Coral skeletons reveal the history of nitrogen cycling in the coastal Great Barrier Reef

The global anthropogenic production of nitrogen (N) now equals the amount generated through biological dinitrogen (N₂) fixation. Human activity has more than doubled pre-industrial N input into the ocean. While nutrient input to coastal ecosystems often leads to algal blooms and eutrophication, other systems show resilience. Terrestrial nutrient and organic matter discharge can accelerate N loss through denitrification and anammox. Phosphorus (P), not lost through microbial activity, can accumulate, leading to N limitation (low inorganic N:P ratios). In tropical systems, N limitation promotes N₂ fixation, restoring N:P ratio balance. The response of coastal ecosystems to nutrient enrichment isn't always a simple cause-and-effect relationship. On the Great Barrier Reef (GBR), nutrient inputs have increased since European settlement in the 1850s. This has led to the conclusion that the coastal GBR is N-replete, perhaps even eutrophic in some areas. However, this contradicts nutrient budgets showing strong N deficits (more N leaving than entering). This discrepancy might be due to over/underestimation of denitrification and N₂ fixation rates. Other evidence suggests the N deficit isn't solely a methodological artifact. Inorganic N concentrations are consistently low in coastal GBR waters, except during floods. This might be explained by rapid uptake by phytoplankton, whose abundance is thought to have increased due to terrestrial nutrient enrichment, although this is hard to confirm due to high seasonal variability and a lack of long-term data. Phytoplankton eventually settle, becoming incorporated into sediment organic matter, efficiently processed and lost via denitrification. Denitrification is a major N loss pathway in the coastal GBR, exceeding N input through N₂ fixation. Enhanced terrestrial discharge promoting greater denitrification should manifest as increased P availability, supported by coastal water column and sediment nutrient data showing N:P ratios well below the Redfield ratio and the average of terrestrial runoff. This suggests N limitation in the GBR lagoon most of the year, with high P availability promoting N₂ fixation. Early reports suggested increased N₂ fixation due to anthropogenic activity, but this has been difficult to prove without long-term pre-European records. This study uses the ¹⁵N/¹⁴N ratio of N in coral skeletons to understand how N cycling in the coastal GBR has responded to anthropogenic nutrient discharge since European settlement.

Previous research has shown increasing particulate and dissolved nutrient inputs to the GBR since European settlement in the 1850s, evidenced by numerical estimates and coral skeleton proxy data. This has led to the conclusion that the coastal GBR is nitrogen-replete and possibly eutrophic in some areas. However, this contrasts with modern nutrient budgets which indicate strong nitrogen deficits. The discrepancy might stem from underestimation of nitrogen fixation and overestimation of denitrification rates. Other studies support the notion of nitrogen limitation in the GBR lagoon for much of the year, except during periods of significant flooding. The role of increased nitrogen fixation as a response to anthropogenic activity has been debated, lacking long-term pre-European data to validate the hypothesis. Prior studies utilizing coral skeletal data have shown mixed results, with some suggesting no significant changes in nitrogen levels while others have indicated alterations.

This study analyzed the ¹⁵N/¹⁴N (δ¹⁵N) ratio of nitrogen trapped in the organic skeletal matrix of four *Porites lutea* coral cores from two locations (Havannah Island and Pandora Reef) in the central inshore GBR. An additional short core from Geoffrey Bay (Magnetic Island) was also analyzed. To constrain the δ¹⁵N of nitrogen in terrestrial runoff, δ¹⁵N of nitrogen in coastal waters was measured before and after a major flood event in 2019. Coral carbonate powders were collected, cleaned, and analyzed using persulfate oxidation of liberated organic material to NO₃⁻, followed by measurement of its δ¹⁵N after conversion to N₂O using a Thermo Delta V Plus Isotope Ratio Mass Spectrometer (IRMS). Water samples were analyzed for total dissolved nitrogen concentration and δ¹⁵N of total dissolved nitrogen. Statistical analysis included Mann-Kendall trend tests, Student's t-tests, and change point analysis to detect significant trends and shifts in δ¹⁵N values over time. Isotope mixing models were developed to estimate the contribution of terrestrial and nitrogen fixation sources to the observed δ¹⁵N changes in coral skeletons.

The composite δ¹⁵N records for Havannah Island and Pandora Reef showed a significant positive correlation (r = 0.56, p < 0.001) over the 143-year period of overlap (1863–2005). Shifts in δ¹⁵N within each core reflect ecosystem-level changes in external δ¹⁵N. Patterns of δ¹⁵N variability matched reconstructed Burdekin River flow, indicating that river runoff increased δ¹⁵N, while dry periods decreased it. The correlations were better when short-term variability and the period before 1940 were excluded. This suggests that N enrichment following runoff events increases δ¹⁵N. Two lines of evidence suggest altered N cycling patterns since the 1850s: a significant decreasing trend in δ¹⁵N from 1680 to 2012 and a shift in the response of δ¹⁵N to rainfall in the latter half of the 20th century. The mean δ¹⁵N between 1680-1780 (6.4 ± 0.5‰) was significantly higher than the last 100 years (5.8 ± 0.4‰). The mean between 1940 and 2012 was even lower (5.6 ± 0.4‰). A similar decreasing trend was not initially observed in a previous long core from Magnetic Island (1820-1987), but an updated record including a 25-year short core (1987-2011) revealed a significant decrease in δ¹⁵N between 1860 and 2011. The timing of the decrease in the Magnetic Island record is later than in the other records. After 1940, the δ¹⁵N records are poorly or inversely correlated with Burdekin River flow, unlike the period between 1860 and 1940 and 1680 and 1770. Change point analysis showed an increase in δ¹⁵N minima after 1940 and a reduction in the magnitude of positive change points. The decrease in δ¹⁵N suggests a new source of N with lower δ¹⁵N, either increased N₂ fixation or modified terrestrial N with low δ¹⁵N from fertilizer use. The δ¹⁵N of total dissolved N (δ¹⁵N-TDN) in the Burdekin River flood plume ranged from 6.2‰ to 4.6‰, while before the flood, it was 4.1‰. Using an isotope mixing model, the δ¹⁵N-TDN of terrestrial N was calculated to be 8.1 ± 1.1‰, too high to explain the observed decrease in δ¹⁵N in coral records. A second isotope mixing model suggests that increased N₂ fixation is the most likely explanation for the observed decrease in δ¹⁵N in coral records.

The long-term decline in coral skeletal δ¹⁵N, coupled with the analysis of water samples, strongly suggests increased nitrogen fixation in the coastal GBR since European settlement. The observed decrease in δ¹⁵N cannot be attributed to the direct input of anthropogenically modified terrestrial nitrogen based on the isotopic values obtained from the flood plume analysis. The findings highlight the complexity of nutrient cycling in coastal ecosystems and emphasize the significant role of nitrogen fixation in response to enhanced nutrient inputs, particularly phosphorus. While increased terrestrial nitrogen inputs may contribute to overall nitrogen loads, the relatively high δ¹⁵N values of terrestrial nitrogen render it an unlikely primary driver of the observed decrease in skeletal δ¹⁵N. The strong inverse relationship between δ¹⁵N and rainfall after 1940 suggests that the increased nitrogen fixation during dry periods has been amplified by changes in the ecosystem triggered by anthropogenic activities.

This study provides compelling evidence for a significant increase in nitrogen fixation in the coastal Great Barrier Reef since European settlement. The long-term decrease in coral skeletal δ¹⁵N, coupled with isotopic analysis of water samples from flood plumes, strongly supports this conclusion. The results indicate that phosphorus reduction strategies should be prioritized to mitigate the risk of future eutrophication. Future research should focus on quantifying N₂ fixation rates through independent measurements and investigating the interplay between nitrogen fixation and other nitrogen cycle processes under varying environmental conditions. Further investigation into the changes in phytoplankton communities and their contribution to the observed isotopic patterns is also warranted.

The study focuses on a limited region of the Great Barrier Reef, and the findings may not be fully generalizable to the entire reef system. The interpretation relies on isotopic mixing models which involve inherent assumptions and uncertainties, particularly regarding the isotopic signature of terrestrial nitrogen sources and the proportion of fixed nitrogen incorporated into the coral skeleton. The sample size of the water analysis related to flood plumes is limited, and additional data may refine estimates of the isotopic signature of terrestrial nitrogen.