Coastal acidification alters estuarine sediment nitrous oxide and methane fluxes

The impact of coastal acidification on sediment nitrous oxide (N2O) and methane (CH4) fluxes is largely unknown. We exposed temperate estuarine sediments to moderate (pH 7.3) and extreme (pH 6.3) acidification. Sediments were collected from two sites—one exposed to high and the other to low nitrogen loading. We demonstrate that low pH has a strong effect on greenhouse gas fluxes. The response, in terms of both magnitude and direction, was site specific. Sediments from the high nitrogen loading site exhibited increased N2O fluxes and decreased CH4 fluxes under moderate and extreme acidification. In contrast, sediments from the low nitrogen loading site exhibited decreased N2O fluxes under moderate and extreme acidification while CH4 fluxes both decreased (moderate) and increased (extreme). This study highlights the dynamic response of sediment N2O and CH4 fluxes to low pH and emphasizes the need for deeper understanding of ofcoastal acidification impacts on sediment biogeochemistry.

Coastal acidification is driven by excess nitrogen (N) loading and subsequent stimulation of ecosystem metabolism (Duarte et al. 2013).These conditions are further exacerbated by increased temperatures, which enhance metabolic rates, and precipitation, which drives runoff (Cai et al. 2021).Many estuaries experience pH swings of ≥0.5 pH units daily (Sunda and Cai 2012).Therefore, estuaries are natural laboratories to study how acidification alters biogeochemical cycles.Although impacts of acidification on water column processes have been assessed (Wannicke et al. 2018;Taucher et al. 2021), the effects of acidification on sediment biogeochemistry are understudied (Braeckman et al. 2014).
Sediments are key players in ecosystem function, decomposing organic matter (OM), and recycling or filtering nutrients.Sediments also play an important, yet poorly constrained role in the production and consumption of nitrous oxide (N 2 O) and methane (CH 4 ) (Pratt et al. 2014;Mazur et al. 2021)-two potent greenhouse gases (GHGs) with sustained global warming potentials 32 and 263 times that of carbon dioxide (CO 2 ), respectively (Neubauer and Megonigal 2015).Recent global GHG budgets from coastal systems report CH 4 and N 2 O emissions between 5-28 Tg C yr À1 (Rosentreter et al. 2021) and 0.15-0.91Tg N 2 O-N yr À1 (Murray et al. 2015).In marine sediments, anthropogenic influences (e.g., nutrient loading, hypoxia, warming) increase sediment N 2 O and CH 4 emissions (Foster and Fulweiler 2019;Myllykangas et al. 2020).Yet, few studies examine the role of acidification in altering these fluxes.Without understanding the effect of coastal acidification on sediment N 2 O and CH 4 fluxes, we are unable to develop an accurate GHG budget for marine ecosystems under current and future climate change scenarios.This limits our ability to effectively manage coastal ecosystems under a changing climate.
N 2 O is produced biotically (e.g., nitrification, denitrification, dissimilatory nitrate reduction to ammonium [DNRA], anaerobic ammonium oxidation [anammox]) and abiotically (e.g., chemo-denitrification, hydroxylamine decomposition; Heil et al. 2014Heil et al. , 2015;;Zhu-Barker et al. 2015;Mctigue et al. 2016;Wankel et al. 2017).N 2 O is also consumed biotically via denitrification, nitrifier-denitrification, and DNRA (Giblin et al. 2013;Foster and Fulweiler 2016).Evidence from the terrestrial and coastal literature describes pH as a controller of N 2 O dynamics (Quick et al. 2019;Su et al. 2022).For example, in acidic soils, low pH inhibited nitrous oxide reductase (nosZ) and the conversion of N 2 O to N 2 in denitrifiers (Liu et al. 2014).Similarly, in coastal sediments, acidification altered the diversity of the microbial community and reduced the abundance of the nosZ gene by 43%, favoring N 2 O production (Wu et al. 2022).Results from rate studies are mixed.In the open ocean and in a coastal sediment study, acidification reduced the available NH þ 3 ions for NH þ 4 oxidation, decreased nitrification rates, and thus potentially N 2 O production (Hutchins et al. 2009;Beman et al. 2011).Yet in coastal waters, higher nitrification rates (which could increase N 2 O production) under low pH were attributed to low pH adapted nitrifier communities (Fulweiler et al. 2011).Finally, enhanced availability of CO 2 could increase nitrification rates and subsequent N 2 O production (Amaral et al. 2021).
CH 4 is produced in organic-rich sediments by methanogenesis and is oxidized via a variety of pathways (e.g., ammonium oxidation, sulfate reduction, iron reduction; Jorgensen and Kasten 2006;Bray et al. 2017;Wallenius et al. 2021).pH appears to mediate the abundance of microbes driving CH 4 dynamics with methanotrophs described as capable of functioning across a wide pH range in both soils and coastal sediments (Chen et al. 2018).Although some studies have reported methanotrophic communities dominated by acidophilic species (Yanagawa et al. 2013;Sherry et al. 2016).In addition, when iron oxyhydroxide minerals are abundant, alkaline conditions favor methanogenesis and acidic conditions promote iron reduction (Marquart et al. 2019).
Here, we experimentally examined the effect of low pH on estuarine sediment N 2 O and CH 4 fluxes.Our overall study hypothesis was that low pH exposure would increase sediment N 2 O fluxes and decrease CH 4 fluxes.We chose to test this hypothesis at two sites within the same larger estuaryone site is an open basin site which experiences higherN loading and large summer macro-algae blooms (Foster and Fulweiler 2014).The other site is a coastal lagoon which experiences less N loading and is surrounded by salt marsh.Despite these differences, long-term and seasonal water column pH is similar at both sites.Thus, we anticipated that the response to our experimental pH treatments at both sites would be similar but given that excess N loading increases GHG we anticipated a stronger response at the N impacted site.

Site description
Waquoit Bay is a small (6.3 km 2 ), shallow ($ 3 m depth) estuary located on the southwestern shore of Cape Cod, Massachusetts (D'Avanzo et al. 1996).Two sites were selected for this study (Fig. 1).Metoxit Point (MP) is an open basin located south of the Quashnet River which delivers high N loads from residential septic systems (50.0 Â 10 3 kg N km À2 yr À1 ; Table 1) and supports summer large macroalgae mats (Valiela et al. 1997(Valiela et al. , 2000;;Foster and Fulweiler 2014).Sage Lot Pond (SLP) is a small semi-enclosed coastal lagoon, surrounded by salt marsh, tidally flushed, and located further away from residential communities, thus receives a lower nutrient load compared to MP (2.1 Â 10 3 ; Table 1; Newell et al. 2016).Both sites are rich in OM, with MP having lower C : N ratios and higher concentrations of LOI, sediment chlorophyll a (Chl a), and phaeophytin compared to SLP (Table 1).Despite the differences in N load and OM sources, pH conditions across the sites are similar (Supporting Information Fig. S1; Table S1).

Sediment coring and field methods
We collected 12 sediment cores (10 cm in diameter and $ 25 cm in height) from each site using a pull-corer, which maintains the vertical structure of the sediment cores (Foster and Fulweiler 2019).We also collected filtered bottom water (0.2 μm) in acid-washed plastic carboys for use during flowthrough incubations.We transported the cores back to Boston University within 6 h of collection and placed them in an environmental chamber set to bottom water temperatures (30 [AE1] C).

Incubation set-up and sampling
We aimed for experimental conditions representing control (pH 8.0), moderate (pH 7.3), and extreme (pH 6.3) pH conditions (Baumann and Smith 2018).Because of logistical constraints we separated the experiment into two incubationsextreme and moderate (Supporting Information Figs.S2, S3).
To reach experimental conditions, we divided the filtered site water into two 30-gal acid-washed plastic reservoir tanks.One reservoir tank was acidified to the appropriate pH treatment by bubbling it with 10% CO 2 and controlled using a Qubit pH/CO 2 controller system (Qubit Systems Inc.).The second tank was bubbled with compressed air to maintain a pH of 8.0.
In preparation for each incubation, we siphoned off the overlying water of six sediment cores and replaced it with a $ 4 cm headspace of bottom water from the reservoir tanks (McCarthy et al. 2013).Three cores were filled with water from the acidified tank while the remaining three were filled with water from the control tank.Cores and the overlying water were capped with gas-tight plungers fitted with gas-tight PEEK inflow and outflow tubing above the sediment-water interface (overlying water volume $ 314 mL).The inflow tubing was connected to a peristaltic pump (Masterflex L/S, Coleman-Palmer) with Viton (gas-impermeable) tubing to deliver 1.5 (AE 0.2) mL min À1 of treated bottom water (overlying water residence time $ 209 min).The cores were incubated in the dark under continuous flow-through conditions overnight ($ 12 h) to prevent photosynthetic activity from occurring until sampling began the next morning (Newell et al. 2016;Song et al. 2020;Li et al. 2021).We collected two water samples from inflow and outflow ports at 0, 6, 12, and 24 h in 12 mL Exetainers (Labco) for dissolved GHG samples and preserved with 25 μL of saturated ZnCl 2 .Samples were kept in the refrigerator at 4 C and analyzed within 1 month of sampling.

Microprofile measurements of pH
Before and after each incubation, we collected triplicate sediment pH profiles from each core in the lab using a dualmounted, motorized micromanipulator-profiler and the Sensor Trace Pro v.3.0.6 control software (Unisense) to determine how acidification affected sediment surface pH (Supporting Information Fig. S4).We used a Unisense pH 100 microelectrode (100 μm diameter) to collect pH measurements at 500 μM increments with a maximum depth of 2 cm.We calibrated the pH probe daily with a two-point method using standard buffers at pH 4.01 and pH 7.00 (Hach).

N 2 O and CH 4 measurements
Concentrations of dissolved N 2 O and CH 4 were measured directly from the sample using headspace equilibration and measured on a gas chromatograph (GC; $ 4 mL; GC2014; Foster and Fulweiler 2016).The GC was equipped with a flame ionization detector to measure CH 4 and an electron capture detector to measure N 2 O.We estimated concentrations by comparing the area under the produced peak against a standard curve.We determined standard curve concentrations using a linear regression of custom gas concentrations (CH 4 5084 ppb; N 2 O: 495 ppb in N 2 ) made by Airgas.Detection limits during sample analysis were 83.21 ppb for CH 4 and 16.83 ppb for N 2 O.

Flux calculation
We calculated benthic fluxes of GHG using the following equation (Miller-Way and Twilley 1996): The inflow and outflow concentrations represent the average value of the two replicate samples collected at each time point.We calculated the mean AE standard error (SE) flux for each treatment by averaging the fluxes measured across all time points (Supporting Information Table S2).Detection limits for the flux measurements were defined as when the SE's did not overlap with zero (Gardner and McCarthy 2009).

Statistical analyses
All statistical analyses were conducted in R Studio version 4.2.1 (R Core Team 2020).Results of sediment pH measurements are presented as mean (AE SE).Rather than reporting differences between control and acidified treatments using statistical significance, which can often have misleading influence due to sample size (Greenland et al. 2016;Wasserstein and Lazar 2016;Amrhein et al. 2017), we describe the impact of acidification on GHG fluxes using the size of the effect (Fritz et al. 2012;Smidt et al. 2022).We measured the effect  Valiela et al. (1997Valiela et al. ( , 2000) ) and subestuary surface areas from D'Avanzo et al. (1996).
size using the effsize package by comparing all the acidified fluxes to their corresponding control fluxes within each experiment (Torchiano 2016; Supporting Information Table S3).It was necessary for us to compare our acidified cores to their corresponding controls because we were unable to run both experimental incubations at the same time.We used the following equation to calculate Cohen's d for each experiment (Cohen 1988): Cohen's d values above 0.2 considered "small," 0.5 considered "medium," and 0.8 considered "large" effects.

Benthic fluxes of N 2 O
Control cores from the moderate treatment had a mean N 2 O flux of À5.4 (AE3.9) and a zero flux in the extreme treatment (Fig. 3; Supporting Information Table S2; Mazur and Fulweiler 2022b).When exposed to moderate acidification, fluxes ranged from À17.3 to 35.6 nmol m À2 h À1 and had a mean flux of 11.5 (AE5.0)nmol m À2 h À1 .Under extreme acidification, N 2 O fluxes ranged from À2.5 to 22.5 nmol m À2 h À1 and had a mean flux of 8.1 (AE3.4) nmol m À2 h À1 .In MP, sediments transitioned from being a net sink of N 2 O to a net source when exposed to acidification with the moderate treatment having a large effect (d = 1.11) and the extreme treatment having a small effect (d = 0.41) on N 2 O fluxes (Fig. 3; Supporting Information Table S3).
In SLP control cores, mean N 2 O flux in the moderate treatment was zero and 10.8 (AE 3.1) nmol m À2 h À1 in the extreme.After moderate acidification N 2 O fluxes ranged À32.9 to 8.3 nmol m À2 h À1 and had a mean of À13.2 (AE 2.7) nmol m À2 h À1 .In the extreme treatment, N 2 O fluxes ranged from À31.2 to 8.3 nmol m À2 h À1 and had a mean flux of 8.3 (AE 3.4) nmol m À2 h À1 .Here, sediments switched from being a net source to a net sink of N 2 O when exposed to acidification with both acidification treatments having large effects (moderate: d = À1.91;extreme: d = À1.69).

Benthic fluxes of CH 4
Both sites were a source of CH 4 to the overlying water column under control conditions (Fig. 3; Supporting Information Table S3; Mazur and Fulweiler 2022b).Effluxes of CH 4 were $ 36Â higher in MP control cores than those measured in SLP.In MP, CH 4 flux from control cores in the moderate treatment was 13,900 (AE 2,777) nmol m À2 h À1 and 23,546 (AE 5,582) nmol m À2 h À1 in the extreme (Supporting Information Table S2).CH 4 fluxes from the moderate treatment ranged from 3,194 to 12,825 nmol m À2 h À1 with a mean flux of 8,131 (AE957) nmol m À2 h À1 .In the extreme treatment,  CH 4 fluxes ranged from 4,561 to 14,242 nmol m À2 h À1 with a mean flux of 9,478 (AE 917) nmol m À2 h À1 .Acidification had a large and medium effect on dampening CH 4 fluxes in MP (moderate: d = À1.04;extreme: d = À0.79).
Mean CH 4 fluxes in SLP moderate treatment were 568 (AE 65) nmol m À2 h À1 and 362 (AE 40) nmol m À2 h À1 in the extreme.CH 4 fluxes exposed to moderate acidification were half compared to the control (range: 87-1,034 nmol m À2 h À1 ) and had a mean of 295 (AE 79) nmol m À2 h À1 .Fluxes after extreme acidification were higher than those measured in the control, ranging from 232 to 1,460 nmol m À2 h À1 , with a mean of 819 (AE 123) nmol m À2 h À1 .Acidification had a large effect on CH 4 fluxes in SLP (moderate: d = À1.12;extreme: d = 1.44).

Discussion
Coastal ecosystems are important in the global GHG budget, accounting for 60% of marine N 2 O emissions and 75% of marine CH 4 emissions (Bange et al. 1996;Hamdan and Wickland 2016).Our study reveals that the response of sediments to low pH is dynamic and that two sites within the same estuary responded differently to acidification.
When exposed to acidification, N 2 O fluxes at MP increased by $ 4,218%, while fluxes at SLP decreased by $ 275%.Studies in coastal sediments report enhanced N 2 O release under acidification like those measured at MP. Su et al. (2021) observed an increase in N 2 O emissions between 223% and 438%, under acidic conditions (7.5-5.0 pH), while Wu et al. (2022) measured a 49% increase in N 2 O emission rates when pH declined by only 0.3.
Microbial sediment communities involved in N 2 O dynamics are correlated to pH.Enhanced N 2 O production can cause nosZ inhibition, changes in functional microbial community, and decreases in denitrification efficiency (Wu et al. 2022).Under these circumstances denitrification can gradually shift from N 2 to N 2 O (Su et al. 2021).Given these results, the reduced N 2 O production under acidified conditions in SLP was unexpected.Enhanced N 2 O consumption under acidification is described in seawater (Gu et al. 2021), but until this study, had yet to be observed in estuarine sediments.Alternatively, low pH could have enhanced consumption-the net fluxes we measure cannot discern whether production declined, consumption increased, or both occurred.One driver of the differences in N 2 O response can be the quantity, quality, and source of OM (Guo et al. 2020;Chen et al. 2022;Stuchiner and von Fischer 2022).Perhaps the labile OM from macroalgae MP drives N 2 O production under acidification while the more recalcitrant, salt marsh OM at SLP lessened the effect on N 2 O under low pH (Table 1).The availability of inorganic N can fuel N 2 O production (Murray et al. 2015) and a recent study reported that acidification increased sediment DIN fluxes (Simone et al. 2022).However, in our study water column concentrations were similarly low at each site (Table 1) and in most cases we observed no relationship between N 2 O flux and inorganic N concentrations (Supporting Information Tables S4, S5).The one exception would be for SLP where N 2 O and NH þ 4 concentrations were negatively correlated in the control and moderate treatment.We do not have enough information to understand the mechanism behind this correlation.Given that we do not see it for the other fluxes we are hesitant to put much weight behind it.Another explanation could lie within the microbial communities.The microbes and genes responsible for N 2 O dynamics vary throughout estuaries and are dictated by the long-term physical environment (e.g., nutrient loading, oxygen availability; De Bie et al. 2002;Kearns et al. 2015;Su et al. 2022).It is likely the microbial community of MP and SLP is different, which could explain the differences in response to acidification.Untangling why there is a sitespecific response will be a challenging but critical next step to further our understanding of the impact of acidification on coastal biogeochemistry.
Under low pH, CH 4 fluxes from MP were 75% lower compared to the control fluxes.Yet, acidification in SLP sediments resulted in decreased CH 4 fluxes in the moderate treatment and increased CH 4 flux in the extreme treatment.Published observational studies of CH 4 flux in marine sediments are limited, and those available also show a mixed response to pH.In tropical mangrove sediments, CH 4 flux was positively correlated with sediment pH (Chauhan et al. 2015), while in a subtropical estuary both positive and negative relationships between pH and sediment CH 4 flux were reported (Martin et al. 2020).
pH may drive CH 4 fluxes in coastal sediments by altering CH 4 cycling microbial communities (Reshmi et al. 2015;She et al. 2016) and by increasing oxidation due to heavy metal mobilization (Brocławik et al. 2020).We suggest OM quality, methanogen communities and metal availability can help explain differences in CH 4 response to in MP and SLP (Li et al. 2022).Methanogen communities can vary spatially within an estuary and are regulated by anthropogenic nutrients, OM, and metal distribution (Guo et al. 2019;Kharitonov et al. 2021).Due to the different N loads, it's likely the methanogen communities are different between sites (Table 1).In iron and OM-rich sediments, like those in Waquoit Bay, iron reduction is favored over methanogenesis when exposed to acidification, resulting in decreased CH 4 fluxes (Charette and Sholkovitz 2002;Testa et al. 2002;Marquart et al. 2019).The varying response of SLP to pH treatment is intriguing and harder to explain.Perhaps, it highlights a threshold, where the microbial community at SLP is resilient to low pH until an extreme pH condition is present (Emery et al. 2019).
This study begins to shed light on how low pH conditions alter N 2 O and CH 4 benthic fluxes in estuaries.Importantly, our study indicates that future acidification might not have uniform impacts on GHG fluxes across estuaries.Thus, future studies could link net fluxes or isotope labeling experiments with the microbial community composition and sediment properties to untangle these interactive effects.Efforts to improve our understanding of pH impacts on sediment GHG production and consumption will create better informed coastal management practices and more accurate GHG budgets for these systems.

Fig. 1 .
Fig. 1.Map of sampling locations in Waquoit Bay, Cape Cod, Massachusetts, USA.Sediment cores for nitrous oxide and methane fluxes were collected at Metoxit Point (MP) and at Sage Lot Pond (SLP).

Fig. 2 .
Fig. 2. Sediment pH profiles of sediment cores from MP (A) and SLP (B).Each point represents the mean (AE SE) of triplicate measurements collected at

Fig. 3 .
Fig. 3. Nitrous oxide (N 2 O; A) and methane (CH 4 ; B; note the different scales) fluxes (mean AE SE) across the sediment-water interface from MP and SLP.Gray bars represent fluxes from control cores incubated during each treatment (moderate [pH 7.3], solid; extreme [pH 6.3], striped).Bars highlighted in blue are fluxes from moderate-treated cores and bars highlighted in red from extreme-treated cores.The dotted line at zero represents the sediment surface.A positive flux (above this line) represents mean production of GHG or efflux from the sediment, while the negative values represent mean sediment consumption or influx of GHG into the sediment, respectively.Effect size (Cohen's d; C) of pH treatments on N 2 O and CH 4 flux (absolute value) was determined with values above 0.2 considered to be "small" (solid line), 0.5 to be "medium" (dash line) and 0.8 to be "large" (dotted-dash line) effect sizes.Bars with plus values indicate a positive Cohen's d value while bars with a negative sign indicate negative Cohen's d values.

Table 1 .
Site water column and sediment characteristics (mean AE SE), as well as long-term pH trends of the two sampling stations in Waquoit Bay, Massachusetts: MP and SLP.Sediment characteristic data are from 0 to 1 cm depth.
*NOAA National Estuarine Research Reserve System (NERRS).System-wide Monitoring Program.Data accessed from the NOAA NERRS Centralized Data Management Office website: http://www.nerrsdata.org,accessed 28 August 2022.† Loads calculated by Foster and Fulweiler 2014, watershed N Loads from