Advertisement

Hydrogeology Journal

, Volume 26, Issue 8, pp 2739–2751 | Cite as

Dynamics of greenhouse gases in the river–groundwater interface in a gaining river stretch (Triffoy catchment, Belgium)

  • Anna JuradoEmail author
  • Alberto V. Borges
  • Estanislao Pujades
  • Pierre Briers
  • Olha Nikolenko
  • Alain Dassargues
  • Serge Brouyère
Paper
  • 254 Downloads

Abstract

This study investigates the occurrence of greenhouse gases (GHGs) and the role of groundwater as an indirect pathway of GHG emissions into surface waters in a gaining stretch of the Triffoy River agricultural catchment (Belgium). To this end, nitrous oxide (N2O), methane (CH4) and carbon dioxide (CO2) concentrations, the stable isotopes of nitrate, and major ions were monitored in river and groundwater over 8 months. Results indicated that groundwater was strongly oversaturated in N2O and CO2 with respect to atmospheric equilibrium (50.1 vs. 0.55 μg L−1 for N2O and 14,569 vs. 400 ppm for CO2), but only marginally for CH4 (0.45 vs. 0.056 μg L−1), suggesting that groundwater can be a source of these GHGs to the atmosphere. Nitrification seemed to be the main process for the accumulation of N2O in groundwater. Oxic conditions prevailing in the aquifer were not prone for the accumulation of CH4. In fact, the emissions of CH4 from the river were one to two orders of magnitude higher than the inputs from groundwater, meaning that CH4 emissions from the river were due to CH4 in-situ production in riverbed or riparian zone sediments. For CO2 and N2O, average emissions from groundwater were 1.5 × 105 kg CO2 ha−1 year−1 and 207 kg N2O ha−1 year−1, respectively. Groundwater is probably an important source of N2O and CO2 in gaining streams but when the measures are scaled at catchment scale, these fluxes are probably relatively modest. Nevertheless, their quantification would better constrain nitrogen and carbon budgets in natural systems.

Keywords

Greenhouse gases Indirect emissions Groundwater monitoring Gaining stream Belgium 

Dynamique des gaz à effet de serre à l’interface rivière–eau souterraine dans un tronçon drainant de la rivière (bassin versant du Triffoy, Belgique)

Résumé

Cette étude examine la présence de gaz à effet de serre (GES) et le rôle de l’eau souterraine comme voie indirecte des émissions de GES dans les eaux de surface dans un tronçon drainant du bassin versant agricole de la rivière Triffoy (Belgique). À cette fin, les concentrations d’oxyde nitreux (N2O), du méthane (CH4) et de dioxyde de carbone (CO2), les isotopes stables du nitrate et les ions majeurs ont été surveillés dans la rivière et l’eau souterraine pendant 8 mois. Les résultats indiquent que l’eau souterraine est fortement sursaturée en N2O et en CO2 par rapport à l’équilibre atmosphérique (50.1 vs. 0.55 μg L−1 pour le N2O et 14,569 vs. 400 ppm pour le CO2), mais seulement de manière marginale pour le CH4 (0.45 vs. 0.056 μ g L−1), ce qui suggère que l’eau souterraine peut être une source de ces GES vers l’atmosphère. La nitrification semblait être le processus principal d’accumulation de N2O dans les eaux souterraines. Les conditions oxiques prévalant dans les eaux souterraines n’étaient pas la cause de l’accumulation de CH4. En fait, les émissions de CH4 depuis la rivière étaient d’un à deux ordres de grandeur plus élevés que les entrées depuis les eaux souterraines, ce qui signifie que les émissions de CH4 depuis la rivière étaient dues à la production in situ du CH4 dans le lit ou dans les sédiments de la zone riparienne. Pour le CO2 et le N2O, les émissions moyennes depuis les eaux souterraines étaient respectivement de 1.5 × 105 kg CO2 ha−1 année−1 and 207 kg N2O ha−1 année−1 dans les cours d’eau drainants mais quand les mesures sont évaluées à l’échelle du bassin versant, ces flux sont sans doute relativement modestes. Néanmoins, leur quantification permettrait de mieux contraindre les bilans de l’azote et du carbone dans les systèmes naturels.

Dinámica de los gases de efecto invernadero en la interfaz río–agua subterránea en un tramo de río ganador (Cuenca Triffoy, Bélgica)

Resumen

Este estudio investiga la presencia de gases de efecto invernadero (GEIs) y el papel del agua subterránea como una vía indirecta de emisiones de GEIs a las aguas superficiales en un tramo ganador de la cuenca agrícola del río Triffoy (Bélgica). Con este fin, se midireron las concentraciones de óxido nitroso (N2O), metano (CH4) y dióxido de carbono (CO2), los isótopos estables del nitrato y los iones mayoritarios en el río y las aguas subterráneas durante 8 meses. Los resultados indicaron que las aguas subterráneas estaban fuertemente sobresaturadas en N2O y CO2 con respecto al equilibrio atmosférico (50.1 vs. 0.55 μg L−1 para N2O y 14,569 vs. 400 ppm para CO2), pero solo marginalmente para el CH4 (0.45 vs. 0.056 μg L−1), sugiriendo que las aguas subterráneas pueden ser una fuente de estos GEIs a la atmósfera. La nitrificación fue el proceso principal para la acumulación de N2O en las aguas subterráneas. Las condiciones óxicas del acuífero no fueron favorables para la acumulación de CH4. De hecho, las emisiones de CH4 del río fueron de uno a dos órdenes de magnitud más altas que las entradas de las aguas subterráneas, lo que significa que las emisiones de CH4 del río se debieron a la producción in situ de CH4 en los sedimentos del lecho del río o de la zona ribereña. Para el CO2 y el N2O, las emisiones promedio de las aguas subterráneas fueron 1.5 × 105 kg CO2 ha−1 año−1 y 207 kg N2O ha−1 año−1, respectivamente. Las aguas subterráneas son probablemente una fuente importante de N2O y CO2 en ríos ganadores, pero cuando las medidas se escalan a nivel de toda la cuenca, estos flujos son probablemente relativamente modestos. Sin embargo, su cuantificación limitaría mejor los balances de nitrógeno y carbono en los sistemas naturales.

(比利时Triffoy流域)一个袭夺河段中河水–地下水界面温室气体动力学

摘要

本研究调查 了(比利时)Triffoy河农业流域一个袭夺河段温室气体的产生以及地下水作为温室气体排放进入地表水的间接途径的作用。为此,对河流和地下水中的一氧化二氮、甲烷和二氧化碳浓度、硝酸盐的稳定同位素以及主要离子进行了8个月的监测。结果显示,针对大气平衡(一氧化二氮50.1 vs. 0.55 μg L−1 及二氧化碳14,569 vs. 400 ppm)来说,地下水中的一氧化二氮和二氧化碳过度饱和,但只有甲烷接近饱和(0.45 vs. 0.056 μg L−1),表明地下水可能是这些温室气体进入大气的来源。硝化疏忽是地下水中一氧化二氮积累的主要过程。含水层中盛行的氧化条件不是甲烷积累的原因。事实上,河中的甲烷排放高于地下水排放一到两个数量级,意味着甲烷从河中的排放归因于河床中或河岸带沉积物中甲烷的现场产生。对于二氧化碳和一氧化二氮,地下水中的平均排放量分别为1.5 × 105 kg CO2 ha−1 year−1 and 207 kg N2O ha−1 year−1。地下水可能是袭夺河中二氧化碳和一氧化二氮的重要来源,但在测量结果为流域尺度时,这些通量可能相对保守。然而,二氧化碳和一氧化二氮的量化可以更好地约束自然系统中的氮和碳。

Dinâmica de gases de efeito estufa na interface rio - água subterrânea em um trecho de ganho fluvial (bacia de Triffoy, Bélgica)

Resumo

Este estudo investiga a ocorrência de gases de efeito estufa (GEEs) e o papel das águas subterrâneas como um caminho indireto de emissões de GEE nas águas superficiais em um trecho de ganho da bacia agrícola do Rio Triffoy (Bélgica). Para este fim, as concentrações de óxido nitroso (N2O), metano (CH4) e dióxido de carbono (CO2), os isótopos estáveis de nitrato e os íons principais foram monitorados em águas fluviais e subterrâneas durante 8 meses. Os resultados indicaram que a água subterrânea estava fortemente supersaturada em N2O e CO2 em relação ao equilíbrio atmosférico (50.1 vs. 0.55 µg/L para N2O e 14.569 vs. 400 ppm para CO2), mas apenas marginalmente para CH4 (0.45 vs. 0.056 µg/L), sugerindo que as águas subterrâneas podem ser uma fonte destes GEE para a atmosfera. A nitrificação pareceu ser o principal processo para o acúmulo de N2O nas águas subterrâneas. Condições oxidativas predominantes no aquífero não foram a causa do acúmulo de CH4. De fato, as emissões de CH4 do rio foram de uma a duas ordens de magnitude mais altas que os insumos das águas subterrâneas, significando que as emissões de CH4 do rio estavam relacionadas à produção in situ de CH4 nos sedimentos de leito de rio ou de zonas ripárias. Para CO2 e N2O, as emissões médias das águas subterrâneas foram de 1.5 × 105 kg CO2 ha−1 ano−1 e 207 kg N2O ha−1 ano−1, respectivamente. A água subterrânea é provavelmente uma fonte importante de N2O e CO2 nas correntes de ganho, mas quando as medidas são dimensionadas na escala de captação, esses fluxos são provavelmente relativamente modestos. No entanto, suas quantificações restringiriam melhor os balanços de nitrogênio e de carbono nos sistemas naturais.

Introduction

Anthropogenic application of organic and inorganic fertilisers of nitrogen (N) in agricultural landscapes and livestock wastes have a negative impact on groundwater resources quality due to leaching of N species into aquifers (Glavan et al. 2017). Agricultural practices represented up to one third of anthropogenic emissions of greenhouse gases (GHGs; Gilbert 2012), such as nitrous oxide (N2O), methane (CH4) and carbon dioxide (CO2), which all contribute to climate change and N2O to stratospheric ozone destruction (IPCC 2014). Therefore, aquifers below agricultural landscapes can be an indirect source of GHG emissions to the atmosphere because groundwater is generally oversaturated in these GHGs with respect to atmospheric equilibrium (Bell et al. 2017; Jurado et al. 2018; McAleer et al. 2017).

Dynamics of GHGs in groundwater are complex because their occurrence depends on the geochemical conditions (e.g., nitrate NO3, ammonium NH4+, dissolved oxygen DO, organic carbon OC, bicarbonate HCO3, pH, among others) that control nitrogen (N) and carbon (C) cycles (Nikolenko et al. 2018; Jahangir et al. 2013). Denitrification is considered to be the main process of NO3 attenuation under anaerobic conditions in groundwater but N2O is an intermediate product (Rivett et al. 2008). When NO3 is nonlimiting and at intermediate DO concentrations, N2O is not reduced to N2 and it can accumulate in shallow groundwater (Deurer et al. 2008). Nitrification also contributes to the N2O production in groundwater, in which case N2O is a byproduct that can be produced during the oxidation of nitrite (NO2) to NO3 (e.g., Vilain et al. 2012). In addition, hydrogeological parameters (e.g., water table, rainfall periods and aquifer permeability) also play a major role on the dynamics of N2O in groundwater (Jahangir et al. 2013)—for instance, Deurer et al. (2008) suggested that during high-intensity precipitation events, denitrification might be inhibited in the Fuhrberger Feld aquifer (Germany) by the transport of DO with the infiltrating water. This situation promoted variable geochemical conditions leading to “cold” and ‘hot” spots of N2O in near-surface groundwater. Concerning C species, the presence of CH4 in shallow groundwater is associated with strongly anaerobic environments such as wetlands and landfills and comes from a biogenic origin (Bell et al. 2017)—for example, Cheung et al. (2010) reported that dissolved CH4 in shallow groundwater of Alberta (Canada) was of biogenic origin via CO2 reduction. Likewise, CO2 is also produced and consumed by several processes in groundwater such as plant root respiration, oxidation of organic matter and the precipitation and dissolution of carbonate minerals (Wang et al. 2015).

Several studies have assessed the indirect GHG emissions in aquifers below agricultural landscapes (Hasegawa et al. 2000; Jahangir et al. 2012; McAleer et al. 2017; Minamikawa et al. 2010; Vilain et al. 2012; von der Heide et al. 2009), but the contribution of groundwater as a source of GHGs via surface-water bodies such as streams and rivers has received less attention. Groundwater discharge to river (base flow) has been recognized as a potential pathway of N2O into streams and rivers, which generally are net sources of N2O in N-rich environments (Beaulieu et al. 2010; Fox et al. 2014, Gardner et al. 2016; Werner et al. 2012) but can be sinks of N2O in N- and DO-poor environments (Borges et al. 2015, 2018). Groundwater has also been recognised as an important source of CO2 in riverine systems (Worral and Lancaster 2005), especially in small streams and headwaters (Hotchkiss et al. 2015; Johnson et al. 2008). Recently, Borges et al. (2018) have reported that surface waters of the Meuse River network (Belgium) act as a source of CO2, CH4 and N2O to the atmosphere. The authors pointed out that the extremely high concentrations of N2O and CO2 in groundwater might indicate that part of these GHGs could come from groundwater in the Meuse basin, although the actual fraction remains to be quantified.

To date, studies that have simultaneously quantified the contribution of groundwater as a potential source of N2O, CH4 and CO2 in rivers are scarce. Moreover, several authors have recently stated that groundwater-surface water interactions and groundwater hydrology require further analysis to better estimate the contribution of GHGs dissolved in groundwater into atmospheric fluxes at a local scale (Hinshaw and Dahlgren 2016; Jurado et al. 2018; Vidon and Serchan 2016). The objectives of this study are to (1) investigate the occurrence and examine the sources of GHGs in the river–groundwater interface and (2) evaluate the contribution of indirect GHG emissions from groundwater into surface waters. To this end, GHGs, major and minor ions and stable isotopes were sampled over 8 months in a small river catchment (Triffoy) located in the Walloon Region (Belgium).

Materials and methods

Study area

The Triffoy River catchment, with an area of 30.31 km2, is in the natural region of Condroz in Wallonia (Belgium; Fig. 1) and is an agricultural catchment where land use is dominated by cropland (48%) and grassland (38%). The remaining territory is occupied by urban areas (7%), forests (6%) and natural environments (1%). There are no industries in the whole catchment but NO3 concentrations can exceed the limit of good status during winter due to leaching of agricultural soil NO3 residue by infiltrating water (Brouyère et al. 2015, 2017). The average annual rainfall is of 900 mm and the average annual temperature is of 10 °C.
Fig. 1

Location and main aquifers of the Triffoy River catchment (Belgium). NL The Netherlands, LU Luxemburg, DE Germany, FR France

The Triffoy River intersects geological formations of Palaeozoic age, from Devonian to Carboniferous (Briers et al. 2016a). It flows through a Carboniferous limestone syncline located between two Frasnian-Famennian sandstone crests. At the base of Carboniferous limestone, the Hastarien shales constitute impermeable hydrogeological barriers separating the Carboniferous limestone aquifer from the Famennian sandstone aquifer. The sandstone aquifer is limited in extension and capacity, while in contrast, the Carboniferous limestone aquifer is an important groundwater reservoir that belongs to one of the most productive groundwater bodies of Wallonia (RWM021, Fig.1). The limestone aquifer is exploited by two water catchments: Jamagne (2,600,000 m3 year−1) and the Compagnie Intercommunale Liégeoise des Eaux (CILE, 700,000 m3 year−1).

Previous studies carried out in this basin reported two different types of river–groundwater interactions (Briers et al. 2016b): (1) gaining streams where water level is higher in the groundwater, feeding river and helping to maintain its base flow and (2) losing streams where river water recharges the aquifer. The stretch of river monitored in this study is a gaining stream (Fig. 1), and therefore it is suitable to quantify the groundwater contribution to GHGs emissions from rivers. On average, it was estimated that 92% of the Triffoy River baseflow comes from groundwater recharge (Briers et al. 2016c).

A river segment of 2 km (from Jamagne to State river sampling locations, Fig.1) was monitored over 8 months using river gauging and pressiometric and temperature probes installed in piezometers and in the river (MPZ river sampling location, Fig. 1). The monitoring network for the analysis of GHGs is composed by three river sampling locations (Jamagne, MPZ and State) and seven groundwater observation points—five shallow piezometers (MP-4, MP2-3, MP2-6, MP3-3 and MP3-6) and two springs (S1 and S2). The location of these points and the characteristics of the piezometers are summarized in Fig. S1 and Table S1 of the electronic supplementary material (ESM).

Groundwater and river sampling

A total of six field campaigns were carried out from October 2016 to May 2017 (October (C1) and December (C2) in 2016 and January (C3), February (C4), March (C5) and May (C6) in 2017). In all, 40 samples were collected from groundwater and 18 from the Triffoy River at different locations (Fig. 1). Before sampling, the piezometers were purged by pumping three well volumes to remove the stagnant water and samples were collected when field parameters were stabilised. Temperature (°C), electrical conductivity (EC, μS/cm), pH and DO (mg/L) were measured with a portable multi-probe (YSI 556 MPS) within a flow-through cell and samples were stored in a field refrigerator and taken to the laboratory at the end of the sampling day.

Groundwater samples were collected through tubing, avoiding any contact with the atmosphere. Sampling in surface waters was carried out using a 1.7-L Niskin bottle (General Oceanics) and samples for CH4 and N2O were transferred with tubing from the Niskin bottle to 50-ml borosilicate serum bottles that were poisoned with a saturated solution of HgCl2 (200 μl), sealed with a butyl stopper and crimped with an aluminium cap. Four polypropylene syringes of 60 ml for measurements of the partial pressure of CO2 (pCO2) were filled from each sampling point. The pCO2 is expressed in parts per million (i.e., ppm by volume and it is equivalent to micro atmospheres). For the general chemistry (major and minor ions), groundwater samples were collected in polypropylene bottles of 180 ml for major and minor ions and 125 ml for metals—iron (Fe) and manganese (Mn). Metal samples were filtered through a 0.45-μm polyethersulphone and micro-quartz fibre filter and acidified with 1 ml of HCl 12 N for sample preservation. Samples for NO3 isotopes were collected in polypropylene bottle of 60 ml and filtered through 0.22-μm nylon filter. Samples to determine dissolved organic carbon (DOC) were filtered through 0.22-μm nylon filter and stored in 40-ml borosilicate vials with polytetrafluoroethylene (PTFE) coated septa and poisoned with 100 μl of H3PO4 (85%).

Analytical methods

The dissolved concentrations of N2O and CH4 were analysed with the headspace equilibration technique (25 ml of N2 headspace in 50-ml serum bottles) and measured by gas chromatography (GC) fitted with electron capture detection (ECD, SRI 8610C) for N2O and flame ionization detection (FID) for CH4. The SRI 8610C GC-ECD-FID was calibrated with certified CH4:CO2:N2O:N2 mixtures (Air Liquide Belgium) of 0.2, 2 and 6 ppm N2O and of 1, 10 and 30 ppm CH4. The pCO2 was measured in the field using an infrared gas analyser (Li-Cor Li-840) a few minutes after sampling by creating a headspace with ambient air in the polypropylene syringes (1:1 ratio of air and water; Abril et al. 2015). The Li-840 was calibrated with a suite of CO2:N2 mixtures (Air Liquide Belgium) with mixing ratios of 388, 813, 3,788, 8,300 and 19,150 ppm CO2. The reproducibility of the measurements was ±3.2, ±3.9 and ± 2.0% for N2O, CH4 and pCO2, respectively. Major ions (Na+, Mg2+, K+, Cl, SO42− and NO3) and minor ions (NO2 and NH4+) were measured by ion chromatography via a specific ion exchange resin and a conductivity detector. Calcium (Ca2+) concentrations and alkalinity were obtained by potentiometric titration in the laboratory, whereas Fe and Mn concentrations were obtained by atomic absorption spectrometry. Nitrogen (δ15NNO3) and oxygen (δ18ONO3) isotope analyses of NO3 were determined by a mass DELTA V plus spectrometer plus a GasBench II from Thermo using the denitrifier method that convert all sampled NO3 to N2O (Sigman et al. 2001; Casciotti et al. 2002). The notation was expressed in terms of delta (δ) per mil (‰) relative to the international standards for the environmental isotopes (V-SMOW for δ18O and AIR-N2 for δ15N of NO3). The reproducibility of NO3 isotope samples was ±0.4‰ for δ15N and ± 1.6‰ for δ18O of NO3. The NO3 isotope results represent the mean value of true double measurements of each sample. DOC concentration was determined with a wet oxidation total organic carbon analyser (IO Analytical Aurora 1030 W) coupled with an EA-IRMS (ThermoFinnigan DeltaV Advantage).

Indirect GHG emissions from groundwater

The indirect GHG emissions from groundwater to the river (EGHG-Gw) were evaluated using hydrogeological data and the dissolved concentrations of GHGs measured in the groundwater as follows:
$$ {E}_{\mathrm{GHG}\hbox{-} \mathrm{Gw}}=\frac{Q_{\mathrm{dis}}\times \left[{C}_{\mathrm{GHG}-\mathrm{Gw}}-{C}_{\mathrm{GHG}\hbox{-} \mathrm{Eq}}\right]}{A} $$
(1)
where Qdis is groundwater discharge into the Triffoy River (m3 day−1), CGHG-Gw is the measured concentration of a given GHG in groundwater observation points (μg L−1), CGHG-Eq is the GHGs air-equilibrated water concentration and A is the area of the river between the upstream and the downstream river sampling locations (0.51 ha). Groundwater discharge into the Triffoy River (Qdis) was estimated by the difference in stream flow rate between the upstream (Jamagne, Qin) and the downstream (State, Qout) river sampling locations (Fig. 1). Note that groundwater was considered the only recharge source of the river because baseflow conditions prevailed during all the monitoring period. Hence, Eq. (1) represents the maximal flux of GHGs from groundwater to the river.
The fluxes of GHGs from groundwater to the river (Eq. 1) were compared with those from the river surface to the atmosphere (EGHG-Riv). The latter were computed according to:
$$ {E}_{\mathrm{GHG}\hbox{-} \mathrm{Riv}}=k\times \Delta G=k\times \left[{C}_{\mathrm{GHG}\hbox{-} \mathrm{Riv}}-{C}_{\mathrm{GHG}\hbox{-} \mathrm{Eq}}\right] $$
(2)
where k is the gas transfer velocity and ΔG is the air-water gradient (Δ) of a given gas (G). The air water gradient is the difference between the measured concentration of a given GHG in river water (CGHG-Riv, μg L−1) and the GHG air-equilibrated water concentration (CGHG-Eq, μg L−1). k was calculated from the gas transfer velocity normalised to a Schmidt number of 600 (k600) with the Schmidt numbers of N2O, CH4 and CO2, computed from in-situ water temperature according to Wanniknhof (1992). k600 (cm h−1) was computed with the parameterisation of Raymond et al. (2012) as a function of stream velocity (v in m s−1) and slope of the river channel (S is 0.0135, unitless):
$$ {k}_{600}=2.02+2,841\times v\times S $$
(3)

This parameterisation was derived from a compilation of gas tracer experiments in small- to medium-sized rivers and streams, and is then adequate to compute k600 in the Triffoy River. Note that the fluxes computed using Eqs. (1) and (2) should be similar if groundwater is the only source of GHGs to the river (i.e., there are no processes that consume or produce these GHGs in the river–groundwater interface).

Finally, the indirect groundwater N2O emissions were also estimated at catchment scale using the Intergovernmental Panel on Climate Change methodology (IPCC 2006) as follows:
$$ {E_{{\mathrm{N}}_2\mathrm{O}}}_{-\mathrm{GW}}=0.3\times \mathrm{NLeach}\times \mathrm{E}{\mathrm{F}}_{5\mathrm{g}}=0.3\times \mathrm{NLeach}\times \frac{C_{{\mathrm{N}}_2\mathrm{O}-\mathrm{N}}}{C_{\mathrm{N}{\mathrm{O}}_3^{-}-\mathrm{N}}} $$
(4)

This method considers that 30% of fertiliser and manure N applied to soils in agricultural areas is leached to groundwater (NLeach). The EF5g is the emission factor from groundwater and it is defined as the mass ratio of the dissolved concentrations of N2O (CN2O−N) and NO3 (CNO3−N) in groundwater.

Results and discussion

Climatic conditions, water levels and groundwater discharge

Data regarding weather conditions and water levels help to understand the water dynamics in the river–groundwater interface. Figure 2 shows rainfall (mm), temperature (°C) and water levels (in meters above sea level, m.a.s.l.) from December 2016 to May 2017. Total rainfall was 311 mm from October 2016 to May 2017. This value is low compared to the average monthly precipitation for the period 2012–2015 (311.5 mm vs. 561.6 mm; Table S2 of the ESM). All sampled months, except March, presented a lower amount of precipitation than the previous years. The driest months were April 2017 and December 2016 with total precipitations of 15.7 and 21.4 mm, respectively (Fig. 2a). In December 2016, the amount of precipitation was 5 times lower than the average monthly precipitation for 2012–2015 (21.4 mm vs. 98.4 mm). Conversely, November was a relatively wet month with 60 mm of precipitation. During the studied period, daily air temperature ranged from −9 °C (January 2017) to 19.6 °C (March 2017) with an average value of 4.5 °C (Fig. 2b). Diurnal air temperature variation turned out to be large. In contrast, the temperatures of river water and, especially, of groundwater were more constant. Groundwater temperatures ranged from 7.9 to 9 °C with an average temperature of 8.2 °C. River water temperatures varied from 2.3 to 12.8 °C with an average temperature of 7.9 °C, which was similar to the average groundwater temperature.
Fig. 2

a Daily rainfall (mm), b air, river water and groundwater temperature (T) and (c) water level (m.a.s.l.) of the river water in MPZ sampling location and of groundwater at observation points MP2-6, MP3-6 and MP-4. Daily rainfall (mm) was measured by the Walloon Public Service at Modave station

Figure 2c shows the evolution of water levels (m.a.s.l.) in the river (MPZ sampling location) and in groundwater (piezometers MP-4, MP2-6 and MP3-6). River and groundwater levels were relatively constant during the sampling period. Water levels slightly increased from January to March 2017 after rain events and progressively decreased due to the scarcity of rain in April and May. It is also important to point out that river water level was always lower than the groundwater levels, indicating a continuous groundwater discharge to the river. This observation is also supported by temperature measured in the river because it followed the same pattern as groundwater temperature, although it was also partly influenced by air temperature (Fig. 2b).

As pointed out before, groundwater was considered the only source of recharge to the Triffoy River (i.e., 100% groundwater) and the contribution of runoff was likely to be insignificant due to the scarcity of rain events during the sampling period (Fig. 2a). The average groundwater discharge for the sampling period was 5,870 ± 1,310 m3 day−1 and it was higher during the colder months (January and March, being 7,450 and 7,040 m3 day−1) compared to the most temperate ones (May 2017 being 3,840 m3 day−1).

Hydrochemistry of the Triffoy River basin

General hydrochemistry

Understanding the interactions between groundwater and surface water is a key issue to quantify the contribution of groundwater as an indirect source of GHGs via rivers, especially in gaining rivers where groundwater is the main source of river recharge. Figure 3 shows the average concentrations for major ions, metals, redox indicators and GHGs in the Triffoy River versus the average concentrations in the aquifer from October 2016 to May 2017. It can be observed that major ions presented similar concentrations in the river and in the aquifer, indicating that groundwater clearly controlled the chemical composition of the Triffoy River.
Fig. 3

Average concentrations for major ions (mg L−1), redox indicators (mg L−1), metals (μg L−1) and GHG (μg L−1 for N2O and CH4 and ppm for pCO2) in the Triffoy River and in the aquifer

The hydrochemical conditions of groundwater and river water are described using the in-situ parameters measured in the field and major ions (Table 1; Fig. 4). Groundwater pH values ranged from 7 to 7.8 (average is 7.4 ± 0.2). River pH values were slightly higher than those from groundwater with an average value of 8.0 ± 0.2. Average EC values were similar in groundwater and river water, being 673 ± 35 μS/cm and 665 ± 53 μS/cm, respectively. Groundwater concentrations of DO and DOC displayed lower values than river water (Table 1; Fig. 3). Averages DO and DOC concentrations were 4.8 ± 1 and 0.94 ± 0.47 mg L–1 in groundwater and 9.2 ± 1.1and 1.4 ± 0.70 mg L–1 in the river.
Table 1

Range and mean groundwater (seven observation points including MP–4, MP3–6, MP3–3, MP2–6, MP2–3, S1 and S2) and river water (three sampling locations known as Jamagne, MPZ and State) concentrations for some major ions (mg L−1), GHG (μg L–1 for N2O and CH4 and ppm for pCO2), DOC (mg L–1) and in-situ parameters in the Triffoy River basin

 

HCO3 (mg L–1)

Ca2+ (mg L–1)

Mg2+ (mg L–1)

NO3 (mg L–1)

N2O (μg L–1)

CH4 (μg L–1)

pCO2 (ppm)

DOC (mg L–1)

DO (mg L–1)

T (°C)

pH

MP-4

377.2–399.2 (384.3 ± 8)

94.4–98.8 (97 ± 1.5)

31.3–32.8 (32 ± 0.64)

14.8–19.2 (16.9 ± 1.9)

30.9–41.3 (35 ± 4)

0.07–0.16 (0.10 ± 0.04)

17,659–19,530 (18,312 ± 667)

0.47–2 (0.98 ± 0.59)

4.1–6.2 (5.3 ± 0.95)

6.5–12 (10.2 ± 1.9)

7.1–7.7 (7.4 ± 0.21)

MP3-6

338.1–343 (340.1 ± 1.7)

96.8–97.6 (97.2 ± 0.26)

28.1–29.5 (28.7 ± 0.59)

20.84–24.1 (22.3 ± 1.2)

71.8–87.6 (79.1 ± 5.2)

0.02–0.19 (0.09 ± 0.06)

12,004–13,437 (12,793 ± 582)

0.39–1.5 (0.90 ± 0.46)

3.9–7.2 (4.7 ± 1.2)

7.4–10.8 (10.1 ± 1.4)

7–7.7 (7.4 ± 0.22)

MP3-3

350.2–356.3 (353.6 ± 2.2)

97.8–98.3 (98.1 ± 0.26)

29.5–31.4 (30.1 ± 0.79)

19.9–22.9 (21 ± 1.1)

55.3–71.5 (61.6 ± 6.1)

0.01–0.22 (0.11 ± 0.09)

13,137–15,066 (14,107 ± 784)

0.45–1.7 (0.91 ± 0.52)

4.3–4.7 (4.5 ± 0.19)

7.8–10.6 (9.6 ± 1.3)

7.2–7.7 (7.4 ± 0.21)

MP2-6

371.1–381.9 (374.9 ± 4.4)

98.5–99.3 (98.7 ± 0.28)

29.6–31.4 (30.7 ± 0.63)

16.7–21.2 (19.1 ± 1.7)

37.7–46.7 (43.5 ± 3.1)

0.1–0.18 (0.15 ± 0.03)

15,594–18,539 (17,365 ± 1,164)

0.57–1.6 (1.1 ± 0.43)

3.7–7.9 (5.1 ± 1.5)

7–10.8 (9.6 ± 1.5)

7–7.7 (7.4 ± 0.26)

MP2-3

385.8–393.2 (391 ± 3)

99.7–100.6 (100.2 ± 0.33)

30.6–31.6 (31.2 ± 0.41)

14.2–16.4 (15.6 ± 0.89)

26–34.1 (30.9 ± 3.1)

0.12–2.61 (0.98 ± 1.1)

19,020–21,897 (20,000 ± 1,138)

0.56–1.9 (1.1 ± 0.50)

3.3–5.5 (4.3 ± 0.79)

7–10.5 (8.8 ± 1.4)

7.3–7.7 (7.3 ± 0.28)

S1

336.4–331.6 (334.5 ± 1.8)

102.8–103.6 (103.2 ± 0.28)

24.3–25.2 (24.8 ± 0.27)

21.4–24.7 (22.9 ± 1.3)

34.5–44.4 (39.5 ± 3.9)

0.64–4.8 (1.7 ± 1.6)

9,353–8,285 (8,833 ± 400)

0.37–1.9 (0.87 ± 0.57)

4.7–6.2 (5.3 ± 0.63)

8.3–10.7 (9.7 ± 0.9)

7.1–7.8 (7.5 ± 0.29)

S2

334.2–337.9 (336 ± 1.4)

102.6–103 (102.7 ± 0.17)

25.1–26.1 (25.5 ± 0.36)

22–25.2 (23.4 ± 1.4)

58.1–63.9 (59.5 ± 2.2)

0.03–0.14 (0.10 ± 0.04)

11,023–11,965 (11,399 ± 343)

0.50–1.6 (0.80 ± 0.45)

3.2–5.1 (4 ± 0.63)

9.3–10.1 (9.8 ± 0.38)

7.4–7.8 (7.5 ± 0.15)

Jamagne

292.9–382.2 (326.5 ± 32.3)

85.8–119 (100.2 ± 11.3)

23.2–29.2 (24.8 ± 2.3)

15.5–29.6 (22.3 ± 4.8)

0.58–19.5 (7.6 ± 4.1)

3–71.5 (18.9 ± 26)

859–2,081 (1,612 ± 488)

1.3–2.8 (2.1 ± 0.70)

8.1–11.3 (9.3 ± 1.40)

0.45–12.9 (6.6 ± 5.2)

7.6–8.5 (8.2 ± 0.34)

MPZ

330.8–349.1 (337 ± 7.5)

97.3–105.1 (100.9 ± 2.8)

24.6–27.2 (25.7 ± 1)

19.8–25.2 (22.2 ± 2.3)

12.8–18 (15.1 ± 2.1)

0.28–0.65 (0.42 ± 0.13)

3,812–4,975 (4,390 ± 459)

0.7–2.2 (1.3 ± 0.60)

8–10 (9.2 ± 0.78)

5.9–10.1 (7.9 ± 1.8)

7.8–8 (7.9 ± 0.09)

State

332–338.2 (334.3 ± 2.3)

100.1–102.5 (101.3 ± 0.86)

24.8–26.5 (25.4 ± 0.57)

19.8–22.9 (21.5 ± 1.3)

9.7–12.4 (10.9 ± 1)

0.86–2.1 (1.4 ± 0.48)

3,281–3,973 (3,502 ± 253)

0.6–1.5 (94 ± 0.34)

7.5–11.1 (9 ± 1.3)

7.1–11.3 (9.3 ± 1.8)

7.8–8.1 (7.9 ± 0.10)

Fig. 4

Spatial and temporal distribution of some major ions, redox indicators and GHG in river water (SW) and groundwater (GW) for the samplings campaigns carried out in December 2016 (ab) and March 2017 (cd). Note that the name of the sampling points is on top of the x-axis

Major ion compositions showed that groundwater and river water were of Ca-(Mg)-HCO3 type accounting for all sampling campaigns (see Fig. S2 of the ESM). The range and average concentrations and standard deviations for bicarbonate (HCO3), Ca2+, Mg2+ and NO3 for the groundwater observation points and the three river locations are shown in Table 1. Note that NH4+ concentrations are not included because they were below detection limit. The concentrations of these tracers did not present large variation, either spatially or temporarily in groundwater and river water samples (Fig. 4 and Table S3 of the ESM)—for instance, average NO3 concentrations ranged from 18.6 to 22.4 mg L–1 and average HCO3 concentrations ranged from 353.3 to 361.7 mg L–1 in groundwater (Table S3 of the ESM).

Occurrence of GHGs

Average GHG concentrations in groundwater and river water are summarized in Table 1. Groundwater was largely oversaturated in N2O and pCO2, while only slightly oversaturated in CH4 compared with the atmospheric equilibration concentrations (0.55 μg L–1 for N2O, 400 ppm for CO2 and 0.056 μg L–1 for CH4). N2O concentrations ranged from 26 to 87.6 μg L–1 (average concentration of 50.1 ± 16.7 μg L–1, CH4 concentrations ranged from 0.01 to 4.8 μg L–1 (average concentration of 0.45 ± 0.89 μg L–1 and pCO2 values varied from 8,285 to 21,897 ppm (average of 14,569 ± 3,843 ppm). Average N2O concentrations in groundwater were higher in temperate months (October 2016 and May 2017 being 55.2 and 54.1 μg L–1, respectively) than in winter months (minimum average concentration was 46.1 μg L–1 in January 2017; Table C of the supplementary material). Average pCO2 concentrations were constant from October 2016 to February 2017 (around 14,200 ppm) and the highest value was detected in May 2017 (15,402 ppm; Table S3 of the ESM).

Average GHG concentrations in river water were 10 ± 6.3 μg L–1 for N2O, 6.9 ± 16.6 μg L–1 for CH4 and 3,168 ± 1,253 ppm for pCO2. The concentrations of N2O and pCO2 in groundwater were systematically higher than those found in river water (Figs. 3 and 4b,d). In contrast, dissolved CH4 concentrations were lower in groundwater than in the river (average concentrations were 0.45 ± 0.89 μg L–1 vs. 6.9 ± 16.6 μg L–1 respectively). This observation shows that groundwater was not a source of CH4, as also concluded by Borges et al. (2018) based on large-scale analysis in the Meuse basin in Wallonia.

Stable isotopes

Figure 5 shows δ15NNO3 – δ18ONO3 compositions for the groundwater (red dots) and river samples (blue dots) and boxes representing the isotopic compositions of possible NO3 sources (Kendall 1998; Mayer 2005). The isotopic compositions for δ15NNO3 varied from +4.9 to +7.3‰ (average composition +6‰ ± 0.57) for groundwater samples and from +6.2 to +9.9‰ (average composition +7.5 ± 0.85) for river samples. The isotopic compositions for δ18ONO3 ranged from +1.1 to +6.8‰ (average composition +2.9 ± 1.7) for groundwater samples and from +1.9 to +6.9‰ (average composition +3.3‰ ± 1.5) for river samples. All groundwater and river samples agreed with the isotopic values of organic N from soil and/or the lightest values of δ15NNO3 coming from manure or sewage water.
Fig. 5

δ15N versus δ18O values of nitrate for river water (blue dots) and groundwater (red dots). The isotopic compositions for the nitrate sources are taken from Kendall (1998) and Mayer (2005)

Processes that produce and/or consume GHGs in groundwater

Nitrous oxide

The occurrence of N2O depends on geochemical conditions prevailing in groundwater. Oxic conditions observed in groundwater might indicate that N2O resulted from nitrification rather than denitrification since the latter is generally associated with low concentrations of DO. The positive correlation between NO3 and N2O (r = 0.62, Fig. S3a of the ESM) also suggests that nitrification was the main process for the accumulation of N2O in groundwater. Such positive correlation was also observed in other aquifers located below agricultural catchments where nitrification was the main N2O production mechanism (Gardner et al. 2016; Hiscock et al. 2003; Vilain et al. 2012). In addition, the positive correlation between Cl (conservative tracer) and NO3 (r = 0.96, Fig. S3b of the ESM) might indicate that NO3 was not affected by denitrification because their concentrations remained constant during the sampling campaigns (Fig. 4a,c).

The values of NO3 stable isotopes also suggest N2O was produced by nitrification because all groundwater samples fell in the box of soil N (Fig. 5). Values of δ15NNO3 found in groundwater (δ15NNO3 = +6‰) are much lower than those expected from denitrification processes which usually present δ15NNO3 > +15‰ (Otero et al. 2009; McAleer et al. 2017). Experimental studies (e.g., Andersson and Hooper 1983; Mayer et al. 2001) have pointed out that δ18ONO3 generated by nitrification can be calculated as follows:
$$ {\updelta}^{18}{\mathrm{O}}_{\mathrm{NO}3}=2/3\times {\updelta}^{18}{\mathrm{O}}_{\mathrm{water}}+1/3\times {\updelta}^{18}{\mathrm{O}}_{\mathrm{atmos}} $$
(5)

Equation (5) shows that two oxygens come from water and one from atmospheric oxygen during the conversion of NH4+ to NO3. For the Triffoy River catchment, using an isotopic value for δ18Owater of −7.3‰ obtained from a previous study (Briers et al. 2016d) and an isotopic value for δ18Oatmos of +23.5‰ (Kroopnick and Craig 1972), the evaluated δ18ONO3 is equal to +3‰. This value is very close to the average value for δ18ONO3 observed in the collected groundwater samples (+2.9 ± 1.7‰); hence, N2O found in groundwater seems to be produced due to nitrification in the unsaturated zone.

Methane

The oxic conditions that prevailed underground in the Triffoy River basin were not favourable for the accumulation of CH4 in groundwater. The average concentrations in river water were higher than those in groundwater (6.9 μg L–1 vs. 0.45 μg L–1, Fig. 3), suggesting that groundwater was an insignificant source of CH4 in the river.

Carbon dioxide

CO2 enrichment in groundwater might occur when rainwater percolates through the soil, where CO2 is produced by processes such as microbial decomposition of organic matter (heterotrophic respiration) and root respiration (autotrophic respiration; Tan 2010), and subsequent leaching of CO2 to groundwater. These processes produce an enrichment of CO2 and groundwater pCO2 values are typically between 10 to 100 times higher than atmospheric pCO2. When the oversaturated groundwater is discharged in the Triffoy River, CO2 degassing into the atmosphere takes place. This situation leads to an increase of pH in the river water and the progressive precipitation of carbonate minerals. In fact, average saturation indexes (SIs, see Text S1 of the ESM) of carbonate minerals were higher in river water than in groundwater being 0.79 vs. 0.24 for calcite and 0.43 and −0.55 for dolomite, indicating that river water was slightly oversaturated with respect to calcite and dolomite (Table S4 of the ESM).

Other processes that might produce CO2 in groundwater are redox processes such as aerobic respiration and denitrification. Nevertheless, these processes were not likely to occur in the aquifer because of the presence of DO and NO3 in groundwater (see previous explanation that supports the occurrence of nitrification).

Evaluation of GHG emissions from groundwater

In this section, the importance of groundwater as an indirect source of GHGs to the atmosphere was assessed at local scale (per area of the river from Jamagne to State river sampling locations, see the following subsection ‘Local scale’). Afterwards, to place the groundwater GHG emissions in a broader context, the resulting average emissions were upscaled by dividing them by the total agricultural area of the Triffoy River basin (see section ‘Catchment scale’).

Local scale

The maximal contribution of GHG emissions from groundwater to the river was assessed using Eq. (1). Average GHG fluxes from groundwater resulted in 207 kg N2O ha−1 year−1, 1.6 kg CH4 ha−1 year−1 and 1.5 × 105 kg CO2 ha−1 year−1. These fluxes should be similar to those from the river to the atmosphere unless that there are other processes that consumed or produced N2O, CH4 and CO2 in the river–groundwater interface. Average fluxes evaluated from river surface to the atmosphere (Eq. 2) were similar to those evaluated with Eq. (1) for N2O and CO2 (126.9 kg N2O ha−1 year−1 and 9.7 × 104 kg CO2 ha−1 year−1, respectively) but much higher for CH4 (105 kg CH4 ha−1 year−1).

Monthly flux estimates using Eq. (1) for N2O (EN2O-Gw) and CO2 (ECO2-Gw) were systematically higher than those computed with Eq. (2) (EN2O-Riv and ECO2-Riv, except for N2O in May; Fig. 6). This observation indicates that groundwater contributed to the emissions of these two gases to the atmosphere but part of the N2O and CO2 concentrations might had been consumed in the river–groundwater interface. If these GHGs were not consumed before reaching the river, their average concentrations should have been similar to those observed in groundwater. However, groundwater concentrations for N2O and pCO2 were 5 times higher than those measured in the river (50.1 μg L–1 vs. 10 μg L–1 for N2O and 14,569 ppm vs. 3,168 ppm for pCO2). The biggest difference in N2O and CO2 emissions (using Eqs. 1 and 2) occurred in January 2017 when groundwater discharge into the river was maximum. It is important to mention that N2O emissions from the river to the atmosphere (Eq. 2, EN2O-Riv) were higher than those from groundwater (Eq. 1, EN2O-Gw) in May 2017 (Fig. 6). This observation might be explained by the low groundwater discharge into the river compared to other months (3,840 m3 day−1) and the slightly higher concentration of N2O found in river water in May 2017 (Table S3 of the ESM) but also it could indicate an inflow of N2O produced in the river from upstream. The opposite situation was observed for CH4, whose emissions from the river to the atmosphere (ECH4-Riv) were always one to two orders of magnitude higher than the input of CH4 from the groundwater (ECH4-Gw; Fig. 6). This implies that the emission of CH4 from the river to the atmosphere was almost exclusively sustained by in-situ production most probably in riverbed sediments or riparian areas.
Fig. 6

Flux (E) of a N2O, b CO2 and c CH4 from the aquifer to the river (Egw) and from the river to the atmosphere (Eriv) from October 2016 to May 2017. Fluxes are expressed in kg ha−1 year−1 (per surface of river). Note the logarithmic scale for CH4 fluxes

Catchment scale

To evaluate the GHGs emissions at catchment scale, the average EGHG-Gw (Eq. 1) were divided by the agricultural area of the Triffoy basin (26.1 km2) instead of the surface of the river (5.1 × 10−3 km2). This resulted in average fluxes of 0.040 kg ha−1 year−1 for N2O, 3.0 × 10−4 kg ha−1 year−1 for CH4 and 29.8 kg ha−1 year−1 for CO2. Note that these fluxes were evaluated considering a river stretch of 2 km but the total length of the Triffoy River is 12 km (Fig. 1).

Indirect groundwater N2O emissions at catchment scale were also evaluated applying the Intergovernmental Panel on Climate Change (IPCC) method (Eq. 4) that requires the evaluation of the emission factor for groundwater (EF5g). The EF5g coefficient evaluated in this study is 3 times higher than the default value proposed by the IPCC (0.0069 ± 0.0018 vs. 0.0025). Considering that N leaching to groundwater was estimated to be 5.4 kg N ha−1 year−1 in the aquifers of the Condroz region (SPW 2010), the resulting indirect N2O emissions from groundwater were 0.037 kg N2O-N ha−1 year−1 (0.058 kg N2O ha−1 year−1). This value is similar to the one evaluated using groundwater discharge in the river (0.040 kg N2O ha−1 year−1) and other N2O fluxes from groundwater evaluated in aquifers located below agricultural lands—for example, similar estimates of indirect N2O fluxes from groundwater were obtained using the IPCC methodology in the Orgeval catchment in France and major UK aquifers, being 0.035 and 0.04 kg N2O-N ha−1 year−1, respectively (Vilain et al. 2012; Hiscock et al. 2003). The IPCC approach presents some limitations because N leaching to groundwater varies from one site to another—for instance, Jahangir et al. (2013) reported that N leached varied from 8 to 38% of the total N input in four different agricultural settings, resulting in indirect N2O groundwater fluxes ranging from 0.07 to 0.24 kg N2O-N ha−1 year−1. Slightly lower fluxes (0.004 kg N2O-N ha−1 year−1) were evaluated in the Choptank Basin in the USA (Gardner et al. 2016). The authors pointed out that groundwater was a minor source of total biogenic N2O emissions (15% on average) from strongly gaining agricultural streams but it was the primary source of N2 (3.5 kg N2-N ha−1 year−1). Similarly, von der Heide et al. (2009) evaluated that N2O fluxes from the shallow groundwater of the Fuhrberger Feld aquifer (Germany) were 1–2 orders of magnitude lower than N2O flux at soil surface (0.044 vs. 1 kg N2O-N ha −1 year−1), and thus groundwater was a negligible pathway of atmospheric emissions.

CO2 and CH4 indirect fluxes from groundwater in agricultural areas have been less studied than those of N2O. For instance, when Jahangir et al. (2012) evaluated the dissolved C delivery to surface water through groundwater in selected agricultural aquifers of Ireland, groundwater CO2 export was up to 314 kg C ha−1 year−1 (1,151 kg CO2 ha−1 year−1), whereas CH4 export was low (from 0.013 to 2.30 CH4 ha−1 year−1). The authors concluded that the dissolved C loss to surface waters via groundwater was not significant compared to total carbon (TC) content of the topsoil (0.06–0.18% of TC). Similarly, Wang et al. (2015) evaluated that CO2 lost via groundwater to the stream was approximately 73 kg CO2 ha−1 year−1 in the Hongfeng Lake catchment (China, 1,596 km2), which was insignificant compared with soil CO2 emission.

To sum up, groundwater is likely to be an important source of N2O and CO2 in gaining streams but when measures are upscaled at the catchment scale, these fluxes are probably relatively modest. Thus, indirect GHG emissions from groundwater seem to be a minor pathway of GHG atmospheric emissions but their quantification would help to better evaluate the C and N budgets in agricultural catchments.

Conclusions

As GHG concentrations have significantly increased in the atmosphere, studying their dynamics from natural systems remain a major concern. This study investigated the occurrence of N2O, CH4 and CO2 and quantified the contribution of groundwater as an indirect source of these GHGs via river water in the agricultural catchment of the Triffoy River (Belgium). Average groundwater concentrations for N2O and pCO2 were higher than those found in the river samples (50 vs. 10 μg L–1 and 14,569 vs. 3,168 ppm, respectively), suggesting that groundwater could be an indirect source of GHGs to the atmosphere. Nitrification was likely to be the main source of N2O in groundwater, an observation supported by the positive relationship between N2O and NO3, the presence of DO and NO3 and the absence of NH4+ in groundwater. The oxic conditions found in groundwater were not suitable for the accumulation of CH4 in the aquifer and it might be generated in riverbed or riparian zone sediments.

The role of groundwater as an indirect source of GHGs in the river–groundwater interface was evaluated through the net groundwater discharge into the river (Eq. 1) and compared to the inputs from the river to the atmosphere (Eq. 2). Average fluxes obtained for N2O and CO2 using both approaches were similar (207 vs. 126.9 kg N2O ha−1 year−1 and 1.5 × 105 vs. 9.7 × 104 kg CO2 ha−1 year−1), showing that groundwater was a source of release of these GHGs into the atmosphere. The opposite situation was observed for CH4, whose average emissions from groundwater were two orders of magnitude lower than those evaluated from the river to the atmosphere (1.6 vs. 105 kg CH4 ha−1 year−1). This observation indicates that groundwater was an insignificant source of CH4 to the atmosphere. Overall, groundwater in the studied gaining stream was a source that contributed to N2O and CO2 atmospheric emissions, but when these emissions were up-scaled (from the river surface to the catchment area), the resulting fluxes seemed to be insignificant compared to other sources (i.e., direct N2O and CO2 emissions from soils). Nevertheless, their quantification would better constrain N and C budgets in natural systems.

It is suggested that future research efforts should be devoted to investigating the dynamics of GHGs in groundwater, soil and river water over long time periods (i.e., hydrological year) and a wide range of flow conditions (wet and dry periods) to better understand the relative importance of each compartment as a source of GHGs to the atmosphere at a stream scale. Particular effort should be directed to improve the understanding of GHGs production and consumption in the groundwater–river transition zone (e.g., streambed hyporheic sediments). This point will allow researchers to better constrain global N2O, CH4 and CO2 budgets at the river–groundwater interface and thus the N and C budgets.

Notes

Acknowledgements

We thank Marc-Vincent Commarieu for help in gas chromatograph (GC) measurements. Further, we wish to thank the editor and the two anonymous reviewers for their comments and suggestions, which helped to improve the quality of the paper.

Funding information

A. J. and E. P. gratefully acknowledge the financial support from the University of Liège and the EU through the Marie Curie BeIPD-COFUND postdoctoral fellowship programme (2015–2017 and 2014–2016 fellows from FP7-MSCA-COFUND, 600405). A. V. B. is a senior research associate at the Fonds National de la Recherche Scientifique (FNRS). This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 675120. A GC was acquired with funds from FNRS (FNRS, 2.4.598.07).

Supplementary material

10040_2018_1834_MOESM1_ESM.pdf (453 kb)
ESM 1 (PDF 453 kb)

References

  1. Abril G, Bouillon S, Darchambeau F, Teodoru CR, Marwick TR, Tamooh F, Omengo FO, Geeraert N, Deirmendjian L, Polsenaere P, Borges AV (2015) Technical note: large overestimation of pCO2 calculated from pH and alkalinity in acidic, organic-rich freshwaters. Biogeosciences 12(1):67CrossRefGoogle Scholar
  2. Andersson KK, Hooper AB (1983) O2 and H2O are each the source of one O in NO2 produced from NH3 by nitrosomonas: 15N evidence. FEBS Lett 164:236–240CrossRefGoogle Scholar
  3. Beaulieu JJ, Tank JL, Hamilton SK, Wollheim WM, Hall RO, Mulholland PJ, Peterson BJ, Ashkenas LR, Cooper LW, Dahm CN, Dodds WK, Grimm NB, Johnson SL, McDowell WH, Poole GC, Valett HM, Arango CP, Bernot MJ, Burgin AJ, Crenshaw CL, Helton AM, Johnson LT, O’Brien JM, Potter JD, Sheibley RW, Sobota DJ, Thomas SM (2010) Nitrous oxide emission from denitrification in stream and river networks. Proc Nat Acad Sci USA 108:214–219CrossRefGoogle Scholar
  4. Bell RA, Darling WG, Ward RS, Basava-Reddi L, Halwa L, Manamsa K, Ó Dochartaigh BE (2017) A baseline survey of dissolved methane in aquifers of Great Britain. Sci Total Environ 601–602:1803–1813.  https://doi.org/10.1016/j.scitotenv.2017.05.191 CrossRefGoogle Scholar
  5. Borges AV, Darchambeau F, Lambert T, Bouillon S, Morana C, Brouyère S, Hakoun V, Jurado A, Tseng HC, Descy JP, Roland FAE (2018) Effects of agricultural land use on fluvial carbon dioxide, methane and nitrous oxide concentrations in a large European river, the Meuse (Belgium). Sci Total Environ 610:342–355CrossRefGoogle Scholar
  6. Borges AV, Darchambeau F, Teodoru CR, Marwick TR, Tamooh F, Geeraert N, Omengo FO, Guérin F, Lambert T, Morana C, Okuku E, Bouillon S (2015) Globally significant greenhouse gas emissions from African inland waters. Nat Geosci 8:637–642CrossRefGoogle Scholar
  7. Briers P, Jamin P, Ruthy I, Orban P, Brouyère S (2016a) Hydrogéologie du bassin du Hoyoux [Hydrogeology of the Hoyoux basin]. In: Atlas du Karst Wallon-Bassins versants du Hoyoux et de la Solières. Service Publique de la Wallonie, Belgium, pp 41–47Google Scholar
  8. Briers P, Orban P, Brouyère S (2016b) Caractérisation complémentaire des masses d’eau dont le bon état dépend d’interactions entre les eaux de surface et les eaux souterraines (ESO–ESU): délivrable D3.5—quantification des échanges nappe–rivière pour les bassins tests [Complementary characterization of water bodies whose good status depends on interactions between surface water and groundwater (GW–SW): deliverable D3.5—quantification of the nappe–river exchanges for the test basins]. University of Liege, BelgiumGoogle Scholar
  9. Briers P, Orban P, Brouyère S (2016c) Caractérisation complémentaire des masses d’eau dont le bon état dépend d’interactions entre les eaux de surface et les eaux souterraines (ESO–ESU): délivrable D4.1—développement d’indicateurs des interactions entre eaux souterraines et eau de surface [Complementary characterization of water bodies whose good status depends on interactions between surface water and groundwater (GW–SW): deliverable D4.1—development of indicators between groundwater and surface water interactions]. University of Liege, BelgiumGoogle Scholar
  10. Briers P, Orban P, Brouyère S (2016d) Caractérisation complémentaire des masses d’eau dont le bon état dépend d’interactions entre les eaux de surface et les eaux souterraines (ESO–ESU): délivrable D3.4—caractérisation hydrochimique et campagne isotopique pour la discrimination des sources de pollution ponctuelles et diffuses du nitrate [Complementary characterization of water bodies whose good status depends on interactions between surface water and groundwater (GW–SW): deliverable D3.4—hydrochemical characterization and isotopic campaign for the discrimination of point and non-point sources of nitrate pollution]. University of Liege, BelgiumGoogle Scholar
  11. Brouyère S, Briers P, Descy JP, Schmit F, Degré A, Orban P (2017) Caractérisation complémentaire des masses d’eau dont le bon état dépend d’interactions entre les eaux de surface et les eaux souterraines: délivrable D1. 8—rapport final [Complementary characterization of water bodies whose good status depends on interactions between surface water and groundwater: deliverable D1. 8—final report]. University of Liege, BelgiumGoogle Scholar
  12. Brouyère S, Briers P, Schmit F, Orban P, Hallet V, Delloye F, Descy JP (2015) Case study 2: groundwater–surface water interaction in limestone areas of the GWB BE_Meuse_RWM021 (Belgium). Technical report on groundwater associated aquatic ecosystems, Liege University, Belgium, pp 47–49Google Scholar
  13. Casciotti KL, Sigman DM, Hastings MG, Böhlke JK, Hilkert A (2002) Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method. Anal Chem 74:4905–4912CrossRefGoogle Scholar
  14. Cheung K, Klassen P, Mayer B, Goodarzi F, Aravena R (2010) Major ion and isotope geochemistry of fluids from coal bed methane and shallow groundwater wells in Alberta. Canada Appl Geochem 25:1307–1329CrossRefGoogle Scholar
  15. Deurer M, von der Heide C, Böttcher J, Duijnisveld WHM, Weymann D, Well R (2008) The dynamics of N2O near the groundwater table and the transfer of N2O into the unsaturated zone: a case study from a sandy aquifer in Germany. Catena 72(3):362–373CrossRefGoogle Scholar
  16. Fox RJ, Fisher TR, Gustafson AB, Jordan TE, Kana TM, Lang MW (2014) Searching for the missing nitrogen: biogenic nitrogen gases in groundwater and streams. J Agric Sci 152(S1):96–106CrossRefGoogle Scholar
  17. Gardner JR, Fisher TR, Jordan TE, Knee KL (2016) Balancing watershed nitrogen budgets: accounting for biogenic gases in streams. Biogeochemistry 127(2–3):231–253CrossRefGoogle Scholar
  18. Gilbert N (2012) One-third of our greenhouse gas emissions come from agriculture. Nat News, 31 October 2012.  https://doi.org/10.1038/nature.2012.11708
  19. Glavan M, Jamšek A, Pintar M (2017) Modelling impact of adjusted agricultural practices on nitrogen leaching to groundwater. In: Water quality. InTech, Richmond, BCGoogle Scholar
  20. Hasegawa K, Hanaki K, Matsuo T, Hidaka S (2000) Nitrous oxide from the agricultural water system contaminated with high nitrogen. Chemosphere-Global Change Sci 2(3):335–345CrossRefGoogle Scholar
  21. Hinshaw SE, Dahlgren RA (2016) Nitrous oxide fluxes and dissolved N gases (N2 and N2O) within riparian zones along the agriculturally impacted San Joaquin River. Nutr Cycl Agroecosys 105(2):85–102CrossRefGoogle Scholar
  22. Hiscock KM, Bateman AS, Mühlherr IH, Fukada T, Dennis PF (2003) Indirect emissions of nitrous oxide from regional aquifers in the United Kingdom. Environ Sci Technol 37(16):3507–3512CrossRefGoogle Scholar
  23. Hotchkiss ER, Hall RO Jr, Sponseller RA, Butman D, Klaminder J, Laudon H, Rosvall M, Karlsson J (2015) Sources of and processes controlling CO2 emissions change with the size of streams and rivers. Nat Geosci 8(9):696–699CrossRefGoogle Scholar
  24. IPCC (2014) Climate Change 2014: synthesis report. In: Core Writing Team, Pachauri RK, Meyer LA (eds) Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. IPCC, Geneva, Switzerland, 151 ppGoogle Scholar
  25. IPCC (2006) IPCC guidelines for national greenhouse gas inventories. Prepared by the National Greenhouse Gas Inventories Programme. IGES, JapanGoogle Scholar
  26. Jahangir MM, Johnston P, Khalil MI, Hennessy D, Humphreys J, Fenton O, Richards KG (2012) Groundwater: a pathway for terrestrial C and N losses and indirect greenhouse gas emissions. Agr Ecosys Environ 159:40–48CrossRefGoogle Scholar
  27. Jahangir MM, Johnston P, Barrett M, Khalil MI, Groffman PM, Boeckx P, Fenton O, Murphy J, Richards KG (2013) Denitrification and indirect N2O emissions in groundwater: hydrologic and biogeochemical influences. J Contam Hydrol 152:70–81CrossRefGoogle Scholar
  28. Johnson MS, Lehmann J, Riha S, Krusche AV, Richey JE, Ometto JPHB, Couto EG (2008) CO2 efflux from Amazonian headwater streams represents a significant fate for deep soil respiration. Geophys Res Lett 35:L17401CrossRefGoogle Scholar
  29. Jurado A, Borges AV, Pujades E, Hakoun V, Otten J, Knoeller K, Brouyère S (2018) Occurrence of greenhouse gases in the aquifers of the Walloon region (Belgium). Sci Total Environ 619–620:1579–1588CrossRefGoogle Scholar
  30. Kendall C (1998) Tracing sources and cycling of nitrate in catchments. In: Kendall C, McDonnell JJ (eds) Isotope tracers in catchment hydrology. Elsevier, Amsterdam, pp 519–576CrossRefGoogle Scholar
  31. Kroopnick P, Craig H (1972) Atmospheric oxygen: isotopic composition and solubility fractionation. Science 175(4017):54–55CrossRefGoogle Scholar
  32. Mayer B (2005) Assessing sources and transformations of sulphate and nitrate in the hydrosphere using isotope techniques. In: Isotopes in the water cycle. Springer, Dordrecht, The Netherlands, pp 67–89Google Scholar
  33. Mayer B, Bollwerk SM, Mansfeldt T, Hütter B, Veizer J (2001) The oxygen isotope composition of nitrate generated by nitrification in acid forest floors. Geochim Cosmochim Acta 65(16):2743–2756CrossRefGoogle Scholar
  34. McAleer EB, Coxon CE, Richards KG, Jahangir MMR, Grant J, Mellander PE (2017) Groundwater nitrate reduction versus dissolved gas production: a tale of two catchments. Sci Total Environ 586:372–389CrossRefGoogle Scholar
  35. Minamikawa K, Nishimura S, Sawamoto T, Nakajima Y, Yagi K (2010) Annual emissions of dissolved CO2, CH4, and N2O in the subsurface drainage from three cropping systems. Glob Chang Biol 16(2):796–809CrossRefGoogle Scholar
  36. Nikolenko O, Jurado A, Borges AV, Knӧller K, Brouyѐre S (2018) Isotopic composition of nitrogen species in groundwater under agricultural areas: a review. Sci Total Environ 621:1415–1432CrossRefGoogle Scholar
  37. Otero N, Torrentó C, Soler A, Menció A, Mas-Pla J (2009) Monitoring groundwater nitrate attenuation in a regional system coupling hydrogeology with multi-isotopic methods: the case of Plana de Vic (Osona, Spain). Agric Ecosyst Environ 133(1):103–113CrossRefGoogle Scholar
  38. Raymond PA, Zappa CJ, Butman D, Bott TL, Potter J, Mulholland P, Laursen AE, McDowell WH, Newbold D (2012) Scaling the gas transfer velocity and hydraulic geometry in streams and small rivers. Limnol Oceano: Fluids Environ 2(1):41–53CrossRefGoogle Scholar
  39. Rivett MO, Buss SR, Morgan P, Smith JWN, Bemment CD (2008) Nitrate attenuation in groundwater: a review of biogeochemical controlling processes. Water Res 42(16):4215–4232CrossRefGoogle Scholar
  40. SPW, DGARNE (2010) État des lieux de la masse d’eau souterraine RWM021 « Calcaires et Grès du Condroz » [State of play of the groundwater body RWM021 “Limestones and Sandstone of Condroz”]. Service Public de Wallonie, Wallonie, BelgiumGoogle Scholar
  41. Sigman DM, Casciotti KL, Andreani M, Barford C, Galanter M, Böhlke JK (2001) A bacterial method for the nitrogen isotopic analysis of nitrate in seawater and freshwater. Anal Chem 73(17):4145–4153CrossRefGoogle Scholar
  42. Tan KH (2010) Principles of soil chemistry. CRC, Boca Raton, FLGoogle Scholar
  43. Vidon P, Serchan S (2016) Impact of stream geomorphology on greenhouse gas concentration in a New York mountain stream. Water Air Soil Pollut 227(12):428CrossRefGoogle Scholar
  44. Vilain G, Garnier J, Tallec G, Tournebize J (2012) Indirect N2O emissions from shallow groundwater in an agricultural catchment (Seine Basin, France). Biogeochemistry 111(1–3):253–271CrossRefGoogle Scholar
  45. von der Heide C, Böttcher J, Deurer M, Duijnisveld WH, Weymann D, Well R (2009) Estimation of indirect nitrous oxide emissions from a shallow aquifer in northern Germany. J Environ Qual 38(6):2161–2171CrossRefGoogle Scholar
  46. Wang S, Yeager KM, Wan G, Liu CQ, Liu F, Lü Y (2015) Dynamics of CO2 in a karst catchment in the southwestern plateau, China. Environ Earth Sci 73(5):2415–2427CrossRefGoogle Scholar
  47. Wanninkhof R (1992) Relationship between wind speed and gas exchange over the ocean. J Geophys Res 97:7373–7382CrossRefGoogle Scholar
  48. Werner SF, Browne BA, Driscoll CT (2012) Three-dimensional spatial patterns of trace gas concentrations in baseflow-dominated agricultural streams: implications for surface–ground water interactions and biogeochemistry. Biogeochemistry 107(1–3):319–338CrossRefGoogle Scholar
  49. Worral F, Lancaster A (2005) The release of CO2 from river waters: the contribution of excess CO2 from groundwater. Biogeochemistry 76:299–317CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  • Anna Jurado
    • 1
    • 2
    Email author
  • Alberto V. Borges
    • 3
  • Estanislao Pujades
    • 1
    • 4
  • Pierre Briers
    • 1
  • Olha Nikolenko
    • 1
  • Alain Dassargues
    • 1
  • Serge Brouyère
    • 1
  1. 1.Urban & Environmental Engineering, Hydrogeology and Environmental GeologyAquapôle, University of LiègeLiègeBelgium
  2. 2.Institute for Groundwater ManagementTechnische Universität DresdenDresdenGermany
  3. 3.Chemical Oceanography UnitUniversity of LiègeLiègeBelgium
  4. 4.Department of Computational HydrosystemsUFZ - Helmholtz Centre for Environmental ResearchLeipzigGermany

Personalised recommendations