Nitrogen status of functionally different forage species explains resistance to severe drought and post-drought overcompensation

Agriculture, Ecosystems and Environment 236 (2017) 312–322

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment journal homepage: www.elsevier.com/locate/agee

Nitrogen status of functionally different forage species explains resistance to severe drought and post-drought overcompensation Daniel Hofera,b , Matthias Sutera , Nina Buchmannb , Andreas Lüschera,* a b

Agroscope, Institute for Sustainability Sciences ISS, Reckenholzstrasse 191, CH-8046 Zürich, Switzerland ETH Zürich, Institute of Agricultural Sciences, Universitätstrasse 2, CH-8092 Zürich, Switzerland

A R T I C L E I N F O

Article history: Received 9 June 2016 Received in revised form 24 November 2016 Accepted 30 November 2016 Available online xxx Keywords: Drought stress Intensively managed grassland Nitrogen nutrition index Post-drought recovery Root growth Symbiotic dinitrogen fixation

A B S T R A C T

Forage species of intensively managed temperate grassland differ substantially in their drought responses. We investigated whether differences in resistance and resilience, based on biomass yield, are related to species nitrogen (N) acquisition and drought-induced N deficiency. A three-factorial field experiment was established with monocultures of four species (first factor) that differed in functional traits regarding N acquisition and rooting depth: Lolium perenne L. (shallow-rooted non-legume), Cichorium intybus L. (deep-rooted non-legume), Trifolium repens L. (shallow-rooted legume), and Trifolium pratense L. (deep-rooted legume). A ten-week summer drought was simulated (second factor) and compared to a rainfed control during two regrowths under drought and one regrowth during a subsequent six-week post-drought period. The distribution of applied fertiliser N (200 kg ha
1 year
1 in total) was manipulated (third factor) with plots receiving no N during drought or 60 kg N ha
1. Soil water availability during drought became increasingly restricted over time. Plant-available soil N was reduced up to 4- and 12-fold during the first and second regrowths under drought, respectively, but was increased up to 4-fold during the post-drought regrowth, compared to rainfed control conditions. Legumes were consistently less N-limited than non-legumes (P < 0.001). Nitrogen derived from the atmosphere (Ndfa) in the legume T. repens was 72% under severe drought (first regrowth under drought). Here, legumes were rather drought-resistant (biomass yield under drought was
22% compared to the rainfed control), while non-legumes were not (
41%). Further, N fertilisation mitigated the negative drought effect on biomass yield of non-legumes from
41% (no N under drought) to 23% (N under drought). Under extreme drought (second regrowth under drought), all species were strongly impaired, irrespective of N fertilisation (
75% on average); yet, Ndfa in T. repens was still 56%. During the postdrought regrowth, former drought-stressed non-legumes overcompensated and revealed +53% higher yield than the control. The interspecific differences in plant species responses to drought suggest a shift from N limitation under severe drought to water limitation under extreme drought. Because legumes were able to compensate for drought-induced restrictions in yield through symbiotic N2 fixation, and non-legumes overcompensated during post-drought, cropping selected legumes in mixtures with non-legumes could improve resistance and resilience of forage swards against severe drought events. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Drought events are projected to increase in frequency and severity in many regions worldwide due to climate change (Seneviratne et al., 2012; Trenberth et al., 2014). Droughts are

* Corresponding author at: Reckenholzstrasse 191, 8046 Zürich, Switzerland. E-mail addresses: [email protected] (M. Suter), [email protected] (N. Buchmann), [email protected] (A. Lüscher). http://dx.doi.org/10.1016/j.agee.2016.11.022 0167-8809/© 2016 Elsevier B.V. All rights reserved.

expected to threaten widespread, grassland-based livestock farming by impairing forage production (Olesen et al., 2011), especially where biomass yield is currently high (Wang et al., 2007). Most studies to date have found that drought events reduce forage yields in extensively (e.g. Grant et al., 2014; Hoover et al., 2014), as well as intensively managed grassland (e.g. Zwicke et al., 2013; Hofer et al., 2016). However, the degree of drought impairment differs substantially depending on management type (Deléglise et al., 2015) and intensity (Gilgen and Buchmann, 2009), cutting frequency (Vogel et al., 2012), species richness (Isbell et al., 2015), and species identity (Hofer et al., 2016).

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

Drought induces not only pure water stress but can severely constrain the availability of soil nutrients, particularly soil nitrogen (N) by restricting mineral N fluxes (Cassman and Munns, 1980; Durand et al., 2010). Indeed, soil N uptake by forage species was found to be reduced due to drought (Dijkstra et al., 2015; Hoekstra et al., 2015). However, it remains open whether plant N uptake is reduced due to drought-induced N limitation in the soil or due to low N demand of the plants resulting from growth limitation in response to water scarcity. Because legumes do not depend solely on soil N (due to their benefit from symbiotic dinitrogen (N2) fixation; Hartwig, 1998), they may increasingly rely on N2 fixation with an increasing limitation of soil N availability under drought. Cropping forage legumes in grassland mixtures could thus promote drought resistance in biomass yield (resistance defined as the degree of impairment during a drought event; Pimm, 1984). The longer N2 fixation is maintained under persisting drought, the longer legumes could continue to produce biomass. Drought resistance in legumes may therefore depend on whether water or N is the primary limiting resource. Cropping deep-rooted species that reach deeper and moist soil layers have been suggested as a drought mitigation option in temperate grassland (e.g. Kemp and Culvenor, 1994; Skinner et al., 2004). Although the main uptake of soil water and nutrients in drought-stressed, intensively managed grassland occurs within the most superficial soil layer down to 30 cm (Hoekstra et al., 2014, 2015; Prechsl et al., 2015), forage species could increase resource uptake by short-term root growth in the course of the drought event, as indicated by increased root biomass (Dreesen et al., 2012) or higher proportion of root biomass at deeper soil layers under drought conditions (Wedderburn et al., 2010). While such evidence comes mainly from rhizotrons or container experiments, root growth data from forage species in the field is rare (but see Prechsl et al., 2015). Fertiliser application by farmers is a common strategy to counteract nutrient deficiency in managed grassland (Bélanger et al., 1992). Nitrogen fertilisation could therefore also help to overcome drought enhanced N limitation, although the growth response to N fertiliser might decline with decreasing soil moisture (Colman and Lazenby, 1975). In a multisite field experiment, we recently found drought-stressed forage species to be significantly impaired despite of N fertilisation. However, species were highly resilient after the drought event, and formerly drought-stressed non-legumes even overcompensated by producing more aboveground biomass than the non-stressed controls (Hofer et al., 2016). The underlying cause of such overcompensation remains unknown. Measuring plant-available mineral N in the soil during drought and post-drought periods could reveal to which degree soil and fertiliser N is accessible to plants and whether N resources not taken up during drought would become plant-available during the post-drought period given adequate water supply. Understanding the drought response of high-yielding and functionally different forage species can promote the development of farming options to adapt forage production to future climate conditions. To this aim, we simulated a ten-week summer drought event on monocultures of four key species of intensively managed temperate grassland, which differ in their functional traits regarding N acquisition and rooting depth. The species’ biomass response was examined under increasing drought severity during the course of a ten-week drought period and a subsequent sixweek post-drought period with ample water supply. We were primarily interested in the interacting effects of water scarcity and soil N availability on plant N status and biomass yield. We also investigated whether drought-induced N deficiency could be overcome by symbiotic N2 fixation of legumes or additional N fertilisation of non-legumes, and whether species’ root growth could increase resource acquisition. We are aware that drought

313

may also limit C assimilation through stomatal closure (Bollig and Feller, 2014); however, here, we focus on the feedback effect of water stress on N availability in the soil and its significance for the plants responses to drought. The following hypotheses were addressed: i Resistance of forage species’ biomass to drought depends on their functional traits to overcome various degrees of soil water and N limitation. ii Symbiotic N2 fixation in legumes promotes drought resistance by preventing plant N limitation. iii Nitrogen fertiliser application during drought increases drought resistance, especially in non-legumes, and/or contributes to a legacy effect to the post-drought regrowth, leading to increased resilience. iv Increased root growth especially at deeper soil layers improves drought resistance, particularly for deep-rooted species.

2. Material and methods 2.1. Site conditions and experimental setup The field experiment was carried out in the North-East of Switzerland at Zürich-Reckenholz (47 260 1200 N, 8 310 5100 E, 479 m a.s.l.). The soil at the site is classified as brown earth, with a top soil composition of 32% sand, 42% silt, 26% clay, containing 2.8% humus, 46 mg kg
1 phosphorus (P), 125 mg kg
1 potassium, 185 mg kg
1 magnesium, and with a pH of 6.9. Experimental plots were established in August 2011, and data presented here refer to 2013, the second year after sowing. In 2013, mean annual temperature was 9.4  C and annual precipitation was 1068 mm. A three-factorial experiment was carried out. Monocultures of four key forage species of intensively managed temperate grassland widely used in ruminant production were selected for investigation (first factor). Species differed in their N acquisition (non-N2-fixing for non-legumes and N2 fixing for legumes) and rooting depth: Lolium perenne L. (shallow-rooted non-legume, cultivar (cv.) Alligator), Cichorium intybus L. (deep-rooted nonlegume, cv. Puna II), Trifolium repens L. (shallow-rooted legume, cv. Hebe), and Trifolium pratense L. (deep-rooted legume, cv. Dafila). Species were sown into plots of 5 m  3 m and two further treatments were established: precipitation was manipulated, consisting of sheltered and rainfed control plots (second factor), and N fertiliser was varied, consisting of plots that were N fertilised during drought, and plots not fertilised during the drought period (third factor; see Section 2.2 for details on the treatments). All plots that received N during drought were replicated three times, while plots that received no N during drought were replicated twice. This resulted in a total of 40 plots that were arranged in an incomplete block design. 2.2. Drought and N fertilisation treatments An extraordinarily strong summer drought event with complete rain exclusion was simulated for ten weeks from June 5th to August 14th (see Table B.1., Appendix B in Supplementary file). Precipitation was excluded by placing rainout shelters (5.5 m  3 m) on the plots of the drought treatment, which were covered by a transparent, ultraviolet light-transmissible plastic foil (Gewächshausfolie UV5, 200 mm, Folitec Agrarfolien-Vertrieb, Germany) (see Hofer et al., 2016 for technical details of rainout shelters). The drought treatment excluded 184 mm of precipitation, which resulted in a simulated summer precipitation of 220 mm during June, July, August (see Table B.2. and Fig. B.1., Appendix B in Supplementary file, for further climatic data related to the

314

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

experiment). The resulting 54% of natural precipitation was below the 0.1th percentile of the probability distribution of summer precipitation of the previous 30 years. As usual for intensively managed grassland systems, plots were cut several times per year. The drought spanned over two entire regrowth periods (see Table B.1., Appendix B in Supplementary file), with the first regrowth lasting five weeks from the beginning of the drought, followed by a harvest and a second regrowth, which encompassed the final five weeks of the drought period. To test species’ resilience in biomass yield following a drought event (see Pimm, 1984 for a definition of resilience), a post-drought regrowth was also evaluated, lasting for six weeks. All plots were fertilised with a total amount of 200 kg N ha
1 year
1 in the form of ammonium-nitrate (Ammonsalpeter 27.5%, Lonza, Switzerland). Nitrogen was applied in several dressings in the spring and after each harvest, except the final harvest of the growing season. The N amount per application was varied within the season: plots were either not N fertilised during the drought period, or received 30 kg N ha
1 at each of the two regrowths. Given an equal total amount of N for all plots, this resulted in higher N application at the two dressings prior to the drought period in the no N treatment (see Table B.1., Appendix B in Supplementary file for details). Phosphorus, potassium and magnesium were applied once in the spring to all plots in nonlimiting amounts for intensively managed grassland according to local fertilisation recommendations (Flisch et al., 2009). 2.3. Measurements 2.3.1. Soil and plant water status Soil moisture content was measured weekly with permanently installed sensors (EC-5 soil moisture sensor, Decagon, USA) during the drought period at depths of 5 cm and 40 cm for the control and drought plots that were N fertilised (a total of 24 plots were recorded). Soil desorption curves (the relation between soil water content and soil matric potential) were determined for both soil depths from six representative plots using a standardised pressure plate method (Agroscope Reckenholz-Tänikon ART, 2012). This provided a metric for the physical soil environment to quantify water stress. A soil matric potential of
1.5 MPa was used as a reference, because it represents the approximate threshold of plant-accessible soil water. To evaluate water stress in the plants, predawn leaf water potential of L. perenne and T. repens was measured in N fertilised plots at the end of the two drought regrowths using a Scholander pressure chamber (Plant Moisture System SKPM 1400, Skye Instruments, UK). A total of 12 plots were evaluated, including control and drought conditions, and 8 leaf samples were measured per plot, whose water potentials were averaged for further analyses. 2.3.2. Aboveground biomass yield Aboveground biomass was harvested after each of six regrowths, while this paper presents data from the two regrowths of the drought period and the subsequent post-drought regrowth (see Table B.1., Appendix B in Supplementary file). Aboveground biomass was harvested using an experimental plot harvester (Hege 212, Wintersteiger, Austria) by cutting the central stripe (1.5 m wide) of each plot at 7 cm height and measuring the fresh weight. Dry matter content was determined by oven-drying a subsample of the fresh biomass at 100  C for 24 h. Drought induced change in aboveground biomass (CAB) per species for each N fertiliser treatment and regrowth was calculated as: CAB (%) = ((DMYdrought/DMYcontrol)
1)  100

(1)

with DMY being the dry matter yield of the drought and rainfed control condition. 2.3.3. Plant N acquisition Plant-available soil N was measured during the two regrowths of the drought period and the post-drought regrowth for all species and treatments. Plant Root Simulator (PRS)TM-probes (Western Innovations, Canada) were used to imitate root nitrate and ammonium sorption (Qian and Schoenau, 2005) and four anion and four cation probes per plot were installed at 5 cm soil depth for seven days towards the end of each regrowth (see Appendix A in Supplementary file for further information on the application of the PRSTM-probes). Plant N concentration was determined from subsamples of harvested aboveground biomass (see Appendix A in Supplementary file for details on the analysis of plant N concentration). The nitrogen nutrition index (NNI) was calculated following Lemaire and Gastal (1997): NNI = Nmeasured/Ncritical

(2)

where Nmeasured denotes the N concentration in the plant and Ncritical is the critical N concentration needed for N unlimited growth of C3 forage species (grass and non-grass species including legumes) depending on DMY, calculated as Ncritical = 4.8  DMY
0.32. Ncritical was kept constant at 4.8% for DMY < 1 t to consider the absence of competition between individual plants under low biomass yield (Poirier et al., 2012). Symbiotic N2 fixation in the legume T. repens was determined under rainfed control and drought conditions by isotope dilution (Chalk, 1985): double-labelled 15N-enriched ammonium-nitrate was applied on a permanently defined, central part of each T. repens plot, with the atom percent excess 15N of 15NH415NO3 being 30% (see Appendix A in Supplementary file for details on 15N application and analysis). Nitrogen isotopes 15N and 14N in aboveground biomass were analysed at the end of each regrowth and values of excess 15N relative to atmospheric N2 were compared to those of a non-fixing reference plant to calculate the percentage N derived from atmosphere (Ndfa) in T. repens following McAuliffe et al. (1958): Ndfað%Þ 15

¼

1


N atom% excess in legume

100

15

! N atom% excess in reference

ð3Þ

with the legume being T. repens and in the reference plant L. perenne. 2.3.4. Root growth and root biomass Root growth during the drought period was measured to examine the plants’ short-term drought response (Chen et al., 2016), potentially enhancing resource uptake and drought resistance. Cylindrical ingrowth bags with a diameter of 7 cm (polyamide, 1800 mm mash opening; SEFAR NITEX 06-1800/61, Sefar, Switzerland) were installed at the beginning of the drought period in all plots that were N fertilised during the drought period (3 ingrowth bags per plot, control and drought conditions). They were placed at a 45 angle to catch representative root growth throughout the soil from 0 to 30 cm deep. Bags were filled with fresh soil from the experimental site, with the soil compressed to approximately the same bulk density after passing through a 3 mm sieve to remove major roots. Ingrowth bags were excavated at the end of the drought period and divided in three sections, corresponding to soil depth ranges of 0–10 cm, 10–20 cm and 20–30 cm. Roots were cleaned from the soil under rinsing water

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

a) 5 cm soil depth 0.5

SMC (m3 m-3)

using a 0.63 mm sieve, dried at 100  C for 24 h, and weighed. Data from the three ingrowth bags per plot (split into the three soil layers) were pooled for further analyses to avoid pseudoreplication. Standing root biomass was measured to highlight changes in total root mass under drought. Soil cores were collected at the end of the drought period from 0 to 30 cm deep in the same plots used for root growth sampling. The soil cores also had a 7 cm diameter and were collected with the same alignment and number per plot as described for ingrowth bags. Collection and processing of the harvested material were conducted as with samples from ingrowth bags.

3. Results 3.1. Very low plant-available water and N during drought, but ample supply during post-drought During the first regrowth of the drought period, soil water limitation under drought became severe and the critical soil matric potential of
1.5 MPa, the approximate threshold of plantaccessible soil water, was reached at both soil depths (Fig. 1). Soil water scarcity affected plant xylem water pressure: predawn leaf water potential of the non-legume L. perenne was
0.27 MPa (1SE: 0.059) under control, but was
0.80 MPa (0.040) under

Control Drought

0.4 0.3 0.2 0.1 0.0

2.4. Data analysis

st

nd

1 regrowth 2 regrowth during drought during drought

Jun

Jul

Post-drought regrowth

Aug

Sep

b) 40 cm soil depth 0.5

SMC (m3 m-3)

Changes in species’ aboveground biomass due to drought and the N fertilisation treatment were analysed by linear mixed regression (Pinheiro and Bates, 2009). Aboveground biomass was regressed on the fixed variables species (factor of four levels), drought (factor of two levels: rainfed control, drought), fertiliser (not N fertilised during drought, N fertilised during drought), and regrowths (factor of three levels: first and second regrowth during drought, post-drought regrowth), including all interactions. To account for the repeated regrowths over time, plot was specified as a random factor (random intercept), thereby accounting for potential correlation of biomass data over time. Aboveground biomass was natural log transformed prior to analysis; therefore, the inference from this regression is equivalent to analysing directly the closely related log response ratio (ln(DMYdrought/ DMYcontrol); see Hedges et al., 1999; Hofer et al., 2016 for details). Inference among fixed factor levels was derived from the models’ contrasts, and where appropriate, the two non-legume and two legume species were pooled for testing (and likewise the two shallow- and two deep-rooted species). Evaluation of plantavailable soil N, the NNI, and Ndfa was conducted using the same model specification as described above. Plant-available soil N was natural log transformed prior to analysis to achieve homoscedasticity and normal distribution of residual variance. Generally, we focused the analysis on specific contrasts, but provide the summary tables of all regressions in Appendix B in Supplementary file. Root growth during the drought period and standing root biomass at the end of the drought period were analysed in a similar way as described. Here, the model included species (factor of four levels), the drought treatment (factor of two levels), and the different soil layers (factor of three levels), including all interactions. Plot was specified as a random factor to account for the consideration that roots from different soil layers originated from the same plot, and root data were natural log transformed. Differences in leaf water potential of L. perenne and T. repens between rainfed control and drought conditions were evaluated by t-tests. All analyses were performed with the statistical software R (R Core Team, 2016).

315

0.4 0.3 0.2 0.1 0.0

st

nd

1 regrowth 2 regrowth during drought during drought

Jun

Jul

Post-drought regrowth

Aug

Sep

Month Fig. 1. Soil moisture content (SMC) at 5 cm (a) and at 40 cm soil depth (b) under rainfed control and drought conditions, measured in plots of all four species during the first and second regrowth of the drought period (grey and dark grey rectangles, respectively, at the bottom of the panels) and the subsequent post-drought regrowth (light grey rectangle). Displayed are means  1 SE per drought treatment (n = 12). Non-visible SEs are due to small values. The dashed horizontal line (



) is the SMC corresponding to a soil matric potential of
1.5 MPa, which is the approximate threshold of plant-accessible soil water.

drought conditions (t = 4.08, P = 0.004 for the difference). Likewise, leaf water potential of the legume T. repens was
0.40 MPa (0.089) and
1.19 MPa (0.159) under control and drought conditions, respectively (t = 6.13, P < 0.001). During the second regrowth, drought stress became extreme and soil moisture was persistently below
1.5 MPa at both soil depths (Fig. 1), meaning that soil water uptake by the plants was no longer possible. Leaf water potential of L. perenne was
0.44 MPa (0.055) and
1.31 MPa (0.146) under control and drought conditions, respectively (t = 6.72, P < 0.001 for the difference), while the respective values for T. repens were
0.39 MPa (0.048) and
0.84 MPa (0.050) (t = 3.48, P = 0.008). Under rainfed control conditions, plant-available soil N was low for the two non-legume species L. perenne and C. intybus, while it was comparably high for the two legume species T. repens and T. pratense in all three regrowths (Fig. 2a, c, and e). This indicates strong soil N uptake by the non-legumes but incomplete uptake by the legumes. In contrast, during both regrowths under drought, plant-available soil N was generally low for all four species. Specifically for legumes, plant-available soil N was 4.4 (t = 4.90, P < 0.001) and 11.6 times lower (t = 7.82, P < 0.001, legumes pooled) under drought as compared to rainfed control conditions (first and second regrowths, respectively). However, during the postdrought regrowth, much more soil N (approx. 3.9-fold, t = 4.54, P < 0.001) was plant-available for former drought-stressed legumes than for the non-stressed controls, indicating high availability of soil N resources not being taken up during the drought period (Fig. 2e). At both regrowths under drought conditions, N fertilisation did not enhance plant-available soil N for any of the four species (Fig. 2a–d). In contrast, during the post-drought regrowth, N

316

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

Not N fertilised during drought

N fertilised during drought

1st regrowth during drought a) 15

b) 15

Control Drought

10

10

5

*** ns

***

*

ns

0

***

ns

0

2nd regrowth during drought c)

d)

15

15

*** 10

***

10

*** ***

5

5

**

**

ns

0

ns

0

Post-drought regrowth e)

f)

50

50

40

40

30

30

*** **

20 10

ns

L. p SR ere -N nn L e C . D inty R b -N u L s T. SR rep -L en T. E s p D rate R n -L se E

0

***

20 10

ns

***

†

ns

0

L. p SR ere -N nn L e C .i D nty R b -N u L s T. SR rep -L en T. E s p D rate R n -L se E

Plant-available soil N (µg cm-2 7 days-1)

***

5

Fig. 2. Plant-available soil N of the four species for the first and second regrowth during the drought period and the post-drought regrowth under rainfed control and drought conditions. Plant-available soil N was measured with Plant Root Simulator (PRS)TM-probes at 5 cm soil depth towards the end of each regrowth. Displayed are means  1 SE (n = 2 if not N fertilised during drought, n = 3 if N fertilised during drought). Non-visible SEs are due to small values, and inference is based on natural log transformed data. ***P  0.001, **P  0.01, *P  0.05, yP  0.1, ns: not significant. SR: shallow-rooted, DR: deep-rooted, NL: non-legume, LE: legume. Note that the y-axis is different in panels e) and f).

fertilisation increased soil N availability for former droughtstressed legumes by 107% (t = 2.43, P = 0.021, legumes pooled, Fig. 2e and f), while soil N availability for former drought-stressed non-legumes remained at low levels (t = 1.80, P = 0.082, nonlegumes pooled).

3.2. Functionally different species react differently to fluctuating water and N availability By the end of the first regrowth during the drought period, in which drought stress became severe, change in aboveground

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

biomass (CAB) was on average
22% for the two legumes, but was
41%, i.e. twice as large, for the two non-legumes (Fig. 3a). For the least impaired species, T. pratense, CAB was only
16%, while for the most impaired species, L. perenne, CAB was
55%, i.e. three times as large (t = 1.85, P = 0.076 for the difference between these two species). Thus, L. perenne was the only species that was significantly

317

impaired by drought (t = 3.28, P = 0.003; Fig. 3a). During the second regrowth of the drought period, when drought stress became extreme, the two deep-rooted species were less affected (-61% CAB, Fig. 3c) than the two shallow-rooted species (
81% CAB) (t = 2.26, P = 0.031 for the difference between pooled shallow- and deeprooted species), but there was no difference in CAB between non-

Not N fertilised during drought

N fertilised during drought

1st regrowth during drought a)

b)

5

Control Drought

ns

4

3

ns

**

ns

ns

ns

**

2

1

1

0

0

2nd regrowth during drought c)

d)

5

5

4

**

3

3

2

***

***

2

***

1 0

***

4

1

***

*** ***

0

Post-drought regrowth e)

f)

5

5

4

4

3

3

2 1

*

ns

2

*

L. p SR ere -N nn L e C .i D nty R b -N u L s T. SR rep -L en s T. E p D rate R n -L s E e

0

ns

1

†

†

* ***

0

L. p SR ere -N nn L e C . D inty R b -N u L s T. SR rep -L en s T. E pr D ate R n -L s E e

Aboveground biomass yield (t DM ha-1)

ns

4

3 2

5

Fig. 3. Aboveground biomass yield of the four species for the first and second regrowth during the drought period and the post-drought regrowth under rainfed control and drought conditions. Displayed are means  1 SE (n = 2 for species not N fertilised during drought, n = 3 for species N fertilised during drought). Non-visible SEs are due to small values, and inference is based on natural log transformed data. ***P  0.001, **P  0.01, *P  0.05, yP  0.1,ns: not significant. SR: shallow-rooted, DR: deep-rooted, NL: nonlegume, LE: legume.

318

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

legumes and legumes. During the post-drought regrowth, former drought-stressed non-legumes outperformed the rainfed control (+53% CAB, Fig. 3e), whereas former drought-stressed legumes did not reach the level of the controls (
32% CAB). Nitrogen fertilisation under rainfed control conditions led to 1.5-fold (t = 1.17, P = 0.250) and 1.8-fold increases (t = 1.90, P = 0.066) in the non-legume biomass yields during the first and second regrowths, respectively (Fig. 3a-d, non-legumes pooled). Yet, no such response was observed for the two legumes (t = 0.27, P = 0.792). Moreover, during the first regrowth, N fertilisation reduced the drought impairment in the non-legumes to the half of the no N treatment (-23% CAB with N fertilisation vs.
41% CAB without N, Fig. 3a and b) (t = 2.18, P = 0.046, for the difference). Thus, CAB of N fertilised non-legumes under drought (-23%) was very similar as CAB of the non-N fertilised legumes under drought (
22%). During the second regrowth, N fertilisation did not enhance biomass yield of drought-stressed species, neither for non-legumes (t = 0.52, P = 0.605), nor for legumes (t = 0.29, P = 0.773).

intybus under all drought and N fertilisation combinations (t = 25.21, P < 0.001, for the difference across all regrowths). Drought generally enhanced plant N limitation, indicated by lower NNIs (Fig. 4), particularly during the second regrowth (extreme drought) when all four species had significantly lower NNIs under drought than under rainfed control conditions (t = 8.98, P < 0.001, across all species, Fig. 4a and c). During the post-drought regrowth, all former drought-stressed species reached or outperformed the NNI of species under rainfed control conditions. Nitrogen fertilisation during the drought period never significantly increased the NNI of drought-stressed species (and thus did not mitigate plant N limitation): not during the first regrowth (t = 1.53, P = 0.136, across all species), nor the second regrowth (t = 0.54, P = 0.594, across all species). In contrast, during the postdrought regrowth, N fertilisation resulted in an increased NNI in former drought-stressed species (t = 3.29, P = 0.003, across all species). The N:P ratio of all four species was below 15 under rainfed control and drought conditions, in both fertilisation treatments, and for all regrowths, indicating no P limitation (see Fig. B.2., Appendix B in Supplementary file).

3.3. Plant N limitation during drought is not affected by N fertilisation 3.4. Symbiotic N2 fixation in T. repens is maintained under drought Plant N limitation was consistently lower in the legumes T. repens and T. pratense than in the non-legumes L. perenne and C.

a)

b)

Nitrogen derived from the atmosphere (Ndfa) in the legume T. repens was not affected by drought during the first regrowth when drought stress was severe (t = 0.50, P = 0.637) and was around 70% under rainfed control and drought conditions (Fig. 5a). Importantly, Ndfa in T. repens was still 56% during the second regrowth under extreme drought conditions, which corresponded to the higher NNI of legumes as compared to non-legumes (Fig. 4). During the post-drought regrowth, Ndfa was considerably reduced in former drought-stressed T. repens compared to the non-stressed control (t = 3.96, P = 0.008), which is likely due to high plant-available soil N (Fig. 2e). N fertilisation reduced Ndfa under rainfed control and drought conditions for all three regrowths during drought and post-drought (Fig. 5b; t = 3.61, P = 0.009, across all regrowths). 3.5. Drought stimulated root growth of non-legumes at deeper soil layers and enhanced total root biomass

c)

d)

Root growth was generally greater under drought than under control conditions at the deeper soil layers (10–20 cm and 20– 30 cm), in particular for the two non-legume species at the deepest layer (t = 3.05, P = 0.007, non-legumes pooled; Fig. 6a and b). Across all species, root growth increased under drought by 44% at the deepest layer, while it was reduced by 12% in the top soil layer (0–

a) Not N fertilised during drought

Ndfa (%)

100

Fig. 4. The nitrogen nutrition index (NNI) in aboveground biomass of the two shallow-rooted (a, b) and the two deep-rooted species (c, d) under rainfed control (Ctr) and drought conditions (Drt) at two N fertilisation treatments. The drought period is indicated by grey and dark grey rectangles at the bottom of each panel (first and second regrowth, respectively), the post-drought regrowth by the light grey rectangle. Values below 1.0 indicate plant N limitation. NNI = Nmeasured/Ncritical with Ncritical = 4.8  DM
0.32 (Lemaire and Gastal, 1997). Displayed are means  1 SE (n = 2 if not N fertilised during drought, n = 3 if N fertilised during drought). Nonvisible SEs are due to small values. SR: shallow-rooted, DR: deep-rooted, NL: nonlegume, LE: legume.

Control Drought

b) N fertilised during drought

100

75

75

50

50

25

25

0

0 Jun

Jul

Aug

Month

Sep

Jun

Jul

Aug

Sep

Month

Fig. 5. Nitrogen derived from the atmosphere (Ndfa) in aboveground biomass of T. repens under rainfed control and drought conditions not N fertilised (a) and N fertilised during drought (b). The drought period is indicated by grey and dark grey rectangles at the bottom of the panels (first and second regrowth, respectively), the post-drought regrowth by the light grey rectangle. Displayed are means  1 SE (n = 2 if not N fertilised during drought, n = 3 if N fertilised during drought).

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

L. perenne SR-NL

C. intybus DR-NL

319

T. pratense DR-LE

T. repens SR-LE

Root growth (g DM dm-3, log scale)

0.3

0.5

1.0

1.0

†

(c)

0.2

0.3

ns ns

0.5

1.0 (d)

ns ns

20-30

†

1.0

ns

20-30

20-30

20-30

*

0.5

10-20

ns

0.3

0-10

(b)

0.2

10-20

10-20

ns

0.5

0-10

†

0.3

10-20

(a)

0.2 0-10

Control Drought

0-10

Soil depth (cm)

0.2

ns

Total root mass (g DM dm-3, log scale)

0.3

1.0

3.0 7.0

3.0 7.0

†

1.0

3.0 7.0 (g)

ns

* ns

0.1

20-30

20-30

*

0.3

0.3

1.0

3.0 7.0 (h)

ns

10-20

*

0.1

0-10

(f)

10-20

20-30

20-30

*

1.0

10-20

10-20

*

0.3

0-10

ns

0.1 0-10

(e)

0-10

Soil depth (cm)

0.1

†

*

Fig. 6. Root growth (into ingrowth bags) during the drought period (a–d) and total root biomass at the end of the drought period (e–h) of the four species under rainfed control and drought conditions. Samples were taken at different soil depths and all species were N fertilised during drought. Displayed are means  1 SE (n = 3); non-visible SEs are due to small values, and inference is based on natural log transformed data. **P  0.01, *P  0.05, yP  0.1, ns: not significant. SR: shallow-rooted, DR: deep-rooted, NL: non-legume, LE: legume.

10 cm). Total root biomass was greater under drought than under control conditions for all species and soil layers (Fig. 6e–h; t = 5.40, P < 0.001, across species and soil layers). Specifically, total root biomass increased under drought by 79% at the deepest soil layer, while it increased by only 47% at the top-most layer (average across species). 4. Discussion We assessed the role of N acquisition during and after a single drought event in intensively managed temperate grassland. The results highlight: (i) the high importance of soil N availability as a key driver of drought effects on plants, which could explain differences in drought resistance and resilience among forage species, (ii) the relevance of symbiotic N2 fixation of legumes and of additional N fertilisation for non-legumes to increase resistance to severe drought, and (iii) the limited value of the ‘deep rooting’ trait and root growth to improve drought resistance. 4.1. Limitation in plant-available soil N affected plant response to severe drought while limitation of soil water was determinant under extreme drought Plant-available soil N, as measured in this study, is the balance between the N sources from fertiliser and mineralisation of soil organic matter, and the N uptake by the plants. For legumes and non-legumes this balance clearly differed under the ample water supply of rainfed conditions (Fig. 2a, c, and e). It can be assumed that the size of soil N sources was comparable in legume and nonlegume plots and, thus, that the distinct differences in N

availability derived from differences in the N uptake, which must have been large and complete in non-legumes but incomplete in legumes. As a consequence, plant-available soil N under legumes has the potential to indicate the size of the soil N source. Under drought, this source was strongly decreased and was below the level of the rainfed control, but was above the rainfed conditions during post-drought. Low soil N sources under drought are confirmed by the small aboveground biomass yields of the nonlegumes (Fig. 3a and c). These findings suggest that soil N sources are highly sensitive to variations in soil water conditions (Burke et al., 1997). Reasons for restricted plant-available soil N under drought can be (i) insufficient dissolution of the applied solid fertilisers, (ii) reduced litter decomposition (Sanaullah et al., 2012; Walter et al., 2013), (iii) restricted N mineralisation (Cassman and Munns, 1980; Borken and Matzner, 2009), and (iv) inhibited physical transport of N in the soil solution towards the roots (Durand et al., 2010). In the course of the drought period, water became increasingly scarce. The water stress ultimately reached extreme levels, with soil moisture being persistently below the critical threshold of
1.5 MPa at both soil depths (Fig. 1), and the increasing limitation in soil water had different implications for the drought resistance of our forage species. 4.2. Symbiotic N2 fixation improved resistance to severe drought The drought resistance of our forage species depended on their functional traits, with N2 fixing legumes having approximately half the yield losses than non-legumes under severe drought (Fig. 3a). This result corresponds to previous findings on drought responses of the same species growing at three experimental sites under

320

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

different drought severities, where legumes had a significantly lower drought-induced yield loss (
8% and
24%) than the nonlegumes (
51% and
68%) at two of the sites under severe drought (Hofer et al., 2016). We therefore compiled data from this study and Hofer et al. (2016) to reveal drought effects on aboveground biomass yields as affected by increasing drought stress for a total of four site-years. This analysis confirmed that legumes’ yield resists drought stress under severe drought distinctly better than nonlegumes (Fig. 7, around 30 days with soil matric potential <
1.5 MPa). The combined data suggest a linear decrease in drought resistance in non-legumes with increasing water scarcity, whereas legumes are able to maintain biomass yield under severe drought at the cost of collapsing completely when drought stress becomes extreme. Fig. 7 also shows that the drought response of these forage species might not differ markedly in the initial phase of a drought event when the drought stress and the species’ drought response is still small. We argue that the legumes’ superior drought resistance at severe drought is a result of symbiotic N2 fixation, which for T. repens was clearly not influenced by severe drought (Fig. 5). This is – to our knowledge – the first time that symbiotic N2 fixation in a forage legume is shown to sustain at field conditions under projected drought stress, and the result can well explain the consistently higher NNIs in legumes than in non-legumes (Fig. 4). The decrease in Ndfa from 72% under severe drought to 56% under extreme drought is most probably due to severe limitations of other growth resources than N, such as soil water and P (compare increased N:P ratios under extreme drought, Fig. B.2.) (Almeida et al., 2000; Divito and Sadras, 2014), and concurrent downregulation of symbiotic N2 fixation due to low N demand of the plant (Hartwig, 1998). Under extreme drought, the benefit of N2 fixation in T. repens could therefore not be transformed to increased resistance in biomass yields (Fig. 3c). The shallow root morphology of T. repens (Evans, 1977) may have accelerated the exposure to water stress and probably led to severe physiological damage towards the end of the drought period (Sanderson et al., 2003). Consequently, our results indicate that in legumes N limitation

under drought may be circumvented by symbiotic N2 fixation as long as water scarcity does not become extreme. This is a clear advantage compared to non-legumes, as indicated by the shallowrooted non-N2-fixing L. perenne, which already showed a significant biomass loss during the first regrowth under severe drought (Fig. 3a). For the response of symbiotic N2 fixation to such shortterm changes in drought stress, the drought sensitivity of freeliving Rhizobia sp. seems irrelevant, but rather the short-term regulation of the N2 fixation activity of the nodulated plant. This regulation of symbiotic N2 fixation through N feedback mechanisms can occur on many scales from the level of gene expression (such as regulation of nitrogenase activity in the nodules) to whole ecosystems (e.g. restrictions in mineral N fluxes will constrain availability of soil N and plant N uptake) (Hartwig, 1998). Plant N limitation might also affect water – and carbon (C) uptake, and the process of photosynthesis. There is convincing evidence that plant growth is inhibited by drought earlier and more severely than photosynthesis itself (Muller et al., 2011). Given that both processes, photosynthesis and growth, are strongly linked to N metabolism, this would indicate that, at early stages of water scarcity, N limitation should primarily inhibit organ expansion (and thus C sink; Muller et al., 2011), rather than C uptake and photosynthesis (C source). In our study, an additional analysis of the water use efficiency (WUE, as measured by the carbon isotope ratio of leaf tissue, d13C) revealed that the two legume species had also consistently higher WUE than the two non-legume species under rainfed control and drought conditions (Fig. B.3., Appendix B in Supplementary file), indicating also a relative advantage of these forage legumes compared to nonlegumes regarding the water-carbon balance under drought (Adams et al., 2016). Further research should clarify the role of N limitation on the plant water- and C balance, which would require to measure all parameters simultaneously under increasing water deficit. Legumes provide many benefits to forage production under adequate water supply (reviewed in Dumont et al., 2014; Lüscher et al., 2014; Phelan et al., 2015), including overyielding in mixtures with grasses. Hofer et al. (2016) demonstrated that such mixture overyielding was evident even under severe drought. Here, we provide further support for cropping forage legumes in mixtures with non-legumes as a drought mitigation option against severe drought events. The legumes’ sustained ability of symbiotic N2 fixation and comparably higher WUE under moderate to severe drought could overcome drought-induced limitations in soil N. Non-legumes grown in mixtures with legumes might profit from the increased N acquisition also under drought (Hoekstra et al., 2015; Hofer et al., 2016), as well as from high soil N availability during post-drought (Fig. 2c) due to enhanced litter and root decomposition with high N content. 4.3. Nitrogen fertilisation improved non-legumes’ resistance to severe drought, but did not affect post-drought overcompensation

Fig. 7. Drought effects on aboveground biomass (expressed as the log response ratio and as change in aboveground biomass due to drought) in relation to increasing soil water scarcity (expressed as number of days with a soil matric potential (SMP) below
1.5 MPa across 5 cm and 40 cm soil depth [sum of both depths divided by two]). Trend lines are based on linear (non-legumes, dashed line) and non-linear (legumes, continuous line) least-square optimisation. Adding a nonlinear term to non-legumes’ fit proved to be not significant (P > 0.2). Data are compiled from Fig. 3 and Hofer et al. (2016); they were established under similar conditions regarding design and technical facilities, consist of two harvests during a drought period of comparable length, and represent three site-years from Switzerland and one from Ireland. Displayed are means  1 SE. SR: shallow-rooted, DR: deep-rooted, NL: non-legume, LE: legume.

Additional N fertilisation during the drought period could partly mitigate aboveground biomass losses in non-legumes under severe drought (Fig. 3b), and the benefit of N fertilisation for the non-legumes led to a similar CAB as for the non-fertilised legumes under drought (approx.
20%). This mitigation effect of N fertilisation clearly points to the relevance of drought-induced limitations in soil N sources for these forage species. The nonlegumes were able to incorporate all additional N into biomass and could have utilised even more N, as indicated by the NNI that remained unchanged by N fertilisation (Fig. 4). Yet, N fertilisation had no effect on biomass losses under extreme drought, which we ascribe to the increased, direct water limitation under prolonged drought stress. Other studies that combined levels of soil water and

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

321

N fertilisation found similarly declining efficiency of N fertilisation under persisting drought (e.g. Colman and Lazenby, 1975; Gonzalez-Dugo et al., 2005). Our results thus suggest that additional N fertilisation under drought might partly mitigate aboveground biomass losses in non-legumes as long as water stress is not extreme. All species except T. repens were highly resilient following the extreme drought event and the formerly drought-stressed nonlegumes even outperformed the non-stressed controls during the post-drought period (Fig. 3e). This excellent resilience and the outperformance are in agreement with previous findings from three other experiments (Hofer et al., 2016) and occurred independent of N fertilisation during drought (Fig. 3e and f). Rather, N availability and acquisition were probably high as a result of resources not taken up during drought, and more favourable soil processes during post-drought, which is supported by the high soil N availability for legumes (Fig. 2e) and by NNIs at least as high in formerly drought-stressed as in non-stressed species (Fig. 4a and c). There is evidence that drought enhances plant and microbial necromass in the soil (Borken and Matzner, 2009) and dryrewetting events increase soil microbial activity (Gordon et al., 2008). As a result, drought events can induce a post-drought pulse in N mineralisation (Borken and Matzner, 2009) and a short-term increase in soil fertility (Bloor and Bardgett, 2012). Moreover, in our experiment a priming effect in response to N fertilisation (Léon et al., 1995) may have further amplified resilience, as there was distinct N fertilisation during the post-drought period (see Table B.1., Appendix B in Supplementary file).

symbiotic N2 fixation, until water scarcity becomes extreme. Thus, legumes might be used to improve drought resistance in mixed forage swards and might therefore serve as an adaptation option to more frequent summer drought events as projected under climate change.

4.4. Short-term root reactions had limited potential to mitigate drought impairment

Adams, M.A., Turnbull, T.L., Sprent, J.I., Buchmann, N., 2016. Legumes are different: leaf nitrogen, photosynthesis, and water use efficiency. Proc. Nat. Acad. Sci. 113, 4098–4103. Agroscope Reckenholz-Tänikon ART, 2012. Referenzmethoden der Forschungsanstalten, Band 2: Bodenuntersuchungen zur StandortCharakterisierung. Agroscope Reckenholz-Tänikon ART, Zürich, Switzerland. Almeida, J.P.F., Hartwig, U.A., Frehner, M., Nösberger, J., Lüscher, A., 2000. Evidence that P deficiency induces N feedback regulation of symbiotic N2 fixation in white clover (Trifolium repens L.). J. Exp. Bot. 51, 1289–1297. Bélanger, G., Gastal, F., Lemaire, G., 1992. Growth analysis of a tall fescue sward fertilized with different rates of nitrogen. Crop Sci. 32, 1371–1376. Bloor, J.M.G., Bardgett, R.D., 2012. Stability of above-ground and below-ground processes to extreme drought in model grassland ecosystems: interactions with plant species diversity and soil nitrogen availability. Perspect. Plant Ecol. Evol. Syst. 14, 193–204. Bollig, C., Feller, U., 2014. Impacts of drought stress on water relations and carbon assimilation in grassland species at different altitudes. Agric. Ecosyst. Environ. 188, 212–220. Borken, W., Matzner, E., 2009. Reappraisal of drying and wetting effects on C and N mineralization and fluxes in soils. Glob. Change Biol. 15, 808–824. Burke, I.C., Lauenroth, W.K., Parton, W.J., 1997. Regional and temporal variation in net primary production and nitrogen mineralization in grasslands. Ecology 78, 1330–1340. Cassman, K.G., Munns, D.N., 1980. Nitrogen mineralization as affected by soil moisture, temperature, and depth. Soil Sci. Soc. Am. J. 44, 1233–1237. Chalk, P.M., 1985. Estimation of N2 fixation by isotope dilution: an appraisal of techniques involving 15N enrichment and their application. Soil Biol. Biochem. 17, 389–410. Chen, S., Lin, S., Loges, R., Hasler, M., Taube, F., 2016. Comparison of ingrowth core and sequential soil core methods for estimating belowground net primary production in grass-clover swards. Grass Forage Sci. 71, 515–528. Colman, R.L., Lazenby, A., 1975. Effect of moisture on growth and nitrogen response by Lolium perenne. Plant Soil 42, 1–13. Deléglise, C., Meisser, M., Mosimann, E., Spiegelberger, T., Signarbieux, C., Jeangros, B., Buttler, A., 2015. Drought-induced shifts in plants traits, yields and nutritive value under realistic grazing and mowing managements in a mountain grassland. Agric. Ecosyst. Environ. 213, 94–104. Dijkstra, F.A., He, M., Johansen, M.P., Harrison, J.J., Keitel, C., 2015. Plant and microbial uptake of nitrogen and phosphorus affected by drought using 15N and 32 P tracers. Soil Biol. Biochem. 82, 135–142. Divito, G.A., Sadras, V.O., 2014. How do phosphorus, potassium and sulphur affect plant growth and biological nitrogen fixation in crop and pasture legumes? A meta-analysis. Field Crop. Res. 156, 161–171. Dreesen, F.E., De Boeck, H.J., Janssens, I.A., Nijs, I., 2012. Summer heat and drought extremes trigger unexpected changes in productivity of a temperate annual/ biannual plant community. Environ. Exp. Bot. 79, 21–30. Dumont, B., Andueza, D., Niderkorn, V., Lüscher, A., Porqueddu, C., Picon-Cochard, C., 2014. A meta-analysis of climate change effects on forage quality in grasslands:

Few short-term reactions of root growth to drought were found during the ten-week drought period (Fig. 6), although root activity is known to change quickly under drought (Hoekstra et al., 2014, 2015; Dijkstra et al., 2015). The general, drought-induced increase in total root biomass of all species and across all soil depths is explained by constricted decomposition in soils during drought (Sanaullah et al., 2012; Walter et al., 2013) and because we did not differentiate between living and dead roots. Inconsistent results are reported regarding drought-induced changes in the entire standing root biomass (e.g. Skinner and Comas, 2010; Dreesen et al., 2012) and root masses at distinct soil layers (e.g. Wedderburn et al., 2010; Prechsl et al., 2015). High root biomass due to increased root growth under drought in combination with a lower decomposition rate might, however, promote fast resource uptake during post-drought as soon as soil resources are again available (Padilla et al., 2013). 5. Conclusions As a result of the different drought responses in biomass yield between non-legumes and legumes at the first regrowth, it can be inferred that soil N deficiency impaired plant N status and biomass yield at these stages more than direct water deficiency. However, during the second regrowth when drought stress became extreme, soil water limitation must have become so strong that all species were heavily impaired, irrespective of their N acquisition and rooting depth. This leads us to infer that a shift from N- to water limitation occurred during the ten-week drought period. This is in agreement with Hooper and Johnson (1999) who state that colimitation of water and N occurs above a certain amount of water supply, but that water limits plant growth primarily when precipitation is extremely low. We further conclude that legumes resist increasing drought stress to some degree (Fig. 7) because they can overcome the initial, predominant N limitation through

Acknowledgements The authors thank Cornel Stutz and Rafael Gago for technical support and Marianne Leuzinger, Désirée Bäder, Lukas Vögeli and Lenny Weber for their help with field work and sample processing. Nyncke Hoekstra is kindly acknowledged for valuable inputs during experimental planning, Peter Weisskopf and Marlies Sommer for determining the soil desorption curves and Katharine Seipel for linguistic revision of the manuscript. Meteorological data were thankfully provided by the Federal Office for Meteorology (MeteoSwiss), and the use of Plant Root Simulator (PRS)TM-probes was partly subsidised by Western AG Innovations. We acknowledge financial support by the Animal Change project from the European Union’s Seventh Framework Programme (FP7/20072013; grant agreement no. 266018). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.agee.2016.11.022. References

322

D. Hofer et al. / Agriculture, Ecosystems and Environment 236 (2017) 312–322

specificities of mountain and Mediterranean areas. Grass Forage Sci. 70, 239– 254. Durand, J.-L., Gonzalez-Dugo, V., Gastal, F., 2010. How much do water deficits alter the nitrogen nutrition status of forage crops? Nutr. Cycl. Agroecosys. 88, 231– 243. Evans, P.S., 1977. Comparative root morphology of some pasture grasses and clovers. New Zeal. J. Agric. Res. 20, 331–335. Flisch, R., Charles, R., Richner, W., 2009. Grundlage für die Düngung im Acker-und Futterbau. Agrarforschung 16, 1–100. Gilgen, A.K., Buchmann, N., 2009. Response of temperate grasslands at different altitudes to simulated summer drought differed but scaled with annual precipitation. Biogeosciences 6, 2525–2539. Gonzalez-Dugo, V., Durand, J.L., Gastal, F., Picon-Cochard, C., 2005. Short-term response of the nitrogen nutrition status of tall fescue and Italian ryegrass swards under water deficit. Aust. J. Agric. Res. 56, 1269–1276. Gordon, H., Haygarth, P.M., Bardgett, R.D., 2008. Drying and rewetting effects on soil microbial community composition and nutrient leaching. Soil Biol. Biochem. 40, 302–311. Grant, K., Kreyling, J., Dienstbach, L.F.H., Beierkuhnlein, C., Jentsch, A., 2014. Water stress due to increased intra-annual precipitation variability reduced forage yield but raised forage quality of a temperate grassland. Agric. Ecosyst. Environ. 186, 11–22. Hartwig, U.A., 1998. The regulation of symbiotic N2 fixation: a conceptual model of N feedback from the ecosystem to the gene expression level. Perspect. Plant Ecol. Evol. Syst. 1, 92–120. Hedges, L.V., Gurevitch, J., Curtis, P.S., 1999. The meta-analysis of response ratios in experimental ecology. Ecology 80, 1150–1156. Hoekstra, N.J., Finn, J.A., Hofer, D., Lüscher, A., 2014. The effect of drought and interspecific interactions on depth of water uptake in deep- and shallowrooting grassland species as determined by d18O natural abundance. Biogeosciences 11, 4493–4506. Hoekstra, N.J., Suter, M., Finn, J.A., Husse, S., Lüscher, A., 2015. Do belowground vertical niche differences between deep- and shallow-rooted species enhance resource uptake and drought resistance in grassland mixtures? Plant Soil 394, 21–34. Hofer, D., Suter, M., Haughey, E., Finn, J.A., Hoekstra, N.J., Buchmann, N., Lüscher, A., 2016. Yield of temperate forage grassland species is either largely resistant or resilient to experimental summer drought. J. Appl. Ecol. 53, 1023–1034. Hooper, D.U., Johnson, L., 1999. Nitrogen limitation in dryland ecosystems: responses to geographical and temporal variation in precipitation. Biogeochemistry 46, 247–293. Hoover, D.L., Knapp, A.K., Smith, M.D., 2014. Resistance and resilience of a grassland ecosystem to climate extremes. Ecology 95, 2646–2656. Isbell, F., Craven, D., Connolly, J., Loreau, M., Schmid, B., Beierkuhnlein, C., Bezemer, T. M., Bonin, C., Bruelheide, H., de Luca, E., Ebeling, A., Griffin, J.N., Guo, Q., Hautier, Y., Hector, A., Jentsch, A., Kreyling, J., Lanta, V., Manning, P., Meyer, S.T., Mori, A.S., Naeem, S., Niklaus, P.A., Polley, H.W., Reich, P.B., Roscher, C., Seabloom, E.W., Smith, M.D., Thakur, M.P., Tilman, D., Tracy, B.F., van der Putten, W.H., van Ruijven, J., Weigelt, A., Weisser, W.W., Wilsey, B., Eisenhauer, N., 2015. Biodiversity increases the resistance of ecosystem productivity to climate extremes. Nature 526, 574–577. Kemp, D.R., Culvenor, R.A., 1994. Improving the grazing and drought tolerance of temperate perennial grasses. New Zeal. J. Agric. Res. 37, 365–378. Léon, M., Laine, P., Ourry, A., Boucaud, J., 1995. Increased uptake of native soil nitrogen by roots of Lolium multiflorum Lam. after nitrogen fertilization is explained by a stimulation of the uptake process itself. Plant Soil 173, 197–203. Lüscher, A., Mueller-Harvey, I., Soussana, J.F., Rees, R.M., Peyraud, J.L., 2014. Potential of legume-based grassland-livestock systems in Europe: a review. Grass Forage Sci. 69, 206–228. Lemaire, G., Gastal, F., 1997. N uptake and distribution in plant canopies. In: Lemaire, G. (Ed.), Diagnosis of the Nitrogen Status in Crops. Springer, Berlin, pp. 3–43. McAuliffe, C., Chamblee, D.S., Uribe-Arango, H., Woodhouse, W.W., 1958. Influence of inorganic nitrogen on nitrogen fixation by legumes as revealed by 15N. Agron. J. 50, 334–337. Muller, B., Pantin, F., Génard, M., Turc, O., Freixes, S., Piques, M., Gibon, Y., 2011. Water deficits uncouple growth from photosynthesis, increase C content, and modify

the relationships between C and growth in sink organs. J. Exp. Bot. 62, 1715– 1729. Olesen, J.E., Trnka, M., Kersebaum, K.C., Skjelvåg, A.O., Seguin, B., Peltonen-Sainio, P., Rossi, F., Kozyra, J., Micale, F., 2011. Impacts and adaptation of European crop production systems to climate change. Eur. J. Agron. 34, 96–112. Padilla, F., Aarts, B.J., Roijendijk, Y.A., de Caluwe, H., Mommer, L., Visser, E.W., de Kroon, H., 2013. Root plasticity maintains growth of temperate grassland species under pulsed water supply. Plant Soil 369, 377–386. Phelan, P., Moloney, A.P., McGeough, E.J., Humphreys, J., Bertilsson, J., O’Riordan, E. G., O’Kiely, P., 2015. Forage legumes for grazing and conserving in ruminant production systems. Crit. Rev. Plant Sci. 34, 281–326. Pimm, S.L., 1984. The complexity and stability of ecosystems. Nature 307, 321–326. Pinheiro, J.C., Bates, D.M., 2009. Mixed-Effects Models in S and S-PLUS. Springer, New York. Poirier, M., Durand, J.-L., Volaire, F., 2012. Persistence and production of perennial grasses under water deficits and extreme temperatures: importance of intraspecific vs. interspecific variability. Glob. Change Biol. 18, 3632–3646. Prechsl, U.E., Burri, S., Gilgen, A.K., Kahmen, A., Buchmann, N., 2015. No shift to a deeper water uptake depth in response to summer drought of two lowland and sub-alpine C3-grasslands in Switzerland. Oecologia 177, 97–111. Qian, P., Schoenau, J.J., 2005. Use of ion-exchange membrane to assess nitrogensupply power of soils. J. Plant Nutr. 28, 2193–2200. R Core Team, 2016. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project. org. Sanaullah, M., Rumpel, C., Charrier, X., Chabbi, A., 2012. How does drought stress influence the decomposition of plant litter with contrasting quality in a grassland ecosystem? Plant Soil 352, 277–288. Sanderson, M.A., Byers, R.A., Skinner, R.H., Elwinger, G.F., 2003. Growth and complexity of white clover stolons in response to biotic and abiotic stress. Crop Sci. 43, 2197–2205. Seneviratne, S.I., Nicholls, N., Easterling, D., Goodess, C.M., Kanae, S., Kossin, J., Luo, Y., Marengo, J., McInnes, K., Rahimi, M., Reichstein, M., Sorteberg, A., Vera, C., Zhang, X., 2012. Changes in climate extremes and their impacts on the natural physical environment. In: Field, C.B., Barros, V., Stocker, T.F., Qin, D., Dokken, D.J., Ebi, K.L., Mastrandrea, M.D., Mach, K.J., Plattner, G.-K., Allen, A.P., Tignor, M., Midgley, P.M. (Eds.), Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. A special report of working groups I and II of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge and New York, pp. 109–230. Skinner, R.H., Comas, L.H., 2010. Root distribution of temperate forage species subjected to water and nitrogen stress. Crop Sci. 50, 2178–2185. Skinner, R.H., Gustine, D.L., Sanderson, M.A., 2004. Growth, water relations, and nutritive value of pasture species mixtures under moisture stress. Crop Sci. 44, 1361–1369. Trenberth, K.E., Dai, A., van der Schrier, G., Jones, P.D., Barichivich, J., Briffa, K.R., Sheffield, J., 2014. Global warming and changes in drought. Nat. Clim. Chang. 4, 17–22. Vogel, A., Scherer-Lorenzen, M., Weigelt, A., 2012. Grassland resistance and resilience after drought depends on management intensity and species richness. PLoS One 7, e36992. Walter, J., Hein, R., Beierkuhnlein, C., Hammerl, V., Jentsch, A., Schädler, M., Schürings, J., Kreyling, J., 2013. Combined effects of multifactor climate change and land-use on decomposition in temperate grassland. Soil Biol. Biochem. 60, 10–18. Wang, Y., Yu, S., Wang, J., 2007. Biomass-dependent susceptibility to drought in experimental grassland communities. Ecol. Lett. 10, 401–410. Wedderburn, M.E., Crush, J.R., Pengelly, W.J., Walcroft, J.L., 2010. Root growth patterns of perennial ryegrasses under well-watered and drought conditions. New Zeal. J. Agric. Res. 53, 377–388. Zwicke, M., Alessio, G.A., Thiery, L., Falcimagne, R., Baumont, R., Rossignol, N., Soussana, J.F., Picon-Cochard, C., 2013. Lasting effects of climate disturbance on perennial grassland above-ground biomass production under two cutting frequencies. Glob. Change Biol. 19, 3435–3448.