Isotopic composition of skeleton-bound organic nitrogen in reef-building symbiotic corals: A new method and proxy evaluation at Bermuda

Abstract

The skeleton-bound organic nitrogen in reef-building symbiotic corals may be a high-resolution archive of ocean nitrogen cycle dynamics and a tool for understanding coral biogeochemistry and physiological processes. However, the existing methods for measuring the isotopic composition of coral skeleton-bound organic nitrogen (hereafter, CS-δ¹⁵N) either require too much skeleton material or have low precision, limiting the applications of this relatively new proxy. In addition, the controlling factors on CS-δ¹⁵N remain poorly understood: the δ¹⁵N of source nitrogen and the internal nitrogen cycle of the coral/zooxanthellae symbiosis may both be important. Here, we describe a new (“persulfate/denitrifier”-based) method for measuring CS-δ¹⁵N, requiring only 5 mg of skeleton material and yielding a long-term precision better than 0.2‰ (1σ). Using this new method, we investigate CS-δ¹⁵N at Bermuda. Ten modern Diploria labyrinthiformis coral cores/colonies from 4 sampling sites were measured for CS-δ¹⁵N. Nitrogen concentrations (nitrate + nitrite, ammonium, and dissolved organic nitrogen) and δ¹⁵N of plankton were also measured at these coral sites. Among the 4 sampling sites, CS-δ¹⁵N shows an increase with proximity to the island, from ∼3.8‰ to ∼6.8‰ vs. atmospheric N₂, with the northern offshore site having a CS-δ¹⁵N 1–2‰ higher than the δ¹⁵N of thermocline nitrate in the surrounding Sargasso Sea. Two annually resolved CS-δ¹⁵N time series suggest that the offshore-inshore CS-δ¹⁵N gradient has persisted since at least the 1970s. Plankton δ¹⁵N among these 4 sites also has an inshore increase, but of only ∼1‰. Coral physiological change must explain the remaining (∼2‰) inshore increase in CS-δ¹⁵N, and previous work points to the coral/zooxanthellae N cycle as a control on host tissue (and thus carbonate skeletal) δ¹⁵N. The CS-δ¹⁵N gradient is hypothesized to result mainly from varying efficiency in the internal nitrogen recycling of the coral/zooxanthellae symbiosis. It is proposed that, in more productive inshore waters, greater food uptake by the coral causes a greater fraction of its low-δ¹⁵N regenerated ammonium to be excreted rather than assimilated by zooxanthellae, raising the δ¹⁵N of the inshore corals. If so, coral tissue- and CS-δ¹⁵N may prove of use to reconstruct and monitor the state of the coral/zooxanthellae symbiosis over space and time.

Introduction

Because reef-building corals are widespread in the tropical and subtropical ocean and have a relatively fast growth rate (0.5–2 cm/year), coral skeleton represents one of the most promising recorders of climate variability in the low-latitude ocean. Applications to date include the reconstruction of high-frequency climate variability associated with the El Niño-Southern Oscillation (ENSO) (Dunbar et al., 1994, Gagan et al., 2000, Watanabe et al., 2011, Cobb et al., 2013), the Pacific Decadal Oscillation (Linsley et al., 2000), and the North Atlantic Oscillation (Kuhnert et al., 2005, Goodkin et al., 2008). Previous coral-based paleoceanographic studies have focused on the inorganic components of coral skeleton, with attention to the organic matter within coral skeleton only recently (e.g., LaVigne et al., 2008), primarily due to the analytical challenges of working with this relatively organic-poor material.

The coral’s aragonite forms at the interface of coral tissue and the underlying skeleton. Seawater containing calcium ions and dissolved inorganic carbon is transported via paracellular pathways (Cohen and McConnaughey, 2003) to an extracellular space between the calicoblastic epithelium and the existing skeleton, where crystals are nucleated and grown. The organic matrix is synthesized in the calicoblastic epithelium and then transported into the space below (Muscatine et al., 2005, Drake et al., 2013), and it has been hypothesized to serve as the calcification nucleus and control the crystallographic direction of the aragonite precipitation (Cuif et al., 1999, Mass et al., 2013). The skeleton-bound organic matter reflects the carbon and nitrogen sources of the coral-zooxanthellae symbiotic system (Drake et al., 2013) and may thus provide insights into past ocean biogeochemical conditions and/or the physiological and ecological history of corals.

Several studies have explored the isotopic composition of coral skeleton-bound organic nitrogen (hereafter, CS-δ¹⁵N) in reef-building symbiotic corals, using a variety of methods. Muscatine et al. (2005) did a CS-δ¹⁵N survey in the modern ocean, finding that the CS-δ¹⁵N in symbiotic coral (4.09 ± 1.51‰, n = 24) is much lower than CS-δ¹⁵N in non-symbiotic corals (12.28 ± 1.81‰, n = 17). They suggested that CS-δ¹⁵N is a proxy for the coral host/zooxanthellae symbiosis, using their data to argue that this symbiosis may have started as early as the Triassic. Marion et al. (2005) generated CS-δ¹⁵N downcore records at Bali, Indonesia. They interpreted CS-δ¹⁵N as a proxy for the δ¹⁵N of the fixed nitrogen sources to the reef and found that low-δ¹⁵N synthetic fertilizer has been affecting coastal coral reefs since the 1970s. Yamazaki et al. (2011) also interpreted CS-δ¹⁵N as a proxy for the δ¹⁵N of nitrogen sources. On a subtropical island near Taiwan, where the riverine nitrate is the dominant nitrogen source to the reef, they observed a correlation between an onshore-offshore CS-δ¹⁵N gradient and the gradient in surface water nitrate δ¹⁵N.

At this point, two major issues limit the utility of CS-δ¹⁵N. First, the methods so far employed have had major drawbacks. At least three different methods have been used to measure CS-δ¹⁵N (Muscatine et al., 2005, Uchida et al., 2008, Yamazaki et al., 2013), and these methods either require large quantities of skeletal material (e.g., >1 g; Marion et al., 2005, Muscatine et al., 2005) and/or have very low precision (e.g., 1 SD greater than 0.75‰; Uchida et al., 2008, Yamazaki et al., 2013). Because of the difficulty of the measurements, there has also been limited method testing (e.g., of cleaning protocols). Second, beyond the methodological issues, the controlling factors on CS-δ¹⁵N remain poorly understood. The δ¹⁵N of the nitrogen sources to the reef is expected to affect CS-δ¹⁵N, and there are now some data in support of this expectation (Yamazaki et al., 2011). However, coral physiology, including the symbiosis with zooxanthellae, may also be important (Muscatine et al., 2005, Swart et al., 2005).

In this study, we seek to make inroads on both of these problems. First, we report the development of a robust, highly sensitive method for measuring CS-δ¹⁵N, based on previous foraminifera-bound δ¹⁵N studies (Ren et al., 2009, Ren et al., 2012). This method requires ∼5 mg of coral skeleton material and yields a long-term precision better than 0.2‰. Second, with this method, we measured CS-δ¹⁵N from 10 cores/colonies from 4 sites on the pedestal of Bermuda, as well as nutrient concentrations and plankton δ¹⁵N from the same sites. We find a ∼3‰ inshore-offshore gradient in CS-δ¹⁵N, at least ∼2‰ of which requires a physiological (rather than environmental N isotopic) explanation. We propose that variation in the efficiency of the internal nitrogen recycling between the coral host and its zooxanthellae is responsible for the observed CS-δ¹⁵N difference among the sites.