Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Urine effects on grass and legume nitrogen isotopic composition: Pronounced short-term dynamics of δ15N

  • Bettina Tonn ,

    Roles Conceptualization, Formal analysis, Investigation, Supervision, Visualization, Writing – original draft

    btonn@gwdg.de

    Affiliation Department of Crop Sciences, Grassland Science, University of Goettingen, Goettingen, Germany

  • Ina Porath,

    Roles Formal analysis, Investigation, Writing – review & editing

    Affiliation Department of Crop Sciences, Grassland Science, University of Goettingen, Goettingen, Germany

  • Fernando A. Lattanzi,

    Roles Conceptualization, Writing – original draft

    Affiliations Lehrstuhl für Grünlandlehre, Technische Universität München, Freising-Weihenstephan, Germany, INIA La Estanzuela, Instituto Nacional de Investigación Agropecuaria (INIA Uruguay), Colonia, Uruguay

  • Johannes Isselstein

    Roles Conceptualization, Supervision, Writing – review & editing

    Affiliations Department of Crop Sciences, Grassland Science, University of Goettingen, Goettingen, Germany, Centre for Biodiversity and Sustainable Land Use, Goettingen, Goettingen, Germany

Urine effects on grass and legume nitrogen isotopic composition: Pronounced short-term dynamics of δ15N

  • Bettina Tonn, 
  • Ina Porath, 
  • Fernando A. Lattanzi, 
  • Johannes Isselstein
PLOS
x

Abstract

Nitrogen stable isotope (15N) natural abundance is widely used to study nitrogen cycling. In grazed ecosystems, urine patches are hot-spots of nitrogen inputs, losses, and changes in δ15N. Understanding δ15N dynamics in urine-affected vegetation is therefore crucial for accurate inferences from 15N natural abundance in grasslands. We hypothesized that leaf δ15N following urine deposition varies with time and plant functional group. Specifically, we expected (i) short-term decreases in δ15N due to foliar absorption of 15N-depleted volatilized ammonia, (ii) followed by increases in δ15N due to uptake of 15N-enriched soil inorganic nitrogen, and (iii) that the magnitude of these changes is less in legumes than in grasses. The latter should be expected because ammonia absorption depends on leaf nitrogen concentration, which is higher in legumes than grasses, and because biological nitrogen fixation will modify the influence of urine-derived nitrogen on δ15N in legumes. We applied cattle urine to a mixture of Lolium perenne and Trifolium repens in a pot experiment. Nitrogen concentration and δ15N were determined for successive leaf cohorts and bulk biomass either 17 (early) or 32 (late) days after urine application. Early after urine application, leaves of L. perenne were 15N-depleted compared to control plants (δ15N 0.1 vs. 5.8‰, respectively), but leaves of T. repens were not (-1.1 vs. -1.1‰, respectively). Later, both species increased their δ15N, but T. repens (4.5‰) less so than L. perenne (5.9‰). Vegetation sampled within and outside urine patches in the field further supported these results. Our findings confirm that foliar ammonia uptake can substantially decrease grass foliar δ15N, and that in both grass and legume the direction of the δ15N response to urine changes over time. Temporal dynamics of plant δ15N at urine patches therefore need to be explicitly addressed when 15N natural abundance is used to study nitrogen cycling in grazed grasslands.

Introduction

The effect of grazing animals on nitrogen cycling is a topic of interest both in natural and agricultural ecosystems. One approach to studying this complex process involves using the natural abundance of the rare stable nitrogen isotope 15N as an integrator of the nitrogen cycle [1]. Advances in recent decades have improved our understanding of the link between the isotopic composition of soil, plant and animal nitrogen and the underlying ecological processes [1, 2, 3]. For instance, ecosystem enrichment in 15N has been interpreted as the consequence of higher nitrogen loss rates [4, 5], or as an indicator for higher rates of nitrogen cycling [6, 7], because pathways of nitrogen loss discriminate strongly against 15N, particularly gaseous losses through ammonia volatilization and denitrification.

In grazed grasslands, animal excreta play a central role in nitrogen turnover [8, 9] creating hot-spots with high rates of gaseous nitrogen losses and nitrate leaching [1013]. Thus, grazing has been associated with 15N enrichment of plants and soil [6, 7, 14]. Close positive relationships (r2 of 0.55–0.86) between 15N abundance in soils, plants or animal tissue and stocking rates have been reported in some managed grasslands [4, 15, 16], but not in others [1720]. This shows how the complexity of the nitrogen cycle in these systems can prevent straightforward, simple inferences between nitrogen cycling and 15N abundance.

A little understood aspect of animal excreta is their effect on the temporal δ15N dynamics of grasslands. The strong fractionation of ammonia volatilization (ε ~ 40 to 60‰, [1]) results in the emission of highly 15N-depleted gaseous ammonia and the retention of highly enriched NH4+ in the soil solution [1, 3, 13, 21]. The ability of plants to take up gaseous ammonia from the atmosphere by diffusion through stomata [22] has the potential to decrease foliar 15N abundance at urine patches. 15N depletion through assimilation of gaseous ammonia derived from animal excreta has indeed been reported in other systems [21, 23]. In managed temperate grasslands, up to ~ 25% of the grazing area is affected by urine in a given year [24]. Therefore, if considerable 15N depletion regularly occurs at urine patches in grazed grasslands, this may have important implications for the interpretation of 15N natural abundance patterns in these systems: hotspots of high nutrient losses might, at least temporarily, be associated with 15N depletion rather than enrichment. To gauge the relevance of this process for 15N natural abundance studies in grazed grasslands, it is necessary to know over which time period the urine-induced plant biomass depletion in 15N may persist.

Further, gaseous ammonia uptake may differ between grasses and legumes, as ammonia uptake is controlled by the ammonia concentration gradient between ambient and substomatal air. The latter is affected by nitrogen nutrition status [25, 26, 27], which is frequently higher in legumes than in grasses [28]. An initial ammonia-induced 15N-depletion would be followed by increases in 15N abundance caused by uptake of 15N-enriched soil nitrogen. Again, this may differ between grasses and legumes because legumes can access atmospheric nitrogen and are thus less dependent on soil-derived nitrogen [29].

The aim of this study was to understand the short-term dynamics of 15N abundance in grasses and legumes affected by urine deposition. Our hypotheses were (i) that plant 15N abundance would decrease shortly following urine application because the plants take up 15N-depleted gaseous ammonia. We expected (ii) that plant 15N abundance would subsequently increase, reflecting the uptake of enriched inorganic nitrogen from the soil solution. Finally, we expected (iii) that variation in 15N abundance would be lesser in the nitrogen-fixing legume Trifolium repens L. than in the grass Lolium perenne L. These hypotheses are conceptually summarized in Fig 1.

thumbnail
Fig 1. Conceptual graph of hypothesized urine deposition effects on nitrogen uptake and isotopic composition of grass and legume species.

Assumed contribution of different nitrogen sources to the total plant nitrogen uptake of (a) a grass, (b) a legume, following urine deposition (vertical dotted line) over time. Nmin soil: mineral nitrogen from soil nitrogen pool, Nmin urine: mineral nitrogen derived from urine, NH3(g): nitrogen taken up as gaseous ammonia, BNF: nitrogen from biological nitrogen fixation. The predicted changes of isotopic composition of currently assimilated nitrogen in both grass and legume are shown in (c). The time covered by the figure is approximately three weeks. For more details on the underlying assumptions, see S1 File.

https://doi.org/10.1371/journal.pone.0210623.g001

Materials and methods

Pot experiment

The experiment was conducted in a greenhouse without artificial lighting. The two experimental factors, urine application (urine applied / untreated control) and harvest date (early / late), were assessed in a completely randomized design with four replications (pots). The greenhouse had one open and three closed sides. Pots were placed on tables that could be moved outside the greenhouse when stable weather conditions allowed.

The species used were the grass Lolium perenne cv. ‘Signum’ and the legume Trifolium repens cv. ‘Rivendel’. Seven to ten days after sowing, from 17 to 20-Apr-2012, fifteen seedlings of L. perenne and ten seedlings of T. repens each were planted together into square pots measuring 18x18x18 cm. A nitrogen-poor growing substrate was produced by mixing potting soil (sandy loam, pH (CaCl2) 6.8, 0.45 g kg-1 extractable phosphorus (calcium acetate lactate), 1.13 g kg-1 extractable potassium (CaCl2), 97 g kg-1 organic matter, 62 g kg-1 carbon, 4.5 g kg-1 total nitrogen, δ15N of total nitrogen 5.1‰) with sand in the ratio of 1:6 (weight-based). To ensure uniform establishment of seedlings, the top 1.5 cm layer consisted entirely of potting soil. Before the experiment started, plants were cut to a height of 5 cm twice, 27 and 45 days after sowing.

Three weeks later, on 15-Jun-2012, cattle urine was applied to half of the pots. Before urine application, pots of the urine treatments were separated from those of the control treatment by a distance of at least 10 m, to avoid uptake of gaseous ammonia by plants in the control treatment. Per pot, 300 ml of cattle urine was then added in the urine treatment, while the control received an equal amount of distilled water. Both liquids were applied close to the soil surface using a wash bottle and were spread uniformly over the pot surface without wetting aboveground plant parts. After four days, pots of the two treatments were joined again.

The urine that was applied was collected from five Simmental cows of the university-owned stocks at the Research Farm Relliehausen (University of Goettingen), located at the Solling uplands (51° 47’ N, 9° 42’ E). Care and housing conditions of the cattle were in line with German regulations: animal welfare act (TierSchG) and Tierschutz-Nutztierhaltungsverordnung (TierSchNutztV). The University’s Animal Welfare and Ethics Committee approved the use of urine samples from cattle in this study (approval no. E1-18). The animals’ diet during the weeks preceding the sampling partly consisted of grazed pasture and partly of grass silage. To collect urine, the animals were temporarily kept in place using the headlock of the feed fence of their barn, with regular access to feed and water. When animals voluntarily started to urinate, urine was captured with a bucket. It was then immediately stored at cool temperatures and frozen before being applied. At application, a subsample was taken for analysis. An amount of 180 ml taken from the urine subsample was acidified by adding nitric acid, freeze dried and analyzed for its nitrogen concentration and isotopic composition (see below). δ15N of the urine applied was determined as 2.0‰, while nitrogen concentration was 5.6 g l-1. The nitrogen application rate in the urine treatment was therefore 1.68 g per pot, corresponding to 52 g m-2, which can be considered as typical for naturally occurring urine patches [9].

Two pots were set up for each treatment. One was used to harvest total aboveground biomass either 17 days (early harvest) or 32 days (late harvest) after the urine application (2-Jul-2012 and 17-Jul-2012, respectively). Plants were cut at soil surface level and sorted into L. perenne and T. repens. The second pot was used to monitor and harvest successive leaf cohorts, i.e. leaves of comparable developmental stage. For this, five tillers of L. perenne and five leaves of T. repens were marked using coloured wire on 12 to 13-Jun-2012. Tillers of L. perenne were chosen so that the lamina of the youngest growing leaf had a visible length of 2–4 cm. The next older leaf was permanently marked with a tip of correcting liquid. The leaves marked in this way made up the first leaf cohort of L. perenne that was sampled later. Leaves of T. repens were marked on stolons whose youngest leaf was in the developmental stage 0.6 on the zero-to-one scale according to Carlson [30]. The next older leaf of that stolon was marked with wire and was part of the first leaf cohort of T. repens. During the course of the experiment marked tillers and stolons were monitored in intervals of three to seven days. The visible lamina length of marked L. perenne leaves and subsequently developing leaves of the same tiller was measured. Developmental stage according to Carlson [30] of marked T. repens leaves and leaves subsequently developing on the same stolon were recorded. Leaves of newer cohorts were marked as older leaves became senescent. Using these measurements, the main period of growth for the leaf cohorts harvested during the experiment could be estimated (S1 Fig). At the early harvest, laminae of the first three leaf cohorts of L. perenne were sampled from all marked tillers. Additionally, entire leaves of the third leaf cohort of marked stolons of T. repens were collected. At the second harvest date, laminae of leaf cohorts three and four were collected from L. perenne as well as T. repens leaves from the fifth leaf cohort. All harvested samples were dried for 48 h at 60°C, and their dry mass was determined. Both L. perenne and T. repens remained in a vegetative state throughout the experiment.

A temperature sensor was suspended 40 cm above the soil surface of the experimental plots and temperature was logged in 15-min intervals. During the period of 48 h following urine application, average temperature was 21.3°C, with maximum temperatures of 26.6°C within the first 24 h after urine application, and 39.1°C within the following 24 h. Average temperature was 21.5°C between urine application and first harvest, and 21.3°C between first and second harvest.

Field sampling

Field samples were taken at the Grünschweige Grassland Research Station (48° 23′ N, 11° 50′ E, 435 m above sea level) near Freising, Germany, on 23-Aug-2010. The sampled paddock was part of a long-term grazing experiment and was kept at a target sward height of 5 cm through continuous cattle grazing. For details on experimental site and grazing management see [31]. The δ15N of topsoil total nitrogen was 4.5 ‰. For this, in spring several soil cores (5 cm internal diameter, 10 cm depth) were taken from the paddock. All aboveground plant biomass was clipped and the soil cores were pooled. Soil samples were sieved through a 2 mm sieve, separating all visible roots, and then air-dried. Dried soil was analysed for δ15N as described below. At the time of sampling, recent urine patches could be distinguished by a distinctly darker green colour and higher sward height. Six urine patches and additionally two control patches without visible influence of animal excrements were selected. In each of the eight patches, ten vegetative tillers of L. perenne were chosen randomly. The laminae of the youngest fully developed leaf and the most recent dead leaf of these tillers were collected and immediately dried at 60°C for 24 hours. T. repens had low cover in fresh urine patches, and few dead leaves could be found. When available, youngest dead leaf and first fully extended leaf of ten stolons were sampled in the same patches as L. perenne, and treated in the same way. L. perenne was in a vegetative state at the time of sampling, while T. repens was flowering.

Sample preparation and analysis

All plant samples were dried to constant weight at a temperature of 60°C in a forced-air oven. Soil samples were dried at room temperature over several days. All samples were ground using a ball mill. Samples from the pot experiment were analyzed for their nitrogen concentration and nitrogen isotope abundances using an isotope ratio mass spectrometer (Delta Plus, Finnigan MAT), linked with a Conflo III-Interface (Finnigan MAT, Bremen, Germany) to an elemental analyzer (Carlo Erba NA 1110, Milan, Italy), with Acetanilid as an internal standard. Analysis of the samples from the field experiment followed the same procedure, except that a Conflo II-Interface (Finnigan MAT, Bremen, Germany) was used and that the internal standard was flour. In both cases, long-term precision for δ15N of the internal lab standards was better than ± 0.2‰. Nitrogen isotope abundances are presented using δ notation [δ15N (‰) = (Rsample/Rstandard)– 1, where R is the isotope ratio 15N/14N, and the standard is atmospheric N2].

Statistical analysis

The urine application response of leaf nitrogen concentration and δ15N of the different leaf cohorts in the pot experiment was analyzed using linear mixed effects models. Fixed effects were leaf cohort, urine application, and their interaction. Experimental pot was treated as a random effect. The results from the two leaf sampling dates were analyzed separately. The factor leaf cohort had four levels at the first sampling date (first, second and third leaf cohort of L. perenne, third leaf cohort of T. repens) and three levels at the second harvest date (third and fourth leaf cohort of L. perenne, fifth leaf cohort of T. repens).

Two separate types of analysis were performed for the total aboveground biomass harvest of the pot experiment. Firstly, linear mixed effects models were performed with the single and interactive effects of urine application, harvest date and species as fixed effects and pot as a random effect in order to assess the response of biomass nitrogen concentration and δ15N. Secondly, values for total harvested biomass, species-weighted mean nitrogen concentration and δ15N as well as yield percentage of T. repens were calculated for each pot. The response of these parameters to urine application, harvest date and their interaction was tested using generalized least squares models.

Analyses were performed with the statistical software R 3.2.3 [32] and the package “nlme” [33]. In all cases, residuals were inspected for homoscedascity and normality. Where necessary, variance structure was adapted as indicated in footnotes to Tables 1, 2 and 3. Minimum adequate models were selected by sequentially removing fixed effects to reach a minimum value of the Second-order Akaike Information (AICc) as calculated by the package “MuMIn” [34]. Significance of the fixed effects remaining in the model was tested using sequential Wald tests. For significant predictors, post-hoc tests were performed as Tukey tests using the package “lsmeans” [35]. A significance level of P < 0.05 was chosen throughout.

thumbnail
Table 1. Effects of urine application and harvest date on yield, botanical and chemical composition of the Trifolium-repens-Lolium-perenne mixture.

https://doi.org/10.1371/journal.pone.0210623.t001

thumbnail
Table 2. Effects of urine application and leaf cohort on nitrogen concentration and nitrogen isotopic composition of Lolium perenne and Trifolium repens leaves.

https://doi.org/10.1371/journal.pone.0210623.t002

thumbnail
Table 3. Effects of urine application on nitrogen concentration and nitrogen isotopic composition of Lolium perenne and Trifolium repens aboveground bulk biomass.

https://doi.org/10.1371/journal.pone.0210623.t003

Results

Apparent nitrogen recovery and productivity response to urine application

An increased aboveground dry matter yield in response to urine application was only observed at the late harvest (Table 1). Aboveground nitrogen yield showed a non-significant numerical increase by 58% at the early harvest, and a significant increase by 87% at the late harvest in the urine compared to the control treatment (Table 1). Apparent recovery of urine nitrogen was calculated as the difference between nitrogen yield with and without urine application, divided by amount of nitrogen applied via urine. It equalled 0.11 g g-1 at the early and 0.26 g g-1 at the late harvest date. The biomass proportion of T. repens was decreased by urine application, yet increased from the early to the late harvest date (Table 1).

Nitrogen concentration and isotopic composition of leaf cohort harvests

At both harvest dates, the urine treatment response of leaf nitrogen concentration differed between leaf cohorts (Table 2). Urine application increased the nitrogen concentration in all leaf cohorts of L. perenne, but in neither of the two harvested leaf cohorts of T. repens (Fig 2). Nitrogen concentrations were lower in L. perenne than in T. repens leaves in the control treatment (Fig 2). In the urine treatment, however, L. perenne leaves generally had the same nitrogen concentrations as the simultaneously harvested leaves of T. repens (Fig 2).

thumbnail
Fig 2. Nitrogen concentration and nitrogen isotope composition of Lolium perenne and Trifolium repens leaves as affected by urine application.

Laminae of four successive leaf cohorts of L. perenne (Lp1, Lp2, Lp3, Lp4) and leaves of two leaf cohorts of T. repens (Tr3, Tr5) were harvested at two dates. Grey: urine, white: control; error bars: standard error of the mean. Different letters mark significant (p < 0.05) differences between mean values within each harvest date.

https://doi.org/10.1371/journal.pone.0210623.g002

Like nitrogen concentration, δ15N also consistently varied between leaves of L. perenne and T. repens in the control treatment. With an average value of 5.8 and 4.4‰, L. perenne leaves were more enriched in 15N than leaves of T. repens with values of -0.1 and -0.7‰ in the early and late harvests, respectively (Fig 2). Again, there was a significant interaction between leaf cohort and urine treatment (Table 2). The second and third leaf cohort of L. perenne, collected at the early harvest date, had a significantly lower δ15N than the equivalent leaves of the control (Fig 2). On average, δ15N of urine-treated L. perenne leaves harvested at that date was as low as 0.1‰. The isotope composition of T. repens leaves was not affected by urine application at the first harvest (Fig 2). A different response was observed in the leaves collected at the second harvest, two weeks later. Urine application lead to an increased δ15N in the fourth leaf cohort of L. perenne and in the sampled T. repens leaves, and no longer lead to a difference between urine and control treatment in the third leaf cohort of L. perenne (Fig 2).

Nitrogen concentration and isotopic composition of aboveground bulk biomass of L. perenne and T. repens

Results obtained from the bulk biomass harvests were qualitatively similar to those of the leaf cohort harvests (Fig 3, S2 Fig). In the control treatment, nitrogen concentrations of bulk biomass were always lower in L. perenne than in T. repens (Fig 3). Urine application raised nitrogen concentrations in L. perenne to the same level as those of T. repens, which were unaffected by urine (Fig 3). This resulted in an increased nitrogen concentration in total aboveground biomass in the urine treatment, with no effect of harvest date (Table 1).

thumbnail
Fig 3. Nitrogen concentration and nitrogen isotope composition of Lolium perenne and Trifolium repens bulk biomass as affected by urine application and harvest date.

Total aboveground biomass of the Lolium-perenne-Trifolium-repens mixture. Grey: urine, white: control; error bars: standard error of the mean.

https://doi.org/10.1371/journal.pone.0210623.g003

For δ15N, the interactions of urine treatment with species and with harvest date were both significant (Table 3). Averaged over harvest dates, δ15N means of the urine treatment × species combinations increased significantly in the order of Control T. repens < Urine T. repens = Urine L. perenne < Control L. perenne. Averaged over species, δ15N means of urine treatments increased significantly in the order of Urine Early < Control Late < Urine Late < Control Early. As a result of these two interactions, the urine treatment lead to a much lower δ15N of L. perenne at the first harvest, compared to the corresponding control treatment or to both treatments at the late harvest (Fig 3). Conversely, in the second harvest, T. repens in the urine treatment was enriched in 15N compared to the control treatment or to both treatments at the earlier harvest (Fig 3). These species-specific reactions combined with the response of the species’ relative contribution to aboveground biomass resulted in a significant interaction between urine treatment and harvest date on the level of total biomass δ15N (Table 1). The lowest mean δ15N of 1.1 and 2.4‰, respectively, were observed in the urine treatment at the early and in the control treatment at the late harvest (Table 1). These values were significantly lower than the δ15N of the control treatment at the early and the urine treatment at the late harvest (4.0 and 5.1‰ respectively, Table 1). Both nitrogen concentrations and isotopic composition varied more strongly between replications in the urine than in the control treatment (Fig 3).

Nitrogen concentration and isotopic composition of L. perenne and T. repens in the field

Nitrogen concentration was generally higher in live L. perenne leaves sampled at urine (4.78 ± 0.48%, mean ± standard deviation) compared to control patches (2.25 ± 0.27%, Fig 4). While δ15N of live L. perenne leaves ranged between 0.3 and 1.8‰ in control patches and the majority of urine patches, at two urine patches strongly negative values of -8.2 and -8.4‰ were found (Fig 4). Dead L. perenne leaves had lower nitrogen concentrations and tended to be slightly depleted compared to the corresponding live leaves (Fig 4). Again, at two urine patches δ15N was conspicuously less (-3.9 and -4.2‰) than the range observed at the remaining patches (-1.5 to 0.6‰, Fig 4). Leaves of T. repens showed a lower variability both in terms of nitrogen concentration and δ15N (Fig 4).

thumbnail
Fig 4. Nitrogen concentration and isotopic composition of Lolium perenne and Trifolium repens leaves sampled on a pasture.

Live and dead laminae of L. perenne or leaves of T. repens were sampled at six urine patches (five in case of T. repens) and two control patches not recently affected by urine. Values of live and dead leaves collected at the same patch are connected with a dotted line.

https://doi.org/10.1371/journal.pone.0210623.g004

Discussion

δ15N dynamics in a grass and a nitrogen-fixing legume following urine application

Urine deposition more than doubled nitrogen concentration in bulk biomass of L. perenne 17 days after application, and decreased its δ15N relative to the untreated control (Fig 2). Fifteen days later, δ15N of the bulk biomass increased, returning to control levels (Fig 2). This supports the idea that uptake of volatilized ammonia in urine patches decreases δ15N of shoot biomass shortly after urine application, but then nitrogen taken up from the soil becomes more enriched as time after urine deposition passes. This agrees with results from individual leaves: while the initial depletion was similar in all three leaf cohorts, subsequent enrichment was most marked in the youngest leaf cohort sampled, which started to emerge between one and two weeks after urine application (Fig 3, S1 Fig).

Ammonia volatilization strongly fractionates against 15N (ε of ~ 40 to 60‰, [1]). As in any Rayleigh process, the actual isotopic composition of volatilized ammonia may vary considerably, depending on the completeness of the reaction. Average δ15N values of -36 to -18‰ have been observed in ammonia lost from urine patches [13, 21]. Eriksen and Høgh-Jensen [36] attributed decreases in δ15N in shoots of L. perenne two weeks after the application of artificial urine to either exchange between urine-derived ammonium and depleted adsorbed soil ammonium or to fractionation during urea hydrolysis. Given the high proportion of sand in our substrate, exchange of adsorbed ammonium is unlikely to have played a quantitatively important role in our experiment. Further, as urea hydrolysis would have been relatively fast [37], effects of fractionation would have become negligible since they are equalized once the reaction is completed. Thus, uptake of depleted ammonia best explains the 15N depletion we observed in L. perenne and might also have contributed to the results of Eriksen and Høgh-Jensen [36].

As expected, the dynamics of δ15N in T. repens differed from those in L. perenne. Firstly, urine deposition had almost no effect on nitrogen concentration (Fig 2, Fig 3) of T. repens. Secondly, neither bulk aboveground biomass nor individual leaf cohorts of T. repens showed a δ15N response to urine deposition at the first harvest date (Figs 2 and 3). Thirty-two days after urine application, however, δ15N did increase markedly in T. repens (Figs 2 and 3).This indicates that uptake of soil nitrogen had overridden biological nitrogen fixation as a source of nitrogen. Such substantial reductions in the contribution of biological nitrogen fixation to nitrogen uptake of T. repens have previously been observed to last for two to five months after urine deposition [22, 23].

Although the interpretation of changes in δ15N of T. repens is complicated by the unknown proportion and isotopic composition of the biologically fixed nitrogen [2, 3], the results suggest that uptake of gaseous ammonia in T. repens was less than in L. perenne. In our experiment, as in most grasslands, T. repens had higher nitrogen concentrations than L. perenne (Fig 3). This would have minimized net stomatal ammonia uptake, because higher nitrogen nutrition status increases ammonia concentration in substomatal air [25, 26, 27]. Additionally, a smaller leaf area and therefore absorption surface, or also lesser stomatal conductivity of T. repens during the first days after urine application, may have further contributed to species differences in ammonia assimilation [22].

Relationship between field and greenhouse results, and implications for the use of 15N natural abundance in studies of grazed systems

Plant biomass of grazed grasslands is highly variable in nitrogen isotopic composition [36, 38]. A controlled-environment experiment allowed testing hypotheses about short-term variations in δ15N of grassland plants in response to urine deposition. Both the microclimate in the greenhouse and the substrate used in the pot study may have potentially affected ammonia emissions and assimilation. Temperatures were higher in the greenhouse than outside, which would have somewhat increased cumulative ammonia volatilization [39]. Absence of wind may have led to lesser air exchange within the plant canopy and therefore to increased ammonia concentrations, favouring assimilation but at the same time reducing cumulative ammonia volatilization [39]. To maximize the importance of urine-derived or biologically fixed nitrogen for the nutrition of our experimental plants, we used a substrate with low organic matter concentration. In soils with higher organic matter concentration, a smaller proportion of ammonia might be volatilized, and higher rates of mineralization might dampen the 15N enrichment subsequent to urine deposition [13]. While field conditions therefore might have decreased ammonia assimilation and consequently the size of the early urine addition effect on biomass δ15N compared to our results, they would not have changed the pattern of rapid depletion and subsequent enrichment that was observed in L. perenne (Figs 2 and 3) and in the mixed culture biomass (Table 1).

Our field data support that ammonia assimilation may affect foliar nitrogen isotopic composition following urine deposition. Live leaves of L. perenne sampled from urine patches had highly variable δ15N values, some leaves as low as -8.4‰ (compared to a mean of about 1‰ in non-affected patches; Fig 4). This is very similar to the maximum difference in δ15N observed between the most depleted leaves of the urine treatment (-3.1‰) and the mean value of the corresponding control leaves (6.0‰) in the pot experiment (Fig 2). Notably, the most depleted leaves sampled in the field were those with the highest nitrogen concentration. Based on the results of the pot experiment, this suggests more recent urine deposition: at the first harvest of the pot experiment, nitrogen concentration and δ15N of L. perenne leaves in the urine treatment were negatively correlated (Pearson’s r = -0.67, df = 6, p = 0.001), while the correlation tended to be positive (Pearson’s r = 0.71, df = 6, p = 0.051) fifteen days later, at the second harvest. T. repens leaves in the field showed neither clear depletion nor enrichment in plants growing in urine vs. non-urine patches (Fig 4). Experiments sampling urine patches of known age would be helpful to assess magnitude and temporal dynamics of the nitrogen isotopic composition changes that occur after urine deposition in the field in greater detail.

Our study shows that the time-lag following urine deposition as well as plant functional group affect nitrogen isotopic composition. Being aware of the temporal sensitivity of biomass δ15N to urine deposition can be important when making inferences about nitrogen cycling in grazed grasslands. For instance, urine deposition disrupted the usual relationship observed between legume biomass proportion and bulk biomass nitrogen concentration (positive) and isotopic composition (negative). This is because urine affects nitrogen concentration and isotopic composition of grasses much more than that of legumes, and because the magnitude of this effect varies with time (S3 Fig). Likewise, a grazing study [18] observed differences in δ15N of up to -8‰ between successive leaf cohorts of several grassland species in Inner Mongolia. Given that biomass was sampled in continuously stocked pastures, urine patches might have been responsible for part of that variability.

Consequently, in addition to the non-trivial problem of representatively sampling urine and non-urine patches [24, 40], time lag between probable urine deposition and sampling date has to be considered as well in grazing studies. The observed temporal dynamics of nitrogen isotopic composition after urine deposition might, however, also be utilized as temporal tracers of urine deposition events. Results from our pot experiment (Fig 2) indicate that nitrogen isotopic composition of consecutive leaf cohorts, linked to leaf appearance rates, might be used to retrospectively determine time lag to urine deposition in visually determined recent urine patches.

Conclusions

In line with our initial expectation, urine application led to strong temporal dynamics in biomass nitrogen isotope composition, with different responses in the investigated grass and legume species. Depending on the time lag between urine deposition and biomass sampling, plant δ15N at urine patches may show very different responses, with the short-term response diametrically opposed to assumptions generally made in nitrogen cycling studies of grazed grasslands. The generally expected 15N enrichment was only observed one month after urine application, most clearly in the legume. Two weeks earlier, the grass species was found to be strongly depleted in 15N, and isotopic composition of the legume to be unaffected by urine addition. We therefore conclude that not only spatial, but temporal heterogeneity in nitrogen cycling caused by urine patches needs to be explicitly addressed when using 15N natural abundance in grazing studies.

Supporting information

S1 Fig. Development of leaf length of successive leaf cohorts of Lolium perenne over time.

https://doi.org/10.1371/journal.pone.0210623.s001

(PDF)

S2 Fig. Relationship between leaf and bulk aboveground biomass values of nitrogen concentration and isotopic composition.

https://doi.org/10.1371/journal.pone.0210623.s002

(PDF)

S3 Fig. Relationship between the proportion nitrogen contained in Trifolium repens and mixture nitrogen concentration / isotope composition.

https://doi.org/10.1371/journal.pone.0210623.s003

(PDF)

S2 File. Original data pot experiment–leaf cohorts.

https://doi.org/10.1371/journal.pone.0210623.s005

(CSV)

S3 File. Original data pot experiment–bulk biomass.

https://doi.org/10.1371/journal.pone.0210623.s006

(CSV)

S4 File. Original data field observations.

https://doi.org/10.1371/journal.pone.0210623.s007

(CSV)

Acknowledgments

We thank Barbara Hohlmann, Camille Bienvenue and the technical staff of the Department of Crop Sciences at the University of Goettingen for their contribution to establishing and maintaining the experiment; Barbara Hohlmann specifically for the collection of urine samples; Dr. Rudi Schäufele for analysis of field samples, and the Centre for Stable Isotope Research and Analysis at the University of Goettingen for analysis samples from the pot experiment. Field data are derived from the ‘Summer School on Stable Isotopes in Environmental Sciences, Ecology and Physiology 2010’ at the Technical University of Munich; our thanks go to Laura Breitsameter and Oguzhan Angay, who participated in data acquisition and interpretation.

References

  1. 1. Robinson D. δ15N as an integrator of the nitrogen cycle. Trends Ecol Evol. 2001;16: 153–162. pmid:11179580
  2. 2. Craine JM, Brookshire ENJ, Cramer MD, Hasselquist NJ, Koba K, Marin-Spiotta E, et al. Ecological interpretations of nitrogen isotope ratios of terrestrial plants and soils. Plant Soil. 2015;396: 1–26.
  3. 3. Högberg P. Tansley Review No. 95. 15N natural abundance in soil-plant systems. New Phytol. 1997;137: 179–203.
  4. 4. Watzka M, Buchgraber K, Wanek W. Natural 15N abundance of plants and soils under different management practices in a montane grassland. Soil Biol Biochem. 2006;38: 1564–1576.
  5. 5. Kriszan M, Schellberg J, Amelung W, Gebbing T, Pötsch EM, Kühbauch W. Revealing N management intensity on grassland farms based on natural δ15N abundance. Agric Ecosyst Environ. 2014;184: 158–167.
  6. 6. Frank DA, Evans RD. Effects of native grazers on grassland N cycling in Yellowstone National Park. Ecol. 1997;78: 2238–2248.
  7. 7. Rui Y, Wang S, Xu Z, Wang Y, Chen C, Zhou X, et al. Warming and grazing affect soil labile carbon and nitrogen pools differently in an alpine meadow of the Qinghai-Tibet Plateau in China. J Soils Sediments. 2011;11: 903–914.
  8. 8. Haynes RJ, Williams PH. Nutrient cycling and soil fertility in the grazed pasture ecosystem. Adv Agron. 1993;49:119–199.
  9. 9. Whitehead C. Nutrient elements in grassland: Soil-plant-animal relationships. Wallingford, UK: CABI, 2000.
  10. 10. Decau ML, Simon JC, Jacquet A. Fate of urine nitrogen in three soils throughout a grazing season. J Environ Qual. 2003;32: 1405–1413. pmid:12931896
  11. 11. Wachendorf C, Lampe C, Taube F, Dittert K. Nitrous oxide emissions and dynamics of soil nitrogen under cow urine and dung patches on a sandy grassland soil. J Plant Nutr Soil Sci. 2008;171: 171–180.
  12. 12. Cameron KC, Di HJ, Moir JL. Dicyandiamide (DCD) effect on nitrous oxide emissions, nitrate leaching and pasture yield in Canterbury, New Zealand. NZ J Agric Res. 2014;57: 251–270.
  13. 13. Wells NS, Baisden WT, Clough TJ. Ammonia volatilisation is not the dominant factor in determining the soil nitrate isotopic composition of pasture systems. Agric Ecosyst Environ. 2015;199: 290–300.
  14. 14. Li C, Hao X, Willms WD, Zhao M, Han G. Effect of long-term cattle grazing on seasonal nitrogen and phosphorus concentrations in range forage species in the fescue grassland of southwestern Alberta. J Plant Nutr Soil Sci. 2010;173: 946–951.
  15. 15. Schwertl M, Auerswald K, Schäufele R, Schnyder H. Carbon and nitrogen stable isotope composition of cattle hair: ecological fingerprints of production systems? Agric Ecosyst Environ. 2005;109: 153–165.
  16. 16. Aranibar JN, Anderson IC, Epstein HE, Feral CJW, Swap RJ, Ramontsho J, et al. Nitrogen isotope composition of soils, C3 and C4 plants along land use gradients in southern Africa. J Arid Environ. 2008;72: 326–337.
  17. 17. Han G, Hao X, Zhao M, Wang M, Ellert BE, Willms W, et al. Effect of grazing intensity on carbon and nitrogen in soil and vegetation in a meadow steppe in Inner Mongolia. Agric Ecosyst Environ. 2008;125: 21–32.
  18. 18. Wittmer MHOM, Auerswald K, Schönbach Ph, Bai Y, Schnyder H. 15N fractionation between vegetation, soil, faeces and wool is not influenced by stocking rate. Plant Soil. 2010;340: 25–33.
  19. 19. Xu Y, He J, Cheng W, Xing X, Li L. Natural 15N abundance in soils and plants in relation to N cycling in a rangeland in Inner Mongolia. J Plant Ecol. 2010;3: 201–207.
  20. 20. Wrage N, Küchenmeister F, Isselstein J. Isotopic composition of soil, vegetation or cattle hair no suitable indicator of nitrogen balances in permanent pasture. Nutr Cycl Agroecosyst. 2011;90: 189–199.
  21. 21. Frank DA, Evans RD, Tracy BF. The role of ammonia volatilization in controlling the natural 15N abundance of a grazed grassland. Biogeochem. 2004;68: 169–178.
  22. 22. Sparks JP. Ecological ramifications of the direct foliar uptake of nitrogen. Oecologia. 2009;159: 1–13. pmid:18975011
  23. 23. Erskine PD, Bergstrom DM, Schmidt S, Stewart R, Tweedie CE, Shaw JD. Subantarctic Macquarie Island–a model ecosystem for studying animal-derived nitrogen sources using 15N natural abundance. Oecologia 1998;117: 187–193. pmid:28308485
  24. 24. Moir JL, Cameron KC, Di HJ, Fertsak U. The spatial coverage of dairy cattle urine patches in an intensively grazed pasture system. J Agric Sci. 2011; 179: 473–485.
  25. 25. Schjoerring JK, Husted S, Mattson M. Physiological parameters controlling plant-atmosphere ammonia exchange. Atmos Environ. 1998;32: 491–498.
  26. 26. Massad RS, Nemitz E, Sutton MA. Review and parameterisation of bi-directional ammonia exchange between vegetation and the atmosphere. Atmos Chem Phys. 2010;10: 10359–10389.
  27. 27. Farquhar GD, Firth PM, Wetselaar R, Weir B. On the gaseous exchange of ammonia between leaves and the environment: determination of the ammonia compensation point. Plant Physiol. 1980;66: 710–714. pmid:16661507
  28. 28. Parsons AJ, Orr RJ, Penning PD, Lockyer DR, Ryden JC. Uptake, cycling and fate of nitrogen in grass-clover swards continuously grazed by sheep. J Agric Sci. 1991;116: 47–61.
  29. 29. Ledgard SF, Saunders WMH. Effects of nitrogen fertiliser and urine on pasture performance and the influence of soil phosphorus and potassium status. NZ J Agric Res. 1982;25: 541–547.
  30. 30. Carlson GE. Growth of clover leaves − developmental morphology and parameters at ten stages. Crop Sci. 1966;6: 293–294.
  31. 31. Schnyder J, Schwertl M, Auerswald K, Schäufele R. Hair of grazing cattle provides an integrated measure of the effects of site conditions and interannual weather variability on δ13C of temperate humid grassland. Global Change Biol. 2006;12: 1315–1329.
  32. 32. R Core Team. R: A language and environment for statistical computing. R, Foundation for Statistical Computing, Vienna, Austria. 2015. Available from: https://www.R-project.org.
  33. 33. Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team. nlme: Linear and nonlinear mixed effects models. R package version 3.1–121. 2015. Available from: http://CRAN.R-project.org/package=nlme.
  34. 34. Barton K. MuMIn: Multi-Model Inference. R package version 1.15.6. 2016. Available from: https://CRAN.R-project.org/package=MuMIn.
  35. 35. Lenth RV, Hervé M. lsmeans: Least-squares means. R package version 2.18. 2015. Available from: https://CRAN.R-project.org/package=lsmeans.
  36. 36. Eriksen J, Høgh-Jensen H. Variations in the natural abundance of 15N in ryegrass/white clover shoot material as influenced by cattle grazing. Plant Soil. 1998;205: 67–76.
  37. 37. Thomas RJ, Logan KAB, Ironside AD, Bolton GR. Transformations and fate of sheep urine-N applied to an upland U.K. pasture at different times during the growing season. Plant Soil. 1988;107: 173–181.
  38. 38. Craine JM, Towne EG, Ocheltree TW, Nippert JB. Community traitscape of foliar nitrogen isotopes reveals N availability patterns in a tallgrass prairie. Plant Soil. 2012; 356:395–403
  39. 39. Whitehead DC, Raistrick N. Effects of some environmental factors on ammonia volatilization from simulated livestock urine applied to soil. Biol Fertil Soils. 1991;11: 279–284.
  40. 40. Dennis SJ, Moir JL, Cameron KC, Edwards GR, Di HJ. Measuring excreta patch distribution in grazed pasture through low-cost image analysis. Grass Forage Sci. 2013;68: 378–385.