads to higher coral and spat survival (Birrell et al. 2008; Arnold et al. 2010; Smith et al. 2023). In addition, macroalgae (including CCA) have a different suite of photosynthetic pigments than Symbiodiniaceae which may lead to different responses to light intensity or spectra than corals (Barott et al. 2009; Cikoš et al. 2022). The success and feasibility of large-scale coral restoration programs that produce larvae will rely on careful consideration of light intensity and spectral quality to ensure that optimal conditions, and maximum benefits, are provided to spat prior to deployment on the reef (Randall et al. 2020; Banaszak et al. 2023).

This study assessed the effects of light intensity and spectral quality on the performance of coral spat over a 12-week period. Acropora sp. nov. aff. kenti larvae were settled onto conditioned substrate with a live benthic community typical of coral aquaculture facilities, including several species of CCA, and exposed to combinations of light quantity (i.e., intensity) and quality (i.e., spectrum) to identify optimal light conditions for spat survival and growth.

Materials and methods

Coral collection, larval cultures, and settlement

Coral colonies were surveyed around Magnetic Island, QLD, Australia (19°07′43.2″S 146°52′38.5″E), and the Palm Island Group, QLD, Australia (18°45′54.8″S 146°31′36.1″E), between October 9 and 20, 2021. Gravid colonies of a species most similar to Acropora kenti (A. sp. nov. aff. kenti hereafter abbreviated as A. kenti) were identified via the presence of pigmented eggs. Until recently, this species was thought to be related to A. tenuis (Bridge et al. 2023). Gravid colonies (40–50 cm diameter) were collected using a hammer and chisel on SCUBA (1–9 m depth; Great Barrier Reef Marine Park Authority Permit G21/45348.1), transported to the Australian Institute of Marine Science National Sea Simulator (SeaSim; Townsville, QLD, Australia), and held in outdoor aquaria (2800 L; aquarium dimensions: 280 × 100 × 44 cm) for 5 days until spawning. The aquaria were maintained at reef temperature (~ 27.2 °C) and under natural sunlight (maximum photosynthetically active radiation (PAR): 200 µmol m−2 s−1). The holding aquaria were semi-recirculating systems that received three turnovers of filtered seawater per day (FSW; 0.1 µm nominal).

When setting of gamete bundles was observed, indicating that spawning is imminent, colonies were individually isolated in six 60 L aquaria. Buoyant bundles were collected from 6 colonies that fully spawned within 1 h of release, gently agitated and filtered through a 106 µm mesh to separate eggs and sperm, and then the eggs were washed thoroughly with FSW (> 4 × 300 mL). Similar volumes of eggs from the six colonies were pooled into 60 L FSW with ~ 1 × 106 sperm mL−1 for fertilization. After 1 h under static conditions, embryos were gently rinsed in FSW to remove excess sperm and transferred to 500 L culture tanks at a density of 0.3 larvae mL−1 (100 L h−1 flow through; ~ 27.2 °C). Culture tanks were gently aerated to keep embryos suspended for 24 h after which point aeration was increased to promote water circulation.

Concrete tiles were used as a settlement substratum and were conditioned prior to larval settlement. Dunlop concrete resurfacer (Ardex Australia Pty Ltd.) was used to make 28 × 28 cm tiles following manufacturer instructions. Tiles were conditioned in a semi-recirculating indoor aquarium containing a mixed community of crustose coralline algae (CCA) to develop a biofilm (Supplementary Fig. 1; 2800 L; aquarium dimensions: 280 × 100 × 44 cm; semi-recirculating with 3 turnovers d−1). Custom LED aquarium lights (blue and white dominated) and water temperature were controlled using a programmable logic controller (PLC, Siemens PCS7, Supervisory Control and Data Acquisition System (SCADA); max. PAR: 120 µmol m−2 s−1; temp.: ~ 27.2 °C; Luter et al. 2021). After 6 weeks of conditioning, tiles were 43 ± 2% covered with a thin layer of CCA (Supplementary Fig. 2), which induces settlement of Acropora spp. (Harrington et al. 2004). Conditioned tiles were cut into smaller (4.2 × 4.2 cm) experimental tiles which were used for larval settlement (N = 213).

For larval settlement, experimental tiles were distributed between two 50 L aquaria with FSW. Approximately 2,000 9-d-old A. kenti larvae were added to each aquarium and allowed to settle on tiles overnight. Water quality parameters were within normal range during this period (27.5 ±  < 0.1 °C, 8.0 ±  < 0.1 mg O2 L−1, 33.2 ±  < 0.1 PSU; HQ30D, Hach, USA). Tiles with more than five spat were transferred to a holding tank and larval settlement was repeated up to three times until all tiles had ≥ 5 spat to provide sufficient tile replication. Then, 10,000 cells mL−1 of cultured Cladocopium proliferum (Butler et al. 2023), originally isolated in 2010 from coral A. kenti near Magnetic Island in Queensland, Australia (Australian Institute of Marine Science ID: SCF 055-01.10), were added to the settlement tanks overnight in order to induce uptake of Symbiodiniaceae (Quigley et al. 2020).

Experimental setup and light treatments

Twenty-four hours after symbiont inoculation, tiles were distributed to 16 flow-through aquaria (50 L) where they were held for 12 weeks (0.8 L min−1 FSW; 27.3 ±  < 0.1 ˚C SE). Each aquarium contained 8–9 tiles and had an internal pump for water circulation (Tunze turbelle, Penzberg, Germany).

To investigate the independent and combined effects of light spectrum and light intensity, we crossed two light spectra (blue or full spectrum) with four intensities (Fig. 1a, Table 1; midday max. nominally: 5, 15, 60, or 160 µmol m−2 s−1), resulting in eight light treatments. Each light treatment was replicated in two aquaria with the highest intensity aquaria placed directly under the light source and lower intensity aquaria at increasing distance from the light source. Light was provided by LED panels developed by the National Sea Simulator and controlled by a SCADA system following a 12 h photoperiod with sinusoidal intensity profile (Fig. 1b). The blue spectrum was dominated by wavelengths of blue light and simulated artificial lighting common to some indoor aquarium systems and can be efficiently utilized by Symbiodiniaceae in coral (Wangpraseurt et al. 2014); the full spectrum simulated sunlight emission between 400 and 700 nm. Light intensities corresponded to 0.1‒5.4 mol m−2 d−1 daily light integral (DLI; Table 1) which encompass light intensities in the reef habitats preferred by coral larvae (Doropoulos et al. 2016).

Fig. 1
figure 1

a Light spectra for the blue and full spectrum treatments (160 µmol m−2 s−1). b Example diurnal cycle for the 160 µmol m−2 s−1 treatment

Full size image
Table 1 Summary of light intensity treatments. Maximum light intensity represents the average of 10 weekly measurements and estimated daily light integral (DLI)
Full size table

Each spectrum was measured using a Jaz spectrometer at 160 µmol m−2 s−1 (JAZ-EL200, Ocean Insight, USA). Light intensity was measured continuously in the 160 µmol m−2 s−1 treatment via the SCADA (LI-192 underwater quantum sensors, LI-COR Biosciences, USA) and weekly in all tanks manually (LI-192 with LI-250A light meter, LI-COR Biosciences, USA). The DLI for each treatment was calculated by scaling the continuous data from 160 µmol m−2 s−1 to 60, 15, and 5 µmol m−2 s−1 treatments (Table 1).

Spat were supplied daily with a mixed live feed (2000 cells mL−1 of a mix of Tisochrysis lutea, Nannochloropsis oceanica, Pavlova lutheri, Dunaliella sp.; 0.5 nauplii mL−1 of rotifers; 0.5 nauplii mL−1 of Artemia salina). All tile surfaces were uniformly cleaned on a weekly basis with a gentle brushing to remove superficial diatoms (but not green algae, brown algae or CCA). Turbo sp. snails (n = 8‒10, < 10 mm diameter) were added to each tank to help control diatom and non-coralline algal growth.

Algal fouling

The benthic macroalgal community was monitored to assess the effects of light treatments on algal cover and determine whether the light treatments also affected spatial competition with coral spat. Each tile was photographed fortnightly with a Nikon D810 camera with a 60 mm lens and Nikon Speedlight SB-910 flash. The algal composition of 8 tiles per treatment was measured from images using ImageJ (Rasband 2018). Pixels on each tile that represented red (CCA)-, green- and brown-colored algae were manually selected and then used to sample the image for all pixels with matching hue, saturation, and brightness (Image > Adjust > Color threshold > Sample; Supplementary Fig. 2). Thus, the algal categories represent a color group rather than a taxonomic group of algae. Given the taxonomic diversity and difficulty in identifying morpho-anatomically CCA to species level, we have not attempted to quantify different CCA species. However, at the end of the experiment, thick CCA crusts were identified to genus level by microscopic examination. The area of the tile was measured by tracing the outline of the tile and the proportion of settlement tile area occupied by each component of the community was calculated.

Spat survival and size

Spat survival was measured by calculating the proportion of living coral spat at each time point out of the number of spat initially settled on each tile (n = 18 tiles per treatment). All living spat were counted; any clumps of spat where individuals could not be distinguished were counted as one spat. Spat growth (mm2 week−1) was measured by tracing perimeters of spat in ImageJ and calculating the change in area from their initial size (Rasband 2018). Spat size can decrease through time due to tissue retraction or overgrowth by algae. Up to 26 spat (the median number per tile) were measured per tile and only spat that could be identified throughout the entire experiment were included, including chimeras established at settlement but excluding spat that fused during the experiment.

Statistical analysis

All statistical analyses and graphical results were performed in R version 4.3.0 (R Core Team 2023). CCA, green, and brown algal cover were analyzed individually using generalized linear models. CCA cover was analyzed using a beta error distribution, logit link function, and a three-way interaction of spectrum, intensity, and time (including lower-order interactions and main effects). Brown algal cover was analyzed in the same way as CCA cover only using a tweedie error distribution to account for the absence of brown algae on some tiles. Green algal cover was also analyzed using a tweedie error distribution, however a model with a three-way interaction failed to converge, and two-way interactions of spectrum, intensity, and time were used instead. There was no significant difference between a model with a three-way interaction versus a model with three two-way interactions for green algal cover (\(\chi_{(1.18)}^{2}\) = 18.7, p = 0.41).

To determine whether spat survival differed among spectra, intensities, and over time, survival was analyzed using a generalized linear mixed model with a binomial error distribution and log link (Bates et al. 2015). Predictors included light spectrum, light intensity, and time and their higher-order interactions. In addition, Kaplan–Meier survival curves were used to determine final survival estimates and hazard ratios for each treatment (Kassambara et al. 2021; Therneau 2022). The survival as a function of CCA cover was compared between light treatments using 10 and 20% lethal concentrations (LC10, LC20), which is the % CCA cover required to reduce spat survival by 10 or 20%, respectively (Fisher et al. 2022). Spat size was analyzed using a linear mixed model with spectrum, intensity, and time as fixed predictors and aquarium and tile as random intercepts (Bates et al. 2015).

For all responses except LC10 and LC20, tank and tile were used as random effects to account for grouping and repeated measurements. Residuals from model predictions were used to check for uniformity and dispersion (Hartig 2022). Significant interactions (e.g., spectrum*intensity*time) were examined using pairwise comparisons of treatments within each timepoint (Kassambara et al. 2021; Therneau 2022; Lenth 2023). P values for pairwise comparisons were adjusted using the Tukey method (Lenth 2023).

Results

Algal fouling

Initially, settlement tiles exhibited 50 ± 2% (SE) total fouling including 43 ± 2% crustose coralline algae (CCA), 6 ± 1% green-colored algae, and 1 ± 0.2% brown-colored algae (Fig. 2). At the end of the experiment, total fouling had increased to 88 ± 2% SE and consisted of 80 ± 2% CCA, 4 ± 1% green algae, and 5 ± 1% brown-colored algae when averaged across all treatments (Fig. 2, Supplementary Fig. 2). Light treatments did not alter the general structure of the benthic macroalgal community (Fig. 2): CCA was 10 times more abundant than green or brown algae in every light treatment at the end of the experiment (Supplementary Fig. 3). The CCA included Lithophyllum sp., Mesophyllum sp., Porolithon sp., and Titanoderma sp.; however, these genera were not easily distinguished morphologically and were scored as a group (Fig. 3). By the end of the experiment, tiles exposed to high light intensity were almost completely covered by a thick, cream-colored Lithophyllum species while low-light tiles had CCA with thin, purple crusts (Supplementary Fig. 3).

Fig. 2
figure 2

Proportion of tile area occupied by brown algae, crustose coralline algae (CCA), and green algae. Rows distinguish components of the algal community while columns distinguish the two light spectra. Brown algae, CCA, and green algae are represented by brown, black, and green lines, respectively. Light intensities (5, 15, 60 and 160 µmol m−2 s−1) are indicated using color shades, with darker shades representing lower intensities. Points represent means and error bars represent SE

Full size image
Fig. 3
figure 3

Competition with CCA and green algae at 0, 4, and 12 weeks. The top row shows CCA (Lithophyllum sp.) which commonly overgrew coral spat. Note that the appearance of the same CCA changed over the course of the experiment. The bottom row shows overgrowth by green algae. Large patches of green algae were rare, but green algae often occurred around the margin of coral spat

Full size image

CCA cover nearly doubled from 43 to 80% by week 12 (Fig. 2, Supplementary Fig. 2). The increase in CCA cover depended on both the light intensity and spectrum (Table 2; spectrum*intensity*time). Post hoc comparisons of light treatments within each week indicate the main factor influencing CCA cover was light intensity (Supplementary Table 1). Initially, high light intensity had a negative effect on CCA cover, whereby CCA cover was significantly lower under blue 160 µmol m−2 s−1 versus at 5 and 15 µmol m−2 s−1 (both spectra). CCA cover was also significantly lower under blue 160 µmol m−2 s−1 compared to full 160 µmol m−2 s−1, suggesting an initial effect of light spectrum at the highest light intensity. However, there were no other differences in CCA cover between spectra for a given light intensity. In weeks 4, 6, and 8, there were no significant differences in CCA cover among light treatments. By week 12, light intensity had the opposite effect to that observed at week 2, where CCA cover was significantly lower at blue 5 µmol m−2 s−1 than in all other treatments.

Table 2 Algal fouling composition analysis using Wald chi-square tests of generalized linear models for the main effects of spectrum, intensity, and time and higher-order interactions while tank and tile were included as random effects
Full size table

Green algal cover, including Bryopsis sp. and Pedobesia sp., ranged from 2 to 6% and was recorded on 73% of the tiles by week 12 (Fig. 2). Light spectrum and light intensity had independent effects on the abundance of green algae, but the effects of each treatment varied over time (spectrum*time, intensity*time; Table 2). Almost no significant effects of light treatments were observed until week 6, after which point higher light intensities had a negative effect on green algal cover. Beginning at week 6, green algal cover was significantly lower at 60 and 160 µmol m−2 s−1 compared to 5 and 15 µmol m−2 s−1 at nearly all timepoints (Supplementary Table 2). To a lesser degree, light spectrum also affected green algal cover, which was significantly higher under blue versus full spectrum light in weeks 6 through 12 (Supplementary Table 2).

Brown algal cover, including Colpomenia sp., Lobophora sp., and chain-forming diatoms, ranged from 1‒4% and was observed in 61% of tile photographs. While changes in brown algal cover depended on both light intensity and light spectrum, the major driver was a negative effect of high light intensity (Table 2; spectrum*intensity*time): in week 10, brown algal cover was significantly lower at 60 and 160 µmol m−2 s−1 (full spectrum) than 5 µmol m−2 s−1 (blue spectrum; Supplementary Table 3). This negative effect of light intensity on brown algae was reinforced by week 12, when brown algal cover was significantly lower at 60 and 160 µmol m−2 s−1 (full spectrum) compared to 5 µmol m−2 s−1 (both spectra). In addition, in week 12, brown algal cover was significantly lower at 60 µmol m−2 s−1 blue compared to 5 µmol m−2 s−1 full spectrum, further suggesting that intensity rather than spectrum is the primary driver of brown algal cover.

Spat survival and size

Spat survival gradually reduced over the 12-week experimental exposure period in all treatments but was strongly dependent on light treatment (Fig. 4). Higher light intensity had a strong negative effect on survival, but the effect was not apparent until after 6 weeks (intensity*time; Fig. 4a, Table 3), when survival was higher at 5 and 15 µmol m−2 s−1 compared to 60 or 160 µmol m−2 s−1, regardless of light spectrum (Supplementary Table 4). Survival continued to be higher at low light intensity through week 12 (Fig. 4b, Supplementary Table 4). Hazard ratios also describe diminished survival with increasing light intensity: spat under 15, 60, and 160 m−2 s−1 µmol m−2 s−1 were 88, 346, and 285%, respectively, more likely to die than spat at 5 µmol m−2 s−1. In contrast to light intensity, light spectrum had little effect on mortality risk, as spat under full spectrum were 13% more likely to die than spat under blue spectrum light. In pairwise comparisons (spectrum*intensity), the only significant effect of light spectrum was at 60 µmol m−2 s−1, where average survival over the course of the experiment was 7.4% lower under full compared to blue spectrum light (Fig. 4a; Supplementary Table 4).

Fig. 4
figure 4

a Spat survival estimated from Kaplan–Meier survival curves over 12 weeks of exposure to each light treatment. Light spectra are indicated using different colors and light intensity as different shades, with darker shades corresponding to lower light intensities. Highlighted regions indicate ± SE. b Mean survival and 95% confidence intervals after 12-week exposure to each light treatment

Full size image
Table 3 Linear model results comparing spectrum*intensity*time for spat survival and growth
Full size table

Surviving spat exhibited slightly negative growth (due to tissue retraction or overgrowth, see following section) across all experimental treatments (Fig. 5). Changes in spat area depended on light spectrum and light intensity (spectrum*intensity*time), but no two light treatments were significantly different in post hoc comparisons in any given week (Table 3, Supplementary Table 5). The effect of light intensity on spat size also varied over the course of the experiment (intensity*time): there were no significant differences in weeks 0‒4, spat were significantly smaller at 160 compared to 15 µmol m−2 s−1 in weeks 6 and 8, but no significant differences in size were evident in weeks 10 or 12 (Supplementary Table 5).

Fig. 5
figure 5

Spat area (mm2) over 12 weeks of exposure to each light treatment. Area represents the net outcome of growth and partial overgrowth, with only surviving spat included. Light spectra area indicated using different colors and light intensities as different shades (µmol m−2 s−1). Points represent means and error bars represent SE

Full size image

Relationship between CCA abundance and spat survival

High CCA cover was associated with low survival of A. kenti spat, which was exacerbated by high light intensity (Fig. 6). Under low light intensity, spat survival remained high despite high CCA cover; whereas under high light, spat survival decreased as CCA cover increased, suggesting more aggressive overgrowth by CCA (Fig. 6, Table 4). In comparison, the CCA cover associated with 20% spat mortality (LC20) decreased from > 95% CCA cover at 5 µmol m−2 s−1 to 63% CCA cover at 160 µmol m−2 s−1, suggesting that CCA was more detrimental to spat at higher light intensity (Fig. 6). In particular, Lithophyllum sp. completely covered tiles and was observed aggressively overgrowing spat, especially with thick crusts in the 60 and 160 µmol m−2 s−1 treatments (Fig. 6, Supplementary Fig. 3). Mortality at 5 µmol m−2 s−1 only reached 12% at the highest CCA cover recorded of 95%. CCA cover associated with 10% mortality (LC10) of coral spat did not show a clear association with light intensity (Table 4).

Fig. 6
figure 6

Spat survival was highest at low CCA cover but survival at CCA cover depended on the light intensity. Points represent the survival on each tile after 12 weeks. Curved lines and dotted regions represent the model-averaged predicted survival and 95% credible intervals, respectively. Vertical lines represent the lethal concentrations leading to 20% mortality (LC20)

Full size image
Table 4 The lethal concentration of crustose coralline algae (CCA) resulting in 10 and 20% mortality (LC10 and LC20) for coral spat at each light intensity
Full size table

Discussion

To achieve ecologically meaningful scales of coral restoration, aquaculture programs engaging sexual propagation strategies need to produce a high yield of coral spat (Randall et al. 2020), requiring the optimization of larval settlement, and larval and spat survival (Pollock et al. 2017; Hancock et al. 2021; Rahnke et al. 2022). This study demonstrates that rearing spat under low light intensity maximizes their chance of survival on settlement surfaces with competitive algae, leading to a higher potential yield for coral restoration. The poor performance of spat grown at the higher light intensities tested was largely due to the presence of crustose coralline algae (CCA), which rapidly overgrew spat over a period of 12 weeks. In comparison to light intensity, light spectrum (full vs blue) had relatively minor effects on algal growth and spat survival. This study highlights the potential benefits of low-light conditions for rearing coral spat in aquaria and suggests minimal benefits of using full-spectrum light sources that mimic that of sunlight compared to blue-dominated LED lights (McMahon 2018; Kreh 2019; Rahnke et al. 2022).

The conditioning of substrates with biofilms, including CCA, is a common practice to induce settlement of coral larvae in experimental and aquaculture systems (Guest et al. 2010; Randall et al. 2020; Banaszak et al. 2023). In this study, coral larvae were presented with settlement surfaces that were half-covered with algae, the majority of which was CCA. This allowed for the assessment of the effects of light intensity and spectra (full vs blue) on algal growth, spat survival and the interaction between benthic macroalgae and spat in the first 12 weeks post-settlement. The total benthic algal cover increased from 52 to 88% over this period, with CCA dominating the algal community under all light intensity and spectrum treatments. In the first weeks, CCA cover was lowest under the highest intensity blue light, yet by week 12, the lowest CCA cover was evident in the lowest intensity blue light treatment, indicating dynamic responses of the algal community to light intensity as the experiment progressed. Similar but less extreme patterns in growth were also observed for CCA exposed to full spectrum light. After 12 weeks, there was 29 and 5% more CCA in high light than low light treatments in the blue and full spectrum exposures, respectively. High light intensity clearly favors rapid and potentially aggressive algal growth, yet light spectral quality may be important to controlling selective algal growth under some conditions. Algae harvests visible light with a variety of photosynthetic pigments including chlorophyll a (all algae), chlorophyll b (green algae) and chlorophyll c (brown algae), primarily between 420 and 480 nm (blue spectrum) and between 630 and 670 nm (red) (Ke et al. 2021). However, CCA and cyanobacteria also contain the accessory photosynthetic pigments phycoerythrin and phycocyanin, which strongly absorb between 480 and 630 nm (green to red) (Kirk 2010). It is likely that the growth of each alga was at least subtly influenced by spectral quality in this experiment (Haxo and Blinks 1950). For example, CCA growth may be most inhibited by low intensity blue light due to the inability of phycoerythrin and phycocyanin to contribute to light harvesting for photosynthesis under these conditions, although chlorophyll a is also able to efficiently harvest light in the blue spectrum, potentially compensating light requirements under blue light.

Minor differences in cover of green and brown algae were observed among treatments, with both groups generally favored by low light conditions under both spectra. However, these results should not be assessed independently of the dominant CCA and any effects on the growth of the green and brown algal communities is likely to be influenced by both the direct effect of light conditions and competition for space with CCA. The direct effects of light intensity and spectra on each of the three groups of algae could only be assessed in the absence of competitors on the same substratum.

Changes in cover only partially represent the influence of light on CCA, as the composition of CCA species also varied between light treatments over the duration of the exposure. However, identification of CCA to species (or even genera) was not practical due to limited morphological features that can be used for taxonomic purposes, and because CCA crusts changed in appearance with time (e.g. due to ontogenesis) and light treatment. Nevertheless, by 12 weeks, tiles in the high intensity treatments of both spectra were dominated by thick crusts (not measured but > 1 mm) of Lithophyllum sp., while tiles in low light treatments were not dominated by this species and CCA crusts were thinner (~ 0.5 mm) under both spectra. It is unclear, however, whether these thin CCA represent different species or are simply early life stages of the dominant CCA.

Low light intensity was associated with the highest spat survivorship by week 12, where spat survival was three times higher at 5 µmol m−2 s−1 compared to at 160 µmol m−2 s−1 (both spectra). This result could be due to a direct influence of light intensity on spat health or indirect influence via effects on algal communities that then alter competitive dynamics between spat and algae. Light intensity has the potential to influence coral spat health through the effects on symbiont autotrophy. Prior to the establishment of a functional symbiosis, spat rely on lipid reserves and heterotrophy (Figueiredo et al. 2012; Toh et al. 2013a). The A. kenti spat were aposymbiotic until infection with C. proliferum within days of settlement and, although not measured, visual indications of symbiosis establishment (brown color of spat) were observed after 4 weeks. It is not clear in this experiment when autotrophy would provide a significant contribution to energetics of the spat. Other studies have reported that the optimal light intensity for spat growth can change over their early development. For example, 4-week-old spat grew faster at 10 µmol intensity compared to 240 µmol m−2 s−1 (McMahon 2018), while conversely a longer-term study found that 8-week-old species spat benefited from higher light intensity (> 120 µmol m−2 s−1; Kreh 2019). A previous study on A. millepora found that light attenuation had no influence on spat until 8 weeks, after which spat survival, size and symbiont density were all reduced (Brunner et al. 2022). While the timing of increased light requirements likely vary among coral species, these studies demonstrate that exposure to low light during early development appears to benefit a range of coral species.

The total CCA cover increased over 12 weeks across all light intensities; however, the LC20 for spat survival was reached at far lower % CCA cover under high light intensities than under the lower light intensities. While it could be argued that this indicates a direct negative effect of high intensity light on spat health, indirect effects likely play a role, as it was evident that the survival and growth of spat was strongly affected by competition with and overgrowth by macroalgae, particularly the CCA Lithophyllum sp. at higher light intensities. Although CCA is a favored inducer of coral larval settlement, the benefits of Lithophyllum sp. are less understood than for other CCA such as Porolithon or Titanoderma sp. (Heyward and Negri 1999; Harrington et al. 2004; Abdul Wahab et al. 2023). Lithophyllum sp. covered the surface of most settlement tiles by 12 weeks and it rapidly grew over spat in the process, especially at 60 and 160 µmol m−2 s−1 (Supplementary Fig. 3). In addition, the dominance of CCA at high light intensity likely limit the abundance of other competitors, such as green and brown colored algae (Jorissen et al. 2020). Despite the visual evidence of overgrowth of spat by aggressive CCA at higher light intensities, further investigation is required to disentangle possible direct and indirect effects of light intensity on spat survival (Jorissen et al. 2020).

Spat size appeared to increase under 60 and 160 µmol m−2 s−1 between weeks 8 and 12, suggesting that low light intensity may limit the growth of older spat as reported in other studies (McMahon 2018; Kreh 2019; Brunner et al. 2022). However, the survivors experienced little to no net growth over the course of the experiment, potentially due to space limitation. In the absence of competitors, spat growth would be an ideal whole organism response metric to assess the direct effects of light intensity. However, low light intensity might also have promoted higher survival via reduced photoinhibition or photodamage in coral symbionts (Roth 2014; Nitschke et al. 2022). Future studies addressing the direct vs indirect effects of light intensity on spat growth and survival in aquaculture should include coral photophysiological and symbiont density measures and an experimental design pairing light exposures with and without spatial competition.

In contrast to intensity, light spectrum had only minor effects on spat or macroalgae after 12 weeks and did not seem to affect competition or overgrowth processes. Previous studies on the other hand found that light spectrum affects coral growth, although they disagree as to whether blue, white, or red light led to the fastest growth (Kinzie et al. 1984; Rocha et al. 2013; Mendes et al. 2017). Coral symbionts absorb light between 400 and 550 nm (blue) as well as between 650 and 700 nm (red) (Scheufen et al. 2017). In this study, the custom LED panels produced blue-dominated and full spectrum light, potentially producing a different response in comparison to other studies that applied a variety of spectra (LED, LEP, or fluorescent sources). Additionally, the intensity of high energy blue wavelengths here was up to 2.5 times higher in the blue spectrum than full spectrum due to the restricted spectral range of the former light source and required light intensities. On the one hand, this could benefit symbiotic spat through higher intensities of photosynthetically usable radiation, but excess energy may result in photoinhibition and the generation of harmful reactive oxygen species (Nitschke et al. 2022). Adult A. kenti generally occurs in shallow, relatively high-light intensity, full light spectrum, habitat such as upper reef slopes (Veron 2000), while spat survival is often highest in reef crevices with low light intensity (~ 5 µmol m−2 s−1; Doropoulos et al. 2016), indicating that light requirements over the lifecycle of coral are likely to change dramatically (McMahon 2018; Kreh 2019; Brunner et al. 2022). In several studies, the benefits to coral growth from a full-spectrum light source are only reported at higher light intensity (e.g., > 40 µmol m−2 s−1; Wijgerde 2013; Wijgerde et al. 2012), suggesting that the light spectrum (and potentially different light sources) may be of limited significance when rearing spat under low light.

Several studies that reported faster growth rates of spat at higher light intensities have linked this with the additional autotrophic energy acquisition from their symbiont (Hancock et al. 2021; Koch et al. 2022). Both feeding and provision of symbionts have long-term effects on spat development, as early access to both nutritional pathways can increase the survival of spat reared ex situ and transplanted in situ (Suzuki et al. 2013; Toh et al. 2014). Therefore, the energy requirements of spat should be carefully considered (and potentially supplemented further) if low light conditions are applied during rearing in aquaculture to manage algal competition. A. kenti was provided with cultured symbionts in this study, but it is unclear as to when the photosymbiosis would contribute significant nutritional benefits to the coral; previous work suggest benefits are apparent > 8 weeks after settlement (McMahon 2018; Kreh 2019; Brunner et al. 2022). Spat are likely to feed on zooplankton within days following settlement (Cumbo et al. 2012; Toh et al. 2013a, b), and while it is unclear to what extent coral spat rely on autotrophy vs heterotrophy, it is clearly an important future consideration for coral aquaculture toward maximizing spat health and survival prior to deployment for reef restoration.

Ex situ restoration programs using sexually produced corals will require holding spat in aquarium systems prior to deployment (Lippmann et al. 2023), during which conditions can be adjusted to optimize the growth, condition, and survival of spat (Barton et al. 2017). This study demonstrated that low light conditions over the first 12 weeks can reduce mortality in spat by limiting overgrowth by CCA communities (used to induce larval settlement). The application of low light intensity for spat holding could also provide additional benefits for the production facility, including reducing power consumption (if conducted indoors) and minimizing husbandry (including the control of algal growth on tanks) required to maintain aquaria. As this study showed minimal effects of light spectra on algal growth or spat survival, costs could be further reduced by raising spat under shaded natural light prior to deployment, which may in turn facilitate increasing the scale of coral aquaculture for restoration.

While spat survival can be optimized through the control of light intensity ex situ, consideration should be given to the fate of spat following deployment under less controlled conditions. For example, there was very little mortality in spat maintained at ≤ 15 µmol m−2 s−1 for the first 10 weeks, but at 12 weeks, greater mortality was observed, likely due to overgrowth by slowly maturing CCA. This growth of CCA is likely to continue if spat were held longer and following in situ deployment. This ongoing hazard to spat could be reduced by excluding aggressive CCA such as Lithophyllum sp., and using a less competitive, thinner crust CCA species for settlement induction (Heyward and Negri 1999), the application of non-biocidal antifoul coatings (Tebben et al. 2014; Roepke et al. 2022), or by inducing settlement of coral larvae using dead (e.g. frozen or dried) CCA on settlement substrate. For a symbiotic organism that depends on photosynthesis, it is counterintuitive to maintain corals at low light intensity; while low light improves spat survival, it would be logical to increase light as the functional symbiosis with Symbiodiniaceae becomes established prior to deployment, which may help older spat survive through the first months following deployment in situ. This optimization of aquarium and substrate conditions, along with efforts to optimize spat nutrition, and increase spat thermal tolerance have the potential to enable restoration programs to meet the pace and scale of coral decline for effective coral restoration (Barton et al. 2017; McLeod et al. 2022; Suggett and Madeleine J.H. van Oppen 2022; Banaszak et al. 2023).