Introduction

Coral reefs provide critical ecosystem services to millions of people worldwide but are under threat due to multiple anthropogenic stressors acting at a local to global level (Hughes et al. 2017). Recently, turbid nearshore reefs have shown resilience in the face of increasing thermal stress and have been proposed as important ecological refugia (Cacciapaglia and van Woesik 2016; Zweifler et al. 2021). Although traditionally perceived as ‘sub-optimal’ (Kleypas et al. 1999), these reefs sustain high live coral cover and typically show resilience to bleaching during elevated temperatures as suspended particles protect corals from damaging high irradiance (Perry and Larcombe 2003; Waheed and Hoeksema 2013; Browne et al. 2019; Sully and van Woesik 2020). Moreover, recent evidence from the fossil record shows early diversification of ancient coral communities occurred in turbid reefs, with the persistence of high diversity lasting over geologic timescales (Santodomingo et al. 2016). Yet, responses to turbid reefs are confounding among coral taxa (Bainbridge et al. 2018). Although turbidity can cause slower growth (Risk and Sammarco 1991) or loss of photosynthetic endosymbionts (Dallmeyer et al. 1982), some mixotrophic species adapt by utilising resources to offset stress from particle loads (Anthony 2000; Fox et al. 2018). However, aside from corals, little is known about how other key reef builders calcify in turbid reefs. A robust understanding of different organisms and functional groups is required because individual responses result in ecosystem-wide effects (Bainbridge et al. 2018).

Giant clams (Cardiidae: Tridacninae) are large and long-lived symbiont bearing bivalves, distributed throughout the Indo-Pacific (Rosewater 1965). They fulfil critical roles in tropical reef communities due to their contributions to the structure, complexity and ecology of reefs (Neo et al. 2015). Their rapid (daily increment widths of 5–200 µm) and continuous shell growth (Ma et al. 2020) has long been attributed to the photosynthetic rate of their endosymbiotic zooxanthellae (Klumpp and Griffiths 1994). Therefore, changes in growth are expected to be sensitive to external environmental conditions that directly or indirectly alter light exposure (Sano et al. 2012). In some species, translocated photosynthates from endosymbionts satisfy all daily metabolic energy requirements (Jantzen et al. 2008). However, tridacnids are mixotrophs and like many bivalves, take up particulate organic matter (POM) (Klumpp et al. 1992) through filter-feeding. The balance of autotrophy and heterotrophy to total energy requirements changes depending on inter- and intraspecific differences in size, photosynthetic potential (Jantzen et al. 2008), and ontogeny (Klumpp and Griffiths 1994). Compared to most tridacnids, Tridacna squamosa (fluted giant clam) is a ‘true mixotrophic species’ due to its reduced photosynthetic activity, showing a reliance on a heterotrophic strategy to satisfy its metabolic needs within its range (Tedengren et al. 2000; Jantzen et al. 2008). Indeed, the survival of T. squamosa may be comparable or even higher under turbidity relative to clearer water (Guest et al. 2008; Yong et al. 2022). However, how growth and biomineralization processes of T. squamosa vary between turbid and clear waters and external mechanisms controlling these processes are not understood. Increasing understanding of how giant clams respond to local stressors is required for conservation strategies under continued environmental change (Watson and Neo 2021).

Petrographic microscopy, scanning electron microscopy (SEM) and laser scanning confocal microscopy (LSCM) methods have revealed daily growth increments in both the inner- (Watanabe et al. 2004; Aubert et al. 2009; Elliot et al. 2009; Schwartzman et al. 2011; Sano et al. 2012; Ayling et al. 2015; Hori et al. 2015; Warter et al. 2015; Arias-Ruiz et al. 2017; Gannon et al. 2017; Warter and Müller 2017; Ma et al. 2020; Yan et al. 2020; Yan et al. 2021; Zhao et al. 2021; Liu et al. 2022) and outer layers of giant clam shells (Duprey et al. 2015; Komogoe et al. 2018; Killam et al. 2021). Shell growth chronologies constructed from these daily increments are important alongside powerful geochemical techniques (e.g. LA-ICP-MS) in reconstructing highly time-resolved profiles of water temperature (e.g. Arias-Ruiz et al. 2017), productivity (e.g. Elliot et al. 2009), paleoweather (Komagoe et al. 2018; Yan et al. 2020) and diurnal light cycles (Sano et al. 2012), among others. The refinement of identification of growth increments is important in different environments and ensures accuracy of time-series analysis (Warter and Müller 2017). Yet, the environmental controls on variation in daily growth increments are little understood and the potential of tridacnid shell growth chronologies on their own remains largely untapped (Zhao et al. 2021).

In this work, we investigate the growth rate of T. squamosa in turbid and clear reefs on annual, seasonal and daily scales. We focus on measuring daily increment widths and investigate the microstructure of increments in shells collected alive from two contrasting reef sites—a turbid reef near a river outlet and a less turbid reef. We select the Coral Triangle region of northeast Borneo (Sabah, Malaysia) as our study site because turbid reefs occupy about 30% of reefs within the region (Sully and van Woesik 2020). To disentangle potential external environmental drivers of growth between the less turbid and turbid reef, we compare shell growth chronologies and their spectral characteristics with in situ and remote sensing environmental data. We further investigate growth across a gradient of turbidity by exploring growth trends in shells collected dead at four additional reef sites.

Materials and methods

Regional setting

Darvel Bay (4° 5356′ N, 118° 2646′ E) is the largest semi-enclosed bay on the eastern coast of Sabah and is connected to the Pacific Ocean through the Sulu-Celebes Sea (Fig. 1). It is within the northern corner of the Coral Triangle, a region noted for its high levels of marine biodiversity and endemism (DeSilva et al. 1999; Veron 2000). Coral reefs here are dominated by small patch and fringing reefs developed around the coastline and numerous islands (Ditlev et al. 1999). Mangrove ecosystems are associated with small freshwater catchment areas and prevail near Sakar Island and around the estuaries of the Silabukan and Tingkayu rivers, which play important roles in carrying wastewater discharge and land-based pollutants into the coastal area (Saleh et al. 2007). Climatic conditions are controlled by the Indo-Australian monsoon system, divided into the southwest monsoon between May to September (dry season) and northeast monsoon from November to March (wet season) (Saleh et al. 2007). The tidal range is mesotidal (between 2 and 4 m) (Santodomingo et al. 2021) and strong tidal currents control water circulation and runoff (Saleh et al. 2007).

Fig. 1
figure 1

False colour composite map (bands 7/4/2) of sample sites in Darvel Bay (wet season). Enlarged points correspond to key localities Baik and Triangle, where Tridacna squamosa were collected alive. River plume originating from the Tingkayu river in bright blue

Full size image

Study sites and samples

We carried out an in-depth comparison of daily growth rate for T. squamosa at two key reef sites in Darvel Bay—Baik and Triangle (Fig. 1). Baik is proximal to a fish farm, and activities at the site include recreational diving. It has low sediment input and is indicative of a less turbid reef. Triangle is approximately 3.5 km from the mouth of the Tingkayu River, which discharges sediment to the site at low tide. Four shells were extracted alive at Baik and Triangle in April 2019 (SSbaik and SSct, respectively) and February 2020 (NS207 and ZW156, respectively) between depths of 5 and 8.5 m (Tables 1, S1).

Table 1 Coral reef sampling sites in Darvel Bay (East Sabah, Malaysia)
Full size table

In addition, eight shells were collected dead from the seafloor at Baik (SS01BBT), Triangle (SS03BCT, SS03BCT) and four additional locations across a gradient of turbidity (Sakar (SS02ASN), Misan (SS01ASM), THalu (SS01ATH) and Blue Lagoon (SS01ABH)) (Fig. 1; Tables 1, S1). Although these shells appeared to record in situ environmental change and bioerosion assessment of shells collected from specific environments showed similar features, we highlight the caveat that they may have been translocated by local fishermen or wave action post-mortem. Due to these uncertainties, shells collected dead were used only for comparison of overall growth and not employed for further analysis to explore the relationship between growth and environmental variables. To broaden the comparison to a wider region, another two shells from Semporna reefs (SEMA, SEMB) situated south of Darvel Bay were obtained from collections at Amgueddfa Cymru—National Museum Wales (NMW). It is unknown if the shells from Semporna were collected dead or alive.

In situ measurements of water clarity at each site were estimated with a turbidity meter (Nephelometric Turbidity Units: NTU) and Secchi disc in the dry season of 2019, yet prolonged drought experienced throughout southeast Sabah during early 2019 (Payus et al. 2020) meant that measurements were atypically low. Therefore, we define turbidity herein based on the amount of suspended particles using the diffuse attenuation coefficient of spectral irradiance at 490 nm wavelength Kd(490) (Table 1). This choice is supported by studies such as Loila et al. (2019), who used remote sensed Kd(490) values as robust indicators of differences in turbidity on coral reefs. We define low turbidity as mean annual Kd(490) < 0.2 m−1 and the turbidity threshold as mean annual Kd(490) > 0.2 m−1 (e.g. Yu et al. 2016). We interpret this parameter alongside remote sensed chlorophyll-a concentrations, shown to be controlled by river runoff in the wet season (Chen et al. 2007).

All shells were cut into ~ 1–2-cm-thick slices along the axis of maximum growth (transversal from umbo to upper shell margin) (Figs. 2a, b) with either a HC Evans and Son (Eltham) LTD (250 mm blade, 1 mm thickness) or Logitech GTS1 Thin Section Cut-Off diamond saw. Thin sections (60 µm thickness) of the inner layer cut perpendicular to the direction of growth were prepared from slices, ground with 1000 grit sandpaper, and polished with 0.3 µm alumina oxide (Fig. 2c). To improve visibility of features, sections were etched with 0.5% HCL for 15 s. Prior to analysis, dead collected shells were investigated for diagenesis with Raman Spectroscopy and SEM. Evidence of original aragonite was found in all shells and alteration to calcite not detected (Supplemental for more information; Figs. S1, S2). Lengths of shells ranged between 98.59 to 362.62 cm and heights 67.27 to 215.08 cm (umbo to margin).

Fig. 2
figure 2

a Valve of Tridacna squamosa with red vertical line indicating section location. b Transversal section from umbo to upper shell margin highlighting the inner layer (IL), outer layer (OL), pallial line (PL) and hinge (H). c Thin section of the inner shell layer denoting first growth (juvenile growth, J) and last growth (adult growth, A). d Daily growth increments under transmitted light microscopy (20 x magnification) from first (J) and last growth (A)

Full size image

Environmental data

We used in situ and satellite remote sensed data to characterize environmental factors for the last year of growth (LYOG) in shells collected alive from Baik and Triangle (Table S2). Sea surface temperature (SST) and light intensity (lux) were characterized with a HOBO temperature and light logger and data collected every 10–15 min between 2019 and 2020. Tide data were sourced from the Sea Level Monitoring Facility (www.ioc-sealevelmonitoring.org/bgraph.php?code=ms006&output=tab&period=30&endtime=2022-01-31-) at Lahad Datu (stn. ms006). Cloud cover, salinity, chlorophyll-a, Kd(490), rainfall and additional SST data for 2018–2019 were sourced from Google Earth Engine (GEE) or the National Oceanic and Atmospheric Administration (NOAA) database (Table S2).

Shell growth chronologies

Shell growth chronologies were assembled from the daily increments of thirteen T. squamosa shells to determine daily growth rate. We used a mixed-method approach to image increments with a light microscope (Leica DMR) (10–40 × magnification) and a Zeiss Sigma HD field emission gun SEM (1500 × magnification) at the School of Earth and Environmental Sciences, Cardiff University. For SEM, sections were coated with 20 nm gold–palladium alloy using a BIO-RAD SC500 sputter coater and growth increments and their microstructure observed at high vacuum with an acceleration voltage 10 kV and aperture size 30 µm. Images were obtained along the height of the inner layer and montaged on Microsoft Image Composite Editor (ICE) (light microscope) or Oxford Instruments AZtec 6.0 software (SEM).

The number and width of daily increments were determined from light microscopy and SEM images using ImageJ 1.53 (Schneider et al. 2012) and Coorecorder 9.8 (Cybis) (Maxwell and Larsson 2021) software. By using sets of images from both methods and software, multiple counts were obtained for each shell (Table S3). To reduce sampling bias when comparing shells collected alive with environmental data, a second person also counted sets of increments from these shells. Final mean shell growth chronologies were compiled from multiple counts that met certain criteria, which was based on how many increments could be defined (Supplemental for more information; Table S4; Figs. S3, S4).

Data analysis

Data analysis was carried out in RStudio Version 4.0.3 (R Core Team 2020) unless otherwise specified. The R code is available from the Mendeley Data Repository at https://doi.org/10.17632/mbjzc2nbsn.1. We used a dimensionless standardized growth index (SGI) to control for ontogenetic decelerated growth rates (e.g. Jones et al. 1986). Often used in studies of bivalves (e.g. Butler et al. 2013), including giant clams (Zhao et al. 2021), the SGI is an estimation of how growth deviates from the average growth trend (see Schöne 2003). We applied the SGI to our data as ages of samples likely covered life stages from juvenile to adult. We investigated SGI values on daily, seasonal and annual scales. To examine differences in annual growth, a one-way analysis of variance (ANOVA) and post hoc Tukey test were performed between shells pooled into the seven reefs. Assumptions of ANOVA were checked for normality and homogeneity of variance using Shapiro–Wilk's test, QQ (quantile–quantile) plots of standardized residuals and Bartlett's test. On a monthly scale, SGI values were plotted with monthly averaged environmental data for the LYOG in the shells collected alive at Baik and Triangle. These data were explored as added-variable (AV) plots because they control for the variability of other explanatory variables when plotting the effect of x on y. They therefore more accurately show relationships because other variables in the model are adjusted for (Draper et al. 1966). Principal component analysis (PCA) was also performed on environmental data for time periods corresponding to LYOG to explore annual and seasonal relationships, allowing the number of variables under investigation to be reduced. Finally, we used spectral analysis to investigate daily growth with tidal cycles in frequency space in Baik and Triangle because tidal patterns are thought to be expressed in the daily increments of T. squamosa (Evans 1972). Multi-taper method spectral analysis (MTM) (Thompson 1982) was carried out in K-Spectra v.3.9.3 (SpectraWorks) to extract dominant frequencies. Cycles were analysed with significance of spectral peaks at 95–99% compared to red noise background.

Results

Environmental data

Mean SST for 2019–2020 at Baik and Triangle was 29.53 °C and 30.01 °C, respectively (Fig. 3a). Throughout the year, monthly SST varied between 28.11 °C and 30.47 °C at Baik and 28.93 °C and 30.91 °C at Triangle, peaking at the end of the dry season and generally dropping towards the end of the wet season. Light intensity (lux) corresponded with SST in Baik and showed a peak in October, varying between monthly lux of 7122.50 and 16,444.75 throughout the year compared to 7285.83 lux and 13,705.91 lux for Triangle (Fig. 3b). However, it is important to consider that we could not retrieve data for lux between August and October for Triangle, as the HOBO logger was lost. Precipitation showed a bimodal distribution pattern, with maximum rainfall towards the beginning of the dry- and mid-wet season in both sites (Fig. 3f). Chlorophyll-a (Fig. 3d) generally corresponded to Kd(490) (Fig. 3c) in both sites, but maximum monthly values were higher at Triangle (4.19 mg m−3, 0.51 m−1, respectively) than Baik (1.06 mg m−3, 0.14 m−1, respectively). Cloud cover was similar between sites (Fig. 3e), peaking in June and dropping during the late wet season in March. Although patterns were similar in 2018–2019, prolonged drought in early 2019 revealed atypically low cloud cover (51%, both sites) and rainfall (62.98 mm Baik; 43.35 mm Triangle) in February 2019. Sea surface salinity (SSS) showed a partial inverse behaviour with rainfall (Fig. S5).

Fig. 3
figure 3

a Monthly average sea surface temperature (SST), b light (lux), c Kd(490) (m−1), d chlorophyll-a (mg m−3), e cloud cover (%) and f rainfall (mm) from Baik and Triangle reefs in Darvel Bay for 2018–2020. Environmental data shown relate to last year of growth (LYOG) for samples collected live in April 2019 and February 2020

Full size image

We used PCA to further explore relationships between environmental variables at Baik and Triangle. PCA defined two principal components covering 77.1% of the cumulative variance (Fig. 4). PC1 accounted for 43.9% of variance, with highest negative association of SST and cloud cover and highest positive association of salinity and rainfall, while PC2 accounted for 33.2% of variance and showed highest negative association of chlorophyll-a and Kd(490) and positive association of cloud cover and SST. Baik and Triangle generally clustered into two groups regardless of season. However, two points that represented the dry season of 2019, when anomalous drought was experienced, were positioned in the top left quadrant and did not cluster with others.

Fig. 4
figure 4

Principal component analysis (PCA) for environmental data: monthly average sea surface temperature (SST), Kd(490) (m−1), chlorophyll-a (mg m−3), cloud cover (%) and rainfall (mm) from Baik and Triangle reefs in Darvel Bay, showing the loading of each variable (black arrows) and PCA scores (points) of each season for 2018–2020. Superimposed 95% confidence ellipsoids contain group points. Point sizes represent quality of representation of each individual point

Full size image

Growth increment microstructure

The microstructure of daily growth increments in the inner shell layer varied among the thirteen T. squamosa shells depending on collection site. Two distinct microstructures of daily increments were noted: Type 1—Paired growth couplets, consisting of a thicker- and thinner layer, delineated by crystalline structure, similar to the thicker prismatic layer and thinner layer with oblique crystals described by Gannon et al. (2017) (Fig. 5a, c); Type 2—Two adjacent organic rich growth lines delineating a band and intersecting an irregular or cone complex crossed lamellar microstructure (Taylor et al. 1969) (Fig. 5b, d). Type 1 was generally found in shells from the Triangle reef, while Type 2 was common in Baik, Blue Lagoon, THalu, Misan, Sakar and Semporna. Despite these differences, overall observations of the outer shell layer showed an identical microstructure in every specimen that consisted of a crossed lamellar arrangement (Taylor et al. 1969).

Fig. 5
figure 5

SEM images of the microstructure of paired daily increments in Tridacna squamosa from Triangle reef a, c, consisting of a simple prismatic layer (Pr) and layer with smaller crystals (s). b, d In Baik shells, there is a complex crossed lamellar microstructural arrangement with faint growth lines (GL) running perpendicular to first order lamellae

Full size image

The microstructure observed in Type 1 showed a stronger delineation than Type 2. Thus, it was easier to observe, measure and count growth increments and in turn, data from shells with Type 1 had lower errors (calculated by value 1 subtracted from value 2, divided by value 1) between individual counts (0.74–2.37%). In contrast, the recorded variability in counting and measuring the least visible increments of Type 2 demonstrates that these increments were harder to distinguish. A comparison of the relationship between days alive and daily growth (µm) for individual counts that made up final mean growth chronologies of live collected Tridacna squamosa shells SSbaik (Baik, 2019 collected), SSct (Triangle, 2019 collected), NS207 (Baik, 2020 collected) and ZW156 (Triangle, 2020 collected) is shown in Fig. 6. For instance, the SSbaik shell showed a very large offset of 880 increments between petrographic and SEM approaches (error 44.88%). However, data were similar between observer counts in the juvenile region of growth (Fig. 6a).

Fig. 6
figure 6

Comparison of relationship between days alive and daily growth (µm) for individual counts of live collected Tridacna squamosa shells SSbaik (Baik, 2019 collected) a, SSct (Triangle, 2019 collected) b, NS207 (Baik, 2020 collected) c, ZW156 (Triangle, 2020 collected) d PM-LM = primary counter, measurement tool ImageJ, light microscopy images; PM-SEM = primary counter, measurement tool ImageJ, SEM images; SM-LM = secondary counter, measurement tool ImageJ, light microscopy images

Full size image

Annual shell growth

Lifespan of all shells was estimated between 0.72 (262 days) and 7.23 (1726 days) years (Table S4). Seven shells were under 3 years, indicating the juvenile life stage, while six were over 3 years, which may indicate adulthood (Lucas 1994). Shell growth chronologies over 3 years generally showed a concave-down shape and could be divided into two broad stages: Stage 1—Rapid acceleration at start of life and Stage 2—Deceleration in or after 3 years (Fig. 7). Younger shells that presumably had not reached latter growth stages showed no concavity but continued rapid growth or no relationship with days alive.

Fig. 7
figure 7

All mean shell growth chronologies derived from daily growth increment widths of live (a, c) and dead (b, d) collected Tridacna squamosa for comparison of shell growth over lifespan. Measurements are presented by rate of daily growth (μm) (a, b) and detrended growth rates shown in the dimensionless unit of the standardized growth index (SGI) (c, d). Lines are smoothed (span 0.1) and colours represent sample. Shells younger than 365 days are excluded from SGI values

Full size image

Mean shell growth rates over lifespan varied greatly and ranged from 3.91 ± 1.39 to 8.13 ± 3.37 mm/yr (Table S4), corresponding to daily increment widths between 2.02 and 41.40 µm (mean 14.90 µm/day). In all shells, mean annual growth in the first year of life was 6.28 mm/yr (n = 11), 5.51 mm/yr (n = 7) in 0–2 yr, 5.18 mm/yr (n = 2) in 2–4 yr, 3.58 mm/yr (n = 1) in 4–6 yr, following age-related deceleration of growth (Fig. 7). Shells from Semporna (n = 2) showed pooled fastest overall mean annual growth (7.46 mm/yr), followed by those from Triangle (n = 4) (6.16 mm/yr), Misan 5.41 mm/yr (n = 1), Baik 4.61 mm/yr (n = 3), Sakar 4.26 mm/yr (n = 1), and THalu 3.91 mm/yr (n = 1). Among individual shells, ZW156 from Triangle showed fastest mean growth (21.55 µm/day), while SS01ATH from THalu (10.72 µm/day) grew slowest (Fig. 7a, b).

SGI values revealed contrasting results to raw growth rates and showed similar minimum and maximum values between all T. squamosa shells irrespective of site or age (Fig. 7c, d). One-way ANOVA and post hoc Tukey test confirmed no significant differences in SGI over lifespan between shells from different sites (p > 0.05) (Table. S5).

Seasonal shell growth and environmental relationships

Distinct temporal patterns that were most prominent at a seasonal level were revealed in live collected shell chronologies where exact dates of death were known. The most striking difference was acceleration of growth that generally occurred in the wet season for Baik, while a seasonal trend in Triangle was less clear (Fig. 8). Undated SGI chronologies from dead collected shells at the same sites revealed similar seasonal acceleration in Baik, while Triangle showed smaller fluctuations and shorter-term cycles throughout the year (Fig. S6).

Fig. 8
figure 8

Dated mean shell growth chronologies derived from daily growth increment widths for Tridacna squamosa collected from Baik (blue lines) (b, d) and Triangle (red lines) (a, c) reefs in 2019 (SSct, SSbaik) and 2020 (ZW156, NS207). Measurements from raw data are presented in mm/yr (a, b) and detrended growth rates shown in the dimensionless unit of the standardized growth index (SGI) (c, d). Red dashed vertical lines represent one year of growth, and purple and orange bars represent wet and dry seasons, respectively. Raw growth (e) and SGI chronologies (f) showing only last year of growth (LYOG) in Baik (SSbaik) (blue line) and Triangle (SSct) (red line) reveal seasonal variation between the two sites. Raw data are represented by pale lines, and data after smoothing are represented by bold lines

Full size image

Undated SGI chronologies from dead shells collected at additional sites generally showed a similar seasonal range of SGI values to live collected shells (Fig. S7). The shell from Blue Lagoon revealed positive SGI values in the second year of growth, compared to only negative SGI values in the third year, while the shell from Misan showed a distinct peak in the third year of growth. Younger shells at THalu and Semporna that were under 3 years showed a short deceleration in the second year of growth before death.

AV plots revealed significant relationships between monthly SGI and cloud cover (p < 0.05) and SST (p < 0.05) in one Baik shell (SSBaik) (Fig. S8) and salinity (p < 0.01) and rainfall (p < 0.01) in the additional older Baik shell (NS207) (Fig. S9). In contrast, shells from Triangle (ssCT; ZW156) showed no significant correlation between tested environmental variables and SGI values (Figs. S10, S11). Nonetheless, we did note a degree of synchronization with bimodal chlorophyll-a and Kd(490) peaks in ZW156, while growth appeared relatively consistent year-round in ssCT (Fig. S12).

Spectral analysis

Multi-taper method (MTM) spectral analysis was applied to daily SGI values and tidal range to search for similarity of periodicities in frequency space. Daily growth increments in Triangle shells were characterized by high-frequency significant peaks corresponding to periods of ~ 2–3 days and spectra revealed several similarities with daily maximum sea level (Fig. 9). Significant peaks in Baik shells occurred over a wider range of frequencies corresponding to periods of ~ 2–6 days and showed less similarity with tidal spectra (Fig. S13).

Fig. 9
figure 9

Multi-taper method (MTM) power spectra of a daily maximum tide (m) and b daily SGI growth chronology of Triangle reef shell ZW156 in frequency space. AR(1) significance peaks are set to 95 and 99% and represented to a red noise background. Significant peaks labelled with arrows and identified in approximate days in relation to frequency (1/day)

Full size image

Discussion

Multi-method approach to sclerochronology in giant clams

In this study, we established shell growth chronologies for thirteen giant clam shells based on counting and measuring widths of daily growth increments with a mixed-method approach (petrography and SEM). To reduce uncertainty, increments were counted multiple times with different software and live collected shells cross-checked by a secondary person. However, large offsets existed between counts in some samples, which could be attributed to either differences in microstructure that reduced growth increment visibility or visualization of bands between approaches.

Molluscan shell microstructure is underpinned by genetic mechanisms but is further influenced by external environmental factors (Clark et al. 2020). Size and shape of biomineral units may change depending on temperature (Milano et al. 2017; Höche et al. 2020, 2021), light attenuation (Pätzold et al. 1991) and food supply (Clark et al. 2020). Here, we found two distinct microstructures of daily increments for the inner shell layer of T. squamosa: Type 1 in the Triangle reef, which consisted of paired daily growth couplets with a thicker- and thinner layer delineated by crystalline structure and Type 2 in other reef sites, with two adjacent growth lines cutting a complex crossed lamellar microstructural arrangement that was either irregular or cone (Taylor et al. 1969) and running perpendicular to first order lamellae (Agbaje et al. 2017). Delineation of daily increments from Type 1 shells was clearer than Type 2, and this was reflected in the differences between individual counts (Table S3). Distinctly demarcated paired increments that vary in morphology along one growth increment like Type 1 have previously been recorded for T. gigas (Pätzold et al. 1991; Gannon et al. 2017) and are presumably driven by the daily light cycle (Sano et al. 2012). The thicker prismatic layer is thought to be deposited in the daytime when symbiotic algae allow light-enhanced calcification, while the thinner layer may calcify at night when less light is available for photosynthetic processes (Sano et al. 2012). Although Pätzold et al. (1991) described a similar microstructure with faint growth lines like Type 2, they only reported its occurrence near the umbo and not throughout other regions of the shell. We found Type 2 to be common in early stages of growth when growth was faster, presumably a continuation of the simple crossed lamellar microstructural arrangement of the outer shell layer. Although this concurs with reports of less defined crystalline boundaries indicating faster deposition of microstructural components (Gannon et al. 2017), Type 2 continued into adulthood in many shells and Type 1 only recorded in Triangle, which may relate to tidal rhythm (discussed below).

Apart from microstructure, differences in visibility of daily growth increments in giant clam shells have also been attributed to method specific issues. For example, changes in focal length and low fluorescence of shell material under LSCM (Zhao et al. 2021). We used high-resolution SEM imaging across the entirety of the inner layer of each shell because this method allows singular daily increments to be revealed in detail in secondary electron images (Welsh et al. 2011). Shell regions whose daily increments remained unclear under SEM are likely a true feature of the sample instead of an artefact of sampling technique. We additionally found the boundary between the outer- and inner shell layers (i.e. the pallial line) distinct under SEM (Fig. S14), similar to other studies (Gannon et al. 2017; Ayling et al. 2017). Yet, this was sometimes unclear under light microscopy, which may be attributed to variation in focal length due to the presence of a highly topographic complex crossed lamellar microstructural arrangement. These results emphasize the importance of applying a multi-method approach in counting daily growth increments of giant clams because large unveiled offsets could severely obstruct interpretation for high-resolution studies.

Shell growth mediated by ontogeny

A challenge in the study of sclerochronology is to disentangle vital effects and externally forced growth signatures (e.g. temperature) (Schöne 2003). As for many bivalves, physiological processes of tridacnids change with the onset of sexual maturity and energy priorities switch between growth and reproduction (Jones et al. 1986). Over time, increments become tightly packed and rate and year to year amplitude of growth decreases (Romanek and Grossman 1989; Arias-Ruiz et al. 2017; Zhao et al. 2021). In this study, shell growth chronologies accounting for ontogenetic growth were investigated before applying the SGI. We found most shells over approximately three years old demonstrated rapid acceleration at the beginning of life and reduction in growth in or after three years (Fig. 7). Although delayed onset of sexual maturity (approximately 10 years) and subsequent deceleration of growth has been recorded for Tridacna maxima (Jones et al. 1986; Romanek et al. 1987; Chambers 2007) and T. gigas (Klump and Griffiths 1994; Lucas 1994), the switch between juvenile and adult growth phases may occur earlier in other species (Lucas 1994). While our shells are young (~ 1–7 yr) relative to the natural lifespan of Tridacna spp. (several decades) (Rosewater 1964), growth curves in some samples may indicate early onset maturity for T. squamosa.

Environmental influences on shell growth

Our results revealed mean shell growth rates from seven reefs that ranged from 3.91 ± 1.39 to 11.67 ± 3.97 mm/yr, corresponding to daily increment widths between 2.02 and 41.40 µm (mean of 14.90 µm/day). Despite shells being collected from a range of reefs subject to varying turbidity, results were generally consistent with those in the literature for modern and cultured shells, showing mean daily increment widths of 3–32.7 µm for measurements of the inner layer of T. squamosa and other Tridacnidae (Aubert et al. 2009; Elliot et al. 2009; Schwartzmann et al. 2011; Ariaz-Ruiz et al. 2017; Gannon et al. 2017; Yan et al. 2020; Zhao et al. 2021) (Fig. 10). Moreover, removal of ontogenetic growth signals using the SGI revealed no significant differences across annual growth rate from shells at any of our sites over lifespan. However, we do caution the uncertainties of the interpretation of the dead collected shells from the seafloor and it is possible that they were translocated post-mortem.

Fig. 10
figure 10

Mean daily growth rate (μm) of the inner shell layer for Darvel Bay and Semporna Tridacnidae (current study) compared to published growth values. Reef location SCS Southern South China Sea, PNG Papua New Guinea, GBR Great Barrier Reef. Sample zhao_md3 = Zhao et al. (2021); yan_XB10 = Yan et al. (2020); arias-ruiz_GSL14_19 = Ariaz-Ruiz et al. (2017); gannon_K-133 = Gannon et al. (2017); gannon_PT-1 = Gannon et al. (2017); elliot_Tgplam = Elliot et al. (2009); schwartzman = Schwartzmann et al. (2011), aubert = Aubert et al. (2009). Error bars represent standard deviation if available

Full size image

Influences on annual shell growth

One explanation for similar annual SGI values across sites may be related to the trophic plasticity documented for T. squamosa. Unlike some giant clam species who are functional autotrophs, T. squamosa has a photoautotrophic range that is extended by heterotrophy (Jantzen et al. 2008). Although contributions of heterotrophy to total energy requirements in tridacnids change depending on ontogenetic stage and size (Fisher et al. 1985; Klump et al. 1992; Klumpp and Griffiths 1994; Yau and Fan 2012), T. squamosa is unusual because it has significantly lower photosynthetic potential irrespective of age and is more reliant on filter-feeding (Jantzen et al. 2008). It is known to increase rates of filter-feeding with depth (Tedengren et al. 2000) and is common on the deeper fore-reef compared to other species, which are abundant in shallow (< 3 m) reef flat (Roa-Quiaoit 2005). Our results lend weight to T. squamosa surviving and potentially thriving in turbid reefs and concur with other studies (Guest et al. 2008; Yong et al. 2022), who reported the species can survive well and even accelerate growth on impacted reefs.

Influences on monthly shell growth

On the contrary, at a seasonal level we found substantial site-to-site variability between the shells collected alive from Baik and Triangle, indicating that different environmental factors may play some role in regulation of intra- annual growth trends. At the less turbid reef Baik, growth generally accelerated in the wet season and monthly SGI values negatively correlated with temperature, cloud cover, rainfall and salinity. Although in previous studies of giant clams increment widths positively correlate with SST (Aubert et al. 2009; Duprey et al. 2015; Komagoe et al. 2018; Zhao et al. 2021), the negative correlation found between SGI and temperature in Baik may relate to depressed growth above a thermal tolerance threshold of 27 °C (Schwartzman et al. 2011; Killam et al. 2021). As monthly SST varied between 28.11 °C and 30.47 °C at Baik, increment thickness could become erratic and a stress reaction initiated (Schwartzman et al. 2011). Another explanation may be the small variation in SST throughout the year (~ 2 °C), considered negligible in terms of contribution to growth rate (Gannon et al. 2017). As such, SST may play an indirect role in modulation of growth of our shells due to its relationship with light. A positive correlation between light and growth has been observed in several tridacnid species on diurnal—(Sano et al. 2012; Yan et al. 2021) to seasonal scales (Lucas et al. 1989). Yan et al. (2021) found co-variance of daily growth rate with outgoing longwave radiation (OLR), but no significant correlation with SST for Tridacna derasa, indicating local effective solar radiation, is more important than SST in determination of increment width. Our results of an inverse relationship between monthly SGI and cloud cover indicate at least some light dependence on growth in Baik caused by photosynthetic activity of endosymbionts.

Seasonal rainfall within Darvel Bay is enhanced by southwest (dry season) and northeast (wet season) monsoons. Rainfall may indirectly influence growth by washing particulate matter into the region, increasing local turbidity and impeding photosynthetic processes (Gannon et al. 2017). High-resolution paleoweather reconstructions show abrupt decreases in daily increment thickness due to cold upwelling and strong wave action from typhoon and tropical cyclone activity (Komagoe et al. 2018; Yan et al. 2020). Although monthly SGI in Baik shows a negative relationship with rainfall, periods of reoccurring accelerated growth in the mid to late wet season may be attributed to enhanced vertical mixing after increased rainfall, bringing nutrient rich sub-surface waters to the surface, coupled with subsequent increased solar irradiance (Gannon et al. 2017; Yan et al. 2020). While Baik is indicative of a low turbid reef and Kd(490) and chlorophyll-a do not increase to the same extent as Triangle, we propose that a sequence of events similar to that recorded by Gannon et al. (2017) following: (1) higher rainfall during the early wet season, (2) elevated Kd(490) and chlorophyll-a, (3) increased sunlight, may result in favourable growing conditions.

Unlike Baik, shells from Triangle lacked any statistically significant relationship between monthly SGI and environmental variables. Yet, we did note a positive association between SGI and chlorophyll-a and Kd(490) in some months. Triangle is a naturally turbid reef situated approximately 3.5 km from the mouth of the river

Tingkayu, which discharges sediment to the site at low tide. During ebb tide, freshwater carrying sediment loads are directly discharged seaward to the reef at the surface, while density gradients generating two-layer circulation patterns move saltwater landward at the bottom (Saleh et al. 2007). Sediment input is likely land derived POM produced in the river basin by different mechanisms (e.g. freshwater productivity and remains of plants and microorganisms) and mineral sediment (Bainbridge et al. 2018). In similarity to mixotrophic corals (Fox et al. 2018), elevated chlorophyll-a concentrations at Triangle may indicate T. squamosa utilises chlorophyll-a as a key contributor towards metabolic requirements in reduced photic depth. Local irradiance may be of less importance for T. squamosa compared to other giant clam species, and it can sustain growth in turbid reefs if a suitable local food source is available.

Influences on daily shell growth

At a daily scale, spectral characteristics of SGI chronologies revealed similarities with tidal range at Triangle. Significant spectral peaks occurred at around two to three days corresponding closely to peaks in maximum daily tide. Tidal rhythms are known to leave characteristic daily growth or geochemical patterns in many bivalves, such as differentiation of increment widths that relate to spring-neap variability or daily tidal emergence (Goodwin et al. 2001) and trace element cyclicity (Warter and Müller 2017; de Winter et al. 2022). It has recently been suggested that under natural conditions the circadian clock genes of some bivalves may also run at tidal frequency, indicating behavioural patterns relating to a circadian cycle can be modulated by tidal rhythm (Tran et al. 2020). Based on this hypothesis, the presence of paired daily increments in Triangle shells may relate to reef sediment flux controlled by a circadian rhythm paced to tides that result in variable light intensities. The thicker prismatic layer presumably related to higher irradiance may be accreted during flood tide, when clear water is transported offshore to the reef. At ebb tide when suspended sediment is transported to the reef lowering photic depth, smaller crystals could be deposited. However, more work is needed to elucidate the relationship between short-term tidal cycles and daily growth increments in tridacnids.

Conclusion

In this study, we provide the first assessment of growth in giant clams from turbid reefs using a mixed method approach. We investigated daily growth increments in thirteen Tridacna squamosa shells from different coral reef sites with varying degrees of turbidity. Similarities were found in annual growth rate between clear and turbid reefs, while there were differences in seasonal growth trends, indicating that an interplay of different environmental factors may regulate intra-annual growth between reefs. At a daily scale, we found differences in microstructure and spectral characteristics of daily increments, which may relate to tidal variability at the turbid site near a river source. Our results do not indicate that turbidity negatively influences growth but are consistent with sustained growth, which may relate to the trophic plasticity of T. squamosa. This work supports growing evidence that resilient marginal habitats with elevated turbidity may serve as important conservation hotspots and our results are useful in the context of management and conservation for T. squamosa under changing future oceanic conditions. Although encouraging, we cannot ascertain what this may mean for other parameters important for survival, such as skeletal density. More work is needed on different tridacnid species to better understand the multiple environmental and physiological influences of turbidity.