Introduction

Phenological observations of corals describe the influence of cyclic and seasonal physico-chemical factors such as water chemistry, temperature, photoperiod and food availability on coral physiological processes, nutritional composition and energy levels (Harrison 2011; Brooke and Järnegren 2013; Hinrichs et al. 2013a; Abdul Wahab et al., 2014). Throughout an annual cycle, scleractinian corals must partition metabolic energy among key physiological processes, including maintenance, growth, reproduction and competitive interactions (Leuzinger et al. 2012). The energy required to undertake these processes compared to the energy available is known as a coral’s energetic status, and this has implications for survival and fitness in the face of physico-chemical changes (Grottoli et al. 2006; Rodrigues and Grottoli 2006; Anthony et al. 2009). On a macro-scale, a coral’s nutritional profile predominantly comprises protein, carbohydrates and lipids. Lipids are energy dense, yielding at least one-third more energy relative to proteins or carbohydrates (Parrish 2013).

Lipids are a major component of the scleractinian coral composition (10–40% of ash free dry weight), while their constituent classes and fatty acids play important roles in energy storage, cell membrane structure and overall fitness (Patton et al. 1977; Harland et al. 1993; Yamashiro et al. 1999; Bergé and Barnathan 2005; Farre et al. 2010). Coral lipid profiles vary significantly in response to physiological processes such as photosynthesis, respiration, cell replenishment and reproduction, and these processes, in turn, are influenced by external physico-chemical factors such as water chemistry, rainfall, temperature and food availability (Arai et al. 1993; Leuzinger et al. 2003; Imbs 2013). As such, inter-annual variability in environmental conditions can influence important life processes in corals, and this is reflected in coral lipid profiles, making them a valuable and ubiquitous tool for monitoring coral health (Hinrichs et al. 2013a; Seemann et al. 2013).

The total lipid concentration in corals comprises several characterised lipid groups. Structural lipids such as sterols and phospholipids are major cell membrane and intracellular organelle constituents (Lall 2002), while storage lipids such as wax esters and triacylglycerols provide readily catabolised energy compounds (Lin et al. 2013). Meanwhile, fatty acids are essential for energy, growth, reproduction, eicosanoid production and membrane fluidity (Sargent et al. 2003).

The building blocks from which lipids are produced are obtained by corals through both phototrophy and heterotrophy, and a coral’s ability to shift from one mode to the other is poorly understood, though known to be species specific and heavily dependent upon environmental conditions (Anthony and Fabricius 2000; Anthony et al. 2002; Ribes et al. 2003; Palardy et al. 2005). Photosynthetic efficiency varies seasonally, since endogenous zooxanthellae densities fluctuate in response to changes in light intensity and temperature (Costa et al. 2005). Corals are generally considered non-selective, passive feeders, meaning the quantity and nature of heterotrophically derived nutrients depend on seasonal availability as opposed to trophic preference (Ribes et al. 1999). As such, the taxonomic and nutritional composition of the lower aquatic food web, particularly plankton, directly affects the lipid profile of the coral consumer (Treignier et al. 2008; Seemann et al. 2013).

Exploitation of different food sources by corals throughout the year can be inferred via fatty acid trophic markers (Imbs and Latyshev 2011), since dietary fatty acid composition is largely reflected in the tissue of the consumer (Dalsgaard 2003). Fatty acid profiles can also indicate the components of a coral’s cell membrane that are indispensable, as evidenced by selective retention and adjustment of specific fatty acids during unfavourable conditions such as food scarcity or environmental stress, and major life history events, such as reproduction (Izquierdo et al. 1996; Kainz and Fisk 2009; Parrish 2013). This knowledge is fundamental to understanding the environmental and biochemical processes underlying coral health today, and predicting how anomalous physio-chemical events and climate-driven changes will affect coral reef assemblages in the future (Oku et al. 2003a; Hoegh-Guldberg et al. 2007; Dodds et al. 2009; Hinrichs et al. 2013).

Therefore, to elucidate baseline coral energetic status in relation to physico-chemical changes in the environment, the current study examined the seasonal changes in the lipid and fatty acid composition over an annual cycle in the scleractinian coral, Acropora millepora—a common, polytrophic coral that undergoes broadcast spawning each spring (Van Oppen et al. 2011; Bay et al. 2013; Hemond et al. 2014). The coral energetic profiles were correlated with several environmental parameters, including water temperature, salinity, turbidity and PAR.

Materials and methods

Sample collection

Samples were collected bi-monthly from twenty tagged Acropora millepora colonies (all of reproductive size of > 20 cm diameter) between November 2009 and November 2010. This resulted in nine sample dates (early November 2009, late November 2009, January 2010, March 2010, May 2010, July 2010, September 2010, October 2010 and early November 2010). Coral were located in 1–3 m below LAT on a shallow reef flat on the western side of Halfway Island. Halfway Island is 1 of 17 continental islands in the Keppel Island group, in inshore waters of the southern Great Barrier Reef, Australia (23° 11′ 56.7″ S, 150° 58′ 10.1″ E). Field collections were approved by the Great Barrier Reef Marine Park Authority under permit number G12/35236.1. The Keppel Islands represent a relatively small, isolated, inshore fringing reef system with a history of frequent disturbance, principally from flooding from the nearby Fitzroy River (Jones et al. 2011). The island group lies at the interface of the massive Fitzroy River catchment and the southern Great Barrier Reef continental shelf (Ryan et al. 2006). This provided a useful setting to examine endogenous regulation of a reef-building coral’s energetic status in response to environmental and physico-chemical changes. At each sampling point, two replicate, central branches were removed from each colony using a diving knife. Samples were snap frozen in liquid nitrogen within 10 min of collection and stored at −80 °C until analysis.

Lipid analysis

Coral branches were freeze-dried for 96 h (Labconco FreeZone, Kansas City, MO, USA). Due to large organic material deposits in coral skeletons, samples were pre-processed using the in toto crushing method described by Conlan et al. (2017a, b). Briefly, samples were placed inside a stainless steel mortar and pestle (cleaned with methanol), which was placed inside a manual laboratory hydraulic press (Model C, Fred S. Carver Inc., Summit, NJ, USA), and pressurised to 70kN, crushing the corals to a fine powder.

Total lipid content was estimated gravimetrically according to the method described in Conlan et al. (2014). The lipid content was quantified on a microbalance. Total ash was determined by incineration in a muffle furnace (C & L Fetlow, Model WIT, Blackburn, Victoria, Australia) at 450 °C for 12 h. The ash content was subtracted from the total dry weight to obtain ash free dry weight (AFDW), allowing lipid quantification without the calcium carbonate skeleton.

Lipid class analysis was determined using an Iatroscan MK 6 s thin-layer chromatography–flame ionisation detector (Mitsubishi Chemical Medience, Tokyo, Japan) to detect the neutral lipids wax ester, triacylglycerol, free fatty acids and 1,2 diacylglycerol, and the polar lipids sterol, phosphatidylethanolamine, phosphatidylserine/phosphatidylinositol and phosphatidylcholine as per Conlan et al. (2014). Since glycolipids commonly elute with monoacylglycerols and pigments, the term ‘acetone mobile polar lipid’ (AMPL) was used in this study (Parrish et al. 1996). AMPL was quantified using the 1-monopalmitoyl glycerol standard (Sigma-Aldrich Co., USA), which demonstrates an intermediate response between glycoglycerolipids and pigments (Parrish et al. 1996). Following initial lipid extraction, fatty acids were esterified into methyl esters using an acid-catalysed methylation method and analysed by gas chromatography as described in Conlan et al. (2014). The identification of uncommon or obscure fatty acids, such as 10:0, was confirmed via GC-mass spectrometry using a fatty acid methyl ester library.

Environmental parameters

Data for environmental parameters that potentially influence Acropora spp. phenology (Hinrichs et al. 2013a; b) were sourced from AIMS weather stations located within The Keppel Island group and compiled monthly between September 2009 and February 2011 (Australian Institute of Marine Science, AIMS data centre, http://data.aims.gov.au). Water temperature, rainfall and PAR data were obtained from the Square Rocks weather platform (−23° 09′ 86″ S, 150° 88′ 55″ E). Chlorophyll and turbidity data were obtained from AIMS Humpy Island fluorescence and turbidity loggers situated on the reef flat of an adjacent island in the same group (−23° 21′ 63″ S, 150° 96′ 34″ E). Historical discharge data from the Fitzroy River were obtained from the Queensland Government Water Monitoring Information Portal, https://water-monitoring.information.qld.gov.au/host.htm, and compiled monthly between July and October each year from 2005 to 2011. These data provided baseline historical discharge data to allow comparison of the 2010/2011 La Niña event within the context of the past 5 years.

Statistical analysis

Data were analysed using R software version 3.3.3 (R Development Core Team 2018; R Studio Team 2018). Due to inconsistent and nonlinear sampling times, which were necessary to capture the spawning periods, longitudinal statistical analyses were not appropriate to accurately describe the current dataset. Data were also heteroscedastic and non-normal. Therefore, differences between individual biometrics were analysed using a nonparametric Kruskal–Wallis test followed by a post hoc Student’s t test at a significance level of P < 0.05 (de Meniburu 2015). Environmental data were aggregated monthly and correlated against coral samples of the same time period. Correlations were calculated using the Pearson correlation coefficient. Correlation significance was determined by Pearson’s product–moment correlation. All available data were considered and not transformed. Lipid class and fatty acid profiles were visualised with a canonical analysis of principal coordinates (CAP) based on a discriminant analysis (Anderson and Willis 2003) using the CAPdiscrim function (BiodiversityR package) (Kindt and Coe 2005). CAP analyses were performed on untransformed percentage composition data using a nonparametric Bray–Curtis similarity matrix. CAP score plot ellipses show 80% confidence intervals to visualise group separation. Graphs were prepared using the ggplot2 package (Wickham 2016).

Results

Environmental parameters

Chlorophyll concentrations, which represent phytoplankton blooms, peaked twice over the sampling period (Fig. 1a). These occurred in November 2009 and February 2010. The overall trend in chlorophyll over the sampling period correlated positively with both PAR levels and total rainfall (Fig. 2). Between September 2009 and July 2010, PAR levels followed a predictable sinusoidal trend, peaking in spring with the lowest levels recorded mid-winter 2010 (Fig. 1c). After July 2010, PAR levels deviated from this trend into summer 2010, exhibiting substantial and irregular fluctuations. Mean rainfall recorded two major peaks over the sampling period, in February 2010 and December 2010 (Fig. 1d). This was reflected in both Fitzroy River discharge and turbidity levels. Mean discharge peaked in February and March 2010 and again in December 2010 and January 2011 (Fig. 1b). Turbidity levels peaked in February 2010 and January 2011 (Fig. 1e). Water temperature followed a monocyclic, sinusoidal trend over the sampling period, peaking in January 2010 and lowest in August 2010 (Fig. 1f).

Fig. 1
figure 1

Environmental parameters on the Great Barrier Reef from September 2009—February 2011. Values are presented as mean ± SEM

Full size image
Fig. 2
figure 2

Heat map showing the most biologically pertinent correlations between environmental variables and lipid constituents of Acropora millepora from November 2009 to November 2010. Total lipid represents mg g ash free dry weight sample−1. Unless stated otherwise, lipid constituent data represent mg g ash free dry weight lipid−1. Environmental units: chlorophyll = mg g3 −1, PAR = µmol s−1 m−2, rainfall = mm, Fitzroy River discharge = ML day−1, turbidity = NTU, and water temperature =  °C. Values are presented as Pearson correlation coefficients whereby 1 is total positive linear correlation, 0 is no linear correlation, and −1 is total negative linear correlation. Bold values represent statistically significant results (P < 0.05)

Full size image

Substantial differences were recorded between several physico-chemical parameters between the lead up period prior to spawning in 2009 (July–October) and the same period in 2010, reflecting anomalous weather conditions from a La Niña event. Indeed, mean discharge from the Fitzroy River in 2010 was ~ 530% higher from July to October when compared to the average discharge for same period over the past 5 years (2005–2009) (Supplementary Fig. S1). In contrast, average chlorophyll and PAR levels for July–October 2010 were markedly lower compared to the same period in 2009 (Supplementary Figs. 2a and 2b). Average turbidity during this period was also significantly higher in 2010 compared to 2009 (Supplementary Fig. S2c).

Total lipid concentration

The first sampling point occurred on the 2 November 2009, 6 days prior to A. millepora colonies spawning. Total lipid peaked at this time point (Fig. 3), followed by a decrease at the next sampling point, 11 days post-spawn (~ 20% reduction). Following this, no significant differences were detected in total lipids at each sequential sampling point, except for the November 2010 post-spawn. Thus, excepting spawning events, total lipids remained relatively stable throughout the year. At the October 2010 pre-spawn, the accumulated lipids were significantly lower than the November 2009 pre-spawn point. Despite this, total lipid reduction post-spawn in 2010 was ~ 20% greater than in 2009. The subsequent total lipid concentration at the November 2010 post-spawn time point was equivalent to the lowest concentration for the sampling period, which occurred in May 2010.

Fig. 3
figure 3

Total lipid concentration and proportion of storage and structural lipids of Acropora millepora from November 2009 to November 2010. Values are presented as mean ± SEM. Values in the same lipid group that do not share the same letters are significantly different (Student’s t-test, P < 0.05)

Full size image

Lipid class composition

Storage lipids decreased significantly from the first sampling point (November 2009 pre-spawn) to the second (November 2009 post-spawn) (Fig. 3). This was attributable to major reductions in the storage lipids, wax ester and triacylglycerol, with the latter being statistically significant (~ 44% reduction) (Supplementary Table S1). Phosphatidylcholine was the only structural lipid class reduced over the 2009 spawning (~ 20% reduction).

Despite the total lipid reduction in January, triacylglycerol increased at this point (~ 33% increase) (Supplementary Table S1). Indeed, triacylglycerol concentrations in January 2010 were the second highest across the sampling period. Triacylglycerol then decreased significantly in March, reaching its lowest point in May 2010. However, from March onwards, triacylglycerol, along with wax ester, sterol and phosphatidylethanolamine, did not record significant fluctuations in sequential time points, corresponding to the cyclical pattern of the total lipid composition described above.

Generally, structural lipids followed an inverse trend to triacylglycerol (Supplementary Table S1). Structural lipids were the predominant classes during the winter period, especially AMPL, which was second to triacylglycerol as the most prevalent lipid class across the sampling period.

Despite exhibiting relative stability from January to July 2010, free fatty acids recorded two significant spikes over the sampling period. The first occurred in late November 2009 and the second in September 2010 (Supplementary Table S1). Both peaks were ~ threefold greater than the next highest concentration recorded over the sampling period. Otherwise, free fatty acids were relatively stable between January and July 2010.

The CAP score plot (Fig. 4a) illustrated time point separation for the lipid class profile. There was a strong separation in the lipid class profile between the November 2009 pre-spawn and all other time points, attributable to elevated wax ester and triacylglycerol (Fig. 4b). Additionally, the November 2009 post-spawn and September 2010 time points showed strong separation compared to all other time points, yet were strongly correlated with each other, being influenced by free fatty acids and sterol. Consistent with the individual lipid class results, the CAP analysis found little separation between the remaining time points, which were influenced by structural lipids.

Fig. 4
figure 4

Canonical analysis of principal coordinates (CAP) of total lipid class profile (% lipid) of Acropora millepora from November 2009–November 2010, a sample scores and b vectors representing the top ten lipid class contributors to variance in lipid composition. WAX = wax ester, TAG = triacylglycerol, FFA = free fatty acids, AMPL = acetone mobile polar lipids, 1,2DAG = 1,2 diacylglycerol, PSPI = phosphatidylserine/phosphatidylinositol, PE = phosphatidylethanolamine, PC = phosphatidylcholine

Full size image

Fatty acid composition

The highest total fatty acid concentration of the entire sampling period occurred prior to spawning in November 2009 (Fig. 5). Saturated and monounsaturated fatty acids recorded high concentrations at this time point, particularly 10:0, 14:0 and 16:0, and 16:1n-7, 18:1n-9 and 20:1n-9 (Supplementary Table S2), as illustrated by the CAP analysis (Fig. 6). Most of these fatty acids, alongside most individual polyunsaturated fatty acids, decreased significantly after the November 2009 spawning.

Fig. 5
figure 5

Total fatty acid concentration and major fatty acid classes of Acropora millepora from November 2009 to November 2010. Values are presented as mean ± SEM. Values in the same fatty acid (FA) groups that do not share the same letters are significantly different (Student’s t-test, P < 0.05). SFA = saturated fatty acids, MUFA = monounsaturated fatty acids, PUFA = polyunsaturated fatty acids

Full size image
Fig. 6
figure 6

Canonical analysis of principal coordinates (CAP) showing overall fatty acid profile (% fatty acids) of Acropora millepora from November 2009 to November 2010 a sample score plot and b fatty acid vectors representing the top ten contributors to variance

Full size image

Notably, 10:0 was the only fatty acid completely depleted after the 2009 November spawning, despite comparatively high concentrations pre-spawn (Supplementary Table S2), and did not exceed 1 mg g AFDW−1 until September 2010. Only two fatty acids increased quantitatively following the November 2009 spawn, namely 18:0 (+27%) and 22:6n-3 (DHA) (+18% increase). In most cases, changes in the total fatty acids corresponded to changes in triacylglycerol (Fig. 2).

Total fatty acids also exhibited a significant, positive correlation with water temperature, largely due to saturated fatty acids. On the other hand, monounsaturated fatty acids and polyunsaturated fatty acids correlated negatively with water temperature. As such, the highest saturated fatty acids proportions occurred between November 2009 and January 2010, while monounsaturated fatty acids were highest in September 2010, and polyunsaturated fatty acids were highest from March to July 2010 (Supplementary Table S3).

The CAP showed a strong separation between the 2010 October pre-spawn and all other time points (Fig. 6a). Like the November 2009 pre-spawn, this was driven by the saturated fatty acids, 10:0 and 14:0, and the monounsaturated fatty acids, 16:1n-7 and 18:1n-9 (Fig. 6b). Separation January, March and May 2010 was influenced by 18:0, 18:4n-3 and 20:5n-3.

Overall, fatty acid 16:0 was the major contributor to the significant variation between sample points (Fig. 6b), reflecting its quantitative dominance in all samples (Supplementary Table S3).

Discussion

Understanding biochemical changes in scleractinian corals and their respective physico-chemical drivers is fundamental to identifying the key factors influencing present-day coral health. This information helps to identify deviations from ‘baseline’ coral health indices, which will ultimately assist with coral reef management, monitoring programs and conservation efforts (Oku et al. 2003a; Hoegh-Guldberg et al. 2007; Dodds et al. 2009; Hinrichs et al. 2013a).

As demonstrated in numerous taxa (Dridi et al. 2007; Urzúa et al. 2011; Connelly et al. 2015), this study identified significant seasonal trends in the total lipid, lipid class and fatty acid composition of the scleractinian coral, A. millepora. Total lipids were highest at the first of nine sampling points ~ one week prior to the November mass spawning of corals on the GBR. This result is typical of gravid, broadcast spawning corals, which produce lipid-rich eggs that develop into lecithotrophic larvae (Baird et al. 2009; Figueiredo et al. 2012; Viladrich et al. 2016). Indeed, the lipid class composition at this point was dominated by wax ester, triacylglycerol and phosphatidylcholine, which are major constituents of coral eggs and larvae (Arai et al. 1993; Figueiredo et al. 2012). Phosphatidylcholine is a major cell membrane constituent and is prevalent in marine eggs due to its importance in larval development and growth (Coutteau et al. 1997; Tocher et al. 2008). Meanwhile, wax ester and triacylglycerol constitute a fatty acid reserve that provides energy and positive buoyancy in long-distance larval dispersal (Richmond 1987; Harii et al. 2007; Figueiredo et al. 2012). The fatty acid groups, saturated fatty acids and monounsaturated fatty acids, which were also elevated before spawning, provide a major energy source during larval development, supporting tissue formation and providing energy for locomotion (Kamler 2008; Parma et al. 2015; Hauville et al. 2016). The predominant individual saturated fatty acids, 10:0, 14:0, and 16:0, and monounsaturated fatty acids, 16:1n-7, 18:1n-9, and 20:1n-9, are known to be readily catabolised and are thus heavily relied upon for energy in lecithotrophic larvae (Jeffs et al. 2002).

Acropora millepora generally broadcast spawn between late October and early December. This often coincides with nutrient levels in the southern GBR being enhanced by the East Australian Current and Capricorn Eddy (Hatcher 1990; Weeks et al. 2010). As egg lipid profiles are directly affected by maternal provisioning (Sargent et al. 1999), and given the 2009 spawning period coincided with peak chlorophyll concentrations, spring microbiota blooms likely provided a food source for lipid stockpiling prior to spawning. Indeed, the pre-spawn coral samples showed elevated levels of the phytoplankton biomarkers, 18:4n-3 and 18:2n-6 (Dalsgaard et al. 2003). Further, since elevated PAR levels during this period would also promote increased coral phototrophy, these fatty acids likely represent symbiotic zooxanthellae activity as well (Latyshev et al. 1991).

Total lipid decreased at the next sampling point, ~ 11 days after the 2009 November spawning. Predictably, this was due to reduced wax ester, triacylglycerol and phosphatidylcholine–signifying the release of reproductive materials (egg bundles) (Oku et al. 2003a; Connelly et al. 2015; Viladrich et al. 2016). Interestingly, free fatty acids peaked after spawning. Free fatty acids are released when fatty acid chains are cleaved from their molecular backbones, and represent an immediate energy source (Sargent et al. 1999; Tocher 2003; Imbs 2013; Cipriano et al. 2015). This suggests a colony mechanism to quickly overcome reproductive stress (Rossi et al. 2006; Saito et al. 2015; Viladrich et al. 2016).

In contrast to most fatty acids, 18:0 and DHA increased after the 2009 spawn, signifying selective retention of these fatty acids and conforming to previous suggestions of their negligible roles in coral eggs and larvae (Arai et al. 1993; Figueiredo et al. 2012; Conlan et al. 2017a, b). Low DHA concentrations in coral eggs may reflect the more complex oxidation process required for DHA catabolism compared to EPA (which was significantly reduced post-spawn) (Tocher 2003; Laurel et al. 2012). Moreover, conserving certain fatty acids from reproductive output also reflects a biochemical strategy of the parent colony to preserve indispensable membrane components during stress (Izquierdo et al. 1996).

Notably, 10:0 was the only fatty acid completely depleted after spawning, despite comparatively high concentrations pre-spawn. Medium-chain fatty acids, such as 10:0, are readily utilised and have been shown to constitute an important and easily accessible energy source in larval fish (Fontagne et al. 2000; Hillestad et al. 2011). Importantly, 10:0 did not exceed 1 mg g AFDW−1 until September 2010, indicating a negligible role in adult coral metabolism, despite an apparently major role in larval lecithotrophy, where it has been observed previously (Conlan et al. 2017a, b).

Total lipids continued to decline in January, possibly reflecting surplus lipid catabolism for growth, which has been shown to be highest in spring/summer for Acropora spp. (Titlyanov et al. 2006). However, despite this reduction, triacylglycerol and total fatty acids both increased. This may reflect heterotrophic feeding on microbiota, particularly phytoplankton (Ojea et al. 2004), since chlorophyll continued to be high through the summer period. The timing of spawning around seasonal phytoplankton blooms allows polytrophic corals to feed heterotrophically pre- and post-spawning to boost energy stores, while also ensuring a food source for newly settled recruits (Brooke and Järnegren 2013).

Notably, PAR levels showed a strong, positive correlation with triacylglycerol. This relationship likely reflects energy input derived from both endogenous supply through phototrophy (Oku et al. 2003b) and exogenous supply through heterotrophic feeding on phytoplankton, since PAR levels are an important determinant of phytoplankton blooms (Venables and Moore 2010). The latter is supported by the significant correlation of PAR levels with chlorophyll.

Increased total fatty acids in January 2010 were largely due to saturated fatty acids, which demonstrated positive correlations with water temperature and chlorophyll. The correlation between saturated fatty acids and water temperature likely reflects adaptive regulation of cellular lipid melting points to maintain optimal membrane fluidity during the warmer months in both corals and their prey, although this requires further examination (Lin et al. 2013). This would be necessary for efficient diffusion of lipophilic compounds, appropriate lipid molecule geometry and maintaining membrane-bound enzyme activity (Filimonova et al. 2016). Correspondingly, polyunsaturated fatty acids correlated negatively with water temperature, being lowest during January, and highest from March to July.

Between February and March 2010, high rainfall and widespread flooding was recorded in The Keppel Island region (Thompson 2011), and this coincided with markedly increased discharge from the Fitzroy river, one of the two turbidity peaks during the sampling period, and a second chlorophyll peak. This event also coincided with a shift in the coral fatty acid profile, as illustrated by the CAP, which showed strong separation of the fatty acid profiles between January and March 2010 (Fig. 6a). This may reflect a shift in the nature of available exogenous food sources. Indeed, the high rainfall and resulting Fitzroy River discharge in February coincided with a sharp increase in turbidity, implicating suspended particulate matter (SPM) as a potential food source during this period (Anthony 1999). Concurrently, seasonal variations in physico-chemical factors influence the lipid profile of plankton (i.e. fatty acid composition), and different lipid classes often contain different fatty acid moieties within a phytoplankton taxon (Connelly et al. 2015 and references therein). Thus, the second chlorophyll peak in February 2010 may represent a microbiota bloom of different species.

Comparatively low chlorophyll was recorded during the cooler period from March to July 2010. This coincided with further decreases in coral total lipids, storage lipids and total fatty acids, likely indicating reduced energy intake from both phototrophic and heterotrophic sources. Concordantly, structural lipid proportions increased over this period. Indeed, changes to structural lipid classes appeared to be more closely related to changes in storage lipids, particularly triacylglycerol, suggesting limited endogenous modification to membrane lipids in response to seasonal change (Ahn 2000).

This study suggests A. millepora spawned marginally earlier in 2010 compared to 2009. This was evidenced by a peak accumulation of reproductive biomarkers, particularly wax ester, triacylglycerol and 10:0 in October 2010, followed by marked reductions in mid-November. Presumably, this spawn coincided with the full moon which occurred on the 23rd/24th of October, 2010 (MoonConnection.com 2018). Notably, lipid accumulation in October 2010 did not reach the concentrations recorded prior to the November 2009 spawn, likely due to poor exogenous food availability in the lead up to this reproduction event. Indeed, prior to the 2010 spawn, chlorophyll and PAR were far lower than prior to the 2009 spawn, indicating both phototrophic and heterotrophic feeding modes were limited. Nevertheless, the total lipid reduction post-spawn was ~ 20% greater than in 2009, suggesting greater resource allocation to eggs. Previous reports have shown resource limitation and stress may cause increased reproductive investment, risking colony death due to the depletion of vital nutritional reserves necessary to support metabolic functioning (Grottoli et al. 2004; Leuzinger et al. 2012). Subsequently, total lipids in November 2010 were equal to the lowest concentration recorded across the sampling period.

The 2010–2011 spring and summer recorded extreme weather conditions attributable to an exceptionally strong La Niña event and early arrival of the north Australian monsoon (BOM 2011; Abdul Wahab et al. 2014). This study suggests A. millepora colonies entered the 2010–2011 La Niña event with low energy reserves to allocate to stress resistance, which would be necessary to combat the major environmental anomalies during the 2010–2011 summer (BOM 2010, 2011). For instance, in the GBR, the La Niña year saw an increase of up to 933% in total monthly rainfall between July and December 2010 (Abdul Wahab et al. 2014). High dissolved inorganic nutrients and low salinity elicited by high rainfall and terrestrial runoff are related to metabolic stress in corals (Fabricius 2005; Scott et al. 2013). Additionally, whereas the high rainfall and turbidity in February–March coincided with a major chlorophyll peak, which may have buffered the adverse physico-chemical conditions (Hinrichs et al. 2013a), this was notably absent during the spring rainfall, and ~ 38% lower than the same period in 2009.

Thus, the high turbidity in the latter half of 2010 was likely caused by SPM, due to increased discharge from the Fitzroy River. Although some corals may feed on SPM, relative to plankton, it is a poor-quality food source (Anthony and Fabricius 2000; Veron 2000). Furthermore, high SPM concentrations can decrease light availability in reef waters, reducing photosynthetic yield in zooxanthellae (Fabricius 2005). In addition, PAR levels were ~ 20% lower during this period in 2010 compared to the same time in 2009. Low light levels have been shown as detrimental to coral lipid synthesis, particularly wax ester and triacylglycerol (Oku et al. 2003b; Hinrichs et al. 2013a).

It is thus somewhat unsurprising that reduced coral cover was recorded at Halfway Island in 2010 (Thompson 2017). Moreover, coral spat settlement to settlement tiles in late 2010 was lower than previously recorded at all reefs in this region (Davidson et al. 2019)–suggesting a poor reproductive outcome in 2010, despite the apparently greater maternal resource investment.

Extreme weather events are common on the GBR and likely to increase in frequency and intensity, with climate change projections predicting an increase in water temperature, storm severity and rainfall (Solomon et al. 2007; Abdul Wahab et al. 2014). Establishing baseline indices into the key physico-chemical drivers of coral health is therefore necessary to enable recognition of potential periods of coral vulnerability and inform policies on future anthropogenic disturbances in coral reef ecosystems.