Seasonal variations in nitrate isotope composition of three rivers draining into the North Sea

Introduction Conclusions References


Introduction
On a global scale, human activity has resulted in a doubling of the reactive nitrogen pool since pre-industrial times (Galloway, 2003) with more localised hot spots at the land-sea interface (Gruber and Galloway, 2008).In the North Sea area, the reactive nitrogen discharge by rivers has increased tenfold since the 1960's over assumed pristine conditions (Behrendt and Opitz, 1999;BMU, 2001).Today, the German Bight in the South Eastern North Sea receives approximately 350 kt of reactive N (mainly in the form of nitrate) per year from rivers draining densely populated and industrialised areas of NW Europe (Johannsen et al., 2008).The increased reactive nitrogen loads and ensuing eutrophication continue to be an environmental pressure on many coastal seas (Howarth and Marino, 2006).The long average water residence time (Weston et al., 2004) permits severe eutrophication to unfold in the German Bight (Hickel et al., 1993).Most nutrients deriving from Germany enter the Southern North Sea via the Rhine and the Elbe rivers and also from medium-sized rivers, such as the Weser River and Ems River (UBA, 2009).Nitrate inputs from point sources have decreased from their peak in the 1980's in response to environmental legislation (P ätsch and Radach, 1997), but the annual average nitrate loads of these rivers are slow in responding to pollution reduction measures, and thus are not solely determined by inputs from manageable sources.There is a continuing need to understand the nature of this apparent buffering.At present, total nitrogen loads in German rivers are thought to derive mainly from diffuse sources, dominated by a reactive nitrogen surplus stemming from agriculture (fertiliser and manure) (Hamm, 1996;Behrendt et al., 2002;G ömann et al., 2005).Previous studies of nitrate isotope properties (Johannsen et al., 2008;D ähnke et al., 2008) and microbial communities (Herfort et al., 2009)  been overlooked.Additionally, hydrological characteristics of the rivers highly influence transformation processes of riverborne nitrate, and are controlled by stream size and discharge regime.
The natural abundance of δ 15 N-NO − 3 and δ 18 O-NO − 3 has been used to identify sources and to trace transformation processes of riverborne nitrate (Deutsch et al., 2006;Mayer et al., 2002;Sebilo et al., 2006;Wankel et al., 2009;Schlarbaum et al., 2010).Different sources of stream water nitrate (direct fertiliser input, atmospheric deposition, sewage and wastewater inputs, or leaching of nitrate produced by nitrification of ammonia from agricultural soils; Behrendt et al., 2002) are specific in their dual isotopic composition.In precipitation, the δ 15 N-NO − 3 varies depending on location and season between −4‰ and +9‰ and is similar to the δ 15 N-NO − 3 of nitrate produced by nitrification (Kendall, 1998).Added attributes of these two sources are distinct ranges of δ 18 O-NO 3 in nitrate, with atmospheric NOx in precipitation having a δ 18 O value between 20‰ and 60‰ (Kendall, 1998), and nitrate originating from nitrification falling in a range between −10‰ and +16‰ (Kendall, 1998;Mayer et al., 2001;Burns and Kendall, 2002).δ 15 N-NO − 3 of nitrate from organic fertilizers and from municipal waste water is characterized by high δ 15 N values between 4‰ and 9‰, sometimes even higher (Grischek et al., 1998).
The isotopic mixture of nitrate sources is processed within rivers by different biological processes that in turn are associated with specific isotopic fractionation.Isotopic fractionation arises from biological processes, such as assimilation and denitrification, and each process is associated with characteristic fractionation factors ε (the isotopic enrichment/depletion of the product relative to the substrate in ‰) that determine the isotopic composition of residual nitrate and of any products.A useful indicator of nitrate assimilation in rivers is δ 15 N in suspended particulate organic matter (δ 15 N-PN).
Although particulate nitrogen originates from either external sources or from production within the river, analysis of paired δ 15 N-NO − 3 and δ 15 N-PN may shed light on the contribution of assimilation.Introduction

Conclusions References
Tables Figures

Back Close
Full This study expands on a previous investigation (Johannsen et al., 2008) of river nitrate in five rivers that contribute substantially to eutrophication in the German Bight of the SW North Sea.That study was performed over the course of only one year and did not include particulate matter and δ 18 O-H 2 O.By expanding our time series, we wished to answer the following questions: (1) Can we characterize major nitrate sources by means of specific isotopic fingerprints?(2) How does river discharge and water residence time influence nitrate concentration and its isotopic composition?(3) Do we find evidence for biological processes within the rivers that influence nitrate loads and nitrate isotopic variability? 2 Materials and methods

Rivers studied
This study was conducted at monitoring sites of three rivers (Rhine, Weser, Ems) that differ substantially in annual loads.The load of Rhine is approximately 220 kt N/year, whereas the rivers Weser and Ems discharge 40 kt and 6 kt N per year, respectively.All three rivers are highly impacted by human activities, the populations in their watersheds are dense, and agricultural land-use is comparatively high (70-90% of urban and arable land-use).More details are given in Johannsen et al. (2008).

Sampling
Water was sampled by staff of official LAWA (L änderarbeitsgemeinschaft Wasser) monitoring sites located upstream of the estuaries at salinities around 0 (PSU) (Fig. 1).
During the first sampling period (2006/2007), surface water samples were filtered through a PVDF filter (0.45 µm).After 2007, water samples were filtered through a pre-combusted (400 • C, 6 h) GFF filter (0.45 µm).These were used to determine the concentration and δ 15 N of particulate N. Filtrate was frozen in acid rinsed PE bottles Introduction

Conclusions References
Tables Figures

Back Close
Full for analyses of nitrate concentrations and isotopic composition.Aliquots of selected samples were carefully defrosted in February 2010 and sent to the laboratory of the Museum f ür Naturkunde in Berlin for the analysis of δ 18 O-H 2 O.

Nitrate concentrations
Nitrate concentrations were measured using a continuous flow analyzer (Bran and Luebbe, Auto Analyzer 3) according a modified method of Armstrong et al. (1967).The method has a detection limit of 6 µmol NO − 3 /L within the expected range of 50-450 µmol NO − 3 /L for anthropogenically influenced rivers.

Isotopic analyses of nitrate
Isotopic analyses of nitrogen and oxygen of NO − 3 were carried out using the denitrifier method (Casciotti et al., 2002;Sigman et al., 2001) which is based on the isotopic analysis of nitrous oxide (N 2 O) produced by denitrifying Pseudomonas aureofaciens (ATCC#13985) strains.Isotope ratios are expressed in the delta notation (δ 15 N relative to atmospheric N 2 and δ 18 O relative to Vienna Standard Mean Ocean Water, both defined as 0‰): where R is 15 N/ 14 N or 18 O/ 16 O ratio of sample and standard, respectively.Analyses of δ 15 N-NO − 3 were corrected using the international nitrate isotope standard IAEA-N3 (δ 15 N=4.7‰).Analyses of δ 18 O-NO − 3 were corrected for exchange, fractionation, and blank against IAEA-N3 with an assigned δ 18 O-NO − 3 of 22.7‰.We applied this value (which is now accepted to be 25.6‰; B öhlke et al., 2003), because we compare results with those of a previous study of river nitrate (Johannsen et al., 2008) that was based on a δ It should be noted that the denitrifier method measures the isotopic composition of nitrate+nitrite, but the contribution of nitrite to reactive nitrogen is negligible (<0.3%) in our sample set and has therefore been ignored.

Oxygen isotope measurements of water
Stable isotope ratios of oxygen (

Particulate nitrogen
The δ 15 N of particulate nitrogen on filters (δ 15 N-PN, expressed in‰ versus air N 2 ) was determined in duplicate on subsamples of individual GFF-filters with an EA interfaced via a Con-Flo III to an Isotope Mass Spectrometer (Finnigan Delta Plus XP), calibrated using IAEA-N-1 (δ 15 N=0.4‰) and IAEA-N-2 (δ 15 N=20.3‰).The reproducibility for δ 15 N-PN was 0.3‰, but the variability in δ 15 N-PN for sub-samples of the same filter was higher than 0.3‰ (see error bars in Fig. 2), and was caused by heterogeneities of the material on the filters.

Discussion
All three rivers under study are impacted by human activity, but the degree of agricultural land use and specializations differ in the catchment areas.The Ems drainage basin is dominated by intensive animal husbandry, because it is more competitive on the prevailing sandy soils, and the livestock production in the Ems catchment is high compared to the German average.Similarly, Weser drains grassland with intensive cattle farming and milk production (Osterkamp et al., 2001;Statistisches Bundesamt, 1990).The Rhine drains forested and urban areas with higher population density (G ömann et al., 2005;Wendland et al., 1993)  a relatively higher proportion of in-river nitrate consumption.We felt that longer term data sets would substantiate this assumption, and that additional information on the in-river consumption of nitrate should be gleaned from analysis of particulate N. In other studies, significant positive correlations have been observed between δ 15 N-PN and human population density catchment areas (e.g., Nishikawa et al., 2009), which is directly related to the percentage of residential and agricultural area.We observed the highest average δ 15 N-PN of 9.2‰ in the river Ems, which is most impacted in terms of agricultural and urban land use.Average δ 15 N-PN of 6.1‰ in Weser and 6.6‰ in Rhine River are also relatively high and in the same range reported for δ 15 N-PN in nonpristine rivers (Middelburg and Nieuwenhuize, 1998;Sigleo and Macko, 2002;Voss et al., 2006).δ 15 N-PN thus appears to be a good indicator for the degree of human impact on watersheds, and together with mean winter δ 15 N-NO − 3 characterise dominant anthropogenic sources of reactive nitrogen in rivers.
As stated above, agriculture prevails in all three watersheds, and agricultural soils are known to host large amounts of reactive nitrogen.Nitrate in rivers draining these soils may originate from different sources and processes, such as direct elution of atmospheric deposition, inorganic fertiliser and manure, or recycled nitrate from mineralization of soil organic nitrogen or nitrification of ammonia.All these sources are known to contribute to nitrate stocks in surface and groundwater, and each has more or less Introduction

Conclusions References
Tables Figures

Back Close
Full specific isotopic compositions (Kendall, 1998).Synthetically produced nitrate in fertilizer has a δ 18 O-NO − 3 between 18‰ and 22‰, which is close to the atmospheric δ 18 O value (Kroopnick and Craig, 1972;Amberger and Schmidt, 1987;Wassenaar, 1995).Typical mineral fertiliser applied in Germany is between 19.4‰ and 25.7‰ (Deutsch et al., 2005).Ranges of δ 15 N-NO − 3 are typically between −5‰ and 8‰ (Kendall, 1998;Macko and Ostrom, 1994); δ 15 N-NO − 3 between 0.8‰ and 4.4‰ were measured for mineral fertiliser in Germany (Deutsch et al., 2005).The δ 15 N value of fertilized soils is usually higher (between 0‰ and 14‰; Fogg et al., 1998;Kendall, 1998) than that of fertilizers applied (Kendall, 1998;Macko and Ostrom, 1994), because 15 N is enriched by an array of fractionation processes acting on the original isotopic mixture during N-assimilation, harvesting of crops, and denitrification in soils.The relatively low δ 18 O-NO − 3 in Rhine, Weser and Ems compared to synthetic fertilizers (between 18‰ and 22‰) suggest that direct leaching of mineral fertilizer from soils into rivers is only of minor relevance.This reflects increasing popularity of good (and cheaper) farming practise of targeted synthetic fertilisation immediately near the individual crop plants.
But it is known that agricultural soils in Germany are over-saturated with reactive N from decades of poor fertilising practise (UBA, 2009), and the legacy of excessive fertilisation has undergone massive fractionation when finally reaching stream waters, resulting in high δ 15 N of this contribution.In contrast to synthetic fertilizers, organic fertilizers are by-products of animal breeding and discharged directly by free range husbandry or are generously applied as organic fertilizer to field crops to remove waste (Renger, 2002).Nitrate originating from this practise (after nitrification of ammonia) has high δ 15 N values between 8‰ to 25‰ (Kendall, 1998;Macko and Ostrom, 1994) caused by fractionation during ammonia volatilization, which enriches the residual ammonium pool in 15 N (Frank et al., 2004).According to Blume (2004), up to 80% of the ammonia in organic fertilizers is volatilized during the first days after application, if not worked into the soil properly.Similar to the residual of mineral fertilizer, subsequent nitrification of the residual ammonia produces NO 3 that is then collected by rivers draining intensively farmed watersheds (Ahad et al., 2006;Ostrom et al., 1997;Horrigan et al., 1990).

Nitrate from nitrification
A key process in nitrate generation in soils and a significant source of riverine nitrate pools is nitrification of reduced N species (Hales, 2007).Nitrification is the bacterial oxidation of ammonia to nitrate via nitrite, mediated in two steps by aerobic chemolithoautotrophic bacteria and archaea.The first step, oxidation of ammonia into nitrite, is performed by two groups of organisms, ammonia-oxidizing bacteria (AOB) and ammoniaoxidizing archaea (AOA) dominating both marine and terrestrial environments (Francis et al., 2007).The second step, oxidation of nitrite into nitrate, is known to be performed by four genera of bacteria: Nitrococcus, Nitrospina, Nitrospira and Nitrobacter (Casciotti, 2009).Nitrobacter was generally considered the dominant or even the sole nitrifier in soils, but evidence has been lately given that Nitrospira is as well a common soil bacterium (Bartosch et al., 2002;Attard et al., 2010).Nitrate deriving from nitrification has a δ 15 N that is a legacy of the ammonia substrate (with reported fractionation factors ε AMO for the ammonium oxidation process that range from 14 to 38‰; Casciotti, 2009 and the references therein).The δ 18 O of this nitrate is apparently determined by the combined isotope ratios of ambient water and atmospheric (soil) molecular oxygen.Studies in fresh water environments suggest that two of the oxygen atoms originate from ambient water and another one from dissolved oxygen (Hollocher, 1984;Anderson and Hooper, 1983;Kumar et al., 1983), resulting in a δ 18 O-NO − 3 equal to: Atmospheric oxygen has a known δ 18 O value of 23.5‰ (Kroopnick and Craig, 1972).suggested that the contribution to δ 18 O-NO − 3 soil water and dissolved oxygen depends on ammonia concentration and the type of nitrifying organisms present in soils.Heterotrophic nitrifiers (no energy yield) apparently appear in soils with low ammonium concentrations and perform at characteristically low nitrification rates, possibly in a series of oxidation steps of organic-bound amino-groups.In this reaction sequence, two oxygen atoms may originate from the organic nitrogen compound, and only one from water (Mayer et al., 2001).The nitrate thus produced appears to be relatively higher in δ 18 O-NO − 3 than that produced from the second pathway, nitrification by chemolithoautotrophic consortia (Spoelstra et al., 2007) dominating environments with relatively high nitrification rates (Focht and Verstraete, 1977).As mentioned before, AOB and AOA are the prevailing chemolithoautotrophic ammonia oxidizers.Nitrification by AOA is operative over a wide range of environmental conditions (pH, temperature) in soils, where archaeotal cells far outweigh bacterial cells (Leininger et al., 2006), being potentially important actors within the nitrogen cycle in several ecosystems (Erguder et al., 2009).For the archaea, the isotopic systematics of nitrification is unknown, because they are as yet not cultivated for this experimental purpose.Considering the bacterial ammonia oxidizers, at least the genera Nitrobacter is assumed to produce nitrate with a δ 18 O-NO − 3 according to Eq. (1) (Mayer et al., 2001;Spoelstra et al., 2007).Under favourable conditions (e.g., availability of ammonia), nitrification in the water column of the river is likely to be an additional source of stream nitrate.During winter months, when ammonium, nitrate and nitrite concentrations were measurable, Herfort et al. (2009) detected genes of Crenarchaeota Group 1.1a, the major ammonia oxidizer in marine environments, in Rhine samples and reported that they accounted for 0.3% of the total microbial community in surface waters.Crenarchaeotal nitrification potential in the Rhine River in summer was lower, attributed to higher competition from bacteria and phytoplankton.To our knowledge, nothing is known about the oxygen isotopic composition of nitrate deriving exclusively from archeal nitrification, but it is safe to assume that it constitutes an internal nitrate source for the studied rivers during winter, albeit of unknown magnitude and isotopic fingerprint.In any event, based Introduction

Conclusions References
Tables Figures

Back Close
Full on the ammonia availability in agricultural soils and in river water, the contribution to nitrate in rivers from internal nitrification should be smaller than the contribution from soil nitrification, although D ähnke et al. ( 2008) reported significant (>50%) additions to estuarine nitrate pools of the Elbe river from nitrification in the estuarine turbidity maximum.

Atmospheric deposition
The isotopic signature of nitrate deriving from atmospheric deposition is distinct from any other nitrate source.The range of δ 18 O-NO − 3 in precipitation in Europe is between 55‰ and 75‰ (Durka et al., 1994), whereas the δ 15 N-NO − 3 of precipitation varies widely due to various sources of oxidized nitrogen and numerous fractionation processes preceding deposition (Heaton, 1986;Freyer, 1991).The δ 18 O-NO − 3 in river nitrate here are much lower than in atmospheric NO − 3 , so that precipitation cannot be a significant contribution to the nitrate pools of rivers Rhine, Weser and Ems.This agrees with a study by Burns and Kendall (2002), who investigated different nitrate sources in two American watersheds.

Influence of river discharge
Besides their discharge (Rhine has a seven times higher discharge yield than the Weser and a seventy higher discharge than the Ems; 2500 m 3 /s, 370 m 3 /s and 36 m 3 /s, respectively), the rivers differ in their respective discharge regimes.Whereas the discharge regimes of the rivers Weser and Ems is pluvial, the discharge regime of the Rhine River is glacial-nival upstream (the Alpine Rhine) and pluvial in the lower Rhine region.The Alpine Rhine discharge is controlled by the alpine snow-and glacier melt, which depends on temperature, having a minimum in winter and a maximum in summer.The influence of the pluvial regime increases downstream and the discharge becomes compensated by winter precipitation, but high discharges are maintained during summer (Hirschfeld, 2003;Kempe and Krahe, 2005).This mix of water sources buffers Introduction

Conclusions References
Tables Figures

Back Close
Full the Rhine hydrograph with regard to discharge minima, so that discharge varies by a factor of three whereas the discharge in Weser and Ems is six times higher during winter than during summer months (Fig. 3).The variability in δ 18 O-H 2 O in Rhine water is a consequence of these different hydrological regimes.As the δ 18 O-H 2 O of precipitation in central Europe decreases southwards (Buhl et al., 1991), the δ 18 O-H 2 O of surface waters in the Rhine catchment is spatially and temporally quite variable.Variations in δ 18 O-H 2 O values reflect the alpine influence in isotopically lighter δ 18 O-H 2 O compared to the precipitation in the lower Rhine area.According to Buhl et al. (1991) and Hirschfeld (2003), discharge of the Alpine Rhine has a δ 18 O-H 2 O between −12‰ an −13.5‰, whereas δ 18 O-H 2 O of lowland precipitation is around −8‰ (Yurtsever, 1975;Bowen and Revenaugh, 2005).We measured lowest δ 18 O-H 2 O (<−10‰) in Rhine at Bimmen-Lobith between April and August when snow melt takes place (Fig. 2a).
During winter season (October-March), δ 18 O-H 2 O values are higher (>−9‰) due to higher influence of the pluvial regime.Rivers Weser and Ems are solely rain-fed (de Jonge, 1995), and such seasonal variations in δ 18 O-H 2 O were not observed (Fig. 2b     and c).Nitrate concentrations in rivers depend substantially on the time course of precipitation and its influence on discharge amounts.In seasons of high precipitation and during flood events in winter, when biological activity is low, high amounts of rN are leached and transported into rivers (Hamm, 1996).This is reflected in significant positive correlations (R 2 >0.9) between NO − 3 loads and water discharge in all 3 rivers (Table 2).It also underscores that diffuse sources dominate the nitrate pools: If more or less constant point sources (waste water inlets, tributaries) were the dominant source, the NO − 3 loads would decrease at times of higher discharges due to dilution (Hamm, 1996).Assuming an ideal dilution, the nitrate concentration of a point source would behave according to: where c 0 is the constant background concentration of NO − 3 , A is the load of a point 6066 Introduction

Conclusions References
Tables Figures

Back Close
Full source and Q the discharge (Hebbel and Steuer, 2006).In Fig. 4 we display a theoretical dilution curve against mean nitrate loads of the river Rhine, Weser and Ems.The actual nitrate discharge clearly differs from the theoretical dilution curve and indicates massive diffuse nitrate inputs instead.Moreover, nitrate concentrations increase with increasing discharge: Low nitrate concentrations in summer months (Fig. 2a-c) increase towards winter accompanied by higher discharges fed by precipitation events by that season.The correlation between NO − 3 concentrations and riverine discharge is best in the river Ems, where precipitation in the catchment is immediately funnelled into the river through permeable sandy soils.The high infiltration rates of the sandy soils favour nitrogen leaching to the groundwater and their discharge into the river (Liersch and Volk, 2009).
But the nitrate concentrations and isotopic mixtures are not solely determined by diffuse groundwater seepage.Seasonally varying biological productivity and nitrate assimilation in the rivers is clearly a major process.Because assimilation is low in winter (as indicated by the indirect, but significant, negative correlation between NO − 3 concentration and water temperature; Table 2), the more intense soil source and low consumption rates result in higher nitrate loads in winter.

Biological processes in the rivers
Anthropogenic sources of rN external to the rivers determine the base level isotopic composition of NO − 3 in winter, but biological activity alters the initial δ 15 N-NO − 3 and δ 18 O-NO − 3 in the course of the year.Two processes that remove nitrate are associated with preferential removal of δ 14 NO − 3 : denitrification and assimilation of nitrate (Mariotti et al., 1981;Kendall, 1998).Denitrification, a microbial process which is obligate anaerobic supposedly begins below oxygen levels of around 2 mg/L (Painter, 1970) and is unlikely to occur to any significant extent in the well oxygenated water columns of Rhine, Weser or Ems.Denitrification in sediments, according to present knowledge, causes only small fractionation because the rate limiting step is the non-discriminating Introduction

Conclusions References
Tables Figures

Back Close
Full process of nitrate diffusion (Sebilo et al., 2003;Reinhardt et al., 2006;Lehmann et al., 2007).Although sedimentary denitrification is likely to occur in the river beds and in riparian mud flats, where organic matter substrate is available in high amounts, this nitrate sink would go unnoticed in the isotopic mixture of river nitrate.Assimilation of nitrate, on the other hand, is likely to occur when temperature and light conditions are 15 N and δ 18 O combined with decreasing NO − 3 concentrations in Rhine, Weser and Ems have similar trends with maxima in summer and minima in winter for all sampling periods (Fig. 2a-c), suggesting that NO − 3 is consumed during summer months when biological activity increases.Isotope effects 15 ε and 18 ε during nitrate assimilation are similar, because the slope of δ 18 O to δ 15 N in residual nitrate has been observed to follow a 1:1 line in culture studies (Granger et al., 2004).A plot of δ 18 O-NO − 3 against δ 15 N-NO − 3 (Fig. 5) for individual rivers in our study yielded slopes of 0.37, 0.53 and 0.54 for the rivers Rhine, Weser and Ems, respectively.The deviation means that the increases in δ 18 O and δ 15 N in nitrate are not solely due to assimilation, and that other processes and nitrate sources interfere with a single phytoplankton sink for nitrate.This is not surprising, because the enrichment slope of 1:1 during assimilation is typically observed in closed systems, where nitrate is consumed without external replenishment.The dominant external input from soil water nitrate seepage obviously represents an open system, where drainage-and ground water supply continuously and diffusely adds nitrate with a more or less invariant isotopic composition over the entire year.In this case, a less steep slope due to incidentally added nitrate is to be expected.Plotting δ 18 O-NO − 3 versus δ 15 N-NO − 3 and evaluation of differences in slopes Introduction

Conclusions References
Tables Figures

Back Close
Full 3 in the Rhine and no systematic differences between winter and summer suggest that assimilation is not causing strong fractionation.Rhine River has by far highest discharges of the three (Fig. 3) and we assume that the prevailing high current velocity and low water residence time impede phytoplankton growth (Peterson et al., 2001;Wollheim et al., 2001).Conversely, a decrease in current velocity enhances the removal rate of nitrate from the water column via assimilation and results in δ 15 N enrichment in the residual nitrate (Trudeau and Rasmussen, 2003).Because of their high surface to volume ratios, smaller streams are more efficient in processing nutrients relative to transport (Alexander et al., 2000).
In Ems, the range of NO − 3 concentrations and its isotopic composition is wider than in the rivers Weser and Rhine.Its initial NO − 3 concentration, assumed to be the maximum concentration measured in winter and normalized to 1, declines to 0.1 in the river Ems, whereas the initial NO − 3 concentration declines only to 0.5 in Rhine.Comparing the three rivers, we find a strong relationship between 1) stream size, 2) consumption efficiency and 3) fluctuation of δ 15 N and δ 18 O: With higher discharge amounts, the consumption efficiency is lower and summer increase in δ 15 N and δ 18 O of nitrate smaller.
Consumption of nitrate produces particulate nitrogen that is locked to the δ 15 N of nitrate by the fractionation factor ε associated with assimilation.In previous studies, the isotopic fractionation factor (ε) associated with nitrate assimilation was in the range of 4 to 9‰ for field observations (Altabet et al., 1991;Horrigan et al., 1990;Sigman et al., 1999;Wu et al., 1997) and is often assumed as 5‰.Under nitrate-replete conditions, the δ 15 N-PN produced at any time during assimilation should be δ 15 N-NO − 3 −5‰ (Mariotti et al., 1981).In theory, we should thus expect a negative correlation of δ water was between −9.4‰ and −7.3‰ which accounts for 2/3 to the δ 18 O-signal of NO − 3 from nitrification in soils.Besides external nitrate inputs, a potential internal nitrate source during winter when ammonia is abundant was vaguely attributed to water column nitrification.Furthermore, we revealed that assimilative processes are responsible for fractionation of δ  Full   Full in rivers draining into the North Sea suggested that sources and sinks of nitrate within rivers may have Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 2002).Analytical precision (±1σ) based on replicate measurements of samples and standard was generally 0.2‰ for δ 15 than the two other rivers; land use is dominated by grassland, cattle production and intensive cash cropping.In a previous study (Johannsen et al., 2008), we found a robust positive correlation (R 2 =0.7) between the percentage of urban and agricultural land use in catchment areas and δ 15 N-NO − 3 values for a range of several rivers.Furthermore, we observed a significant negative correlation between δ 15 N-NO − 3 values and the nitrate load (R 2 <−0.7) for each seasonal cycle.This suggested that mean annual δ 15 N-NO − 3 over river borne nitrate should generally increase with decreasing annual nitrate loads, as a result of Introduction Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | processes, such as partial denitrification, selective assimilation by and harvesting of crops, this leads to a progressive enrichment of δ 15 N-NO − 3 and δ 18 O-NO −

δ 18 O
-H 2 O ranged between −10.3‰ and −8.5‰ in Rhine, −8.5‰ and −7.6‰ in Weser Introduction Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | for Rhine compared to the patterns observed in the Weser and Ems.The complete lack of correlation between δ 18 and is observed in the river Ems (R 2 =−0.45; p<0.05), but is neither observed in the Rhine River (R 2 =−0.18), nor in the Weser (R 2 =0.098).Discussion Paper | Discussion Paper | Discussion Paper | summer months when biological activity takes place.In all rivers, a decrease in NO − 3 concentrations was inversely correlated with an increase in δ 15 N-NO − 3 values as consequence of an enrichment of heavy isotopes in the residual NO − 3 pool during assimilation.Variations of isotopic values and fractionation range were attributed to the hydrological characteristics of the respective rivers.The Rhine River is an exception in terms of discharge yield; the nitrate consumption efficiency is lower than in the rivers Weser and Ems due to the fact that a higher current velocity decreases the rate of biogeochemical transformation processes.Strong interdependencies between stream sizes, consumption efficiency and fractionation range were found comparing the three rivers of different sizes.Highest NO − 3 values and highest nitrate consumption rates associated with a observed in the smaller river Ems.We conclude from this that sources of rN and the nitrate consumption efficiency interrelated with the current velocity determines mostly the isotopic composition of riverine nitrateDiscussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Slough, California, J. Geophys.Res., 114, G01011, doi:10.1029/2008JG000729,2009.Wassenaar, L. I.: Evaluation of the origin and fate of nitrate in the Abbotsford Aquifer using the isotopes Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Table 2 .
Correlation matrix (Pearson's) of biogeochemical parameters of the rivers Rhine, Weser and Ems.