Introduction

Methane (CH4) and nitrous oxide (N2O) are important atmospheric greenhouse gases, as they absorb infrared radiation at higher efficiencies per molecule than carbon dioxide (CO2) and participate in atmospheric chemical reactions that lead to changes in stratospheric ozone (Myhre et al. 2013; Ravishankara et al. 2009). According to the last Intergovernmental Panel for Climate Change (IPPC) report, in 2011 the global average atmospheric dry mole fractions of CH4 and N2O were ∼ 1803 and ∼324 ppb, which exceeded the pre-industrial levels by about 150%, and 20%, respectively (Ciais et al. 2013). Even though the atmospheric growth rate of both gases shows a marked interannual variability (Nisbet et al. 2016; Thompson et al. 2017), evidence indicates that they will continue rising due to numerous sources.

Major anthropogenic CH4 sources are agricultural practices, emissions from fossil fuel extraction, landfills and waste and the large increase in ruminants. Natural sources include anaerobic aquatic environments where biogenic methane is formed, such as wetlands, freshwater lakes, streams and rivers, estuarine and coastal regions, but also areas of thermogenic methane release, geothermal vents, and natural biomass combustion (Hamdan and Wickland 2016). Recent studies have concluded that the rapid atmospheric methane rise detected over the last decade is a result of increased emissions from biogenic sources (Nisbet et al. 2016; Schaefer et al. 2016), particularly in the tropics, which can be linked to expanding tropical wetlands in response to positive rainfall anomalies (Nisbet et al. 2016). Comparison with historic methane formation suggests that if the methane growth driven by this type of biogenic emissions continues rising, the present increase will be beyond the largest events in the last millennium (Nisbet et al. 2016).Therefore, the assessment of CH4 release from wetlands and other inland waters is particularly relevant to gain insights on production and emission pathways and reduce uncertainty in the global methane budget (Kirschke et al. 2013). Current uncertainties are attributable to the choice of methodology, as top–down and bottom–up methods result in marked differences in the inventories (Saunois et al. 2016), the approach selected to compute the emissions, discrepancies in surface areas and bias in the geographic distribution of the aquatic systems studied so far. Measurements of methane emissions from aquatic environments have been concentrated in wetlands, freshwater lakes, reservoirs, and fluvial systems whereas reports on methane emissions from estuaries are still relatively sparse (Hamdan and Wickland 2016). Several works have proven that estuaries make a significant contribution to CH4 emissions (Borges and Abril 2011; Cotovicz et al. 2016; Gelesh et al. 2016), globally releasing between 1 and 7 Tg CH4 year− 1(Borges et al. 2016), which is comparable, for instance, to methane fluxes arising from the open ocean domain (< 2 Tg CH4 year− 1, Borges et al. 2016). In fact, methane supersaturation higher than 20,000% is a frequent phenomenon in estuarine waters (Abril and Iversen 2002; Ferrón et al. 2007; Middelburg et al. 2002; Upstill-Goddard et al. 2000), where the gas mostly originates from freshwater inputs and in situ sediment methanogenesis (Borges et al. 2015b). Local CH4 production related to high turbidity in the water column has been also suggested in two British estuaries (Upstill-Goddard et al. 2000). On the other hand, a decrease in dissolved CH4 in turbid waters has been reported in some European estuaries, which has been attributed to accelerated CH4 outgassing due to high turbulence or to increased CH4 oxidation by suspended mater particles-attached bacteria (Middleburg et al. 2002, Abril et al. 2007). Currently, there is significant evidence indicating that estuarine CH4 production comes from sediments (inter-tidal or sub-tidal) from where it diffuses to the water column (Borges and Abril 2011 and references therein). Regardless of the methane source, there is a strong need for more methane emission data in estuarine systems to further constrain global methane budget (Saunois et al. 2016).

Estuaries have been also recognized as a significant source of N2O to the atmosphere (Bange 2006; Barnes and Upstill-Goddard 2011; Murray et al. 2015), emitting about 0.25 Tg N year− 1(Bakker et al. 2014). N2O is produced by microbial processes, such as nitrification (oxidation of ammonium to nitrate) and denitrification (reduction of nitrate to N2), with the yield of N2O formation during both mechanisms being strongly dependent upon dissolved oxygen (DO) concentration. These coastal systems receive high loads of nitrogen (N) that fuel both formation pathways and promote eutrophication, which leads to hypoxia or anoxia (Howarth et al. 2011), conditions that ultimately regulate N2O release (Bakker et al. 2014; Naqvi et al. 2010). Eutrophic European estuaries are considered net N2O emitters (Stocker et al. 2013) although additional investigations are needed to allow for more accurate continental and global budgeting of the emissions.

Here, we present the results of the first measurements of dissolved N2O and CH4 along the Guadalquivir river estuary (SW Spain) during 2016 and 2017. The estuary (Fig. 1) has undergone an intense anthropogenic pressure since the second half of the twentieth century that has led to profound alterations in its original morphology and ecological conditions (Ruiz et al. 2015). Agriculture and surrounding large population settlements introduce considerable inputs of nutrients, organic matter and inorganic compounds, which result in a strong limitation in light availability for primary production (Ruiz et al. 2015, 2017). The composition of the drainage basin of a siliceous origin in the north and a carbonate part in the south also contributes to the high concentration of suspended solids and dissolved carbonates characteristic of these estuarine waters (de la Paz et al. 2007). High turbidity promotes net heterotrophy, leading to a strong CO2 supersaturation and the ecosystem behaves as a massive CO2 emitter (Flecha et al. 2015). Accordingly, events of hypoxia are common in the estuary, especially in its inner part, where DO levels as low as 34 µmol kg− 1 have been measured (Flecha et al. 2015).

Fig. 1
figure 1

Location of the Guadalquivir estuary. Sampling stations are indicated by the red dots (Color figure online)

Full size image

Our study was aimed at (1) characterizing N2O and CH4 distribution and their saturations along the Guadalquivir estuary, (2) providing the first estimates of air–water exchange of both gases and (3) identifying potential production pathways. Moreover, our data can be used to complement the overall diagnosis of the ecological status of the Guadalquivir estuary in the light of proposing remediation management actions to recover the ecosystem-services supplied by this environment (Ruiz et al. 2017).

Methods

Study area

The Guadalquivir river is one of the longest fluvial courses in Spain with a total length of 680 km that extends from its source in the Cazorla mountains, approximately 1400 m above sea level, to its mouth at Sanlucar de Barrameda in the Gulf of Cadiz (Fig. 1). The estuary occupies the river´s last 110 km, hosting a population of 1.7 million people that are clustered in large municipalities that rely on the estuary to support different socioeconomic sectors (agriculture, fisheries, tourism) (Ruiz et al. 2015). As a mesotidal system, tidal influence can be noticeable up to 100 km upstream of the river mouth, where a dam is located (Díez-Minguito et al. 2012). Because the climate in the catchment area (63,822 km2) is Mediterranean sub-humid with well defined seasonality, freshwater discharges from the dam exhibit marked seasonal variations in magnitude and duration. During the dry periods (70% of the hydrological year) discharges lower than 40 m3 s− 1 regularly occur, which are nevertheless sufficient to compensate evaporation losses. These conditions are normal for this estuary (Díez-Minguito et al. 2012). During the wet season (usually lasting from October to April) that is usually characterized by relatively short but intense periods of rainfall, freshwater inputs into the estuary may yield levels exceeding 400 m3 s− 1, which interferes with the tidal effect and disrupts tidal dominance (Díez-Minguito et al. 2013). In fact, high to very high discharges (ranging from 500 to 1000 m3 s− 1) induce a regime shift due to the high-suspended sediment concentrations brought by the freshwater inputs (Losada et al. 2017). Elevated concentrations of suspended matter are also found under normal conditions, although several orders of magnitude lower than those under periods of high discharges (Losada et al. 2017). The estuary is heavily dredged on a regular basis to deepen the navigation channel and ensure access for large container ships to the port of Sevilla, which delivers high load of inorganic compounds to the water column (Caballero et al. 2018).

Sampling strategy

Samples were taken at five sites (Fig. 1) during eight cruises conducted between March 2016 and March 2017 using a 4.85 m inflatable boat. Sampling dates and monthly environmental conditions are listed in Table 1. All surveys commenced at the mouth of the estuary during spring and rising tide. This strategy may have had implications for gas distribution, as sampling went along with the net saltwater flux into the system and salinity could affect the concentration gradient. However, being aware of this caveat, the sampling design ensured comparable mixing conditions every month and also allowed minimizing survey duration.

Table 1 Relevant physico-chemical characteristics of the Guadalquivir estuary during the sampling period
Full size table

From site 1 positioned at the river mouth, sites 2, 3 and 4 were located at 10, 15 and 20 km upstream, respectively. Site 5, at 25 km from the river mouth, represented the entry of the estuary into a creek that penetrates the saltmarshes of Doñana National Park, the largest coastal wetland of South Europe. This site was chosen to assess the environmental status of estuarine waters feeding the southern sector of the Park, which had been isolated from the estuary until very recently (Huertas et al. 2017a). At each site, conductivity, temperature and pH were obtained with a Yellow Spring (YSI Incorporate) portable multiparameter probe YS6820v2. Discrete water samples were taken with a Niskin bottle at 1 m depth to determine CH4 and N2O concentrations, inorganic nutrients (NH4+, NO2−, NO3−, PO43−, and Si), DO, total dissolved nitrogen (TDN), dissolved organic carbon (DOC), suspended particulate matter (TSM) and chlorophyll (Chl a). The Guadalquivir estuary is classified as totally mixed mesotidal (de la Paz et al. 2007; Ruiz et al. 2017) and thus, 1 m depth samples were assumed to be representative for the entire water column.

Analytical techniques

Samples for N2O and CH4 were collected in duplicate 120 mL serum vials, sealed and preserved with HgCl2 to inhibit microbial activity. Trace gas samples were stored upside down in the dark until analysis in the laboratory by static-head space equilibration gas chromatography (GC) using an Agilent 7890 GC equipped with an Electron Capture Detector (ECD) for N2O and Flame Ionization Detector (FID) for CH4 as described in de la Paz et al. (2015). Before chromatographic determination, 20 mL of N2 headspace were introduced in each sample and equilibrated for at least 12 h after initial vigorous shaking. The GC system was calibrated using three standard gas mixtures of different origin: a certified NOAA primary standard with composition similar to the atmosphere (324.97 ± 0.13 ppb for N2O and 1863.4 ± 0.3 ppb for CH4), and two additional standard gas mixtures of N2O and CH4 in a N2 matrix provided by Air Liquide (France) with certified concentrations (1020 and 3000 ppb for N2O; 3000 and 5000 ppb for CH4). The precision of the method estimated from the coefficient of variation based on replicate analysis was 0.6% for CH4 and 0.4% for N2O. Saturation values expressed as percentage (%) for CH4 and N2O were computed as the ratio between the gas concentration measured and the calculated equilibrium concentration for both gases. Calculations of the equilibrium concentrations in the water phase were done using the annual averaged atmospheric mixing ratios CH4 (×CH4atm) and N2O (×N2Oatm) provided by the World Data Center for Greenhouse Gases (http://ds.data.jma.go.jp/gmd/wdcgg/) due to lack of data during the study period in the nearest station of the global monitoring network. Such mean values were calculated as 1866 and 328 ppb for ×CH4atm and ×N2Oatm respectively.

Water samples (5 mL, two replicates) for nutrient analyses were filtered immediately (Whatman GF/F, 0.7 µm), and stored frozen (− 20 °C) for later analyses in the shore-based laboratory. Dissolved nitrate, nitrite and ammonium concentrations were measured with a continuous flow auto analyzer (SkalarSan + + 215) using standard colorimetric techniques (Hansen and Koroleff 1999). Analytical precisions were always better than ± 3%.

DO concentrations were fixed immediately and measured within 24 h upon collection in sealed flasks stored in the dark through an automated potentiometric modification of the original Winkler method using a Metrohm 794 Titroprocessor, with an estimated error of ± 5 µmol kg− 1.The saturation values of O2 were calculated with the equation given by Benson and Krause (1984) and the apparent oxygen utilization (AOU) was obtained by subtracting the oxygen concentration at saturation to the observed oxygen concentration.

Total alkalinity (TA) was also determined by titration of samples collected in borosilicate bottles (500 mL) and poisoned with 100 µL of a HgCl2 saturated aqueous solution, according to Mintrop et al. (2000). Accuracy (± 5 µmol kg− 1) was obtained from regular measurements of Certificate Reference Material supplied by Prof. Andrew Dickson, Scripps Institution of Oceanography, La Jolla, CA, USA (Batch #147). Partial pressure of CO2 in water (pCO2) was computed from TA and pHNBS pairs through the CO2SYS.xls program (Lewis et al. 1998) using the Cai and Wang (1998) and Dickson (1990) constants for carbon and sulfate, respectively.

Concentrations of DOC and TDN were determined by catalytic oxidation at high temperature (720 °C) and chemiluminescence, respectively in a Shimadzu Total Organic Carbon analyzer (Model TOC-VCPH/CPN), according to Álvarez-Salgado and Miller (1998). Analytical precision of the methods was always better than 0.015 and 0.03 mg L− 1 for DOC and TDN measurements respectively.

TSM, particulate organic and inorganic matter (POM and PIM respectively) were determined by the loss on ignition (LOI) method. Known volumes of water were filtered (pre-combusted 450 °C Whatman GF/F, 0.7 µm) and filters dried at 60 °C for 48 h. They were subsequently weighed to derive TSM (g L− 1), further combusted again at 450 °C for 5 h, and weighed to derive PIM and POM by difference.

Chl a analysis was conducted by filtering known volumes of water (Whatman GF/F, 0.7 µm) and filters were dipped in 90% acetone overnight in the dark for extraction. Pigment concentrations were obtained by fluorometry with a Turner Designs 10-AU Model fluorometer, which was calibrated using a pure Chl a standard from the cyanobacterium Anacystis nidulans (Sigma Chemical Company). The precision of the method was 0.025 µg L− 1.

Air–water gas exchange

The gas flux (F, µmol m− 2 day− 1) between the atmosphere and estuarine waters was calculated as:

$${\text{F }}=k{\text{ }}({C_w}-{\text{ }}{C_a}),$$
(1)

where k (cm h− 1) is the gas transfer rate as a function of wind speed at 10 m height, Cw is the measured dissolved gas concentration, and Ca is the equilibrium concentration in water based on the molar atmospheric ratio as above. k was computed from k normalized to a Schmidt number of 600 (k600) according to:

$$k={k_{600}}\sqrt {\frac{{600}}{{Sc}}} ,$$

where Sc is the Schmidt number of each gas calculated from water temperature with the formulations given by Wanninkhof (1992), which has been widely applied and facilitates comparison with other systems.

k600 was computed from U10 using the parameterization given by Jiang et al. (2008).

$${k_{600}}=0.314~ \times {{\text{U}}_{10}}^{2} - 0.436 \times {{\text{U}}_{10}}+3.99,$$

where U10 was calculated according to Smith (1988) using monthly averaged wind speed data provided by a nearby automatic meteorological station located in Lebrija (36°58 × 35′′N, 06°07 × 34′′W) from the Junta de Andalucía network (http://www.juntadeandalucia.es/agriculturaypesca/ifapa/ria/servlet/FrontController).

There is no current consensus about the best parameterization of the gas transfer rate to be used for computation of the air–water gas exchange in estuaries. We chose the wind-dependent expression provided by Jiang et al. (2008) that has been formulated specifically for estuaries by refitting the data compiled by Raymond and Cole (2001) with newer gas exchange measurements in estuaries.

Daily discharge data from the Alcala del Rio dam and rainfall from the Lebrija station were obtained from the Sistema Automático de Información Hidrológica for the Guadalquivir basin (http://www.chguadalquivir.es/saih/Inicio.aspx).

All data contained in this work are available for download from Digital.CSIC, the Institutional Repository of the Spanish National Research Council (CSIC), (https://digital.csic.es/handle/10261/160022, https://doi.org/10.20350/digitalCSIC/8528).

Statistics

Statistics were performed with the program language MATLAB. Probability distributions of variables were examined through a Shapiro–Wilk test. Pearson’s product-moment correlation (PPMC) was used to test for significant correlations between variables. Significance levels were set at p < 0.05. When normality was not met, the criterion established by Havlicek and Peterson (1977) was applied.

Results

Dynamics of CH4 and N2O in the estuary

During the sampling period, CH4 and N2O concentrations ranged from 14 to 750 nmol L− 1 and from 3 to 34 nmol L− 1, respectively (Table 2). As a general trend, an increase in the average concentration of both gases was observed upstream, with CH4 mean values of 27 and 167 nmol L− 1 at sites 1 and 5 respectively and N2O mean values of 15 and 17 nmol L− 1 in the same sites (Table 2). Methane levels in the lowest saline samples ranged from 14 to 750 nmol L− 1 whereas concentrations at the highest salinities ranged from 16 to 39 nmol L− 1 (Table 3). In the case of nitrous oxide, concentrations in the samples characterized by the lowest salinities varied from 3 to 29 nmol L− 1, which were similar to those found in samples with the highest salinities and that ranged between 9 and 25 nmol L− 1 (Table 3). For both gases, concentration in freshwater samples exhibited a higher variability than in their marine counterparts, which was particularly evident in the case of methane (Table 3).

Table 2 Average concentrations (± SD) of dissolved CH4 and N2O and concentration ranges in the sites chosen at the Guadalquivir estuary during the sampling period
Full size table
Table 3 Summary of the concentration of methane and nitrous oxide found in samples with the smallest and highest salinity in the Guadalquivir estuary during each survey
Full size table

The salinity range found within the estuary transect during the different surveys also exhibited marked variations (Table 1). The spatio-temporal variability of salinity (Fig. 2) seemed to be related to the tidal amplitude (Table 1). Thus, during periods of maximum rainfall (May and November 2016), higher discharges from the dam occurred (> 45 m3 s− 1, Table 1) and even though the magnitude of the freshwater inputs was equivalent, the salinity gradient in the water course clearly differed between both months (Table 1). In May, a sharp decline in salinity was observed in all sites, dropping to nearly 0, except in the site located closer to the river mouth (site 1), where salinity reached approximately 10 (Fig. 2). In contrast, in November, salinity approached seawater values at sites 1 and 2, also increasing upstream in relation to other months (Fig. 2), which can be attributed to the tidal coefficient (or amplitude of the tide forecast) present at spring tide that was the highest for the entire sampling period (Table 1). The remaining samplings were characterized by slight variations in salinity in each site, although a clear salinity gradient throughout the estuary transect was still detected (Fig. 2). Noticeable changes in the concentration of biogeochemical variables occurred during May and November 2016 in relation to other surveys. The rise in freshwater flow along the estuary in May was accompanied by a decrease in DO levels (particularly in site 5) and increases in the concentrations of DOC, TDN (except in site 5), CH4 and N2O, although a marked reduction in nitrous oxide was observed in site 5 that even declined below saturation (Fig. 2). The drop of N2O in the tidal creek coincided with the highest concentration of methane measured during the sampling period (Fig. 2). In contrast, when the tide propagated further upstream during November 2016, as indicated by the rise in salinity in all sites (Fig. 2), DOC, TDN and N2O decreased along the estuary whereas methane concentration increased in sites 2, 3 and 4 and decreased in sites 1 and 5.

Fig. 2
figure 2

Temporal variation of salinity and concentration of dissolved oxygen, organic carbon, total nitrogen, CH4 and N2O (open symbols) at the five sites sampled along the Guadalquivir estuary during eight surveys conducted between March 2016 and March 2017. Note that Y axis scale for several parameters varies for site 5

Full size image

Sources of CH4 and N2O

Table 4 summarizes the results of the Pearson correlations performed with the complete dataset of the variables measured. Methane concentration was shown to be positively and significantly (p < 0.05) correlated with Chl a, DOC, AOU and pCO2 and inversely correlated with N2O and nitrate. Accordingly, nitrous oxide concentration was significantly (p < 0.05) and directly correlated with NO3− and negatively correlated with salinity. Furthermore, strong and significant relationships were found between pCO2, DOC and AOU, indicating that organic matter degradation played an important role in generation of carbon dioxide within the estuary, which ranged between 522 and 4300 ppm and increased as salinity decreased (Fig. 3 and note the negative and significant correlation between pCO2 and salinity in Table 4). Salinity also correlated quite well with Chl a concentration and DOC (Table 4; Fig. 3), suggesting that the spatial distribution of these biogeochemical variables was governed by the mixing between freshwater and oceanic water. Nevertheless, a higher variability in the concentration of both variables was observed in riverine waters (salinity < 10) in relation to the levels found in saltier waters (Fig. 3), as was observed for CH4 and N2O concentrations (Table 3). The relationship between the levels of both gases in the estuary and salinity was illustrated by plotting the percentage of saturation of dissolved CH4 and N2O as a function of salinity (Fig. 4a). Estuarine waters were always over-saturated in methane and nitrous oxide with respect to the atmospheric equilibrium, with the exception of site 5 in which N2O under-saturation (39%) in relation to the atmospheric N2O level was found in May 2016 at a very low salinity (0.5, Table 3; Fig. 4a). With the exception of that finding, higher N2O over-saturations were measured in the freshwater portion of the estuary (salinity < 10) regardless of the sampling month, with the highest value approaching 400% in May 2016. Hence, a gradual decreasing gradient of N2O saturation with salinity was noticeable, which was especially evident in the surveys conducted in May and December 2016. This pattern was not so clear for methane, which exhibited an oversaturation range between 520 and 30,800%, and even though higher values were observed in the freshwater–saltwater interface a more scattered distribution was found (Fig. 4a).

Table 4 Pearson correlation matrix for biogeochemical variables measured in the Guadalquivir estuary
Full size table
Fig. 3
figure 3

Dissolved carbon dioxide (pCO2), dissolved organic carbon (DOC) and Chlorophyll (Chl a) vs. salinity in the estuary transect

Full size image
Fig. 4
figure 4

a log10 percent dissolved methane saturation (log10%CH4) vs. salinity (upper panel) and percent dissolved nitrous oxide saturation (%N2O) vs. salinity (lower panel) and b log10 percent dissolved methane saturation vs. percent dissolved O2 saturation (%O2) (upper panel) and percent dissolved nitrous oxide saturation vs. percent dissolved O2 saturation (lower panel) for the Guadalquivir estuary

Full size image

In order to gain insights on the effect of ecosystem metabolism on the dynamics of CH4 and N2O within the estuary, the relationship between the trace gas saturation levels and the oxygen availability was also examined. It is worth noting that the N2O outlier measured in site 5 during May 16 (39% N2O saturation at %O2 saturation of 14) was not statistically considered. As shown in Fig. 4b, log10 %CH4 vs. %O2 displayed a weak but significant negative relationship (r2 = 0.33, n = 40) whereas %N2O was strongly and negatively related with %O2 (Fig. 4b; r2 = 0.56, n = 40) (Fig. 5).

Fig. 5
figure 5

Temporal variation of the air–water CH4 (upper panel) and N2O (lower panel) fluxes for the Guadalquivir estuary during the eight surveys conducted between March 2016 and March 2017

Full size image

CH4 and N2O emission fluxes

CH4 emissions varied from ∼ 13 µmol m− 2 day− 1 in site 1 in November 2016 to a maximum of nearly 1100 µmol m− 2 day− 1 at site 5 in May 2016, although no clear seasonal variation in CH4 fluxes could be detected (Fig. 5). A tendency to higher methane effluxes was observed upstream regardless of the sampling month, although emissions remained below 50 µmol m− 2 day− 1 in most sites with the exception of site 5 in which the outgassing markedly increased in April and May 2016 and January 2017.

Fig. 6
figure 6

a Dissolved nitrous oxide concentration vs. nitrate and b dissolved nitrate vs. dissolved ammonium for the Guadalquivir estuary

Full size image

N2O emissions from the estuarine transect were higher in May 2016 (Fig. 5) coinciding with the flushing of the estuary with freshwater (Fig. 2; Table 1), although during this month, site 5 displayed the only negative N2O flux measured during the whole sampling period and equivalent to − 7.1 µmol m− 2 day− 1. Overall, downstream sites exhibited lower N2O emissions, with the lowest N2O effluxes (below 4 µmol m− 2 day− 1) occurring in all sites in November 2016, when the estuary experienced the highest tidal intrusion (Fig. 2; Table 1).

Discussion

Sources of CH4 and N2O in the Guadalquivir estuary

The spatio-temporal distribution of dissolved CH4 and N2O in the Guadalquivir estuary reflects the hydrodynamics of the system, in which tidal-fluvial interaction is a major driver of the ecosystem metabolic status (Losada et al. 2017; Ruiz et al. 2017). The associations found between the salinity gradient and Chl a, DOC, and pCO2 within the estuary are in agreement with previous findings indicating that the patterns of primary production, organic matter degradation and CO2 emissions are tightly coupled to the freshwater discharge/tidal regime (Flecha et al. 2015; Ruiz et al. 2017). Our study now reveals that the levels of CH4 and N2O observed in the Guadalquivir estuary are also closely related to the balance between the intrusion of the saline plume and the magnitude of freshwater inputs. Even though this relationship was seen along the entire estuarine transect, it was especially noticeable in the tidal creek where levels of CH4 and N2O markedly varied in response to drastic changes in salinity (for instance May and November 2016). The effect of the tidal-fluvial interaction on the trace gas distributions in the estuary could proceed either in a direct way (e.g. sulfate inhibition on methanogenesis) or indirectly through the influence of hydrodynamics on the biogeochemistry of the water course (e.g. changes in nutrient supply or oxygen availability). The average CH4 concentration and over-saturations measured in the Guadalquivir estuary fall in the lower portion of the range reported for temperate and tropical estuaries and rivers (Koné et al. 2010; Maher et al. 2015; Middelburg et al. 2002; Sansone et al. 1999; Shalini et al. 2006; Smith et al. 2000; Upstill-Goddard and Barnes 2016; Upstill-Goddard et al. 2017; Zhang et al. 2008). In addition, no clear seasonal signals in methane levels could be discerned, contrary to what has been observed in other fluvial catchments and estuaries (Bouillon et al. 2012; Koné et al. 2010; Middelburg et al. 2002; Shalini et al. 2006; Upstill-Goddard et al. 2017). This is probably the result of several processes influencing water column CH4, as methane in estuaries stems from several sources: (1) microbial production in sediments that fluxes to the water column, (2) microbial production in adjacent wetlands and transport by the river, or (3) in situ microbial CH4 production as a result of anaerobic organic matter decomposition. The interaction of these pathways with seasonal signals in estuarine hydrodynamics may mask any seasonality in biogeochemistry due to the effect of temperature alone, as suggested for other European estuaries (Upstill-Goddard and Barnes 2016).

Furthermore, it is understood that moderate to high salinity aquatic systems typically show much lower surface water methane concentrations and emissions than freshwater habitats. This is partially due to the high concentration of sulfate in seawater that allows sulfate-reducing bacteria to outcompete methanogenic bacteria for energy sources, consequently inhibiting methane production (Bartlett et al. 1987; Borges and Abril 2011). However, the association between salinity and methane formation can be complicated by site-specific conditions and methane can be also produced in saline environments despite the inhibitory effects of sulfate (Borges and Abril 2011). A compilation of methane measurements in 31 temperate tidal marshes revealed an inverse log-linear relationship between salinity and methane emissions (Poffenbarger et al. 2011). This study concluded that the range of methane emissions from saline marshes could be predicted by salinity and those systems characterized by salinity > 18 have negligible methane production. In our work, methane distribution and effluxes along the Guadalquivir estuary suggest that even though salinity was not the primary controlling factor for methane generation, sulfate inhibition must have been proceeding, as higher CH4 oversaturations were measured at riverine waters. Additionally, gas concentrations in samples characterized by the lowest salinities were much higher than those in their marine counterparts. Hence, the magnitude of the tidal intrusion likely affected methane distribution in the estuary.

In addition to the sulfate inhibition of CH4 formation, the decrease of CH4 concentration with salinity could be due to gas loss terms, such as microbial oxidation and emission to the atmosphere. Dissolved methane loss by these processes can be very fast in coastal and estuarine environments, as recently found by Borges et al. (2017) in the southern bight of the North Sea.

Our results also indicate that anaerobic matter degradation in the water column during the sampling period was unlikely, as even though DO was undersaturated most of the time, oxygen concentrations remained above 5 mg L− 1. Under these aerobic conditions, CH4 levels were still moderate, with saturations ranging from ∼ 520 to 6000%. Therefore, methane diffusion from the sediment was probably the major source of CH4 to the water column. However, it is notable that when an event of isolated hypoxia occurred at site 5 during May 2016 (0.69 mg L− 1, corresponding to O2 undersaturation of 14%), methane concentration sharply increased to reach nearly 800 nmol L− 1 (over saturation of 30,000%), which could suggest in situ aquatic microbial CH4 production by anaerobic organic matter degradation. Nevertheless, methanogens are sensitive to low oxygen concentrations and are slow growing organisms unlikely to proliferate in the water column on short time-scales (Bridgham et al. 2013). Moreover, even though there is circumstantial evidence for CH4 production in the water column, it proceeds under very stable conditions, for instance, in stratified oligotrophic lakes (Grossart et al. 2011). Hence, the most plausible explanation for the high CH4 concentration is that methane losses via microbial oxidation would have been minimized, which would result in the local massive efflux detected. This association between methane release in response to dissolved oxygen decay has been found in some African streams (Borges et al. 2015b). This finding may be particularly relevant for methane emission patterns in the Guadalquivir estuary, where prolonged episodes of hypoxia have been observed after the entry of considerable suspended matter loads by high freshwater discharges (Flecha et al. 2015; Ruiz et al. 2015, 2017). The relationship found between log10%CH4 vs. %O2 indicates that when oxygen saturation dropped below 15%, the level of dissolved methane sharply rose. Therefore, a rise in methane accumulation may be expected in the estuary under high freshwater flooding events, which are common during rainy seasons wetter than that of our study (Losada et al. 2017), as they increase turbidity, reduce sulfate inhibition and cause hypoxia.

It should be noted that freshwater inputs do not necessarily always lead to higher CH4 levels. For instance, in the Meuse river network the highest fluvial methane concentrations have been found during low water due to gas accumulation favored by the increase in residence time and temperature (Borges et al. 2018). Our data suggest that in the case of the Guadalquivir estuary high freshwater is more likely to result in CH4 accumulation due to its effect on the aforementioned mechanisms.

During our sampling period, no correlation between CH4 concentration and turbidity (represented by the TSM content) was found, contrary to the trend reported in some British estuaries (Upstill-Goddard et al. 2000). However, this feature does not preclude that such association could occur in the estuary during episodes of greater turbidity (Losada et al. 2017).

Periods characterized by a high DOC content but aerobic conditions, such as those in April 2016 and January 2017 in the creek (site 5) coincided with increased methane levels (∼ 200 nmol L− 1). This suggests contribution of lateral inputs from the adjacent Doñana marshes, which are significant CH4 emitters (Huertas et al. 2017b). In the main channel of the estuary, methane was also significantly and positively correlated with DOC and Chla. As both variables represent the balance between respiration and productivity in aquatic ecosystems, the direct relationships found with CH4 indicate that methane dynamics in the estuary are regulated by a combination of complex biological interactions. The tight couplings between pCO2, CH4, DOC, PO43− and AOU suggest that the dynamics of pCO2, CH4 and DO were driven by net heterotrophy, as described in freshwater systems (Borges et al. 2015a; Lapierre and del Giorgio 2012) and tidal estuaries (Maher et al. 2015).

Despite methanogenesis being carried out by severely O2-limited archaea (Bridgham et al. 2013), the existence of high CH4 concentrations in oxygenated aquatic systems, as occurred in the Guadalquivir estuary during our study period, is a well known phenomenon. Diffusion of the methane produced in the anoxic sediments and inputs from adjacent floodplain soils and wetlands may have contributed to such pattern, but also methanogenesis in anoxic microsites present in oxygenated soils, which is even activated during flooding (Bridgham et al. 2013; Von Fischer and Hedin 2007). Methane production by photoautotroph-attached archaea (Grossart et al. 2011), and non-microbial aerobic formation in plant tissues (Keppler et al. 2006, 2009) and soils (Hurkuck et al. 2012) have also been reported. Therefore, some of these processes or a combination of them, may have been active in the Guadalquivir estuary during our surveys. Clearly, further work is needed to identify the methane production pathways in the river and their control by environmental factors.

Nitrate availability could also play a role by suppressing methanogenesis (Klüber and Conrad 1998), which has been described recently in some streams of North America (Schade et al. 2016). The negative correlation between CH4 and NO3− would fit the conceptual model of Schade et al. (2016). Nevertheless, the influence of nitrate on CH4 dynamics was not clearly identified in a global meta-analysis of riverine CH4 conducted by Stanley et al. (2016), who stated that the dual role of NO3− as both nutrient and transport electron acceptor complicates the assessment of the relationship between methane and nitrogen availability. In a recent work, Borges et al. (2018) did not find any significant correlation between CH4 and NO3− in the Meuse river network. As indicated by these authors, it is uncertain if correlations between these variables are the result of a direct causality or due to a common driver, such as oxygen availability in the case of the Guadalquivir estuary. Moreover, in mesotidal systems, the inhibitory effect of sulfate on methanogenesis must also be incorporated into the regulation pathways and hence, no definitive conclusion on the regulatory effect of nitrate on CH4 dynamics can be drawn from our data set.

N2O cycling in estuaries is also regulated by complex processes involving oxygen availability, nitrogen load, organic matter inputs, groundwater, and mixing (Codispoti 2010; Murray et al. 2015). Our data reveal a connection between N2O levels and the magnitude of freshwater inputs within the estuary. The downstream sites 1 and 2 always exhibited lower N2O concentrations, which also decreased in the whole estuarine transect during the tidal intrusion in November 2016. However, the %N2O vs. salinity plot suggests that dilution alone cannot explain N2O distribution. In fact, the influence of the fluvial-tidal interaction on N2O estuarine levels seemed to occur through the effect of mixing on the nitrogen concentration since when dissolved nitrogen compounds fell as the result of the tidal flushing, N2O concentrations invariably dropped. Therefore, nitrogen loading within the estuary that is heavily regulated by the mixing conditions may be claimed as a major driver of N2O generation patterns in the system.

This association between N2O and the nitrogen content but especially with NO3− points towards nitrification as the main N2O formation pathway. Parallel increases in NO3− and N2O, which were found in our data set (Fig. 6a), are an indication of nitrification (Beaulieu et al. 2010; Silvennoinen et al. 2008). Furthermore, NO3− that was the main form of inorganic nitrogen in the estuary, also rose as NH4+ concentrations decreased (Fig. 6b), providing additional evidence for N2O produced during nitrification. Nitrification-derived N2O has been described in the Schelde estuary (Wilde and de Bie 2000), some British estuaries (Barnes and Upstill-Goddard 2011) and in the Elbe estuary (Brase et al. 2017). The relationship found between %N2O vs. %O2 supports our conclusion on the prevalence of nitrification in the Guadalquivir estuary. The lowest N2O value (below atmospheric equilibrium) of the data series observed at site 5 in May 2016 concurred with the isolated event of hypoxia and the highest pCO2 level registered (∼ 4330 ppm), conditions that were not observed in the rest of surveys, which together suggests removal of nitrous oxide by denitrification. This observation is in agreement with patterns described in the Amazon floodplains (Richey et al. 1988) and in some African rivers (Borges et al. 2015b) and is relevant for N2O dynamics in an ecosystem where prolonged hypoxic episodes have been reported (Ruiz et al. 2015).

In comparison to other European estuaries, N2O concentrations and saturation levels in the Guadalquivir estuary are in the mid range of values reported. The mean N2O saturation of 183% (± 69%) in the estuary is comparable to those measured in the Gironde (Bange 2006), Temmesjoki (Silvennoinen et al. 2008), Tagus (Gonçalves et al. 2010) and Elbe (Brase et al. 2017) estuaries but lower compared to the Schelde (de Bie et al. 2002) and below the overall value estimated for European estuaries (Barnes and Upstill-Goddard 2011).

CH4 and N2O emissions

The mean CH4 flux during our sampling period was 66.20 (± 171) µmol m− 2 day− 1, which falls within the range of air–water CH4 exchange reported in temperate rivers from 0 to 22,000 µmol m− 2 day− 1 (see compilation by Upstill-Goddard et al. 2017). However, considering the broad interval of values, it is evident that our mean estimation is near the low end, being also smaller than CH4 emissions reported in European estuaries and in temperate and boreal rivers where methanogenesis inhibition by sulfate is negligible (Stanley et al. 2016). CH4 studies conducted in Indian estuaries (Shalini et al. 2006), African rivers and streams (Borges et al. 2015b; Koné et al. 2010; Upstill-Goddard et al. 2017) and in the Amazon River and tributaries (Bartlett et al. 1990; Sawakuchi et al. 2014) also provided higher methane emissions than those in the Guadalquivir estuary. Nevertheless, these fluvial catchments are inherently different than ours whose suspended particulate matter is mainly of inorganic nature (de la Paz et al. 2007) (∼ 80% PIM vs. 20% POM during our sampling period, not shown) and where the oxygen levels would favor aerobic organic matter degradation and CH4 losses via oxidation, at least during the study time course. Similar values of methane emissions have been found in the brackish section of a Danish estuary (Abril and Iversen 2002) and in a tidal creek of the Bay of Cádiz that receives large anthropogenic nitrogen loads (Ferrón et al. 2007). Moreover, it is important to acknowledge that methane fluxes were computed here using the approach given by Jiang et al. (2008) and our study did not quantify CH4 ebullition fluxes whose contribution to total CH4 emissions cannot be precluded during low water periods and transition from high tide to low tide (Baulch et al. 2011).

The mean N2O air–water exchange during our sampling period was 8.5 (± 8) µmol m− 2 day− 1, which is almost two-fold lower than the global median of N2O fluxes for open water estuaries and equivalent to 18.2 µmol m− 2 day− 1 (Murray et al. 2015). Barnes and Upstill-Goddard (2011) provided N2O fluxes in seven British estuaries on the order of 43.2 µmol m− 2 day− 1 and gave an average estimate of 45.7 µmol m− 2 day− 1 for European estuaries. Higher N2O emissions have been also reported in the Schelde estuary (33.6 µmol m− 2 day− 1, de Bie et al. 2002) and in the Seine river (96.5 µmol m− 2 day− 1) (Garnier et al. 2006). In contrast, N2O fluxes in the Guadalquivir estuary are more comparable to those computed in the Tagus (5.8 µmol m− 2 day− 1), (Gonçalves et al. 2010) and Tamar estuaries (8.03 µmol m− 2 day− 1), (Barnes and Upstill-Goddard 2011), in some tidal Australian estuaries (between 2.3 and 15.9 µmol m− 2 day− 1), (Musenze et al. 2014; Sturm et al. 2016), in African rivers (from 2 to 28 µmol m− 2 d− 1) (Borges et al. 2015b; Koné et al. 2010) and in the Amazon River and floodplain (from 0.25 to 6.0 µmol m− 2 day− 1), (Guérin et al. 2008).

Hence, the estuary behaved as a small CH4 source and as a moderate N2O source under the environmental conditions present during the monitoring period. Our results also show that the estuary acts as a net exporter of both gases to the continental shelf of the gulf of Cádiz, as previous studies conducted in the basin had suggested (Ferrón et al. 2010a, b; de la Paz et al. 2015). Further research is still needed to fully characterize CH4 and N2O dynamics in this ecosystem, particularly during events of large flooding, which according to our findings, will likely influence the patterns of gas emissions along the estuary and affect the methane and nitrous oxide budgets in the adjacent coastal region.