Temperature dependence and coupling of dissolved oxygen and pH in coastal habitats
Figures
Abstract
Coastal waters experience daily fluctuations in oxygen (O2), carbon dioxide (CO2), and pH driven by respiration and photosynthesis, processes that are highly sensitive to temperature variations. As global temperatures continue to rise, both photosynthetic and respiratory rates are increasing, with respiration accelerating more rapidly. The resulting decline in photosynthetic oxygen production and CO2 uptake relative to respiratory oxygen demand and CO2 release is leading to widespread declines in oxygen availability and heightened acidification of coastal ecosystems, particularly in warmer regions. Here, we investigate the variability and co-variance of O2 and pH in diverse coastal habitats and their temperature dependence, utilizing a comprehensive dataset of 516,705 data points spanning a temperature gradient of 1–39°C. Our in situ data capture a wide range of ecosystems, including seagrass meadows, micro-, macro-, and filamentous algae, mangrove forests, kelp forests, coral reefs, and bare sediments, from polar to tropical regions across both hemispheres. We show that rising temperatures are associated with decreases in O2 and pH, alongside amplified diel fluctuations in these parameters, increasing the prevalence of daytime hypoxia and hyperoxia. Furthermore, our data reveal a strong correlation between O2 and pH, driven by photosynthetic and respiratory activity, with temperature and across diverse ecosystems and geographic regions. These findings are also relevant for the design of laboratory experiments focused on temperature stressors or simulations of marine heatwaves and emphasize the need for establishing localized baseline data on environmental conditions and water chemistry. The temperature-dependent relationship and co-variation between dissolved oxygen (DO) and pH in metabolic hotspots, such as coastal vegetated ecosystems, underscore the importance of diurnal variability in shaping DO and pH thresholds critical for supporting biodiversity.
Citation: Steckbauer A, Hendriks IE, Vaquer-Sunyer R, Saderne V, Anton A, Krause-Jensen D, et al. (2026) Temperature dependence and coupling of dissolved oxygen and pH in coastal habitats. PLoS One 21(11): e0362740. https://doi.org/10.1371/journal.pone.0362740
Editor: Amitava Mukherjee, VIT University, INDIA
Received: November 6, 2025; Accepted: October 28, 2026; Published: November 20, 2026
Copyright: © 2026 Steckbauer et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are provided in the manuscript and its Supporting Information files.
Funding: This research is a contribution to the ‘Water bodies in Europe: Integrative Systems to assess Ecological status and Recovery (WISER)’ project, funded by FP7 (contact number #226273) and project MEDEICG, funded by the Spanish Ministry of Science and Innovation (number CTM2009-07013), project COOPB20366, “Fortalecimiento de las capacidades científico-técnicas para la gestión de los sumideros de carbono en pastos marinos del Caribe Colombiano”, and King Abdullah University of Science and Technology through baseline funding to CMD (Bas/1/1071-01-01) and SA (BAS/1/1072-01-01). AS was funded by a fellowship from the Government of the Balearic Islands (Department on Education, Culture and Universities) and the EU (European Social Fund) as well as King Abdullah University of Science and Technology. IEH was funded by an RCA – Visitors program from the UWA to Ylva Olsen collecting the data in Australia. RVS was funded by the WISER project. DKJ acknowledges support from the Carlsberg Foundation (grant number CF15-0639) and from the Danish Environmental Protection Agency within the Danish Cooperation for Environment in the Arctic (DANCEA). The funder provided support in the form of salaries and travel stipends for authors A.S.. I.E.H., C.M.D and S.A., but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Rising global temperatures and the intensification of marine heatwaves (MHWs) are exerting complex and multifaceted impacts on coastal ecosystems. These impacts extend beyond simple temperature elevation, driving reductions in dissolved oxygen (O2) levels due to the decreased solubility of O2 in warmer waters and the divergence between increasing rates of respiration and production. This imbalance results in a decline in O2 and a concurrent rise in carbon dioxide (CO2) concentrations with increasing temperature [1]. The uptake of atmospheric CO2 further alters seawater chemistry by increasing bicarbonate ion (HCO3−) concentrations while simultaneously reducing carbonate ion (CO32−) availability by approximately 60%, thereby elevating both dissolved CO2 ([CO2]) and partial pressure of CO2 (pCO2) in the ocean [2]. This process contributes to decreasing seawater pH, a phenomenon referred to as ocean acidification (OA) [1]. Moreover, warming disproportionately accelerates nocturnal respiration relative to daytime net production [3], exacerbating temperature-induced reductions in oxygen levels—a process further intensified by reduced O2 solubility [4]. Over the past 60 years, global oxygen levels have already declined by approximately 2% [5], with projections indicating further declines of between 1% and 7% by the end of the century [6]. While ocean deoxygenation (OD) refers to a global decline in oxygen levels in the open ocean, primarily driven by climate change [7–9], coastal deoxygenation (CD) occurs at localized scales due to processes such as watershed inputs and metabolic activity [7] (e.g., Chesapeake Bay [10,11]).
The term ‘hypoxia’ traditionally refers to water masses with dissolved oxygen (DO) levels below a critical threshold, commonly set at 2.8 mg L−1 (equivalent to 2.0 ml L−1 [12,13]). However, the appropriateness of these thresholds has been increasingly debated, as evidence indicates that 90% of marine organisms experience adverse effects at DO concentrations below 4.59 mg L−1, which represents the median lethal concentration [14]. Higher diagnostic thresholds for ‘hypoxia’, such as 3.5 mg L−1, have also been proposed [15]. In parallel, acidification negatively affects vulnerable marine organisms when pH falls below 8.0 [16].
Coastal ecosystems, such as seagrass meadows, kelp forests, mangrove forests, and coral reefs, experience pronounced daily fluctuations in DO, pH, and CO2, primarily driven by their high metabolic activity. In the Red Sea, seagrass meadows can exhibit DO concentrations exceeding 200% saturation at daytime temperatures above 36°C, while dropping to 0% saturation at night when temperatures exceed 31°C [17]. Similarly, in mangrove forests, DO levels can reach 250% saturation during the day at temperatures surpassing 40°C but may decline to below 50% saturation at night when temperatures fall below 30°C [17]. Similarly, diel fluctuations of 0.2 pH and diel flucutations in excess of pCO2 of 100 ppm are often reported in metabolically-intense coastal waters [18].
While much of the existing research has focused on the negative impacts of deoxygenation and hypoxia on marine organisms, relatively little attention has been given to the phenomenon of oxygen supersaturation, or hyperoxia, until recently. The potential mitigating effects of hyperoxia during elevated temperatures or marine heatwaves (MHWs) remain largely unexplored. Hyperoxia naturally occurs on a diel cycle across various marine habitats [19,20], with extreme levels observed, including over 169.8% saturation in Mediterranean seagrass meadows [21], greater than 250% saturation in coral reefs [18], and above 200% saturation in macroalgae beds [3]. These diel fluctuations can potentially alleviate thermal stress, as hyperoxia may reduce mortality rates among marine species inhabiting coastal ecosystems during periods of elevated temperature stress [17,20], only if nightime hypoxia is not worsened. However, hyperoxia can also have adverse effects on marine organisms [21]. In vegetated coastal ecosystems, pH variability can range from 0.3 to over 1 pH unit, with macroalgae beds notably mitigating the adverse effects of acidification [22]. Such variability in coastal ecosystems is significantly greater than predicted changes for the open ocean by the end of the century. Likewise, the pCO2 variations in coastal ecosystems were already exceeding pCO2 levels for end of the century in 2008–2022 [23].
The impacts of concurrent acidification due to excess CO2 have often been overlooked in studies focused on deoxygenation. Despite the intricate link between O2 and CO2 concentrations in the atmosphere and biosphere—owing to the strong coupling of these gases in anthropogenic combustion and metabolic processes [24]—deoxygenation and acidification are typically examined in isolation. However, a recent meta-analysis demonstrated predominantly additive effects when these dual stressors are considered together [25], highlighting the critical need to address the co-variance of O2, CO2, and pH simultaneously. Thus, understanding the co-variance and temperature dependence of O2, CO2, and pH in coastal habitats is essential for predicting the effects of climate change on metabolic rates and the resulting diel oscillations in DO and pH within these ecosystems. We hypothesize that warming amplifies metabolic rates, thereby intensifying diel fluctuations of DO and pH in vegetated coastal ecosystems, and that these warming-induced patterns are evident when comparing such ecosystems along latitudinal gradients from the poles to the equator.
In this study, we investigate the variability of dissolved oxygen (O2) and pH across a wide temperature regime, diverse ecosystem types, ranging from polar to tropical regions. Using multiparametric sensors, we collected measurements in various geographic zones (polar, temperate, and tropical) and ecosystem categories, including coral reefs, seagrass meadows, macro-, micro-, and filamentous algae, mangrove forests, kelp forests, and bare sediments. Our data collection captured temperature, O2, and pH levels, among other parameters. Observational periods varied from several weeks to over a year, with measurement intervals ranging from 5 to 60 minutes, enabling a comprehensive assessment of the significant co-variation between O2 and pH.
Materials and methods
Study sites
Data collection was conducted across a range of locations, spanning from polar to tropical regions, and encompassing diverse ecosystem types, including seagrass meadows, macroalgal beds, micro- and filamentous algal mats, kelp forests, mangrove forests, coral reefs, and bare sediments (S1 Fig; detailed locations and deployment durations are provided in Supplementary S1 Table). The studied ecosystems included Posidonia oceanica and Caulerpa prolifera meadows in the Mediterranean Sea (incl [21,26].); Zostera marina seagrass meadows in the Baltic Sea; Enhalus acoroides seagrass meadows, mangrove forests (Avicennia marina), and coral reefs in the Red Sea (incl [18,27].); Saccharina longicruris or S. latissima kelp forests and micro- and filamentous algal mats in Greenland [28,29]; P. sinuosa, P. australis, and A. griffithi seagrass meadows in Australia; and Thalassia testudinum seagrass meadows in Colombia. For a comprehensive list of locations, species, and sampling intervals, refer to S1 Table.
Multiparametric sensors (Hydrolab DS5X, Exosonde 1, Exosonde 2, or EUREKA Manta) were deployed to measure dissolved oxygen (DO), temperature, conductivity, depth, salinity, pH,and/or CO2. Data were recorded at intervals of 5, 10, 15, 30, or 60 minutes, depending on the specific study site requirements. The type of instrument utilized at each sampling station is detailed in S1 Table, with sensor accuracy specifications for each CTD model provided in S2 Table. All sensors were calibrated both prior to and following deployments, in accordance with the manufacturer’s guidelines, to account for any potential measurement drift. All pH sensors had manufacturer-specified accuracies of ±0.1 or ±0.2 pH units. Pre- and post-deployment verification using a pH 7.00 reference buffer showed maximum deviations of 0.07 pH units, remaining within the manufacturers’ stated accuracy specifications (see S2 Table). To avoid pseudo-replication, a single measurement was recorded per time point. For deployments in the Mediterranean, Colombia, Australia, and Greenland, pH was measured against the pHNBS 7 before and after the deployment to detect any drift occuring during the measurement period. For the Red Sea, SeaFET instruments were used as part of the data collected in the Red Sea (see [27] for detailed procedures. CO2 measurements in the dataset collected in the Baltic Sea, where no pH data were available, were converted to pHTOT to be compareable to the rest of the data, using the Seacarb software package using Salinity, Temperature, CO2 and TA data available from the long-term water quality sampling program at the location of deployment (Kiel Bight).
Multiparametric sensors were mounted on metal support frames and secured to the seafloor. In the Red Sea, sensors were positioned 20–50 cm from seagrass meadows and within 5 cm of mangroves and coral reef tops. The sensors were strategically located at the interface between the benthic boundary layer and open water, in close proximity to biogenic habitats, to assess the influence of in situ metabolic activity on environmental parameters. In mangrove habitats, sensors were moored on the seaward fringe beneath the canopy, while in seagrass meadows, they were placed near the top of the seagrass canopy. Sensors were positioned horizontally in all locations except in coral reefs, where they were vertically attached to a coral wall on the back-reef side, and in the kelp forests of Greenland, where loggers were placed as close as possible to the kelp in various orientations.
In the Mediterranean Sea, Colombia, Australia, and the Baltic Sea, deployments were relatively short, ranging from days to weeks, with repeated measurements conducted at various study sites throughout the year in the Mediterranean. By contrast, data collection in the Baltic Sea and Red Sea spanned approximately one year. While some areas were explored more extensively, it is important to note that only one mangrove location was included in our dataset. This precludes any robust or general conclusion for mangrove ecosystems, which require further attention in the future. Deployment depths ranged from 1.0 to 15.5 meters, with measurement intervals varying from 5 to 60 minutes. For a complete list of deployments, measurement intervals, depths, and ecosystem types, refer to S1 Table. Further details on deployments by location are available in the relevant publications [3,18,21,27–30].
Data preparation and statistical analysis in R-Studio
Given the different sampling frequencies of data collection, we standardized the dataset by averaging the data to 30-minute intervals per sampling deployment (see Fig 1). This standardization was achieved using the ‘dplyr’, ‘lubridate’, and ‘tidyr’ packages. To assess diel variability in oxygen and pH relative to temperature, we determined the daily maximum, minimum, and mean temperature values, using only complete 24-hour datasets to prevent potential under- or overestimation of the maximum and minimum values. The Pearson Correlation was calculated using ‘cor’ (dplyr package). The R code used for data processing is available at https://github.com/steckbauera/Temperature_O2_pH_data and the data set is available at https://doi.org/10.5061/dryad.q2bvq842b.
Data were averaged for 30 minute intervals. Density represents the relative concentration of data points in a region, showing how values are distributed across an axis or within a plot, with higher density areas indicating where data points are more concentrated. The red line represents a linear regression fit to the data, showing the overall trend of the data presented.
Results
Our analysis included a total of 514,813 data points for DO and 407,253 observations for pH from coastal habitats spanning diverse regions, ranging from polar environments such as Greenland to tropical locales like the Red Sea (S1 Fig). These data represent 1,936 and 1,589 full diel cycles and provide insight into the influence of temperature on DO and pH variability, respectively. Mean temperature was 28.49 °C, and ranged from 1.15 °C in micro-and filamentous algae mats in Greenland to 39.15 °C in the mangrove forest in the Red Sea (see S3 Table). Mean DO was 93.10% saturation (6.72 mg L−1), and ranged from 0.4% sat (0.09 mg L−1) to 442.3% sat (28.53 mg L−1), both recorded in a coral reef in the Red Sea (S2 Fig). Mean pHTOT25C was 7.902, ranging from 6.625 in Mediterranean seagrass meadows to 9.099 in a Red Sea coral reef (S2 Fig). The lowest range of DO and pH was observed in polar regions ~ 35.2% saturation range or 3.2 mg L−1 range, pH range of 0.26 units, respectively. Data standardized to 30-min intervals reveal a pronounced temperature-dependent decline in both oxygen saturation and pH levels across the examined dataset, along with a strong coupling between these two parameters (r = − 0.16 and r = −0.42, respectively; Fig 1, S2 Fig).
The distribution of DO and pH showed near-normal patterns centered around 100% DO saturation and a pH of 8.1, though hypoxic and severely hypoxic conditions were observed in mangrove forests, coral reefs and seagrass meadows during periods of high temperature. Concurrent declines in pH, sometimes below 7.5 units, were recorded in mangroves (Fig 2, S3-S4 Fig).
The density plots illustrate the relative frequency of observations within each ecosystem type.
Diel fluctuations (Fig 3, S5-S7 Fig) of DO were smallest in polar regions with a range of < 50% sat at temperatures between 1°C and 5°C, while temperate and tropical regions exhibited the greatest variability, with DO ranging from near-zero to 250% saturation, particularly in the Red Sea at temperatures up to 35°C. pH showed a similar trend, with polar regions experiencing smaller fluctuations (under 0.25 units per day) and tropical regions such as the Red Sea showing pH ranges of up to ~1 pH unit. The extreme variability observed in tremperate and tropical regions underscores the impact of elevated temperatures on metabolic rates and the associated diel shifts in oxygen and pH levels. Conversely, polar regions exhibited reduced diel variability due in part to limited diurnal light changes during midsummer periods, as evidenced by the Disko Bay study, where continuous daylight constrained metabolic cycling.
The daily ranges of DO saturation and pH (adjusted to 25°C) in relation to mean temperature further illustrate this huge variability. The relationship between mean temperature and daily range of DO saturation (Fig 4a) demonstrated a tendency toward increasing daily range of DO with rising temperatures, suggesting heightened metabolic and environmental fluxes in warmer regions. Conversely, the daily range of pH (adjusted to 25°C) showed a less pronounced trend, with most observations clustering at lower ranges (Fig 4b). The smaller variability in pH is consistent with the log scale for the concetration of H+ ions. Indeed, the range of variability is similar when the relationship between DO and pH is examined with DO on a log10 scale (S8 Fig).
Discussion
Our study, leveraging over 516,705 data points across seven different coastal habitats on four continents, from polar to tropical regions, provides robust evidence of declining DO and pH levels with increasing temperature. This dataset spans a temperature range of over 30°C, with DO range from as low as 0.4% sat (0.09 mg L−1) to 442.3% sat (28.53 mg L−1), and pH ranging from 7.3 to 8.6 pH units. These findings are consistent with previous studies [19,26] but amplify their observations across a much broader range of ecosystems and thermal regimes. Unlike prior studies that focused on isolated parameters (e.g.[31],) or confined locations (e.g. [19,28],), our comprehensive approach across multiple habitats and extended temporal scales adds considerable generality to our findings. The results presented highlight the huge variability DO and pH experience within coastal habitats across temporal scales, often dominated by diel variability. The observed patterns across the analyzed regions highlight the interplay between temperature, metabolic activity, and environmental variability in coastal ecosystems. This variability has critical implications for ecosystem health, species survival, and resilience to climate change, especially in tropical regions facing extreme diel fluctuations in DO and pH.
While satellite-derived data provide valuable insights into global-scale oceanographic trends, they lack the spatiotemporal resolution necessary to capture fine-scale diel changes in coastal environments. Our in situ measurements, conducted at high-frequency intervals, reveal dynamic fluctuations in DO and pH that are often missed by satellite observations. This emphasizes the importance of ground-based monitoring to understand localized processes and their rapid changes, particularly in highly variable coastal ecosystems where metabolic activity, temperature, and other factors influence daily cycles.
The coupling between oxygen and pH through metabolic processes such as photosynthesis and respiration and inorganic carbon metabolism was evident across coastal habitats. Seagrass meadows in temperate regions displayed narrower diel fluctuations compared to their tropical counterparts, including coral reefs, mangroves, and macroalgal beds. Coastal ecosystems often experience pH fluctuations exceeding 0.4 units [21,32], potentially mirroring the variability predicted for the open ocean under future ocean acidification scenarios. Our findings identify a tendency for temperature-dependent diel variability, with the smallest fluctuations observed when temperatures were below 5°C and the largest when temperatures exceeded 25°C. Elevated temperatures reduce oxygen solubility [33,34] and increase the likelihood of hypoxic events (DO < 2.8 mg L−1). Exponential temperature-driven increases in respiratory demand among marine organisms exacerbate susceptibility to warming [35]. The faster increase of respiration relative to photosynthesis with warming [36–38] further intensifies the risk of hypoxia, particularly during night-time when photosynthetic oxygen production ceases. Metabolic rates in coastal habitats are often constrained by nutrient supply [39–42], while excess nutrient supply drives eutrophication and, eventually, coastal hypoxia [43,44]. Hence, much of the variability in the data set and patterns presented here may be attributed to differential nutrient regimes.
Regrettably, nutrient input and concentation data were not available across the different coastal environments in our data set, and sufficiently-reliable high-frequency sampling nutrient sensors were not available to be integrated with temperature, DO, pH sensor packages. Future studies examining the relationship between nutrient fluxes and metabolic responses may help provide further mechanistic understanding around the patterns presented here, particularly if achieving high-freuency sampling. In particular, nutrient fluxes and temperature can interact to determine metabolic rates, suggesting that accounting for nutrient inputs can better isolate the effect of temperature on oxygen and pH dynamics in coastal waters.
Despite being frequently overlooked, the phenomenon of oxygen supersaturation during the day deserves more attention. Across all ecosystems, we observed a tendency toward increasing diel variability in DO and pH with rising temperatures, with oxygen levels surpassing 150% saturation. Giomi and colleagues [18] demonstrated that experimentally-induced oxygen supersaturation extended the survival of several species under simulated warming, suggesting that photosynthetic organisms can supply sufficient oxygen to meet the metabolic demands of marine life. This may provide resilience to acute warming events. Moreover, daytime hyperoxia may act as a buffer against nocturnal hypoxia, as the elevated oxygen concentrations generated by photosynthetic activity during daylight hours can help alleviate oxygen depletion at night. This buffering capacity has been argued to be critical for maintaining the stability of coastal ecosystems, especially during periods of high metabolic demand and elevated temperatures. However, our results show that the potential benefits of increased daytime hyperoxia [18] with increasing ocean temperature are likely to be undermined due to the decline in minimum nightime oxygen levels with increasing water temperature. Nightime oxygen decline with increasing warming may trespass critical thresholds for marine organisms, leading to impacts that cannot be compensated by the benefits of daytime hyperoxia. Moreover, hyperoxia is not always benefitial and can have detrimental effects on marine organisms [21].
It is essential to recognize that oxygen and carbon dioxide (pH) are stoichiometrically linked, yet research often addresses deoxygenation and acidification independently. Our findings confirm that O2 and CO2 (or pH) variability is temperature-dependent, emphasizing that studies isolating one factor may underestimate their combined impacts on coastal ecosystems [45–48]. Baseline surveys of environmental conditions and their variability are thus crucial before planning experiments, as these provide necessary context for understanding the interactive effects of warming, deoxygenation, and acidification.
The lower variability in pH relative to O2 is expected given the log10 scale of pH, as well as the fact that O2 equilibration through air-sea exchange is much faster than that of CO2 due to the buffering of the marine carbonate system [49,50]. The O2 air-sea exchange ratio is much faster than that of CO2, often with a corresponding O2 to CO2 air-sea exchange ratio in excess of 10 [51]. The slow equilibration of CO2 air-sea exchange dampens diel variability relative to that on DO, and adds variability to the relationship betwee DO and pH.
In coastal ecosystems, metabolic processes such as photosynthesis, respiration and inorganic carbon metabolism primarily drive DO and pH dynamics, overshadowing the influence of air-seawater gas exchange mechanisms. Benthic primary producers, including seagrasses and algae, significantly influence local seawater chemistry through these activities, resulting in pronounced diel fluctuations in DO and pH [21,52,53]. For example, seagrass meadows exhibit pH changes closely linked to photosynthesis, buffering against ocean acidification [21]. The role of metabolism in modulating these chemical parameters is further emphasized by studies using the Eddy Covariance system, demonstrating that benthic processes can act as both sinks and sources of acidity [54]. Local biological activities thus play a crucial role in shaping DO and pH levels, particularly in highly heterogeneous coastal habitats, making these environments distinct from the open ocean where air-sea gas exchange dominates.
While our findings emphasize the role of metabolic processes such as photosynthesis and respiration in driving diel fluctuations of DO and pH in coastal ecosystems, inorganic carbon metabolism, specifically calcification and carbonate dissolution, play a critical role in modulating CO2 and pH dynamics in environments dominated by calcifying organisms, such as coral reefs [55,56]. Coral reefs, often growing under oligotrophic conditions, undergo significant shifts in carbonate chemistry due to calcification during daylight hours and dissolution at night. These processes can create strong imprints on local pH and CO2 levels, potentially surpassing the influence of primary production and respiration alone. Calcification leads to the uptake of dissolved inorganic carbon (DIC) and total alkalinity (TA) from seawater, causing an increase in dissolved CO2 and a lowering of the pH, while dissolution has the opposite effect [45,46,56]. As a result, healthy coral reef communities that are actively accreting carbonate framework, are expected to maintain CO2 supersaturation and act, therefore, as net weak sources of CO2 evasion to the atmosphere [56]. This decoupling of metabolic and carbonate processes highlights the complexity of reef ecosystems and underscores the need for integrated studies that capture both biological and chemical drivers of pH variability. Indeed organic and inorganic carbon metabolism contribute to change in CO2 and pH in marine ecosystems, particularly shallow-water ecosystems with an important contribution of calcifying organisms [45,46,56]. Experiments in tropical lagoons supporting calcifying red algae showed that pH increase with net photosynthesis during the daytime can enhance calcification rates, while decreasing pH due to respiratory CO2 release during nightime can suppress calcification [57]. Our data, though primarily focused on metabolic-driven changes, provide a foundation for exploring how calcification and dissolution further modulate carbonate system dynamics within these highly sensitive habitats. Recent research is contributing to disentangle the relative contributions of these processes, enabling improved predictions of ecosystem responses to climate change, using approaches such as coupled O2 and pH eddy covariance [58], as recently demonstrated for a Mediterranean seagrass meadow [59] and a coastal bay environment [46].
Deoxygenation and acidification often co-occur in marine environments, as often reported by studies that concurrently examine both these processes, such as Chinese coastal waters. Examining deoxygenation and acidification concurrently is important to avoid overlooking their potential synergistic effects, especially in highly dynamic and heterogeneous coastal ecosystems. Laboratory experiments frequently lack essential baseline data for study areas, which is critical for accurately simulating environmental stressors and assessing their impacts in the context of climate change. Our findings emphasize that the diel fluctuations of DO and pH tend to increase with rising temperatures and that DO levels can reach supersaturation during the day and hypoxia during the night. However, these fluctuations are not merely a curiosity—they offer critical insights for managing and protecting marine ecosystems. The resilience of these ecosystems to warming-induced stressors depends heavily on understanding their natural variability. High-frequency monitoring and baseline data collection are essential for improving models and informing management strategies to support healthy marine ecosystems and the essential services they provide. Moreover, the variability reported here refers to bulk-water values, while the variability is likely even more extreme in the boundary layer on benthic organisms [20,60–62], where diffusion-limitation would then to amplify DO and pH extremes, while strong hydrodynamic forcing would tend to dampen this variability. The role of multiple drivers, including ecosystem structure, hydrodynamic forcing, temperature and nutrient fluxes in driving the patterns of oxygen and pH variability in coastal ecosystems documented here can be explored by formulating parsimonious mechanistic models able to intregrate all these drivers and their interactions [63,64].
This study provides critical insights into the variability in dissolved oxygen (DO), and pH across diverse coastal ecosystems, spanning a vast range of thermal, climatic and environmental conditions. By compiling and analyzing over 500,000 data points from habitats on four continents, we have highlighted the large variability in DO and pH in coastal waters, their characteristic large diel variability, the relationship between DO and pH, and the tendency for the variability to increase with water temperature. The observed coupling of oxygen and pH through metabolic processes underscores the complexity and vulnerability of coastal ecosystems to climate change, including the risks of hypoxia, acidification, and their combined impacts. The large variability around the observed relationships reflect the extraordinary environmental heterogeneity encompassed by our global dataset, which spans multiple ecosystem types, climatic regions, and oceanographic settings. These relationships represent emergent global patterns rather than strong predictive relationships at individual sites.
Unlike large-scale satellite observations, our high-resolution, fine-scale measurements capture dynamic fluctuations that are pivotal for understanding localized responses to environmental stressors. Furthermore, our findings indicate that metabolic activities predominantly drive DO and pH dynamics in these habitats. High-frequency monitoring, baseline environmental surveys, and integrated studies of deoxygenation and acidification and the response of organic and inorganic carbon metabolism to warming are essential for predicting and mitigating the impacts of climate change on coastal ecosystems and the critical services they provide.
Supporting information
S1 Table. Overview of the study area, instruments used and time frames of the collected data.
https://doi.org/10.1371/journal.pone.0362740.s001
(PDF)
S4 Table. Pearson Correlation between oxygen (% Sat) and Temperature, per Ecosystem Type.
https://doi.org/10.1371/journal.pone.0362740.s004
(PDF)
S5 Table. Pearson Correlation between pHTOT (at 25°C) and Temperature, per Ecosystem Type.
https://doi.org/10.1371/journal.pone.0362740.s005
(PDF)
S6 Table. Pearson Correlation between oxygen (% Sat) and Temperature, per Region.
https://doi.org/10.1371/journal.pone.0362740.s006
(PDF)
S7 Table. Pearson Correlation between pHTOT (at 25°C) and Temperature, per Region.
https://doi.org/10.1371/journal.pone.0362740.s007
(PDF)
S8 Table. Summary of linear regression models relating temperature, dissolved oxygen saturation (O2sat), and pHTOT (25 °C).
Regression equations, slope estimates, 95% confidence intervals, coefficients of determination (R2), sample sizes (N), and p-values are reported. Data were averaged by 30 minute intervals.
https://doi.org/10.1371/journal.pone.0362740.s008
(PDF)
S9 Table. Summary of linear regression models relating temperature, dissolved oxygen saturation (O2sat), and pHTOT (25 °C).
Regression equations, slope estimates, 95% confidence intervals, coefficients of determination (R2), sample sizes (N), and p-values are reported.
https://doi.org/10.1371/journal.pone.0362740.s009
(PDF)
S1 Fig. Global map indicating the location of all study sites.
Note that there are less locations as some of the studies were conducted by the same authors at a similar location. The map layout was generated using R version 4.4.2 and the packages ‘sf,’ ‘ggplot2,’ and ‘dplyr.’ The foundational basemap shapefiles were obtained from Natural Earth (http://www.naturalearthdata.com/), which is in the public domain. The shapefile dataset was accessed via the valentinitnelav/RandomScripts repository on GitHub (https://github.com/valentinitnelav/RandomScripts).
https://doi.org/10.1371/journal.pone.0362740.s010
(PDF)
S2 Fig. Overall data collected at the different study sites: temperature versus oxygen % Sat (A), temperature versus pHTOT25°C (B), and oxygen % Sat versus pHTOT25°C (C).
Density represents the relative concentration of data points in a region, showing how values are distributed across an axis or within a plot, with higher density areas indicating where data points are more concentrated. The red line represents a linear regression fit to the data, showing the overall trend of the data presented.
https://doi.org/10.1371/journal.pone.0362740.s011
(PDF)
S3 Fig. Density distribution of (A) Oxygen (mg L−1), and (B) pHTOT values across different ecosystem types.
The density plots llustrate the relative frequency of observations within each ecosystem type.
https://doi.org/10.1371/journal.pone.0362740.s012
(PDF)
S4 Fig. Density distribution of (A) Oxygen (% Sat), (B) Oxygen (mg L−1), (C) pHTOT25°C, and (D) pHTOT values across different regions.
The density plots llustrate the relative frequency of observations within each region.
https://doi.org/10.1371/journal.pone.0362740.s013
(PDF)
S5 Fig. Diel fluctuations of (A) Oxygen (% Sat), and (B) pHTOT25°C, by mean hour of the day (in mean solar time).
Different colours display the different regions.
https://doi.org/10.1371/journal.pone.0362740.s014
(PDF)
S6 Fig. Diel fluctuations of (A) Oxygen (% Sat), and (B) pHTOT25°C, by mean hour of the day (in mean solar time).
Different colours display the different Ecosystem Type.
https://doi.org/10.1371/journal.pone.0362740.s015
(PDF)
S7 Fig. Diel fluctuations of (A) Oxygen (% Sat), and (B) pHTOT25°C, by mean hour of the day (in mean solar time).
Different colours display the different regions.
https://doi.org/10.1371/journal.pone.0362740.s016
(PDF)
S8 Fig. Overall data collected at the different study sites: oxygen % Sat in log-scale versus pHTOT25°C.
Data were averaged for 30 minute intervals. Density represents the relative concentration of data points in a region, showing how values are distributed across an axis or within a plot, with higher density areas indicating where data points are more concentrated. The red line represents a linear regression fit to the data, showing the overall trend of the data presented.
https://doi.org/10.1371/journal.pone.0362740.s017
(PDF)
Acknowledgments
We thank Paloma Carrillo de Albornoz, Veronica Chaidez, Fernando Lazaro, Núria Marbà, Regino Martinez, Juan de la Cruz Martinez Ayala, Ines Mazarrasa, Rosa Garcia Novoa, Ylva Olsen, Felix Rossbach, and Susann Rossbach for assistance with field work and data collection, and Cassandra Roch for help with manuscript preparation.
References
- 1. Doney SC, Fabry VJ, Feely RA, Kleypas JA. Ocean acidification: the other CO2 problem. Ann Rev Mar Sci. 2009;1:169–92. pmid:21141034
- 2. Feely RA, Sabine CL, Lee K, Berelson W, Kleypas J, Fabry VJ, et al. Impact of anthropogenic CO2 on the CaCO3 system in the oceans. Science. 2004;305(5682):362–6. pmid:15256664
- 3. Vaquer-Sunyer R, Duarte CM, Jordà G, Ruiz-Halpern S. Temperature Dependence of Oxygen Dynamics and Community Metabolism in a Shallow Mediterranean Macroalgal Meadow (Caulerpa prolifera). Estuaries Coasts. 2012;35(5):1182–92.
- 4. Deutsch C, Ferrel A, Seibel BA, Pörtner H-O, Huey RB. Climate change tightens a metabolic constraint on marine habitat. Science. 2015;348(6329):1132–5.
- 5. Schmidtko S, Stramma L, Visbeck M. Decline in global oceanic oxygen content during the past five decades. Nature. 2017;542(7641):335–9. pmid:28202958
- 6. Keeling RE, Körtzinger A, Gruber N. Ocean deoxygenation in a warming world. Ann Rev Mar Sci. 2010;2:199–229. pmid:21141663
- 7. Klein SG, Steckbauer A, Duarte CM. Defining CO2 and O2 syndromes of marine biomes in the Anthropocene. Glob Chang Biol. 2020;26(2):355–63. pmid:31637801
- 8. Byrne RH, Mecking S, Feely RA, Liu X. Direct observations of basin-wide acidification of the North Pacific Ocean. Geophys Res Lett. 2010;37(2):L02601.
- 9. Resplandy L, Keeling RF, Eddebbar Y, Brooks MK, Wang R, Bopp L, et al. Quantification of ocean heat uptake from changes in atmospheric O2 and CO2 composition. Nature. 2018;563(7729):105–8. pmid:30382201
- 10. Cai W-J, Huang W-J, Luther GW 3rd, Pierrot D, Li M, Testa J, et al. Redox reactions and weak buffering capacity lead to acidification in the Chesapeake Bay. Nat Commun. 2017;8(1):369. pmid:28848240
- 11. Li Y, Li M, Kemp WM. A Budget Analysis of Bottom-Water Dissolved Oxygen in Chesapeake Bay. Estuaries and Coasts. 2015;38(6):2132–48.
- 12. Diaz RJ, Rosenberg R. Marine benthic hypoxia: A review of its ecological effects and the behavioural responses of benthic macrofauna. Oceanogr Mar Biol Annu Rev. 1995;33:245–303.
- 13. Diaz RJ, Rosenberg R. Spreading dead zones and consequences for marine ecosystems. Science. 2008;321(5891):926–9. pmid:18703733
- 14. Vaquer-Sunyer R, Duarte CM. Thresholds of hypoxia for marine biodiversity. Proc Natl Acad Sci U S A. 2008;105(40):15452–7. pmid:18824689
- 15. Steckbauer A, Duarte CM, Carstensen J, Vaquer-Sunyer R, Conley D. Ecosystem impacts of hypoxia: thresholds of hypoxia and pathways to recovery. Environ Res Lett. 2011;6(2):025003.
- 16. Kroeker KJ, Kordas RL, Crim R, Hendriks IE, Ramajo L, Singh GS, et al. Impacts of ocean acidification on marine organisms: quantifying sensitivities and interaction with warming. Glob Chang Biol. 2013;19(6):1884–96. pmid:23505245
- 17. Giomi F, Barausse A, Duarte CM, Booth J, Agusti S, Saderne V, et al. Oxygen supersaturation protects coastal marine fauna from ocean warming. Sci Adv. 2019;5(9):eaax1814. pmid:31517051
- 18. Duarte CM, Hendriks IE, Moore TS, Olsen YS, Steckbauer A, Ramajo L, et al. Is Ocean Acidification an Open-Ocean Syndrome? Understanding Anthropogenic Impacts on Seawater pH. Estuaries Coasts. 2013;36(2):221–36.
- 19. Melzner F, Thomsen J, Koeve W, Oschlies A, Gutowska MA, Bange HW. Future ocean acidification will be amplified by hypoxia in coastal habitats. Mar Biol. 2013;160(8):1875–88.
- 20. Giomi F, Rastrick SPS, Calosi P, Carey N, Chapman EC, Turner LM. Oxygen dynamics in marine productive ecosystems at ecologically relevant scales. Nat Geosci. 2023;16(7):560–6.
- 21. Nelson HR, Altieri AH. Oxygen: the universal currency on coral reefs. Coral Reefs. 2019;38(2):177–98.
- 22. Hendriks IE, Olsen YS, Ramajo L, Basso L, Steckbauer A, Moore TS, et al. Photosynthetic activity buffers ocean acidification in seagrass meadows. Biogeosciences. 2014;11(2):333–46.
- 23. Reum JCP, Alin SR, Feely RA, Newton J, Warner M, McElhany P. Seasonal carbonate chemistry covariation with temperature, oxygen, and salinity in a fjord estuary: implications for the design of ocean acidification experiments. PLoS One. 2014;9(2):e89619. pmid:24586915
- 24. Bender ML, Battle M, Keeling RF. THE O2 BALANCE OF THE ATMOSPHERE: A Tool for Studying the Fate of Fossil-Fuel CO2. Annu Rev Energy Environ. 1998;23(1):207–23.
- 25. Steckbauer A, Klein SG, Duarte CM. Additive impacts of deoxygenation and acidification threaten marine biota. Glob Chang Biol. 2020;26(10):5602–12. pmid:32583519
- 26. Hendriks IE, Escolano-Moltó A, Flecha S, Vaquer-Sunyer R, Wesselmann M, Marbà N. Mediterranean seagrasses as carbon sinks: methodological and regional differences. Biogeosciences. 2022;19(18):4619–37.
- 27. Saderne V, Baldry K, Anton A, Agustí S, Duarte CM. Characterization of the CO2 System in a Coral Reef, a Seagrass Meadow, and a Mangrove Forest in the Central Red Sea. JGR Oceans. 2019;124(11):7513–28.
- 28. Krause-Jensen D, Duarte CM, Hendriks IE, Meire L, Blicher ME, Marbà N, et al. Macroalgae contribute to nested mosaics of pH variability in a subarctic fjord. Biogeosciences. 2015;12(16):4895–911.
- 29. Krause-Jensen D, Marbà N, Sanz-Martin M, Hendriks IE, Thyrring J, Carstensen J, et al. Long photoperiods sustain high pH in Arctic kelp forests. Sci Adv. 2016;2(12):e1501938. pmid:27990490
- 30. Ziegler M, Anton A, Klein SG, Rädecker N, Geraldi NR, Schmidt-Roach S, et al. Integrating environmental variability to broaden the research on coral responses to future ocean conditions. Glob Chang Biol. 2021;27(21):5532–46. pmid:34391212
- 31. Pezner AK, Silbiger NJ, Takeshita Y, Johnson MD, Putnam HM, Bahr KD, et al. Increasing hypoxia on global coral reefs under ocean warming. Nat Clim Chang. 2023;13(4):403–9.
- 32. Gobler CJ, Baumann H. Hypoxia and acidification in ocean ecosystems: coupled dynamics and effects on marine life. Biol Lett. 2016;12(5):20150976. pmid:27146441
- 33. Carpenter JH. New measurements of oxygen solubility in pure and natural water. Limnol Oceanogr. 1966;11(2):264–77.
- 34. Garcia HE, Gordon LI. Oxygen solubility in seawater—better fitting equations. Limnol Oceanogr. 1992;37:1307–12.
- 35. Vaquer-Sunyer R, Duarte CM. Temperature effects on oxygen thresholds for hypoxia in marine benthic organisms. Glob Change Biol. 2011;17(5):1788–97.
- 36. Jones RI. The importance of temperature conditioning to the respiration of natural phytoplankton communities. Br Phycol J. 1977;12(3):277–85.
- 37. Enquist BJ, Economo EP, Huxman TE, Allen AP, Ignace DD, Gillooly JF. Scaling metabolism from organisms to ecosystems. Nature. 2003;423(6940):639–42. pmid:12789338
- 38. Iriberri J, Undurraga A, Muela A, Egea L. Heterotrophic bacterial activity in coastal waters: functional relationship of temperature and phytoplankton population. Ecol Model. 1985;28(1–2):113–20.
- 39. Spivak AC, Canuel EA, Duffy JE, Richardson JP. Nutrient enrichment and food web composition affect ecosystem metabolism in an experimental seagrass habitat. PLoS One. 2009;4(10):e7473. pmid:19829713
- 40. Welti N, Striebel M, Ulseth AJ, Cross WF, DeVilbiss S, Glibert PM, et al. Bridging Food Webs, Ecosystem Metabolism, and Biogeochemistry Using Ecological Stoichiometry Theory. Front Microbiol. 2017;8:1298. pmid:28747904
- 41. Nixon SW, Oviatt CA, Frithsen J, Sullivan B. Nutrients and the productivity of estuarine and coastal marine ecosystems. J Limnol Soc South Afr. 1986;12(1–2):43–71.
- 42. Kessouri F, McWilliams JC, Bianchi D, Sutula M, Renault L, Deutsch C, et al. Coastal eutrophication drives acidification, oxygen loss, and ecosystem change in a major oceanic upwelling system. Proc Natl Acad Sci U S A. 2021;118(21):e2018856118. pmid:34001604
- 43. Fennel K, Testa JM. Biogeochemical Controls on Coastal Hypoxia. Ann Rev Mar Sci. 2019;11:105–30. pmid:29889612
- 44. Kemp WM, Testa JM, Conley DJ, Gilbert D, Hagy JD. Temporal responses of coastal hypoxia to nutrient loading and physical controls. Biogeosciences. 2009;6(12):2985–3008.
- 45. Duarte CM, Hendriks IE, Moore TS, Olsen YS, Steckbauer A, Ramajo L. Is ocean acidification an open-ocean syndrome? Understanding anthropogenic impacts on seawater pH. Estuar Coast. 2013;36(2):221–36.
- 46. Mayol E, Ruiz-Halpern S, Duarte CM, Castilla JC, Pelegrí JL. Coupled CO 2 and O 2 -driven compromises to marine life in summer along the Chilean sector of the Humboldt Current System. Biogeosciences. 2012;9(3):1183–94.
- 47. Sampaio E, Santos C, Rosa IC, Ferreira V, Pörtner H-O, Duarte CM, et al. Impacts of hypoxic events surpass those of future ocean warming and acidification. Nat Ecol Evol. 2021;5(3):311–21. pmid:33432134
- 48. Lowe AT, Bos J, Ruesink J. Ecosystem metabolism drives pH variability and modulates long-term ocean acidification in the Northeast Pacific coastal ocean. Sci Rep. 2019;9(1):963. pmid:30700764
- 49. DeGrandpre MD, Hammar TR, Wirick CD. Short-term pCO2 and O2 dynamics in California coastal waters. Deep Sea Research Part II: Topical Studies in Oceanography. 1998;45(8–9):1557–75.
- 50. DeGrandpre MD, Hammar TR, Wallace DWR, Wirick CD. Simultaneous mooring-based measurements of seawater CO2 and O2 off Cape Hatteras, North Carolina. Limnol Oceanogr. 1997;42:21–8.
- 51. Zhai WD, Dai M, Cai WJ. Coupling of surface pCO2 and dissolved oxygen in the northern South China Sea: impacts of contrasting coastal processes. Biogeosciences. 2009;6:2589–98.
- 52. Baumann H, Smith EM. Quantifying Metabolically Driven pH and Oxygen Fluctuations in US Nearshore Habitats at Diel to Interannual Time Scales. Estuaries Coasts. 2018;41(4):1102–17.
- 53. Smith JE, Price NN, Nelson CE, Haas AF. Coupled changes in oxygen concentration and pH caused by metabolism of benthic coral reef organisms. Mar Biol. 2013;160:2437–47.
- 54. Long MH, Charette MA, Martin WR, McCorkle DC. Oxygen metabolism and pH in coastal ecosystems: Eddy Covariance Hydrogen ion and Oxygen Exchange System (ECHOES). Limnol Oceanogr Methods. 2015;13(8):438–50.
- 55. Smith SV, Key GS. Carbon dioxide and metabolism in marine environments. Limnol Oceanogr. 1975;20(3):493–5.
- 56. Gattuso JP, Pichon M, Frankignoulle M. Biological control of air-sea CO2 fluxes: effect of photosynthetic and calcifying marine organisms and ecosystems. Mar Ecol Prog Ser. 1995;129:307–12.
- 57. Semesi IS, Kangwe J, Björk M. Alterations in seawater pH and CO2 affect calcification and photosynthesis in the tropical coralline alga, Hydrolithon sp. (Rhodophyta). Estuar Coast Shelf Sci. 2009;84(3):337–41.
- 58. Berg P, Huettel M, Glud RN, Reimers CE, Attard KM. Aquatic Eddy Covariance: The Method and Its Contributions to Defining Oxygen and Carbon Fluxes in Marine Environments. Ann Rev Mar Sci. 2022;14:431–55. pmid:34587456
- 59. Koopmans D, Schaap A, Meyer V, Färber P, Queiss L, Montilla L, et al. Benthic Ecosystem Calcification Measured with Coupled pH and O2 Aquatic Eddy Covariance. ACS ES T Water. 2026;6(5):2719–30. pmid:42131257
- 60. Martins CPP, Ziegler M, Schubert P, Wilke T, Wall M. Effects of water flow and ocean acidification on oxygen and pH gradients in coral boundary layer. Sci Rep. 2024;14(1):12757. pmid:38830941
- 61. Shashar N, Cohen Y, Loya Y. Extreme Diel Fluctuations of Oxygen in Diffusive Boundary Layers Surrounding Stony Corals. Biol Bull. 1993;185(3):455–61. pmid:29300632
- 62. Noisette F, Pansch C, Wall M, Wahl M, Hurd CL. Role of hydrodynamics in shaping chemical habitats and modulating the responses of coastal benthic systems to ocean global change. Glob Chang Biol. 2022;28(12):3812–29. pmid:35298052
- 63. Kroeker KJ, Donham EM, Vylet K, Warren JK, Cheresh J, Fiechter J. Exposure to extremes in multiple global change drivers: Characterizing pH, dissolved oxygen, and temperature variability in a dynamic, upwelling dominated ecosystem. Limnol Oceanogr. 2023;68:1611–23.
- 64. Jarvis BM, Hagy JD 3rd, Melendez W, Simmons CW, Wan Y. Measuring and modeling diel oxygen dynamics in a shallow hypereutrophic estuary: Implications of low oxygen exposure on aquatic life. Sci Total Environ. 2023;882:163474. pmid:37068685