Agriculture, Ecosystems and Environment 213 (2015) 94104

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment

journal homepage: www.elsevier.com/locate/agee

Drought-induced shifts in plants traits, yields and nutritive value under

realistic grazing and mowing managements in a mountain grassland
Claire Delglisea,b,* , Marco Meissera , Eric Mosimanna , Thomas Spiegelbergerc,d,
Constant Signarbieuxb,e, Bernard Jeangrosf , Alexandre Buttlerb,e,g
a

Agroscope, Institute for Livestock Sciences, Forage Crops and Grazing Systems Group, Route de Duillier 50, CH-1260 Nyon, Switzerland
Swiss Federal Institute for Forest, Snow and Landscape Research (WSL), Site Lausanne, Station 2, CH-1015 Lausanne, Switzerland
Irstea, UR EMGR Mountain Ecosystems, 2 rue de la Papeterie, BP 76, F-38402 St-Martin-dHres, France
d
Universit Grenoble Alpes, F-38402 Grenoble, France
e
Ecole Polytechnique Fdrale de Lausanne (EPFL), School of Architecture, Civil and Environmental Engineering, Laboratory of Ecological Systems (ECOS),
Station 2, CH-1015 Lausanne, Switzerland
f
Agroscope, Institute for Plant Production Sciences, Route de Duillier 50, CH-1260 Nyon, Switzerland
g
Laboratoire de Chrono-Environnement, UMR CNRS 6249, UFR des Sciences et Techniques, 16 route de Gray, Universit de Franche-Comt, F-25030 Besanon,
France
b
c

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 17 December 2014
Received in revised form 23 July 2015
Accepted 24 July 2015
Available online 7 August 2015

Extreme drought events can dramatically impact grassland ecosystems, causing potential loss of forage
production for livestock in temperate grasslands. However, knowledge of drought effects on forage
production, nutritive value and plant community stability in the real context of farming management is
scarce. For this purpose, the effect of a simulated summer drought was studied under two realistic
management types on a semi-natural grassland in the Swiss Jura mountains. The rst management
(grazing) consisted in six consecutive utilizations by animals over the growing season, representing a
rotational grazing system as regionally conducted. The second management (mowing) consisted of
three harvests, corresponding to the usual mowing frequency of hay meadows. In both managements,
drought caused minor short-term modications of species composition and almost no persistent effects
were observed. Besides, important short-term changes were observed in community weighted mean of
several key functional traits, reecting a strong decline in community growth during the drought
followed by a partial recovery two months later. Forage yields, and to a lesser extent its nutritive value,
thus declined during the drought period. Both were still affected in the following months, but had
recovered in spring of the following year. Forage loss was twofold higher in the grazing management, but
recovery as well as nutritive value was slightly improved in this management. The results suggest that
rotational grazing can amplify negative drought impacts, compared to traditional mowing, and highlight
the need to adapt such management in dry years. They also demonstrate that, even under a fairly
intensive management, resilience of such mountain grasslands after one extreme drought event can be
expected to be high, with no persistent changes in species and trait compositions.
2015 Elsevier B.V. All rights reserved.

Keywords:
Rain-shelter
Functional traits
Forage quality
Recovery
Grassland management

1. Introduction
In temperate regions, mountain grasslands are at the core of
livestock farming and bear important economic, social and
ecological values. In Switzerland, agricultural policy supports
forage self-sufciency and promotes grassland-based livestock

* Corresponding author at: Agroscope, Route de Duillier 50, CH-1260 Nyon,

Switzerland. Fax: +41 22 362 13 25.
E-mail address: claire.deleglise@agroscope.admin.ch (C. Delglise).
http://dx.doi.org/10.1016/j.agee.2015.07.020
0167-8809/ 2015 Elsevier B.V. All rights reserved.

farming (Barth et al., 2011). However, with climate change,

grasslands will be challenged to meet a growing demand for
providing forage as well as other services (Easterling et al., 2000;
EEA, 2012). Climate change is very likely to cause a rise in
temperatures, to change rainfall patterns and to increase the
magnitude and frequency of extreme climatic events such as
droughts (IPCC, 2007). For Central Europe, summer droughts are
expected to be among the main consequences of climate change
(Calanca, 2007; Fischer et al., 2012; Frei et al., 2006) and may cause
short-term as well as persistent consequences on grasslands
(Jentsch and Beierkuhnlein, 2008).

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

Grasslands of the Swiss Jura mountains have a long history of

livestock grazing. They have proven to be sustainable over past
decades but little is known about their agronomical response
under extreme weather events such as severe drought (Gavazov
et al., 2013). In particular, the extent to which the production and
the nutritive value of the forage will be affected is difcult to
predict. Indeed, if yield losses due to droughts have been reported
in many studies, their magnitude can be very variable, due to
multiple biotic and abiotic factors (i.e. plant diversity, productivity,
altitude, management) that may interact (Gilgen and Buchmann,
2009; Grant et al., 2014; Kahmen et al., 2005; Vogel et al., 2012;
Wang et al., 2007; Zwicke et al., 2013). In addition, existing
information concerning drought effects on the nutritive value of
forage are contradictory. For temperate regions, most of available
information concerns cultivated grass and legume forages and
reports that water scarcity can improve their nutritive value, e.g.
with increased crude protein content and decreased ber
concentration, presumably due to delayed maturity of plants
(Bittman et al., 1988; Buxton, 1996; Halim et al., 1989; Jensen et al.,
2010; Kchenmeister et al., 2013). On the other hand, it can also be
expected that a severe water stress, by causing leaf senescence, will
lead to an increased stem to leaf ratio as well as to nutrient
translocation from leaves to roots, both of which contributing to
low forage quality (Bruinenberg et al., 2002; Buxton, 1996).
Nevertheless, predicting how forage production and quality will
be affected by droughts requires to account for real management
practices undertaken in grassland-based livestock farming. In
particular, grazing experiments in real context with animals are
rare, despite grazing being the main grassland management
worldwide (van Wieren and Bakker, 2008). In drought studies
conducted on permanent pastures, grazing is often simulated
through cuttings, which are, for reasons of feasibility, generally
uniformly applied to the grassland plots (e.g. Gilgen and
Buchmann, 2009; Mariotte et al., 2013). However, grazing is likely
to impact the spatial structure of the vegetation due to feeding
preferences (Adler et al., 2001), and, through defoliation,
dejections, and trampling, to have positive or negative interactions
with above and below-ground components of grassland ecosystems (Frank, 2007; Kohler et al., 2005). In addition, in many
regions, grasslands are usually grazed in a rotational system, as
illustrated for the Jura mountains (Mosimann et al., 2012). This
implies frequent utilizations, unlike what happens in drought
experiments where cuttings are only applied before and after the
drought treatment, allowing sometimes a long vegetation regrowth period between utilizations (e.g. Mariotte et al., 2013).
While frequent utilizations generally reduces the annual dry
matter production, it maintains plants into an active growing and
tillering phase instead of maturing naturally, which then provides
a forage with improved nutritive value (Bruinenberg et al., 2002).
Therefore, it is likely that grassland management through realistic
grazing utilization inuences the response of the vegetation to
drought in a different manner to the more commonly studied
mowing utilization.
The plant functional approach, in which both species composition and plant traits are monitored, is well suited for studying
grassland responses to environmental changes (e.g. Garnier et al.,
2004; Qutier et al., 2007; Suding et al., 2008). Plant functional
traits can be simple measurements that depict the stature, the leaf
economics spectrum or the root properties of plants (Violle et al.,
2007), directly linked to the functions of plant growth, development and resource acquisition (Weiher et al., 1999; Wright et al.,
2004). Variations in key plant traits such as specic leaf area (SLA)
and leaf nitrogen concentration (LNC), linked to water conservation functions during photosynthesis (Wright and Westoby 2002),
have been documented in response to water stress (Jung et al.,
2014). At the community scale, trait values are averaged across all

95

species and weighted by species abundance (e.g. Garnier et al.,

2004). Variations in community plant traits in response to drought
can therefore be due to both modication of species composition
and intraspecic variability in traits, i.e. plasticity in plant
morphology and physiology (Jung et al., 2014; Lep et al., 2011).
Drought-induced shifts in species composition are however likely
to occur on a long-time scale, with potential lasting effect on the
community properties, whereas intraspecic trait variations can
occur rapidly but with potential effects of short (e.g. seasonal)
duration (Jung et al., 2014; Lloret et al., 2012; Suding et al., 2008).
Such changes in community plant traits can ultimately inuence
agronomic services such as forage yields and its nutritive value
(Gardarin et al., 2014; Qutier et al., 2007).
In the present study, we experimentally studied the effect of a
severe single summer drought on the forage yields and its nutritive
value in a mountain grassland in the Swiss Jura. The grassland was
managed according to the two main farming systems practiced in
the study region: rotational grazing (6 utilizations by sheep per
year; each consisting of 2 days grazing followed by 46 weeks
resting) and mowing (3 forage harvests at 810 weeks interval).
Shifts in plant species composition and trait values were
monitored to study mechanisms of community response to
drought and management. In this context, we hypothesized that:
(i) the dry matter production and the nutritive value of the forage
are negatively affected in the short-term by a severe summer
drought, (ii) grazing management further increases the negative
effect of drought due to frequent utilizations, (iii) these short-term
responses are accompanied by shifts in community plant trait
values, while species composition remains stable, (iv) in the
absence of persistent changes in species composition and trait
values, the recovery in forage production and nutritive value after a
single drought event is high.
2. Materials and methods
2.1. Study site
The experiment was conducted from spring 2012 to spring
2013 on a permanent grassland located in the Jura Mountains
(Agroscope research station, La Frtaz, Western Switzerland,
6
330 56.500 E, 46
500 15.500 N, 1200 m a.s.l.). Climate at the site is
suboceanic with annual precipitations of 1333 mm (19812010 30year mean, MeteoSwiss). The snow cover usually lasts from midNovember to mid-April. During the summer months (June
August), precipitation reaches 347 mm and temperature 13.5
C
(30-year mean, MeteoSwiss). In 2012, the annual rainfall reached
1490 mm, with relatively frequent precipitation during summer
months (JuneAugust: 405 mm). The grassland covers a at area on
a relatively deep soil (50 cm in average), classied as Cambisol
eutric (IUSS Working Group WRB, 2006). The topsoil horizon (0
20 cm) contains 24% sand, 44% silt and 32% clay with a mean pH of
6.6. Based on soil texture, the water holding capacity (WHC) of the
soil, the soil water content at eld capacity (SWCfc) and at wilting
point (SWCwp), were estimated, respectively, around 100 mm,
0.35 and 0.15 cm3 water/cm3 soil, over the 50 cm depth (Saxton
et al., 1986). For validation, we used these estimated parameters to
compute the soil water balance, and thus the soil water content
(SWC) over the experiment period, following the FAO methodology
(Allen et al., 1998) and ensured that the adequacy between
computed and observed data of SWC was good. The plant
community of the grassland is dominated by grasses (6080%)
with rather nutrient-rich species such as Poa trivialis,Agrostis
capillaris, Festuca pratensis, Dactylis glomerata and Lolium perenne.
Both legumes (mainly Trifolium repens) and forbs (mainly
Ranunculus acris and Taraxacum ofcinale), account for about 10
20%. Biomass production (57 t ha1) is high relative to the site

96

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

altitude (Mosimann et al., 2012). The grassland has been used, over
the past 20 years, through relatively intensive rotational grazing by
cattle or by sheep with, in average, six utilizations from May to late
October. Fertilization occurs with manure, which is supplied every
year in autumn.
2.2. Experimental design
The design consisted of the combination of two factors (drought
and management) with two levels each (control vs. drought;
grazing vs. mowing management) conducted within a full factorial
design. In early spring 2012, four blocks containing each four large
plots (6.4 m  12 m each) were established on the grassland. Every
factorial combination (i.e. mowing-control; mowing-drought;
grazing-control; grazing-drought) was repeated four times and,
within each block, randomly allocated to one plot. Plots were
positioned with a buffer of at least 2 m between each plot. Their
positions were chosen in order to minimize environmental
heterogeneity within and among plots and to be representative
of the plant species assemblage of the grassland. Plots had to be
large enough to provide sufcient fodder for two sheep during
each grazing utilization (see below Section 2.2.2). Hence, despite
the limited number of plots, their size enabled to cover a
representative area of the grassland within each plot, providing
relatively low initial variability between plots.
2.2.1. Rainfall exclusion and microclimate monitoring
A severe drought was simulated during 10 weeks in summer
(from June 19th to September 2nd) using semi-cylindrical rainshelters (length: 12 m; width: 6.4 m, height: 3 m, Filclair, Numeris
6.40) covered with a transparent plastic lm (180 mm, transparent
M42, Filclair). Plastic lms were set so as to leave an opening of
approximately 80 cm on the sides of the shelters and the two open
ends were oriented towards the dominant wind direction, allowing
natural air ventilation. Rainwater runoff was canalized away from

2.2.2. Grassland management

The two management treatments tested in this study correspond to the main systems of forage utilization found in the study

D rough t

40

(mm)

Precipitations

50

the edge of plots thanks to plastic pipes. Control plots remained

uncovered and received ambient precipitations. During the drought
simulation period, 300 mm of water were excluded from the drought
plots (Fig. 1a), representing a 74% decrease in the JuneAugust
rainfall and a 20% decrease in total annual precipitation. This
corresponds to a drought intensity that might have a 10-year return
period at the end of the 21st century, looking at the intermediate
scenario in the study region (CH2011, 2011). The index of
precipitation minus potential evapotranspiration (PET) accumulated
during the JuneAugust summer months reached 156 mm under
rain-shelters against 128 mm in the control treatment (Fig. 1b). This
represents an extreme drought intensity that had not been reached
during the past 30 years at the study site (30 years mean: 102 mm,
range: 101/+389 mm, MeteoSwiss). Microclimatic variables were
monitored continuously (at 30 min intervals) from May to October
2012 with data loggers (Em50, Decagon Devices Inc., Pullman, WA,
USA). Measurement of the soil water content (SWC) was done at
10 cm soil depth because 80% of ne roots have been observed to be
located within the top 020 cm soil horizon. On average during the
rain-shelter coverage period, SWC was 40% lower in the drought
treatment compared to the control under both managements
(Fig. 1c). At the end of the coverage period, SWC had reached values
of 0.15 cm3 water/cm3 soil (Fig. 1c), which corresponded to the
estimated SWC at wilting point. The slightly more variable soil
moisture in the grazing than in the mowing management before the
drought period can be imputed to a higher drying sensitivity of
shorter height vegetation (i.e. just being grazed) in the grazing
management (Fig. 1c). Below rain-shelters, no warming in air
temperature was detected at 50 cm above ground compared to the
control (Fig. 1d) and the photosynthetically active radiation was
reduced by 20% at its maximum (data not shown).

30
20
10

(mm)

+ 194 mm

+ 128 mm

Control
Drought

156 mm

0.4
0.3

MC
MD
GC
GD

0.2
0.1
0.0
25

20

(C)

Air temperature

Soil moisture
(m3 m3)

Precip. PET

0
50
40
30
20
10
0
10
0.5

15

MC
MD
GC
GD

10
5
0

May

June

July

August

Sept.

Oct.

Nov.

Fig. 1. Microclimatic variables from May to October 2012 for control (C) and drought (D) conditions in the mowing (M) and grazing (G) managements: (a) Daily precipitations
(data from MeteoSwiss station Bullet/La Frtaz), (b) Daily difference between precipitation and potential evapotranspiration (PET, MeteoSwiss station Bullet/La Frtaz) and
cumulative sums over summer and autumn months, (c) Daily mean soil water volumetric content at 10 cm depth, (d) Daily mean air temperature at 50 cm above soil level
(dash and full lines are superposed because of the absence of rain-shelter effects). Grey shading indicates the drought period from June 19th to September 2nd.

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

region. In this context, the purpose is to compare the two systems

taken as a whole, i.e. considering together the differences in (i) the
modality of defoliation (grazing vs. mowing) and (ii) the number of
utilizations (6 vs. 3).
Rotational grazing was conducted with eight pairs of sheep. For
each utilization, each pair of sheep was randomly allocated to one
of the eight plots of the grazing treatment. Sheep remained 36 to
60 h within plots for each utilization to ensure a grazed sward
residue of about 6 cm height. A water point and a sun shelter were
provided for sheep within each grazed plot. Plots were grazed six
times during the season, with a rst utilization at the end of May,
and about 46 weeks intervals between two utilizations, so as to
correspond to local grazing dates (hereafter referred to as G1G6).
G1 and G2 correspond to pre-drought grazing (for practical
reasons, G2 occurred right after the coverage of rain shelter, but
before any water stress could intervene). G3 and G4 correspond to
grazing during drought (G4 occurred at the end of the coverage
period), and G5 and G6 to grazing after drought. The fth
utilization (G5) had to be skipped in drought plots due the absence
of vegetation regrowth the month after the end of the drought.
The mowing management consisted of three forage harvests
during the season, with a rst cut at mid-June (M1), a second cut at
the end of August (M2) and a third at the end of October (M3),
corresponding to the usual use of hay meadows in the study region.
M1 occurred just before the coverage period and corresponds to
the pre-drought period harvest, M2 occurred at the end of the
coverage period and corresponds to the end of the drought period
harvest and M3 to two months after drought harvest. Harvests
were realised with a motor mower with a 4 cm height.
2.3. Data collection and analysis
All measurements were restricted to a central area of 4 m x 9 m
on each plot to avoid edge effects.
2.3.1. Botanical surveys
Species composition was measured four times in total: (1)
before starting experimental treatments, in May 2012 (initial
state); (2) at the end of the drought treatment, in August 2012
(drought phase); (3) 2 months after the drought, in October 2012
(recovery phase); (4) the spring of the following year, in May 2013
(to evaluate resilience).
Botanical surveys were realized with the point-quadrat method
(originally described by Levy and Madden,1933) with 100 sampling
points evenly distributed in the core area of each plot. At each
sampling point, a steel pin had been inserted vertically in the
sward, and each vascular plant species in contact with the pin had
been counted once, irrespective of eventually multiple contacts for
a given species (Buttler, 1992). Only green individuals could
account for a given species. Senescent individuals (i.e. senescence
was observed in all leaves) were recorded as an additional category
in the list. Then, for a given plot, the relative abundance (or specic
contributionButtler, 1992) of each species was determined as its
number of hits among the total number of hits recorded from the
100 sampling points. Similarly, the relative abundance of senescent
plants was estimated as the number of hits by senescent
individuals among the total number of hits recorded at the
100 sampling points.
Species richness, ShannonWeaver diversity and evenness
indices (Legendre and Legendre, 1998) were computed for each
plot at each sampling period.
2.3.2. Plant functional trait measurements
Plant functional traits were measured immediately after the
botanical survey, at four periods in total (see Section 2.3.1). Here
we were interested in plants maintaining their aerial tissues during

97

the drought stress, which are the most important components

from an agricultural point of view. Thus, we targeted aerial
vegetative traits and did not considered traits linked to drought
survival abilities of species, such as the maintenance of basal
meristems integrity (Volaire et al., 2014). We therefore measured
four aerial traits: plant vegetative height (VH), specic leaf area
(SLA), leaf dry matter content (LDMC) and leaf nitrogen content
(LNC). These traits are of interest because of their potential to vary
in response to disturbed water and nutrient availability and to
defoliation, through their link with the functions of plant growth,
development and resource acquisition (Daz et al., 2001; Weiher
et al., 1999; Westoby et al., 2002; Wright et al., 2004). They were
measured for the 14 main plant species whose cumulated
abundance reached at least 90% of total abundance in each plot
at each sampling period. Only adult plants in a vegetative stage
with no sign of senescence were collected, and the younger but
fully expanded leave was chosen for measurements according to
standard protocols (Cornelissen et al., 2003). For each species and
in each plot, VH was assessed as the average across measurements
on 20 individual plants and SLA and LDMC as the average across
10 individuals plants. However, because of the time required by
this method, some of the less abundant species were assessed only
at treatment level (and this concern only the May 2012, October
2012 and May 2013 sampling periods) by sampling 15 individual
plants pooled from the four replicated plots (Lep et al., 2011). SLA
and LDMC were measured after complete rehydration of plants in
deionized water during at least 12 h, to overcome differences
induced by different level of leaf water-saturation in the eld, as
demonstrated by Garnier et al. (2001). Then, leaves were ground
for each species and treatment to assess LNC with an elemental
analyser (Thermo Fisher Scientic Inc., Milan, Italy). For each trait
and each sampling period, the community weighted mean (CWM)
was calculated at the plot scale, as the mean of trait values (i.e. plot
or treatment averaged value) among all species, weighted
according to the relative abundance of species in the plot (Garnier
et al., 2004).
2.3.3. Dry matter yields
In 2012, biomass samples were collected just before each
grazing and mowing utilization (i.e. from G1 to G6 and from M1 to
M3) for comparative yield assessments. A motor mower with a
4 cm cutting height was used in four delineated stretches (total
surface of 4 m2) located in the core area of each plot. In grazed
plots, the sampled areas were changed at each rotation to ensure a
true grazing effect at these locations. In mown plots, sampled areas
were permanent. The four samples from each plot were pooled and
oven-dried at 60
C during 48 h for dry matter (DM) content and
DM yields determination. In spring of the following year (May
2013), the above-ground biomass was collected simultaneously in
all plots at a single date and processed according to the same
protocol.
To compare the results of DM yields between management
treatments, G1 and G2 were cumulated to correspond to M1 (i.e.
the biomass grown before the drought treatment, initial state), G3
and G4 were cumulated to correspond to M2 (i.e. the biomass
grown during the drought period) and G5 and G6 were cumulated
to correspond to M3 (i.e. the biomass grown after the drought,
recovery phase). In addition, 2012 annual yields were computed by
cumulating the DM yields from all cuts.
2.3.4. Nutritive value of the forage
To assess the nutritive value of the forage, i.e. the crude protein
(CP) content and the neutral and acid detergent ber (NDF, ADF)
contents, dried subsamples (containing green and senescent
biomass) from each plot and each sampling date (G1G6 and
M1M3, May 2013) were ground to pass through a 1-mm mesh

98

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

screen (Brabender mill, Brabender, Duisburg, Germany), and

scanned using near-infrared reectance spectroscopy (NIRS, Bchi
NIRFlex N-500). The spectra were analysed using a large dataset of
calibration samples and 10% of the analyses were checked by wet
chemistry (Agroscope Posieux, Switzerland). The CP content is
based on the nitrogen concentration and represents the forage's
ability to meet the protein needs of livestock. The NDF content
represents the cell wall material in the forage and is inversely
related to the voluntary fodder intake. The ADF content broadly
represents the content of cellulose and lignin contained within the
plants and is negatively correlated to energy and digestibility of the
forage. Then, based on the ADF content, the net energy for lactation
(NEL) was calculated, referring to the forage's ability to meet the
energy requirements of dairy cattle (Agroscope, 2013).
Means of nutritive value components were calculated across
G1 and G2 (pre-drought period), G3 and G4 (drought period) and
G5 and G6 (recovery period) so as to be compared to the nutritive
value of the forage in the mowing management (M1 to M3).
2.4. Statistical analysis
Multivariate analysis of species composition was carried out
using redundancy analysis (RDA), a constrained form of principal
component analysis (Lep and milauer, 2003). For each of the four
sampling periods, cumulative abundance of species accounted for
100% after removal of senescent plant abundance. Factors used in
the analyses were blocks (4 levels), drought and management
treatments with two levels each, and their interaction. Factors
were tested using a Monte Carlo test with 999 permutations (ter
Braak and milauer, 1999).
Univariate data were analysed by means of analysis of variance
(ANOVA) to test the effects of drought, management and their
interaction, under constraints of the blocked design. ANOVAs were
performed for species abundances, species richness, diversity,
evenness, abundance of senescent plants, community weighted
mean traits (VH, SLA, LMDC, LNC), DM yields and forage nutritive
value components (CP, ADF, NDF, NEL). When initial differences
were detected in May 2012, then initial value of the variable was
taken into account as a covariable in the following tests.
Homogeneity of variances and normality of residual distributions
were checked prior to analyses. Post-hoc tests (Tukeys Honest
Signicant Difference method) were performed to assess pairwise
comparisons after ANOVAs. All analyses were carried out using R
3.0.1 statistical software (R Core Team, 2013) with the packages
vegan (Oksanen et al., 2013) and nlme (Pinheiro et al., 2013).
3. Results
3.1. Species composition and diversity
At the initial state, in May 2012, there was no signicant effect
of treatments on the species composition (P > 0.05, Table 1). At the

end of the drought period (August 2012), 18.0% of the variability in

species composition between plots was due to the drought
treatment (P < 0.01) and 10.5% to its interaction with management
(P < 0.05, Table 1). Drought plots were characterized by a lower
proportion of Trifolium repens (grazed plots) or Lolium perenne
(mown plots) than control plots (Fig. 2). Two months later
(recovery phase), drought accounted for 15.0% (P < 0.05) of the
variability in species composition but management seemed more
important (17.4%, P < 0.01, Table 1). In spring of the following year,
drought still accounted for 12.3% of the variability (P < 0.05,
Table 1). This is mainly due to the lasting effect of drought on the
proportion of Trifolium repens which remained lower in drought
plots compared to the control plots (Fig. 2).
At the end of the drought period, proportion of senescent plants
was higher in drought than control plots and this response was
more pronounced in the grazing management (F1,9(interaction) = 15.7,
P = 0.0033, Fig. 2). Two months later (recovery phase), the
proportion of senescent plants was still affected by former drought
but inversely according to the management (F1,9(interaction) = 12.7,
P = 0.0060, Fig. 2): grazed-drought plots had lower proportions of
senescent plants than grazed-control plots whereas mowndrought plots had higher proportions than mown-control plots.
No differences in senescent plant proportions were detected
thereafter in spring 2013 of the following year (P > 0.05).
Drought and management did not inuence the species
diversity. At the end of drought period, there was no effect
(P > 0.05) on species richness per plot (overall mean: 16.6  3.2,
range: 1123 species), on Shannon diversity index (overall mean:
2.29  0.14, range: 1.962.52), or on evenness index (overall mean:
0.82  0.03, range: 0.770.90). No effects were detected during the
recovery or in spring of the following year.
3.2. Community weighted mean traits
Community weighted mean (CWM) traits were strongly
affected by management and drought (Fig. 3, see supplementary
material for ANOVAs results). In summer, in absence of drought
(control plots), VH was higher in mown than grazed plots (+30%),
SLA lower (13%), LDMC higher (+12%) and LNC lower (31%).
Compared to control conditions, the drought signicantly reduced
VH by 32 to 52% in the mowing and grazing management
respectively (P < 0.001), it reduced also SLA (17 to 26% respectively,
P < 0.001) and LNC (18 to 12% respectively, P < 0.001), while LDMC
was increased by 23 to 33% respectively (P < 0.001). Grazing
management signicantly amplied drought effects on SLA
(P < 0.05). Two months later (recovery phase), VH was still lower
in drought than in control plots but to a lesser extent in the grazing
than mowing management (by 23 and 31% respectively, P < 0.05),
and LNC was higher in grazed than mown plots regardless of
drought treatment (P < 0.001). In spring of the following year,
almost no effects of drought were found, except for LDMC in the
mowing management but differences were small (Fig. 3).

Table 1
Results of redundancy analysis (RDA) testing the effects of drought, management and their interaction on species composition (relative abundances) at the four sampling
periods.
Species composition

Initial state

Drought

spring

Block
Drought
Management
Drought  Man.
Error

Recovery

summer

Spring of the following year

autumn

Df

SS

SS

SS

SS

3
1
1
1
9

0.401
0.022
0.040
0.064
0.472

2.55
0.43
0.76
1.23

0.003
0.889
0.590
0.295

0.246
0.180
0.057
0.105
0.411

1.80
3.94
1.25
2.31

0.039
0.002
0.286
0.038

0.197
0.150
0.174
0.060
0.419

1.41
3.22
3.75
1.30

0.157
0.015
0.007
0.247

0.216
0.123
0.078
0.046
0.536

1.21
2.07
1.32
0.77

0.232
0.031
0.234
0.640

P-values are derived from randomization tests. Signicant tests results are in bold face

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

MC
MD
GC
GD

Agrostis capillaris

Poa trivialis

40

99

30
20
10
0

Lolium perenne

Trifolium repens

40

Abundance (%)

30

20
10
0

Dactylis glomerata

40

Festuca pratensis

30
20
10
0

Poa pratensis

40

Senescent
plants

30

20
10
0

Initial
state

Drought

Reco
very

Follow
ing year

Initial
state

Drought

Reco
very

Follow
ing year

Fig. 2. Relative abundance (mean  SE, n = 4) of the 7 main plant species and of senescent plants in each treatment (MC: mowing-control; MD: mowing-drought; G-C:
grazingcontrol; GD: grazing drought) at the four sampling periods. Within each period, signicant differences between MC and MD plots are indicated by black stars and
signicant differences between GC and GD plots by grey stars (P < 0.05, Tukeys post-hoc tests).

total (i.e. annual) DM yield (Table 2, Fig. 4). In absence of drought

(control plots), the annual DM yield was 15% lower in the grazing
than in the mowing management (Fig. 4b). The decrease in
annual DM yield due to the drought reached 47% under the

3.3. DM yields
Drought and management signicantly affected DM yields
during the drought and recovery periods, and subsequently the

Initial state

Drought

Recovery

Spring

Summer

Autumn

(cm)

VH

30

a b

b c

a b

c b

a b

c b

ab a

ab b

a b

c b

a b

c d

a a

b b

C D
M

C D
G

C D
M

C D
G

Spring of the
following year

20
10

SLA

(mm mg1)

ab a

ab b

a b

ab a

C D
M

C D
G

30
20
10

LDMC

(mg g1)

0
300
200
100

(%)

LNC

0
4
3
2
1
0

C D
M

C D
G

Fig. 3. Community weighted mean traits in control (C) and drought (D) conditions in the mowing (M) and grazing (G) managements at the four sampling periods. VH:
vegetative height; SLA: specic leaf area; LDMC: leaf dry matter content; LNC: leaf nitrogen content. Bars are means  SE (n = 4). Within each period, signicant differences
between treatments are indicated by different letters above bars (P < 0.05, Tukeys post-hoc tests).

100

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

Table 2
Summary of ANOVAs testing the effects of drought, management and their interaction on dry matter (DM) yields.
DM yields

Initial state

Drought

M1/G1 + G2

Block
Drought
Management
Drought  Man.
Residuals

Recovery

M2/G3 + G4

Annual 2012

M3/G5 + G6

Spring of the following year

M1M3/G1G6

Df

3
1
1
1
9

1.07
0.40
33.47
1.03

0.411
0.542
<0.001
0.337

2.90
208.82
2.35
13.16

0.094
<0.001
0.160
0.006

0.55
50.86
2.98
4.24

0.661
<0.001
0.119
0.070

2.20
59.27
26.91
3.44

0.158
<0.001
<0.001
0.097

2.81
0.23
1.00
0.23

0.100
0.643
0.344
0.643

Signicant test results are in bold face. M1M3 and G1G6 indicate mowing and grazing utilization numbers

3.4. Nutritive value of the forage

grazing management compared to 25% in the mowing one

(P < 0.001, Table 2, Fig. 4b). The annual DM yields produced under
the drought was about 6 t ha1 in the mowing management,
which falls into the range of natural variations of the system at the
studied site (mean 6 t ha1, range 57 t ha1, unpublished data
from a long-term monitoring program started in 1987). In
contrast, annual DM yields were about 3.5 t ha1 in the grazing
management in response to drought, which represents a major
deviation from the estimated range of natural variations. The
biomass produced during the drought (i.e. M2 and G3 + G4) was
decreased under drought by 73% in the grazing compared to 47%
in the mowing management (P < 0.01, Table 2, Fig. 4a). The
biomass produced during the recovery period (i.e. M3 and
G5 + G6) was still negatively affected by the drought but this
time to a lesser extent in the grazing than in the mowing
management (respectively lowered by 61% and 76% compared to
the control, 0.05 < P < 0.1, Table 2, Fig. 4a). At the end of the season
2012, there was no remaining effect of drought in the grazing
management (G6, P > 0.05) in contrast to the mowing management (M3, P < 0.05). No lasting effects of drought and management on DM yields were observed in spring of the following year
(Fig. 4a, Table 2).

(t ha1)

DM Yields

a
5
4
3
2
1
0

For the different sampling periods, the forage nutritive value

was higher in the grazing than in the mowing management, with
crude protein content (CP) being higher and ber contents (ADF,
NDF) often lower, resulting in overall higher net energy for
lactation (NEL) (P < 0.01, Fig. 5, see supplementary materials for
ANOVAs results). During the drought, CP was signicantly reduced
in both managements (P < 0.001) whereas ber contents were not
signicantly affected (Fig. 5). During the recovery period, a positive
effect of drought (P < 0.001) was observed on CP content and net
energy for lactation (NEL) as well as a negative effect (P < 0.001) on
ADF content. In spring of the following year, no signicant effects of
treatments on the nutritive value of the forage were found after
post-hoc tests (P > 0.05, Fig. 5).
4. Discussion
4.1. Small variations in botanical composition
Drought marginally affected the species composition of the
grassland, accounting for only 18% of the total variability in species
composition across plots at the end of the drought (e.g. negative

Initial state

Drought

Recovery

M1 / G1+G2

M2 / G3+G4

M3 / G5+G6

a a

b b

a b

a c

a b

c b

C D
M

C D
G

C D
M

C D
G

C D
M

C D
G

Spring of the
following year

C D
M

C D
G

Annual

(t ha1)

DM Yields

M1M3 / G1G6
10
8
6
4
2
0

a b

ab c

C D
M

C D
G

Fig. 4. Dry matter (DM) yields in control (C) and drought (D) conditions in the mowing (M) and grazing (G) managements at four periods (a) and for cumulated (annual)
2012 production (b). Bars are means  SE (n =4 ). M1M3 and G1G6 indicate mowing and grazing utilization numbers. The horizontal line in G bars indicates the yield
distribution between the two grazing utilizations occurring during each period. Within each period, signicant differences between treatments are indicated by different
letters above bars (P < 0.05, Tukeys post-hoc tests).

CP
NDF
ADF

(g kg1 DM)

NEL

(MJ kg1 DM)

(g kg1 DM)

(g kg1 DM)

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

300
250
200
150
100
50
0
500
400
300
200
100
0
300
250
200
150
100
50
0

Initial state

Drought

Recovery

M1 / G1G2

M2 / G3G4

M3 / G5G6

a b

a b

c b

a bc

c ab

bc c

a b

b c

a a

b b

a b

b b

C D
M

C D
G

C D
M

C D
G

a a

b b

a a

b b

a b

c c

a ab

a a

b b

C D
M

C D
G

c d

101

Spring of the
following year

6
4
2
0

C D
M

C D
G

Fig. 5. Components of the nutritive value of the forage in control (C) and drought (D) conditions in the mowing (M) and grazing (G) managements at four periods. CP: crude
proteins; NDF: neutral detergent bers; ADF: acid detergent bers; NEL: net energy for lactation. Bars are means  SE (n = 4). Within each period, signicant differences
between treatments are indicated by different letters above bars (P < 0.05, Tukeys post-hoc tests). M1M3 and G1 G6 indicate mowing and grazing utilization numbers.

impact on Lolium perenne under the mowing management and on

Trifolium repens under the grazing management). These effects had
then weakened two months later (explaining 15% of the total
variability) and, in spring of the following year, the only remaining
effect of drought was due to the legume Trifolium repens whose
abundance remained lower in plots that had been subjected to
drought compared to control plots, in both managements. The
lasting reduction in abundance of T. repens conrm previous
observations in which T. repens was reported to be droughtsensitive (e.g. Skinner et al., 2004; Signarbieux and Feller, 2012).
However, the drought event did not modify species diversity and
evenness, which proves that although a slight change in some
species relative abundance occurred, no modication of species
assemblage took place. This result is obviously related to the shortterm period of the stress, not likely to cause processes involved in
species extinction and colonisation (e.g. Jung et al., 2014), as could
happen in response to long-term and/or repeated stress (Gilgen
et al., 2010; Stampi and Zeiter, 2004).
Besides, under both managements, 2438% of plants had
senesced at the end of the drought, which is a common signature of
plant response to drought stress (e.g. De Boeck et al., 2010; Gilgen
and Buchmann, 2009; Zwicke et al., 2013). Since relatively low
changes in species abundances were observed in this mountain
grassland along the experiment, most of senesced individual
probably survived the drought through tolerance of meristematic
tissues to dehydration (De Boeck et al., 2010; Volaire et al., 2014).
4.2. Reversible changes of community functional traits at seasonal
scale
We observed strong modications of plant functional traits in
non-senescent individuals, i.e. individuals resisting drought
through tolerance of aerial tissues to dehydration (Volaire et al.,
2014). At the community scale, lower plant height in drought

compared to control plots (32 to 52%, depending on management) and higher LDMC (+23 to +33%), revealed a clear decrease in
plant growth and an increased leaf toughness, linked with a slow
turnover of plant parts (Jung et al., 2014; Westoby et al., 2002;
Wright and Westoby, 2002). These changes were accompanied by
lower SLA (17 to 26%), which can indicate a reduction of
transpiring leaf surface, a mechanism to reduce water requirements observed in species from dry habitats (Wright and Westoby,
2002; Wright et al., 2004). The decrease in LNC (8 to 12%) may
also reect a reduced N uptake during the water stress with, as a
consequence, a reduced photosynthetic capacity of leaves (Gaborcik, 2003; Wright et al., 2004). These changes were affected by
management, towards a signature of more intense slowdown in
the community metabolism in the grazing management as
compared to the mowing management.
These strong changes were followed two months later by a
relatively high recovery, excepted for plant height which remained
lower in drought plots than in control plots (by 23 to 31%).
However, these discrepancies disappeared in spring of the
following year. As recently analysed in a subalpine plant
community (Jung et al., 2014), these reversible variations highlight
the remarkable morphological and physiological plasticity of
species facing a severe drought stress at the seasonal scale, even
with the combination of several defoliations (e.g. grazing) as it
would occur in intensively managed grasslands. These ndings
support our hypothesis that in response to short-term environmental stress, changes are more likely to be dominated by
morphological and physiological plasticity of species than by
changes in species composition (Jung et al., 2014; Lloret et al.,
2012; Suding et al., 2008). These results also supports the view that
selective pressure on trait values (leading to trait xation and
therefore to changes in community functions) may be low in
response to a severe but single drought event, at least for the
considered vegetative traits (Gutschick and BassiriRad, 2003).

102

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

4.3. The decline and recovery in forage production depend on

grassland management
Controversial effects of imposed climatic water decit have
been found in previous experiments, with negative to no
signicant impacts on above-ground productivity (e.g. Gilgen
and Buchmann, 2009; Zwicke et al., 2013). These contrasting
results may have different origins such as the actual intensity of the
drought treatment (Vicca et al., 2012). In our experiment, as
expected in response to a severe drought intensity (i.e. water
decit assessed with PrecipitationsPET not reached in the past
30 years), biomass production was signicantly lowered, but the
magnitude of the response strongly depended on the management. At the annual scale, DM yields were reduced by a quarter in
the mown system, and by half in the rotational grazing system. In
the rst system, reduced annual DM yield under drought was
nevertheless still in the range of estimated natural variation of the
grassland productivity (range: 57 t ha1, mean: 6 t ha1), whereas,
in the second, a clear deviation from the natural range was
observed. The major yield loss occurred during the drought period
(minus 1.6 to 2.7 t ha1 in the mowing and grazing management,
respectively) because grassland production was at its peak at this
period. This can have critical consequences for famers who have to
conciliate grazing by animals and the establishment of forage
stocks at this period of the year in mountainous areas (Mosimann
et al., 2012). Biomass production was still drastically reduced
during the following months after the end of the drought, but at
this later stage, a tendency to less negative impacts in the grazing
was observed compared to the mowing management (reduction by
61% and 76%, respectively). However, this represented a weak yield
loss (minus 0.4 to 0.7 t ha1) because the ambient production is
naturally low in autumn.
The lower grassland resistance under the grazing management
might be explained by the more frequent defoliations endured by
the vegetation, in particular by the defoliation that occurred during
the drought period (G3). Indeed, combination of repeated
defoliations and water stress can severely compromise plant
regrowth and persistence, possibly due to the depletion of plant
carbohydrate reserves, which are essential for plants to support
regrowth after perturbation (Busso and Richards, 1995; Fulkerson
and Donaghy, 2001; Volaire, 1994). In contrast, in the mowing
management, when water stress began, i.e. several weeks after
rain-shelters were installed, the vegetation was already well
advanced in height and, moreover, did not have to support a
regrowth during the water stress. This conrms the higher drought
resistance of undefoliated vegetation observed in an articial
grassland community (Vogel et al., 2012).
Although the absence of vegetation regrowth during the rst
month following the end of the drought in the grazing management did not allow the fth utilization by sheep (G5), the regrowth
of vegetation was fairly vigorous in this treatment two months
later. Indeed, at the nal utilization, dry matter production in
drought plots was restored to the same level as in control plots
under the grazing management (G6), but not under the mowing
one (M3). It can be hypothesized that recovery in biomass
production was stimulated through better availability of nutrients
for plants in grazed-drought plots, compared to mown-drought
plots, as suggested by a higher community LNC and SLA in these
plots. A drought period can indeed temporarily stop the mineralization processes, leading to accumulation of undecomposed
organic matter that will constitute an important source of
nutrients for plants as soon as the soil rewets (Evans and Burke,
2013; Giese et al., 2011; van Meeteren et al., 2008). In the grazing
management, the presence of animal dejections may have
provided a higher source of nutrients and stimulated recovery
in biomass production to a larger extent than in the mowing

management. The lower proportion of senescent plants in grazeddrought plots compared to mown-drought plots in autumn (1%
against 7%) attests to higher decomposition rates in the grazing
treatment, once water limitation has disappeared (De Boeck et al.,
2010).
4.4. Drought has a smaller effect on the nutritive value of the forage, as
compared to management
The general slowdown in plant metabolism during drought
affected the nutritive value of the forage. In particular, the crude
protein (CP) content of the forage decreased during drought,
similarly to the community LNC. This may be linked to a decrease
in N acquisition by plants during the drought, as a result of a
decrease in both the diffusion rate of nutrients in the soil and
transpiration rates of plants (Evans and Burke, 2013). However,
under moderate drought stress, higher CP content in the forage
was often reported because N becomes more concentrated in
plants whose biomass production is reduced under water stress
(Grant et al., 2014; Jensen et al., 2010). But, under severe drought
stress, leaf loss through senescence (Buxton, 1996) and a reduction
in N xation performance (Kchenmeister et al., 2013) are likely to
cause a decrease in CP content in the forage. Fiber concentration in
the forage (i.e. ADF and NDF contents) was not signicantly
affected by drought, which contradicts other works that report a
reduced ber content in plants under water stress, due to delayed
maturity (Bittman et al., 1988; Kchenmeister et al., 2013). Fiber
content in plants increases with the maturity stage because
maturity increases both leaf ber content and stem to leaf ratio,
with stem containing higher ber content than leaves (Bruinenberg et al., 2002; Halim et al., 1989). In the present case, leaf ber
content may have increased under drought conditions as revealed
by an increased community-LDMC (Pontes et al., 2007). Nevertheless, this effect may have been counterbalanced by a decrease of
the stem to leaf ratio (suggested by the lowered community plant
height in the drought situation), which would then explain that
there was no overall drought effect on ber content of the forage.
Overall, we observed that management had a predominant
effect on the nutritive value of the forage compared to drought.
Indeed, the nutritive value of the forage remained higher in the
grazing management compared to the mowing management, even
under drought conditions (higher CP content and lower ADF
content providing higher energy content, NEL). This reects the
effect of frequent defoliations occurring under the grazing
management, that maintained plants in an earlier developmental
stage compared to the mowing management (Bruinenberg et al.,
2002). However, although repeated grazing can reduce the
negative effects of drought on forage nutritive value compared
to a mowing management, this cannot counteract the considerable
negative impact on DM yields.
High recovery in the nutritive value of the forage was observed
in both managements, with higher CP content and lower ber
concentration in drought than in control plots in autumn. It is
likely that a higher nutrient availability in drought plots compared
to controls favored such a response (Evans and Burke, 2013).
Finally, complete resilience was observed for the different
components of the nutritive value in spring of the following year.
5. Conclusions
In livestock agriculture, productive mountain pastures are
rarely managed through mowing alone and it is not realistic to
study their resistance to drought events with forage accumulating
during the summer months. In this study, we show that when
animals are regularly grazing on pastures, the decline in forage
production due to a severe drought can be much larger than the

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

one measured under extensive mowing. We also observed an

extended period without any vegetation growth, which attests that
in dry years, this type of productive mountain grassland cannot
sustain the intensity and frequency of grazing usually applied. In
regions where livestock grazes on pastures, severe drought events
will thus raise serious management concerns to match animal
requirements all along the growing season, and it will become
crucial to consider a range of adaptation strategies at both farm and
policy levels (Hopkins and Del Prado, 2007; Huber et al., 2013).
Besides, our study demonstrates that, even under fairly intensive
management, grasslands dominated by nutrient-rich grass species
can be expected to be resilient to severe but single drought events
in the Jura mountains, where such events are predicted to have a
10 year return frequency at the end of the century. Finally, this also
indicates that longer-term manipulation experiments are needed
to identify thresholds of stress intensity and/or recurrence in
combination with real management practices, above which the
resilience capacity of such temperate grasslands will be exceeded.
This might then lead to persistent shifts in species and trait
compositions, with most likely long-term consequences on forage
supply functions.
Acknowledgments
The authors wish to thank everyone who helped in the
establishment, management and maintenance of the experimental
site, and especially L. Stvenin and D. Frund. Grateful thanks also go
to W. Herren for the management of the sheep herd, as well as all
those who helped in the eld work and data collection. Research
funding was provided by Agroscope, the Swiss Grassland Federation (ADCF), WSL Research Institute, Ecole Polytechnique Fdrale
Lausanne(EPFL) and the Swiss National Science Foundation (SNF).
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.2015.07.020.
References
Adler, P.B., Raff, D.A., Lauenroth, W.K., 2001. The effect of grazing on the spatial
heterogeneity of vegetation. Oecologia 128, 465479.
Agroscope, 2013. Apports alimentaires recommands et tables de la valeur nutritive
pour les ruminants (Livre vert). .
Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop Evapotranspiration.
Guidelines for Computing Crop Water Requirements. FAO Irrigation and
Drainage Paper 56. Food and Agriculture Organization (FAO) of the United
Nations, Rome 300 pp.
Barth, L., Lanz, S., Hofer, C., 2011. Promotion de la production animale base sur les
herbages dans la politique agricole 20142017. Rech. Agron. Suisse 2, 2025.
Bittman, S., Simpson, G.M., Mir, Z., 1988. Leaf senescence and seasonal decline in
nutritional quality of three temperate forage grasses as inuenced by drought.
Crop Sci. 28, 546552.
Bruinenberg, M.H., Valk, H., Korevaar, H., Struik, P.C., 2002. Factors affecting
digestibility of temperate forages from seminatural grasslands: a review. Grass
Forage Sci. 57, 292301.
Busso, C.A., Richards, J.H., 1995. Drought and clipping effects on tiller demography
and growth of two tussock grasses in Utah. J. Arid Environ. 29, 239251.
Buttler, A., 1992. Permanent plot research in wet meadows and cutting experiment.
Vegetation 103, 113124.
Buxton, D.R., 1996. Quality-related characteristics of forages as inuenced by plant
environment and agronomic factors. Anim. Feed Sci. Technol. 59, 3749.
CH2011, 2011. Swiss Climate Change Scenarios CH2011. C2SM, MeteoSwiss, ETH,
NCCR Climate and OcCC, Zurich, Switzerland, 88pp, ISBN: 978-3-033-03065-7.
Calanca, P., 2007. Climate change and drought occurrence in the Alpine region: how
severe are becoming the extremes. Global Planet. Change 57, 151160.
Cornelissen, J.H.C., Lavorel, S., Garnier, E., Daz, S., Buchmann, N., Gurvich, D.E., Reich,
P.B., ter Steege, H., Morgan, H.D., van der Heijden, M.G.A., Pausas, J.G., Poorter, H.,
2003. A handbook of protocols for standardised and easy measurement of plant
functional traits worldwide. Aust. J. Bot. 51, 335380.
Daz, S., Noy-Meir, I., Cabido, M., 2001. Can grazing response of herbaceous plants be
predicted from simple vegetative traits. J. Appl. Ecol. 38, 497508.

103

De Boeck, H.J., Dreesen, F.E., Janssens, I.A., Nijs, I., 2010. Whole-system responses of
experimental plant communities to climate extremes imposed in different
seasons. New Phytol. 189, 806817.
EEA (European Environment Agency), 2012. Climate change, impacts and
vulnerability in EuropeAn indicator-based report. ISSN 1725-9177
(doi:9210.2800/66071). Copenhagen, Denmark.
Easterling, D.R., Meehl, G.A., Parmesan, C., Changnon, S.A., Karl, T.R., Mearns, L.O.,
2000. Climate extremes: observations, modeling, and impacts. Science 289,
20682074.
Evans, S.E., Burke, I.C., 2013. Carbon and nitrogen decoupling under an 11-year
drought in the shortgrass steppe. Ecosystems 16, 2033.
Fischer, A., Weigel, A.P., Buser, C.M., Knutti, R., Knsch, H.R., Liniger, M.A., Schr, C.,
Appenzeller, C., 2012. Climate change projections for Switzerland based on a
Bayesian multi-model approach. Int. J. Climatol. 32, 23482371.
Frank, D.A., 2007. Drought effects on aboveand belowground production of a
grazed temperate grassland ecosystem. Oecologia 152, 131139.
Frei, C., Scho, R., Fukutome, S., Schmidli, J., Vidale, P.L., 2006. Future change of
precipitation extremes in Europe: intercomparison of scenarios from regional
climate models. J. Geophys. Res. 111.
Fulkerson, W.J., Donaghy, D.J., 2001. Plant-soluble carbohydrate reserves and
senescencekey criteria for developing an effective grazing management
system for ryegrass-based pastures: a review. Aust. J. Exp. Agric. 41, 261275.
Gaborcik, N., 2003. Relationship between contents of chlorophyll (a+b) (SPAD
values) and nitrogen of some temperate grasses. Photosynthetica 41, 285287.
Gardarin, A., Garnier, E., Carrre, P., Cruz, P., Andueza, D., Bonis, A., Colace, M.P.,
Dumont, B., Duru, M., Farruggia, A., Gaucherand, S., Grigulis, K., Kerns, E.,
Lavorel, S., Louault, F., Loucougaray, G., Meslard, F., Yavercovski, N., Kazakou, E.,
2014. Plant trait-digestibility relationships across management and climate
gradients in permanent grasslands. J. Appl. Ecol. 51, 12071217.
Garnier, E., Shipley, B., Roumet, C., Laurent, G., 2001. A standardized protocol for the
determination of specic leaf area and leaf dry matter content. Funct. Ecol. 15,
688695.
Garnier, E., Cortez, J., Bills, G., Navas, M.-L., Roumet, C., Debussche, M., Laurent, G.,
Blanchard, A., Aubry, D., Bellmann, A., Neill, C., Toussaint, J.-P., 2004. Plant
functional markers capture ecosystem properties during secondary succession.
Ecology 85, 26302637.
Gavazov, K., Peringer, A., Gillet, F., Buttler, A., Spiegelberger, T., 2013. Dynamics of
forage production in pasture-woodlands of the Swiss Jura Mountains under
projected climate change scenarios. Ecol. Soc. 18, 38.
Giese, M., Gao, Y., Lin, S., Brueck, H., 2011. Nitrogen availability in a grazed semi-arid
grassland is dominated by seasonal rainfall. Plant Soil 340, 157167.
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, 25252539.
Gilgen, A.K., Signarbieux, C., Feller, U., Buchmann, N., 2010. Competitive advantage
of Rumex obtusifolius L. might increase in intensively managed temperate
grasslands under drier climate. Agric. Ecosyst. Environ. 135, 1523.
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, 1122.
Gutschick, V.P., BassiriRad, H., 2003. Extreme events as shaping physiology, ecology,
and evolution of plants: toward a unied denition and evaluation of their
consequences. New Phytol. 160, 2142.
Halim, R.A., Buxton, D.R., Hattendorf, M.J., Carlson, R.E., 1989. Water-stress effects on
Alfalfa Forage quality after adjustment for maturity differences. Agron. J. 81,
189194.
Hopkins, A., Del Prado, A., 2007. Implications of climate change for grassland in
Europe: impacts, adaptations and mitigation options: a review. Grass Forage Sci.
62, 118126.
Huber, R., Briner, S., Peringer, A., Lauber, S., Seidl, R., Widmer, A., Gillet, F., Buttler, A.,
Le, Q., Hirschi, C., 2013. Modeling social-ecological feedback effects in the
implementation of payments for environmental services in PastureWoodlands. Ecol. Soc. 18 41.
IPCC, 2007. Climate Change 2007: The Physical Science Basis. Contribution of
Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA.
IUSS Working Group WRB, 2006. World reference base for soil resources 2006.
World Soil Resources Reports No. 103. FAO, Rome.
Jensen, K.B., Waldron, B.L., Peel, M.D., Robins, J.G., 2010. Nutritive value of herbage of
ve semi-irrigated pasture species across an irrigation gradient. Grass Forage
Sci. 65, 92101.
Jentsch, A., Beierkuhnlein, C., 2008. Research frontiers in climate change: effects of
extreme meteorological events on ecosystems. C.R. Geosci. 340, 621628.
Jung, V., Albert, C.H., Violle, C., Kunstler, G., Loucougaray, G., Spiegelberger, T., 2014.
Intraspecic trait variability mediates the response of subalpine grassland
communities to extreme drought events. J. Ecol. 102, 4553.
Kchenmeister, K., Kchenmeister, F., Kayser, M., Wrage-Mnnig, N., Isselstein, J.,
2013. Inuence of drought stress on nutritive value of perennial forage legumes.
Int. J. Plant Prod. 7, 693710.
Kahmen, A., Perner, J., Buchmann, N., 2005. Diversity-dependent productivity in
semi-natural grasslands following climate perturbations. Funct. Ecol. 19, 594
601.

104

C. Delglise et al. / Agriculture, Ecosystems and Environment 213 (2015) 94104

Kohler, F., Hamelin, J., Gillet, F., Gobat, J.M., Buttler, A., 2005. Soil microbial
community changes in mountain wooded pastures due to simulated effects of
cattle grazing. Plant Soil 278, 327340.
Legendre, P., Legendre, L., 1998. Numerical Ecology. Elsevier Science BV, Amsterdam,
NL.
Lep, J., milauer, P., 2003. Multivariate Analysis of Ecological Data Using CANOCO.
Cambridge University Press, Cambridge.
Lep, J., de Bello, F., milauer, P., Doleal, J., 2011. Community trait response to
environment: disentangling species turnover vs intraspecic trait variability
effects. Ecography 34, 856863.
Levy, E.B., Madden, E.A., 1933. The point method of pasture analysis. N. Z. J. Agric. 46,
267279.
Lloret, F., Escudero, A., Iriondo, J.M., Martinez-Vilalta, J., Valladares, F., 2012. Extreme
climatic events and vegetation: the role of stabilizing processes. Global Change
Biol. 18, 797805.
Mariotte, P., Vandenberghe, C., Kardol, P., Hagedorn, F., Buttler, A., 2013. Subordinate
plant species enhance community resistance against drought in semi-natural
grasslands. J. Ecol. 101, 763773.
Mosimann, E., Meisser, M., Delglise, C., Jeangros, B., 2012. Potentiel fourrager des
pturages du Jura. Rech. Agron. Suisse 3, 516523.
Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., OHara, R.B.,
Simpson, G.L., Solymos, P., Stevens, M.H.H., Wagner, H., 2013. Vegan:
Community Ecology Package. R package version 2.0-7, .
Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., R Development Core Team, 2013. nlme:
Linear and Nonlinear Mixed Effects Models. R package version 3. 1-109.
Pontes, L.d.S., Soussana, J., Louault, F., Andueza, D., Carrre, P., 2007. Leaf traits affect
the above-ground productivity and quality of pasture grasses. Funct. Ecol. 21,
844853.
Qutier, F., Thbault, A., Lavorel, S., 2007. Linking vegetation and ecosystem
response to complex past and present land use changes using plant traits and a
multiple stable state framework. Ecol. Monogr. 77, 3352.
R Core Team, 2013. R: A Language and Environment for Statistical ComputingR
Foundation for Statistical Computing, Vienna, Austria. . ISBN 3-900051-07-0
http://www.R-project.org.
Saxton, K.E., Rawls, W.J., Romberger, J.S., Papendick, R.I., 1986. Estimating
generalized soil-water characteristics from texture. Soil Sci. Soc. Am. J. 50,
10311036.
Signarbieux, C., Feller, U., 2012. Effects of an extended drought period on
physiological properties of grassland species in the eld. J. Plant Res. 125,
251261.
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,
13611369.
Stampi, A., Zeiter, M., 2004. Plant regeneration directs changes in grassland
composition after extreme drought: a 13-year study in southern Switzerland. J.
Ecol. 92, 568576.
Suding, K.N., Lavorel, S., Chapin, F.S., Cornelissen, J.H.C., Daz, S., Garnier, E.,
Goldberg, D., Hooper, D.U., Jackson, S.T., Navas, M.L., 2008. Scaling

environmental change through the community level: a trait-based responseand-effect framework for plants. Global Change Biol. 14, 11251140.
ter Braak, C.F., milauer, P., 1999. Canoco for Windows. Wageningen, The
Netherlands.
van Meeteren, M.J.M., Tietema, A., van Loon, E.E., Verstraten, J.M., 2008. Microbial
dynamics and litter decomposition under a changed climate in a Dutch
heathland. Appl. Soil Ecol. 38, 119127.
van Wieren, S.E., Bakker, J.P., 2008. The impact of browsing and grazing herbivores
on biodiversity. In: Gordon, I.J., Prins, H.H.T. (Eds.), The Ecology of Browsing and
Grazing. Springer, Berlin, pp. 263292.
Vicca, S., Gilgen, A.K., Serrano, M.C., Dreesen, F.E., Dukes, J.S., Estiarte, M., Gray, S.B.,
Guidolott, G., Hoeppner, S.S., Leakey, A.D.B., Ogaya, R., Ort, D.R., Ostrogovic, M.Z.,
Rambal, S., Sardans, J., Schmitt, M., Siebers, M., van der Linden, L., van Straaten,
O., Granier, A., 2012. Urgent need for a common metric to make precipitation
manipulation experiments comparable. New Phytol. 195, 518522.
Violle, C., Navas, M.-L., Vile, D., Kazakou, E., Fortunel, C., Hummel, I., Garnier, E., 2007.
Let the concept of trait be functional!. Oikos 116, 882892.
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 36910.31371/journal.pone.0036992.
Volaire, F., Barkaoui, K., Norton, M., 2014. Designing resilient and sustainable
grasslands for a drier future: adaptive strategies, functional traits and biotic
interactions. Eur. J. Agron. 52, 8189.
Volaire, F., 1994. Effects of summer drought and spring defoliation on carbohydrate
reserves, persistence and recovery of two populations of cocksfoot (Dactylis
glomerata L.) in a Mediterranean environment. J. Agric. Sci. 122, 207215.
Wang, Y., Yu, S., Wang, J., 2007. Biomass-dependent susceptibility to drought in
experimental grassland communities. Ecol. Lett. 10, 401410.
Weiher, E., Van Der Werf, A., Thompson, K., Roderick, M., Garnier, E., Eriksson, O.,
1999. Challenging theophrastus: a common core list of plant traits for functional
ecology. J. Veg. Sci. 10, 609620.
Westoby, M., Falster, D.S., Moles, A.T., Vesk, P.A., Wright, I.J., 2002. Plant ecological
strategies: some leading dimensions of variation between species. Annu. Rev.
Ecol. Syst. 33, 125159.
Wright, I.J., Westoby, M., 2002. Leaves at low versus high rainfall: coordination of
structure, lifespan and physiology. New Phytol. 155, 403416.
Wright, I.J., Reich, P.B., Westoby, M., Ackerly, D.D., Baruch, Z., Bongers, F., CavenderBares, J., Chapin, T., Cornelissen, J.H.C., Diemer, M., Flexas, J., Garnier, E., Groom,
P.K., Gulias, J., Hikosaka, K., Lamont, B.B., Lee, T., Lee, W., Lusk, C., Midgley, J.J.,
Navas, M.L., Niinemets, U., Oleksyn, J., Osada, N., Poorter, H., Poot, P., Prior, L.,
Pyankov, V.I., Roumet, C., Thomas, S.C., Tjoelker, M.G., Veneklaas, E.J., Villar, R.,
2004. The worldwide leaf economics spectrum. Nature 428, 821827.
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. Global Change Biol. 19, 34353448.