Increasing nitrogen loading in the Gulf of Mexico undermines coral reef resilience Authors: Jonathan Jung, Kristine L. DeLong, Paolo Montagna, Erin L. Murphy, Nicolas N. Duprey, Alan D. Foreman, Amy J. Wagner, Kylie Palmer, Julia Schröder, Carlos Alonso-Hernández, Denis Scholz, Daniel M. Sigman, Gerald H. Haug, Alfredo Martínez-García Published: 2026 ABSTRACT The ongoing global decline of coral reefs underscores the need to identify key mechanisms driving reef deterioration. The Flower Garden Banks (FGB) reefs in the northern Gulf are widely regarded as some of the healthiest in the continental United States, but recent disease outbreaks in the FGB have raised concerns about their long-term resilience. As human and agricultural expansion intensified across the Mississippi River basin, the river has become the dominant source of nutrients to the northern Gulf; however, the evolution of this influence is poorly understood. Here, we reconstruct nitrogen inputs to the FGB using annually resolved coral-bound nitrogen isotopes (CB-δ15N) from two coral cores that collectively span the period from 1753 to 2023. Our results show that anthropogenic nitrogen began to rise around 1850, coinciding with the early use of fertilizers, and increased sharply after the 1960s with the onset of the Green Revolution and the use of synthetic fertilizers, suggesting the Mississippi discharge as the source. Fractional endmember modeling further reveals that the accumulation and mobilization of legacy nitrogen stored in Midwestern soils and groundwater have amplified modern riverine fluxes, driving Mississippi-derived nitrate to more than 60 to 80% of the nitrogen reaching the FGB in recent decades. The highest riverine contributions of the past 270 years occurred in 2016 and 2022–2023, coinciding with disease outbreaks and thermal stress events. These findings suggest that the combined effects of unprecedented nutrient loading and extreme heat are key synergetic drivers of the recent deterioration in coral health observed for this historically resilient reef system. INTRODUCTION The marine nitrogen (N) cycle is a fundamental component of ocean biogeochemistry, underpinning the productivity of marine ecosystems and the biological carbon pump (1). Although most oceanic N exists as biologically inaccessible dissolved dinitrogen gas (N2), the smaller pool of biologically available or “fixed” N is continually renewed through N2 fixation—the conversion of N2 to organic N by specialized diazotrophic microorganisms—and is removed predominantly through denitrification, which reduces nitrate to N2 during respiration under oxygen-deficient conditions (2). Once formed, fixed N participates in a dynamic suite of internal transformations: Nitrate is delivered from the subsurface to the euphotic zone, assimilated by phytoplankton, regenerated through recycling, and exported as organic matter to depth, where remineralization and nitrification convert it back to nitrate. Together, these coupled processes govern the spatial and temporal distribution of fixed N throughout the ocean (2, 3). Human activities are increasingly influencing Earth’s biogeochemical cycles (4, 5). As a case point, human activities have altered both the input-output budget and the internal cycle of fixed N (6). Global inputs of anthropogenically fixed N (∼210 Tg N year−1) now exceed those from natural sources (∼203 Tg N year−1), with substantial fluxes delivered to the ocean via rivers (50 to 70 Tg N year−1) and atmospheric deposition (∼30 Tg N year−1) (7, 8). The inputs to the ocean raise urgent questions about long-term consequences for marine ecosystems, especially those at the ocean margins (9). Sustained declines in coral reef diversity and coral cover highlight the need to identify the key factors and mechanisms underlying reef degradation (10, 11). Long before recent marine heatwaves focused global attention on thermal stress, coral reef science had already identified local anthropogenic pressures, such as nutrient or sewage pollution, sedimentation, and overfishing as major drivers of reef decline (12–15). More recently, rising sea surface temperatures (SSTs) have emerged as a dominant global stressor and the principal driver of increasingly frequent mass bleaching events (16). The coral-algae symbiosis reacts sensitively to warming and can break down under greater-than-average summertime warming (e.g., “degree heating weeks,” the number of weeks >1°C above the maximum monthly mean SST), resulting in coral bleaching (17). However, bleaching susceptibility and outcome depends on local environmental conditions (18). In particular, nutrient pollution can modify coral physiological condition, disrupt holobiont nutrient balance, and impair recovery, thereby increasing the likelihood that thermal stress results in severe or fatal bleaching (19). The interaction between rising SSTs and local stressors such as nutrient enrichment may therefore help to explain spatial differences in bleaching severity across tropical and subtropical coral reefs in recent decades (10, 19–21). Identifying why some coral reefs succumb rapidly to these stressors, whereas others appear comparatively resistant is critical for understanding the drivers of reef persistence (18, 22). The Flower Garden Banks (FGB) National Marine Sanctuary (NMS) provides a useful case study in this context because it is an apparent exception to the broader pattern of reef decline in the wider Caribbean (23). The FGB reefs are located in the open ocean on the edge of the Gulf’s wide continental shelf, starting at a water depth of ∼20 m, situated on top of salt domes rising from the seafloor, ∼200 km south of where the Texas-Louisiana border meets the coast (Fig. 1) (24). With over 50% coral cover, the FGB is considered the healthiest reef in the US continental margin (25). The high coral cover of the FGB reefs is often attributed to their isolated location (e.g., low fishing pressure) and to minimal exposure to land-based pollution, such as excess nutrients or sediment (26, 27). However, it remains unclear to what extent nonpoint source nutrient runoff from the Mississippi River system, including the Atchafalaya and Red Rivers, has influenced these reefs over time and how such anthropogenic impacts compare to natural variability in nutrient availability. Map showing the locations of the coral cores from the FGB and Cayo Santa Maria (Cuba) along with the main surface currents of the Wider Caribbean Region (WCR). The extent of the Mississippi River drainage is indicated across the continental United States, including the positions of the Mississippi River and Atchafalaya River. Approximate current positions are labeled as follows: CC, Caribbean Current; AC, Antilles Current; LC, Loop Current; MC, Mexican Current; GS, Gulf Stream. Note the presence of the coastal counter current opposite to the MC in the northwestern Gulf. This question is particularly relevant because the Mississippi and Atchafalaya Rivers’ discharge has profoundly altered biogeochemical conditions in the northern Gulf, where increasing inputs of N and phosphorus (P) since the mid-20th century have contributed to large seasonal algal blooms, coastal hypoxia, and disease outbreaks among shelf-dwelling marine invertebrates (28, 29). Although these impacts are most clearly expressed on the northern shelf, they highlight the broader capacity of river-derived nutrient loading to alter Gulf-wide ecosystems and motivate investigation of whether remote reefs such as the FGB have been affected through time. In response to the well-documented coastal consequences, the Watershed Nutrient Task Force (WNTF) launched a major hypoxia-reduction strategy in 2001, with the goal of decreasing Mississippi River N discharge by 60% and shrinking the hypoxic zone to under 5000 km2 by 2015 (30, 31). Despite substantial spending on conservation efforts, such as wetland restoration and nutrient management, the 2015 hypoxic zone exceeded WNTF’s target by threefold (30). Persistent gaps in long-term nutrient monitoring across the Gulf continue to constrain integrated coastal management efforts aimed at resolving how land-based nutrient loading affects coastal and offshore ecosystems over time. In this context, paleoenvironmental archives can provide critical complementary insight. Coral skeletal archives are well suited to complement ongoing monitoring as coral-based geochemical proxies record a range of environmental parameters (32, 33). The stable isotopes of N in organic matter bound within the biomineral structure of symbiont-bearing corals {expressed as CB-δ15N = [(15N/14N)sample/(15N/14N)air − 1] * 1000 in per mil (‰)} records processes in the marine N cycle (34–38), even in corals that are hundreds of millions of years old (39–41). In regions characterized by high N2 fixation rates, such as the western North Atlantic, N2 fixation introduces newly fixed N into the ocean with a δ15N value of around −1‰ (42), causing the δ15N of shallow subsurface nitrate to be lower than the mean global pycnocline nitrate δ15N (∼6.2‰) (43–45). Conversely, in oxygen-deficient zones in the water column, 14N-nitrate is selectively removed through water column denitrification, thereby increasing the δ15N of the ocean nitrate pool relative to the δ15N of newly fixed N (42, 43). On continental shelves, sedimentary denitrification tends to consume nitrate completely within the sediment pore waters, preventing 15N-enriched residual nitrate from escaping into the overlying water column (46–48). As a result, sedimentary denitrification typically has minimal direct impact on ocean nitrate δ15N (46, 49). In coastal systems, N isotope dynamics are further shaped by anthropogenic N inputs. Fertilizer derived ammonium and nitrate tend to be produced with low-δ15N N [commonly around −3 to +2‰ (50)] but usually arrive in the ocean with elevated δ15N values through ammonium oxidation coupled to denitrification of N sources along riverine and estuarine pathways (51), and groundwater, human sewage, and manure also exhibit higher δ15N values [commonly between +10 to +20 ‰ (50)], mostly due to denitrification. In addition, anthropogenic atmospheric N deposition tends to lower the fixed N in local surface waters δ15N (52). Together, these natural and anthropogenic processes leave isotopic fingerprints, thereby allowing for identification (45), rate estimation (53), and reconstruction (54) of N sources through CB-δ15N measurements (34, 35, 51, 55, 56). To assess these influences over historical timescales, we measured annually resolved CB-δ15N from two Siderastrea siderea corals in the West FGB reef (57), which together record the time period 1753 to 2023, predating and encompassing early European settlement in the Mississippi Basin and the rise of industrial agriculture. We compare these data to a contemporaneous, predominantly naturally forced CB-δ15N record from Cayo Santa María, northern Cuba (1777 to 2015) (Fig. 1) (56). tThe data provide in situ evidence for long-term changes in N delivery to the Gulf, revealing the onset and magnitude of anthropogenic alteration of the N cycle and its consequences for coastal ecosystem health. RESULTS Average measured CB-δ15N values of the coral record range between 1.87 and 6.88‰ for the two regions analyzed here (Fig. 2A). The lowest average values were recorded at the Cuban site (2.85 ± 0.41‰, n = 239), whereas the highest averages are found at the West FGB (4.99 ± 0.52‰, n = 279). The average CB-δ15N values from Cuba align well with previously measured mean wider Caribbean thermocline nitrate δ15N values (3.28 ± 0.76‰, n = 48) (56) and show multidecadal variability (Fig. 2A). Before 1850, CB-δ15N values from the West FGB are consistently higher by sometimes more than 2‰ but exhibit similar multidecadal variability. This higher δ15N may have reflected natural coastal N inputs or a mixing of shallow-thermocline nitrate, both of which tend to be elevated in δ15N (58). Measured Mississippi NO3−-δ15N values (7.50 ± 1.20‰) (59–61) are higher than average Gulf NO3−-δ15N values (4.30 ± 0.50‰, n = 16). Gulf seawater NO3−-δ15N values are intermediate between those measured from the Sargasso Sea (lower NO3−-δ15N) and those from the South Atlantic (higher NO3−-δ15N), consistent with the Gulf reflecting mixing between these two source-water endmembers (58). Coral-bound nitrogen isotope values of the three coral cores. (A) Coral-bound nitrogen isotope values (given as CB-δ15N in ‰ versus air) were measured from the FGB reef and Cayo Santa Maria in northern Cuba shown with the respective average and range of nitrate-δ15N values for the WCR, the Gulf, and the Mississippi River outflow. The two records are indicated in different colors and show considerable overlap. (B) Normalized CB-δ15N values {[(CB-δ15N − average CB-δ15N)/(SD of CB-δ15N)] and are given as z-scores} of the two West FGB cores as a composite record and the core from Cuba, used to identify the largest deviations from natural variability in the FGB. West FGB CB-δ15N values before 1850 (4.42 ± 0.22‰, n = 93) are statistically indistinguishable from average Gulf NO3−-δ15N values (Welch’s t test, P = 0.359). Between 1850 and 1950, the West FGB CB-δ15N increased significantly (4.84 ± 0.28‰, n = 84, Welch’s t test, P = 0.0006) and diverged even further after 1950 (5.54 ± 0.36‰, n = 74), showing a highly significant difference from average Gulf NO3−-δ15N (Welch’s t test, P = 1.77 × 10−8, n = 121). The West FGB CB-δ15N increases starting around 1850 had no clear parallel in the Cuban CB-δ15N record (Fig. 2A). Disentangling shared natural variability from anthropogenic effects in our records was pursued by normalization for the former. In this approach, normalized CB-δ15N = [(CB-δ15N − average CB-δ15N)/(SD of CB-δ15N)], yielding a z-score, which was smoothed using a Gaussian filter (see Materials and Methods). The records reveal shared and repeated oscillations that predate major anthropogenic N inputs and persist throughout the 18th and 19th centuries (Fig. 2B). Spectral characteristics of these fluctuations align with ∼64-year periodicities that are well documented for the Atlantic Multidecadal Variability (AMV) and associated hydrographic variability across the wider Caribbean (40, 62) (fig. S1). The normalized record from the West FGB was then compared against the normalized CB-δ15N from Cayo Santa Maria in Cuba (1777 to 2015) to evaluate the timing and magnitude of anthropogenic effects relative to natural variability over the past 270 years (Fig. 2B). Across the instrumental era, the West FGB CB-δ15N covaries with estimates of US N use (adj. r2 = 0.647, P < 0.01) and also covaries with previously reconstructed Mississippi nitrate loads estimated by incorporating direct measurements into a model of changing pigment concentrations (i.e., zeaxanthin as a proxy for cyanobacterial dominance) in coastal sediments (adj. r2 = 0.593, P < 0.01) (Fig. 3 and fig. S2) (31, 63, 64). Together with the established linkage between Mississippi nitrate loads and Gulf ecosystem response, these patterns support interpretation of the West FGB record as a two-endmember mixture of a relatively low-δ15N open-ocean source and a higher-δ15N riverine outflow, with the fractional contribution of the Mississippi endmember increasing through the 20th and early 21st centuries (64, 65). FGB CB-δ15N tracks anthropogenic and riverine N inputs to the Gulf. (A) Time series (1754 to 2023) of coral-bound δ15N at the West FGB plotted along US N use and modeled Mississippi nitrate (NO3−)-N load. Shaded horizontal bands denote approximate δ15N endmember ranges for the Gulf open ocean (lower band) and the Mississippi outflow (upper band). (B) Scatterplots over the instrumental overlap showing CB-δ15N versus annual US N use (C) and versus modeled NO3−-N load. Solid lines are least-squares fits with 95% confidence envelopes (gray), and the coefficient of determination (R2) is indicated. Modeled nitrogen loads are taken from David et al. (63). Applying these two isotopic endmembers to the West FGB CB-δ15N time series reveals that the fractional contribution of Mississippi-derived nitrate was minimal before 1850, averaging only 3.1 ± 6.5% (n = 93), indistinguishable from zero given endmember uncertainties. Between 1850 and 1950, the Mississippi fraction increased significantly to 28.0 ± 12.5% (n = 84), reflecting the early stages of enhanced lateral N export from the watershed. After 1950, the fractional contribution rose sharply to 50.4 ± 17.5% (n = 74), a highly significant increase relative to the pre-1850 baseline (Welch’s t test, P < 0.001). In the modern period (2000 to 2023), Mississippi-derived N accounts for 64.2 ± 22.1% of the modeled mixture (n = 24), with peak values exceeding 60% in the most recent decade (Fig. 4A). This progressive shift toward higher riverine influence mirrors the documented intensification of agricultural fertilizer use and nitrate export from the Mississippi basin. The results indicate that the anthropogenic N signal is now the dominant driver of CB-δ15N variability recorded at the West FGB. Time series of the modeled fractional Mississippi contribution at the FGB (in % for 1754–2023). (A) Values are derived from a two-endmember isotope mixing model using a lower-δ15N Gulf nitrate endmember (4.30 ± 0.50‰) and a higher-δ15N Mississippi River nitrate endmember (7.50 ± 1.20‰). The solid line shows the annual fraction (light to dark brown denotes increasing Mississippi influence); the shaded envelope is the propagated uncertainty from endmember SDs. Mississippi contributions remain near zero before ∼1850, rise in the early 20th century, accelerate after ∼1950, and exceed 60% in the early 21st century, consistent with intensified watershed N export. (B) Fractional Mississippi contribution plotted with DHW at the FGB. Horizontal lines mark the 4° and 8°C-weeks thresholds. The 2016 and 2022 disease outbreaks and the 2022–2023 bleaching events coincide with periods of both elevated Mississippi influence and exceedance of 8°C-weeks, indicating compound riverine and thermal stress on the reef. At the West FGB, degree heating weeks (DHW) remained low to moderate through most of the coral record, with infrequent exceedance of the bleaching threshold of 4°C-weeks before the 2000s. Beginning in the 2010s, both the frequency and magnitude of heat stress events increased. Notably, the DHW exceeded the severe bleaching benchmark of 8°C-weeks during the 2016 and 2022–2023 warm seasons. When superimposed on years of heightened Mississippi-derived nutrient delivery (>60%), these high-heat years coincided with outbreaks of coral disease (2016 and 2022) and bleaching events at the FGB (2022–2023) (Fig. 4B) (66, 67). DISCUSSION Mississippi N inputs to the coral reef The elevated δ15N baseline of nitrate in the Gulf of Mexico provides an important context for interpreting Mississippi-derived N inputs to the FGB. In contrast to the Caribbean, where shallow thermocline nitrate is relatively depleted in 15N owing to the remineralization of sinking N derived from N2 fixation, the Gulf nitrate pool is more 15N enriched. Caribbean thermocline nitrate averages ∼3.3‰, lowered from equatorial Atlantic thermocline values of ∼5.74‰ by regional N2 fixation and subsequent remineralization, whereas Gulf nitrate δ15N is elevated at ∼4.3‰ (44, 53, 68). This enrichment likely reflects a combination of boundary-driven vertical mixing of higher-δ15N nitrate from the deeper pycnocline, continental-shelf N loss, and riverine N inputs (61, 69). Thus, the Gulf differs from the broader tropical North Atlantic and Caribbean, where warm, stratified, nitrogen-poor but iron-rich surface waters favor N2 fixation and where phosphorus supplied from the South and equatorial Atlantic supports diazotrophy (70–72). Although large tropical rivers such as the Amazon and Orinoco influence nutrient and biogeochemical gradients across parts of the western tropical Atlantic and Caribbean (73), the FGB are more directly affected by Gulf-specific nutrient cycling and Mississippi–Atchafalaya River inputs. This regional contrast is critical for distinguishing the background isotopic composition of Gulf nitrate from additional anthropogenic N signals exported from the Mississippi watershed. Our CB-δ15N record from Cuba averages 2.85 ± 0.41‰, closely matching the mean Caribbean thermocline nitrate δ15N value of 3.28 ± 0.76‰ across the coral record (56) (Fig. 2A). In contrast, CB-δ15N values from the West FGB are consistently higher, averaging 4.99 ± 0.52‰. This offset is consistent with modern thermocline nitrate δ15N measurements from the respective regions, indicating that the Cuban and West FGB coral records reflect regional nitrate baselines rather than a shared Caribbean-Gulf endmember. Thus, the higher CB-δ15N values at the West FGB are expected from the higher regional nitrate δ15N baseline of the northern Gulf and provide the appropriate starting point for evaluating subsequent temporal changes at the site (Fig. 2A). Despite N isotopic differences between the Caribbean Sea (3.28 ± 0.76‰) and the Gulf (4.30 ± 0.50‰), normalized CB-δ15N records from the two regions can reveal shared variability. Our CB-δ15N data from the West FGB and Cuba indicate a shared multidecadal variability with dominant periodicities of ∼64 years, which is characteristic of the AMV (Fig. 2B and fig. S1). The AMV describes multidecadal SST variability across the North Atlantic (74, 75), influencing atmospheric and oceanic circulation patterns, the position of the ITCZ, and associated shifts in surface wind strength over the Atlantic (76). The impact of AMV on the wider Caribbean region has been reconstructed using SST proxies from coral cores (62, 77), speleothems (78), and sediment cores (79). Recent data show that AMV-driven changes in CB-δ15N of the wider Caribbean are best explained by changes in N2 fixation (56) (Fig. 2A). However, the West FGB CB-δ15N record diverges from its natural variability at the time that agricultural practices expanded following European settlement across the Mississippi basin (Figs. 2B and 3A). Significant settlement and widespread agricultural developments occurred within the Mississippi Basin over the past 270 years, which have substantially altered the river’s outflow regimes and nutrient loads (80, 81). Subsistence farming in the 1700s gradually transitioned to mechanized agriculture in the 19th century, accompanied by the development of export economies along and from the Mississippi River. The basin saw a shift toward industrial-scale farming with the Green Revolution starting in the 1960s, as agricultural practices emphasized monocultures (notably corn and wheat) and introducted genetically modified crops, thereby transforming the Mississippi Basin into a major agricultural hub (29). From 1753 to ∼1850, our record from the West FGB shows low CB-δ15N values (4.43 ± 0.22‰) that are similar to Gulf seawater NO3−-δ15N values (4.30 ± 0.50‰) (58) and are consistent with little to no direct impact from riverine N loads, beyond their possible longer-term role in elevating the δ15N of the Gulf’s thermocline nitrate relative to that entering from the Caribbean Sea. Early agricultural practices in the region did not involve large nutrient inputs, relying on natural or farm-derived sources, such as compost and manure. Our data suggest that during this initial stage, nutrient leakage from soil into riverine waters was minimal, exerting little to no influence on nutrients in the Gulf beyond natural inputs (Fig. 4A). In 1856, the US Congress passed the Guano Islands Act, under which the United States claimed 103 islands, rocks, and cays in the Pacific and Caribbean for guano extraction (82). This act marked the start of widespread nonlocal fertilizer use in the greater Mississippi basin. The West FGB record shows an increase in CB-δ15N values (+1‰) after 1856, suggesting that the introduction of guano fertilizers had a measurable impact on the N load to the Gulf (Figs. 3A and 4A). This guano-driven agricultural expansion signal strengthens after the complete removal of the Great Raft, a large log dam that persisted for many centuries along the Red River just south of present-day Shreveport, Louisiana in the mid-1870s (Fig. 4A) (83). The removal caused a fundamental change in the region, reducing inland flooding due to clogging of the Red River, thus allowing more direct riverine flow of the Mississippi and Atchafalaya Rivers into the northern Gulf (84), which would have reduced the time for algal nutrient uptake and dilution in coastal waters. In addition, extensive levee construction along the Mississippi River started in 1882, controlling river floods and channeling the water, along with its associated sediment and chemical load, more directly and, thus, rapidly to the Gulf (85). With the onset of the Green Revolution in the 1960s and the rise of industrial-scale farming, CB-δ15N values increased by an additional 2‰, approaching modern Mississippi NO3−-δ15N values (59, 60) and ultimately matching them in 2022 and 2023 (Fig. 3A). Our isotope data confirm that the Mississippi River system has become the primary source of nutrients to the northern Gulf (65, 80), with a rapid increase since the 1960s. Today, the Mississippi River basin drains ∼41% of the land area of the continental United States, extending westward to Idaho, northward into Canada, and eastward into New York State (Fig. 1). The river’s discharge is regulated, with 30% flowing through to the Atchafalaya River delta and 70% to the Mississippi River birdfoot delta, with about 53% of this discharge moving westward with the coastal currents along the Louisiana-Texas shelf, where mixing allows the N-rich waters to reach the FGB coral reef ecosystems ∼130 km from the mouth of the Mississippi River (Fig. 1). Increasing synchronization between Mississippi and reef N inputs The two phases of agricultural intensification are associated with qualitatively different dynamical responses in the land-reef coupling. Following the late-19th-century expansion, the lead-lag relationship shifts toward positive values, indicating that increases in terrestrial N inputs precede the reef δ15N response by several years (fig. S2). This pattern is consistent with a phase of N accumulation within the watershed and its gradual downstream propagation to the coastal ocean. In contrast, during and after the mid-20th-century Green Revolution, the lead–lag relationship converges toward zero, implying a more synchronous response between terrestrial forcing and reef isotopic signals (fig. S3). This convergence is consistent with a transition toward faster hydrological and biogeochemical throughput, in which intensified fertilizer application, drainage, and river regulation reduce storage times and promote more immediate flushing of nutrients from land to the Gulf. Our fractional endmember mixing analysis further supports this narrative. Before the 20th century, Mississippi-derived N accounted for only a small percentage of the N reaching FGB, consistent with predominantly Gulf-wide conditions. Beginning around 1850 and accelerating after the 1950s, however, the Mississippi fraction increased sharply, reaching 30 to 50% by the late 20th century and exceeding 60% in the past decade (Fig. 4A). This long-term trend documents a growing dominance of river-derived N in offshore Gulf waters, marking a clear departure from the natural multidecadal variability that characterized the preindustrial era (Fig. 3B). The close correspondence between rising fractional contributions and the post-1960s increase in CB-δ15N demonstrates that anthropogenic nutrient delivery from the Mississippi basin—not internal oceanic processes—has become the principal driver of N supply to the northern Gulf. This offshore propagation of watershed-derived N highlights the extensive spatial footprint of continental nutrient loading and the degree to which modern coral reef ecosystems, even those located more than 180 km from shore, are influenced by intensified agricultural production across the Mississippi Basin. Although agricultural runoff is widely recognized as the dominant contributor to basin-scale N export from the Mississippi–Atchafalaya River Basin, municipal wastewater from this densely populated watershed, which includes nearly 30% of the US population, also represents a measurable anthropogenic N source that may contribute to the integrated downstream signal exported to the Gulf. The U.S. Environmental Protection Agency assessments indicate that major sewage treatment facilities account for roughly 8 to 9% of total N loads delivered from the basin to the Gulf. However, this estimate does not include N inputs from the millions of septic systems distributed throughout the watershed, which remain difficult to quantify at the basin scale but would further increase the total contribution of human waste-derived N. Thus, although wastewater and septic sources are not the dominant driver of basin-scale N export, they are not negligible and likely add to the downstream anthropogenic N signal alongside the larger nonpoint inputs associated with agricultural surplus. Together, these observations suggest that the N signal exported to the Gulf is not attributable to a single source but instead reflects the combined influence of dominant nonpoint agricultural inputs, smaller yet measurable wastewater contributions, and long-term hydrologic modifications that have altered nutrient routing and storage of legacy N within the highly engineered Mississippi watershed. N inputs and coral disease dynamics The timing of the two most recent large-scale disease outbreaks (2016 and 2022) and bleaching events (2022 and 2023) aligns with the periods of >60% riverine influence that is unprecedented over the past 270 years, as well as with a breach of 8°C DHW (Fig. 4) (27, 67). Conversely, during years when only the threshold of 8°C-weeks was breached (e.g., 2010) or was nearly reached (e.g., 2005), coral bleaching was observed at the FGB, but widespread disease outbreaks were not recorded (Fig. 4B). A central mechanism linking nutrient enrichment to coral bleaching susceptibility and increased disease outbreak risk is the DDAM [dissolved organic carbon (DOC)–disease–algae–microorganism] feedback model, which describes how elevated nutrient inputs promote shifts in benthic and microbial community structure that amplify disease risk (12, 86, 87). Under nutrient enrichment, turf and fleshy algae increase in abundance and release substantially higher amounts of bioavailable DOC per unit surface area than corals (88). This algal-derived DOC fuels copiotrophic bacterial growth, increasing the prevalence of microbial taxa and virulence-associated genes linked to coral disease (89). These microbes can directly harm corals by creating localized zones of prolonged hypoxia and elevated microbial respiration, leading to tissue damage and mortality (90). The resulting loss of coral cover further promotes algal expansion, reinforcing a positive feedback loop of microbialization that is increasingly recognized as a key mechanism underlying contemporary coral disease and bleaching at local and regional scales (87). Multiple studies support this interpretation, demonstrating that elevated nutrient availability and high temperatures act synergistically to increase coral susceptibility to bleaching and disease (19, 20, 91). Nutrient enrichment, especially nitrate, alters symbiont density, photosynthetic efficiency, and host-symbiont metabolic exchange, potentially shifting the relationship from mutualistic toward parasitic under stress (92–95). High N:P ratios have been shown to disrupt host-symbiont nutrient balance, reduce photosymbiont performance, and impair carbon translocation to the coral host (91, 94, 96, 97). Implications for coral reef management Measured riverine N:P ratios from the Mississippi (N:PMississippi = 22.5 ± 3) have been above the canonical “Redfield ratio” of 16 since the 1970s (65). When these high-N waters reach typically oligotrophic and often phosphorus-poor reef environments offshore, they can exacerbate nutrient imbalances and increase susceptibility to microbialization, disease, and bleaching. Although riverine N:P ratios have remained relatively constant, the absolute magnitude and spatial extent have increased and expanded the area influenced by river-derived nutrients (91, 94, 98). Although Mississippi-derived plume waters are buoyant, their influence on the FGB need not require direct sinking of low-salinity surface water to reef depth. Instead, episodic cross-shelf transport can advect plume waters offshore as a shallow surface layer above the banks. Laterally supplied organic matter and associated remineralization then transfer N to subsurface reef waters below. This mechanism is consistent with observations of a ∼20-m brackish surface layer over the FGB and near-hypoxic subsurface waters immediately beneath it (66, 99–100). This mechanistic interpretation is further supported by independent ecological observations from the Rapid Climate Vulnerability Assessment for the FGB National Marine Sanctuary, which document an increase in turf algal cover in recent years (101). Turf algae are a key functional group in the DDAM framework as they release disproportionately high amounts of labile DOC that fuel copiotrophic microbial growth and enhance pathogen pressure on corals. The observed rise in turf algae therefore provides in situ evidence that nutrient enrichment is not only altering water column chemistry but also restructuring benthic community composition in ways that increase disease likelihood in corals. Disease outbreaks and bleaching events may therefore become more frequent if N loading from the Mississippi basin continues unabated. More broadly, our findings highlight the value of adopting a long-term paleo-oceanographic perspective in coastal ecosystem management, including the integration of spatially distributed coral archives into monitoring frameworks to better distinguish natural variability from anthropogenic change. MATERIALS AND METHODS Coral locations and fieldwork at the FGB, core processing, and annual sampling are described by DeLong et al. (57). Core 23WFGB1 was collected in August 2023 by FGBNMS staff at a water depth of 23 m near the 05WFGB coral and spans the time interval from 1978 to 2023. Core 05WFGB1 spans the interval from 1753 to 2004 for N isotope analysis (Fig. 2A). The coral chronology is accurate to the year back to 1932 and then chronology uncertainty can be ±1 to 2 years per century due to core breaks (57). Analysis of coral-bound N isotopes The CB-δ15N measurements were performed in the Martínez-García Lab at the Max-Planck Institute for Chemistry (MPIC) in Mainz. We used the persulfate oxidation–denitrifier method (45), first applied to corals by Wang et al. (35, 102), with the analytical modifications described by Moretti et al. (103). Five ±1 mg of coarse coral powder was weighed into a 4-ml VWR borosilicate glass vial, filled with 4.25 ml of sodium hypochlorite (NaOCl) and left on a shaker at 120 rpm for at least 24 hours. NaOCl was removed the next day with a precombusted glass pipette attached to a vacuum line set at 500 mbar. Samples were then rinsed three times with 4 ml of Milli-Q (18.2 MΩ cm−1 at 25°C) and left to dry at 60°C overnight. Once fully dried, six ±0.2 mg of powder was weighed inside an in-house clean room to minimize the possibility of contamination. Thereafter, skeletal-bound organic matter was released by dissolving the final amount of material with 45 μl of 4 M HCl. Concurrently, a persulfate oxidative reactant (POR) solution was prepared inside the clean room with a 4-ml spike of 6.25 N NaOH to reach a high pH. One milliliter of POR solution was added to each dissolved sample and to a minimum of 12 empty, cleaned vials serving as blanks. The entire batch was then placed in a custom-built sample rack that was tightly sealed with a PTFE sheet before being autoclaved at 121°C for 65 min (103). A volume of 1 ml of concentrated denitrifying bacteria (Pseudomonas chlororaphis) was injected into 800 ml of growth media and left 4 to 6 days to grow in the dark at room temperature on a shaking rack. Once the bacteria have grown sufficiently, the medium was transferred to autoclaved PE bottles and centrifuged at 7600 rpm for 10 min. The supernatant was then discarded, and the remaining bacteria pellet was resuspended with a buffered (pH 6.3) resuspension medium. From this, 3 ml was pipetted into muffled 12-ml glass vials, capped with a septum and tightly sealed before being placed upside-down on a needle rack with a small extra needle for pressure release. The needle rack supplies a continuous flow of N2 for at least 3 hours to replace the internal atmosphere and dissolved gases with pure N2. Bacteria vials were removed from the rack, and ∼0.5 ml of the oxidized sample was injected into each bacteria vial. Once injected, bacteria vials were placed in the dark for 2 to 3 hours to ensure quantitative transformation of nitrate to nitrous oxide before being frozen at −21°C. On the day of analysis, bacteria were thawed, lysed with at least five drops of 10 N NaOH and placed on a mass spectrometer for isotope analysis. The δ15N of the N2O was determined by a purpose-built inlet system coupled to a Thermo Fisher Scientific MAT 253 Plus stable isotope ratio mass spectrometer (104, 105). Long-term precision was determined by running internal carbonate standards with each sample batch, which yielded an average carbonate standard-δ15N reproducibility of ±0.2‰. Fractional Mississippi River contribution To estimate the changing contribution of riverine N to surface-ocean nitrate in the northern Gulf, we applied a simple two-endmember isotope mixing model using the FGB δ15N time series. The model solves for the fractional Mississippi River contribution fMS based on the isotopic composition of a low-δ15N open-ocean endmember (δ15Nopen = 4.3 ± 0.5‰) and a higher-δ15N Mississippi watershed endmember (δ15NMS = 7.5 ± 1.2‰). The fractional contribution was calculated asfMS=δ15Nmix−δ15Nopenδ15NMS−δ15Nopenwhere δ15Nmix represents the annual FGB coral-bound δ15N measurements. Uncertainty in fMS was propagated from the uncertainties in the isotopic endmembers, using σΔ=(1.2)2+(0.5)2=1.30‰, corresponding to a 41% fractional uncertainty in the denominator. The resulting propagated uncertainty in fMS was estimated as σfMS=0.41·fMS. This approach allows temporal variability in FGB δ15N to be interpreted as changes in the relative balance between open-ocean nitrate and isotopically enriched riverine nitrate associated with anthropogenic N loading to the Mississippi watershed. Degree heating weeks Thermal stress at the FGB (∼27.9°N, 93.8°W) was assessed with NOAA Coral Reef Watch 5-km SST/DHW products (v3.x). The DHW is the 12-week accumulation of daily SST anomalies above the site’s maximum monthly mean. For Fig. 4B, we used the annual maximum DHW per year and referenced standard bleaching thresholds of 4° and 8°C-weeks. Statistical analysis CB-δ15N values were imported to a Python3 Jupyter Notebook (version 5.7.4) using the Pandas library. Data were plotted with Seaborn/Matplotlib and reprocessed for wavelet analysis according to the waipy script (https://github.com/mabelcalim/waipy). The continuous wavelet transform (CWT) significance test was based on Torrence and Compo (1998), and the cross wavelet analysis (CWA) was based on Maraun and Kurths (106). To compare natural variability, time series were normalized {normalized CB-δ15N = [(CB-δ15N − average CB-δ15N)/(SD of CB-δ15N)]} and smoothed with a Gaussian filter using the SciPy package (version 1.11.2). Correlations are expressed as adjusted r2. Unlike r2, adjusted r2 increases only if the new predictor enhances the model more than would be expected by chance. It can also decrease if a predictor improves the model by less than expected by chance. SDs of averages are given as ±1 SD. For the calculated Mississippi N load fraction, errors were propagated with a Mississippi endmember uncertainty of ±1.2% and a Gulf open-ocean endmember uncertainty of ±0.5%.