Introduction

The dynamic nature of nutrient availability is integral to understanding coral reef ecology. Regional and reef scale processes influence dissolved nutrient availability while the coastal environment and marine productivity comprise primary drivers of particulate availability across reefs (Wyatt et al. 2013; Lowe and Falter 2015). Particulate resource availability is a critical component of fish productivity (Morais and Bellwood 2019), and marine primary production is a driver of heterotrophic feeding patterns in corals (Fox et al. 2018). For example, particulate feeding by some corals can positively affect coral growth parameters (Fabricius 2005) and help to meet the daily energy requirements under normal and anomalous ocean temperatures that cause coral bleaching (Grottoli et al. 2006; Houlbrèque and Ferrier-Pagès 2009). Therefore, changes in reef trophodynamics may have consequences for organisms that are dependent on particulate resources.

As heat stress events can increase the death and decay of reef organisms (Leggat et al. 2019), an invigorated detrital cycle may increase the availability of POM as a major source of carbon for consumers in degraded reefs (Morillo-Velarde et al. 2018). Increased levels of ultraviolet radiation, often associated with thermal stress induced coral bleaching, can cause corals to release organic matter that is depleted in nitrogen (Courtial et al. 2018). Although the regeneration of mucus-derived organic matter by bacteria in the reef environment may increase organic nitrogen, it is possible that nitrogen availability may be depleted during this process (Coles and Strathmann 1973; Courtial et al. 2018). Reef macroalgae are also known to release POM (Haas et al. 2010), while sponges play an important role in cycling coral-derived and algal-derived organic matter and releasing it as detrital POM (Rix et al. 2017, 2018). Considering the importance of particulate resources for many reef organisms especially during and following disturbance events such as coral bleaching, it is critical to characterize resource availability in the reef water column.

This study investigated particulate organic availability and isotopic composition following a mass coral bleaching event in the Maldives, Indian Ocean. Carbon and nitrogen concentrations and stable isotope values (δ13C and δ15N) of POM were compared across shallow (10 m) and mesophotic (30 m) reefs one year before and eight months after the 2016 mass coral bleaching event. Further, a surface pump allowed for the deeper sampling (50 m) of POM and comparison of two sampling methods (diver versus pump) post-bleaching. Because plankton are also a food source for corals and fish, plankton δ13C and δ15N values were measured after the bleaching event. Inorganic nutrients, including a subset of samples for nitrogen and oxygen stable isotope analysis (δ15N-nitrate and δ18O-nitrate), were measured to investigate post-bleaching nutrient availability and trophic status of the shallow to mesophotic (10–50 m) water column (Casciotti 2016).

Materials and methods

Field collection

Pre-bleaching (2015) sampling of POM occurred in the boreal spring (intermonsoon) in the Maldives and is further detailed in Radice et al. (2019) while post-bleaching sampling occurred during the same season in 2017 (Supplementary Material Table S1, Fig. S1). To survey shallow (10 ± 2 m) and mesophotic depths (30 ± 3 m; 47 ± 3 m) along oceanic reef slopes (‘fore reef’), seawater was collected by divers (10 and 30 m) and via a surface pump (all depths) paired with a Sensus Ultra depth recorder (ReefNet) (Fig. S2). Given that optimal mesophotic photosynthetically active radiation (PAR) is 7.5% of the surface (Tamir et al. 2019), 30 m depth in these Maldives reefs is mesophotic since only 4% of surface PAR is available (Radice et al. 2019). Sampling of the 50 m mesophotic depth target required repositioning the boat an average of 600 m from shallow reefs. Seawater (min. 3L, mean 7L) was filtered, and frozen (− 20°C): (1) stored in HDPE bottles prior to inorganic nutrient analysis (0.22 µm polyethersulfone, Millex-GP, Merck), and (2) filtered for isotopic analysis of POM (nominal 0.4 μm glass fiber filters [Machery-Nagel], pre-combusted for 5 h at 450°C). Plankton were collected at dusk by towing mesh nets (53 and 250 µm) submerged below the surface for 10 min at a speed of ~ 2 knots. Daily wind data were obtained from the Maldives Meteorological Service (Maldives Meteorological Service 2018).

Inorganic nutrient and isotopic analyses

Inorganic nutrients (nitrate [NO3−], ammonium [NH4+], phosphate [PO43−]) (mean ± SD) were measured by flow injection analysis at the Advanced Water Management Centre, University of Queensland. Nitrite was undetectable for the majority of samples, and phosphate was undetectable for two samples. Potential nutrient limitation was evaluated using the Redfield ratio of dissolved nutrients (([NO3−] + [NH4+])/[PO43−]) (mean ± SD), with values deviating from the 16:1 ratio providing evidence of either nitrogen or phosphorous limitation (Redfield 1958). A subset of samples (n = 21, 7 samples/depth) were analyzed for nitrogen and oxygen stable isotope values (δ15N and δ18O) of nitrate via a bacterial denitrification assay at the Woods Hole Oceanographic Institution (Sigman et al. 2001; Casciotti et al. 2002). Repeated measurements of three references resulted in analytical precision of 0.5‰ for δ15N and 1.6‰ for δ18O. Stable isotope values are reported in the delta (δ) notation (mean ± SD) expressed in per mil (‰): for example, δ15N = [(Rsample/Rstandard) − 1] × 1000 where R is the heavy to light isotope ratio (15N/14N).

Glass fiber filters were briefly treated with HCl to remove any carbonates, neutralized with Milli-Q water, and oven-dried. Carbon and nitrogen concentrations and stable isotopes (δ13C and δ15N) of POM (n = 112) were analyzed with an Elementar Vario EL Cube elemental analyzer and with a PDZ Europa ANCA-GSL for plankton (n = 30), with analyzers interfaced to a PDZ Europa 20–20 isotope ratio mass spectrometer (Sercon) at the University of California Davis. Repeated measurements of laboratory reference materials (G-6: Nylon 6, G-13: Bovine liver, G-20: Glutamic acid, G-21: Enriched alanine) resulted in analytical precision of < 0.2‰ for both δ13C and δ15N. Due to analytical detection limits, samples with very low nitrogen (< 10 µg; n = 21) were removed from analyses. For elemental C and N, blank corrections were applied by subtracting the average of five samples with the lowest C or N contents (Figs. S3 and S4). Particulate organic nitrogen (PON) and carbon (POC) concentrations were calculated as N or C (µg), respectively, per liter of seawater.

Statistical analyses

Statistical analyses and visualization were conducted using the R computing environment (Wickham 2016; Wilke 2019; R Core Team 2020). Two oceanographic regimes were identified during back-to-back sampling in the 2015 cruise, with Cyanobacterial bloom and Upwelling regimes characterized by different POM δ15N (Radice et al. 2019). The monsoon climate dominates the oceanography in the Maldives, with upwelling occurring the majority of the year. Initial evaluation of post-bleaching (2017) POM data showed that low δ15N values, characteristic of cyanobacteria, were absent. Therefore, post-bleaching POM data (Oligotrophic regime) were compared to the pre-bleaching Upwelling regime data. POM δ13C, δ15N, POC, and PON were analyzed using linear mixed effects models (‘nlme’) with year (pre- and post-bleaching) and depth (10 and 30 m) as fixed interactive effects and site as the random effect (Pinheiro et al. 2020). Following model selection based on Akaike information criteria (AIC), depth was removed only in the POM δ15N model. The pre- versus post-bleaching POM analysis only considered data collected by the same method (diver).

Inorganic nutrient data (NO3−, NH4+, PO43−), collected only in 2017, were analyzed using linear mixed effects models with sampling method (diver and pump) and depth (10, 30, 47 m) as fixed effects and site as the random effect. Following model selection based on AIC, sampling method was removed from the model investigating nitrate and phosphate while depth was removed from the model investigating ammonium. The effect of size fraction (53, 250 µm) on plankton δ13C and δ15N values was analyzed using linear mixed effects models with site as the random effect. The δ13C and δ15N values of post-bleaching POM were analyzed using linear mixed effects models with sampling method (diver and pump) and depth (10, 30, 47 m) as fixed effects and site as the random effect. Following model selection based on AIC, depth was removed from the δ15N model. Post-hoc (Tukey) tests of significant main effects by least-squares means were performed (‘emmeans’) (Lenth 2020).

Results and discussion

Oceanographic regimes in the Maldives were defined a posteriori considering biogeochemical characteristics: cyanobacterial bloom (2015; low δ15N), upwelling (2015; δ15N ~ 5‰), and oligotrophic (2017; low PON) (Fig. 1). POM δ13C values increased in both shallow (10 m) and mesophotic (30 m) reefs in the post-bleaching oligotrophic regime (post-hoc p < 0.001 and p < 0.001, respectively), but did not differ between depths in either year (Fig. 2a; Tables 1 and S2). Increased mean δ13C of POM post-bleaching could be due to: (1) an input of isotopically heavier, carbon-rich sources (e.g., reef-derived POM such as mucus or detritus), (2) a reduction in oceanic POM, or (3) changes in POM community composition, with the latter considered the most likely (Tanaka et al. 2008; Kolasinski et al. 2011; Karati et al. 2017). In a nutrient-depleted oligotrophic water column, small nutrient inputs can affect the POM signal; indeed, highly variable δ13C values were observed post-bleaching (mean  − 20.9 ± 2.9‰; range:  − 25.3 to  − 13.4‰). The wide range in post-bleaching POM δ13C values suggests there is greater diversity in community composition, such as Trichodesmium (high δ13C) and small cyanobacteria and/or eukaryotes (low δ13C), while uniformity in δ15N provides evidence of tight recycling of nitrogen across a variety of species (Tchernov and Lipschultz 2008). Post-bleaching, the large range in C:N indicates many unique combinations of production and recycling while low C:N pre-bleaching reflects cyanobacteria but also possible nitrate storage in phytoplankton (Bode et al. 1997; Liu et al. 2019).

Fig. 1
figure 1

Coral reef water column biogeochemistry and elemental ratios show three oceanographic regimes in the Maldives archipelago. Filled symbols depict pre-bleaching POM samples and open symbols depict post-bleaching POM samples in relation to the 2016 mass coral bleaching event. Particulate organic nitrogen (PON) is shown in relation to a nitrogen (δ15N) and b carbon stable isotope values (δ13C) of particulate organic matter (POM), c particulate organic carbon (POC) is shown in relation to POM δ15N values, and d molar C:N ratios are shown in relation to POM δ13C values

Full size image
Fig. 2
figure 2

A comparison of particulate organic matter biogeochemistry from shallow (10 m) and mesophotic (30 m) coral reefs before and after the 2016 mass coral bleaching in the Maldives. a Carbon stable isotope values (δ13C) of POM. Symbols show the differences between years at each depth, but no differences were found between depths. b Particulate organic nitrogen (PON) concentrations, with symbols showing differences pre- versus post-bleaching. The box and whiskers plots display the median and first and third quartiles while the triangle symbols represent the mean values

Full size image
Table 1 Particulate organic matter carbon (δ13C) and nitrogen (δ15N) stable isotope values, carbon (POC) and nitrogen (PON) concentrations, and molar C:N ratios (mean ± SD) from coral reefs across the central Maldives one year before (2015) and after (2017) the oceanic heat stress event that caused mass coral bleaching in 2016
Full size table

POM δ15N did not differ between pre-bleaching (upwelling) and post-bleaching (oligotrophic) regimes across shallow or mesophotic reefs (Table S2). Nitrogen loss due to recycling may lead to higher δ15N while some nitrogen fixation counterbalances with lower δ15N inputs, leading to the relatively more consistent δ15N values (mean 6.2 ± 1.4‰; range: 2.8 to 8.8‰) but variable δ13C post-bleaching. Similar δ15N values of POM (as above) and plankton (mean 6.4 ± 0.8‰) indicate tight coupling between primary and secondary production, revealing low fractionation between these groups and providing further evidence of an oligotrophic environment (Fig. S5). Plankton δ15N (as above) and δ13C (mean – 21.5 ± 0.4‰) did not differ between the size fractions (Tables S3 and S4). Despite the lack of historic temporal data on POM dynamics and isotopes in the region, the POM data presented effectively characterizes the regional oceanographic regimes that are dependent on upper ocean mixing dynamics (Garcia et al. 2018).

The decreased PON concentrations post-bleaching coincided with reduced wind speed and duration that may have affected wind-driven upwelling in the months preceding the sampling (Fig. 2b, Table S5). In addition to low concentrations of PON and phosphate, the oligotrophic water column was also characterized by: (1) the persistence of high residual δ18O–NO3− values due to low nutrient concentrations, (2) high Redfield ratios (21.8 ± 13.0), and (3) δ15N–NO3− values reflecting potential fractionation due to phytoplankton uptake (Table S6). Overall, there was substantial variability in both δ15N–NO3− and δ18O–NO3− values. Isotopic values indicate that nitrate was derived from a combination of marine and atmospheric sources, with the latter enriched in δ18O (Fig. S6) (Kendall et al. 2008). High Redfield ratios suggest phosphate limitation, which could allow atmospheric nitrate (enriched δ18O) to accumulate in surface waters during periods of anthropogenic aerosol loading such as the northeast monsoon season (Ramana and Ramanathan 2006; Das et al. 2011). The signature of marine-derived nitrate (δ15N ~ 5–6‰) is also evident in POM from the upwelling and oligotrophic regimes as well as plankton. Given a fractionation factor of ~ 5‰, the mean δ15N–NO3− value (9.2 ± 4.9‰) likely reflects phytoplankton assimilation of a deep-water nitrate source of approximately 5‰ (Altabet 2001). Previous work has shown that δ15N values of reef-building corals are consistent with deep-water nitrate δ15N (~ 5‰), which indicates coral dependence on inorganic nutrients and particulates from the dominant upwelling regime (Radice et al. 2019).

Some inorganic nutrient concentrations were affected by reef depth, with higher nitrate and phosphate at 47 m compared to 10 and 30 m depths (p < 0.001 and p ≤ 0.001, respectively) (Fig. 3a-c, Tables S3 and S7). The higher concentration of inorganic nutrients at the deepest sampling depth (47 m) aligns with expectations of increased nutrient concentrations at greater depths (Leichter et al. 2007). While there were differences of 1.6‰ and 1.0‰ in POM mean δ13C and δ15N, respectively, between the sampling methods used in 2017, no effect of depth was observed (Fig. 3d-e, Tables 1 and S3). The discrepancy in the δ13C and δ15N values of POM sampled by divers on the reef versus pump (mean 600 m offshore from shallow reef) may be due to the relative proportion of reef-derived versus offshore (pelagic) organic matter, the phytoplankton community composition, and/or oceanographic forcing of inorganic nutrient supply (Wyatt et al. 2013; Karati et al. 2017). Despite the importance of upwelling ecosystems, our understanding of the dynamics of benthic-pelagic nutrient cycling is often hindered by a lack of measurements of nutrient sources in the tropics in particular (MacKenzie et al. 2019), including the understudied central Indian Ocean. Because of the interactive oceanic-atmospheric processes affecting upwelling (Alory and Meyers 2009), it is important to consider how climate change may affect such regional processes that supply essential nutrients to marine ecosystems. For coral reefs ecosystems in particular, it is critical to consider how the supply and cycling of nutrients may change at local scales in relation to oceanic heat stress events and the consequent impact on reef trophic ecology.

Fig. 3
figure 3

Post-bleaching water column nutrient chemistry across shallow (10 m) and mesophotic (30 and 47 m) coral reefs in the Maldives (2017 only). Dissolved inorganic nutrient concentrations from shallow and mesophotic coral reefs include: a nitrate (NO3−), b ammonium (NH4+), and c phosphate (PO4−3), with italic letters showing differences between depths. Post-bleaching particulate organic matter d carbon (δ13C) and e nitrogen stable isotope values (δ15N) are shown in relation to depth and sampling method, with symbols denoting differences between sampling methods. The box and whiskers plots display the median and first and third quartiles while the triangle symbols represent the mean values

Full size image