Ecology, 88(2), 2007, pp. 381–390
Ó2007 by the Ecological Society of America
EARLY ONSET OF VEGETATION GROWTH VS. RAPID GREEN-UP:
IMPACTS ON JUVENILE MOUNTAIN UNGULATES
ACHAZ VON HARDENBERG,
AND STEEVE D. CO
´partement de Biologie and Centre d’e
´tudes nordiques, Universite
´bec G1K 7P4 Canada
´partement de Biologie, Universite
´de Sherbrooke, 2500 Boul. de l’Universite
´, Sherbrooke, QC J1K 2R1 Canada
Division of Biology, Faculty of Life Sciences, Imperial College London, Silwood Park, Ascot, Berkshire SL5 7PY UK
Alpine Wildlife Research Centre, Parco Nazionale Gran Paradiso, via della Rocca 47, 10123 Torino, Italy
Abstract. Seasonal patterns of climate and vegetation growth are expected to be altered
by global warming. In alpine environments, the reproduction of birds and mammals is tightly
linked to seasonality; therefore such alterations may have strong repercussions on recruitment.
We used the normalized difference vegetation index (NDVI), a satellite-based measurement
that correlates strongly with aboveground net primary productivity, to explore how annual
variations in the timing of vegetation onset and in the rate of change in primary production
during green-up affected juvenile growth and survival of bighorn sheep (Ovis canadensis),
Alpine ibex (Capra ibex), and mountain goats (Oreamnos americanus) in four different
populations in two continents. We indexed timing of onset of vegetation growth by the
integrated NDVI (INDVI) in May. The rate of change in primary production during green-up
(early May to early July) was estimated as (1) the maximal slope between any two successive
bimonthly NDVI values during this period and (2) the slope in NDVI between early May and
early July. The maximal slope in NDVI was negatively correlated with lamb growth and
survival in both populations of bighorn sheep, growth of mountain goat kids, and survival of
Alpine ibex kids, but not with survival of mountain goat kids. There was no effect of INDVI
in May and of the slope in NDVI between early May and early July on juvenile growth and
survival for any species. Although rapid changes in NDVI during the green-up period could
translate into higher plant productivity, they may also lead to a shorter period of availability
of high-quality forage over a large spatial scale, decreasing the opportunity for mountain
ungulates to exploit high-quality forage. Our results suggest that attempts to forecast how
warmer winters and springs will affect animal population dynamics and life histories in alpine
environments should consider factors inﬂuencing the rate of changes in primary production
during green-up and the timing of vegetation onset.
Key words: body mass; green-up; NDVI; plant phenology; population dynamics; resource–animal
Predicting the effects of global warming on organisms
of different ecosystems is a major challenge for
ecologists (Walther et al. 2002). In recent decades,
climate change has affected many biological systems
(Crick and Sparks 1999, Post and Stenseth 1999, Inouye
et al. 2000), and much effort is devoted to understand
the consequences of such changes (Hughes 2000, Hulme
2005). Global climate change is altering seasonal
patterns: for example, the average start of the growing
season shifted by eight days from 1989 to 1998 in
Europe (Chmielewski and Ro
¨tzer 2002) and by 5–6 days
from 1959 to 1993 in North America (Schwartz and
Reiter 2000). The life history strategies of species
experiencing seasonal environments have been selected
to match the best environmental conditions. With
seasonal patterns altered, however, the reproduction of
these species may become out of phase with the period
of highest environmental productivity (Thomas et al.
2001, Berteaux et al. 2004). The ultimate consequences
of the timing of birth are expected to depend to a large
extent on the phenology of organisms at other trophic
levels (Visser et al. 2004). In seasonal environments,
large herbivores typically give birth in late spring or
early summer to match the vegetation green-up period
and allow offspring to beneﬁt from the entire vegetation
growing season (Rutberg 1987). By shifting plant
phenology toward an earlier vegetation onset, global
warming could affect juvenile growth and survival of
many species (Inouye et al. 2000, Visser et al. 2004,
Pettorelli et al. 2005a, c). Highly seasonal environments
such as those in arctic or alpine areas are expected to be
strongly affected by climate change (Oechel et al. 1997).
There is much interest in the inﬂuence of global
warming in mountainous regions (Diaz and Bradley
Manuscript received 25 May 2006; revised 3 August 2006;
accepted 24 August 2006. Corresponding Editor: C. M.
Corresponding author. E-mail: firstname.lastname@example.org
1997), where warmer winters are expected to change the
rain/snow ratio. In northern mountains, climate change
may lead to more winter precipitation, resulting in
deeper snowpack at high elevations (Inouye et al. 2000,
Mysterud et al. 2001, Pettorelli et al. 2005a). Increas-
ingly warm winters, however, may augment winter rain
and run-off at the expense of snowpack, as the
rain/snow boundary moves higher in elevation (Beniston
and Fox 1996, Lapp et al. 2005).
The timing of snowmelt should determine the timing
of spring vegetation onset and thereby affect life
histories of alpine ungulates (Rutberg 1987, Kudo
1991). Because plant phenology is the major factor
affecting forage quality (Laycock and Price 1970), it is
frequently described as the driving force in habitat use
by vertebrate herbivores (Fryxell 1991, Albon and
Langvatn 1992). Both plant crude protein content and
digestibility peak early in the growing season, and then
rapidly decline as the vegetation matures: higher forage
quality is thus associated with early phenological stages
where new green leaves dominate biomass (Crawley
1983). Feeding patch choice and forage selection by
ungulates are positively associated with plant quality
(White 1983, Wilmshurst et al. 1995). A shorter period
when high-quality forage is available should thus lower
herbivore performance (Albon and Langvatn 1992,
Langvatn et al. 1996). Because forage quality peaks
during early phenological stages, slow vegetation
growth should prolong access to high-quality forage.
Moreover, spatial heterogeneity in snowmelt may lead
to spatial heterogeneity in the timing of vegetation
green-up onset, which may lengthen the period when
high-quality forage is accessible to herbivores (Mysterud
et al. 2001, Pettorelli et al. 2005a). Rapid temporal
changes in plant productivity might thus correlate both
with fast vegetation growth and reduced spatial
heterogeneity in timing of vegetation onset in alpine
areas, shortening the period of access to high-quality
forage for herbivores.
Many studies have focused on the impact of an early
start of vegetation growth on herbivore performance
(e.g., Portier et al. 1998, Coˆ te
´and Festa-Bianchet 2001a,
Grifﬁth et al. 2002). Few, however, have attempted to
partition the effect of an early start of vegetation growth
from that of a rapid rate of changes in vegetation
phenology, possibly leading to a shorter period of access
to high-quality forage. Here we assess the effects of
annual variations in the timing of vegetation green-up
onset and the rate of change in plant productivity during
green-up on the growth and survival of juvenile
mountain ungulates in North America and Europe.
We indexed vegetation dynamics by the normalized
difference vegetation index (NDVI), a satellite-based
measurement that correlates strongly with aboveground
net primary productivity (Pettorelli et al. 2005b).
Previous studies generally considered climatic variables
as proxies for plant phenology (Portier et al. 1998, Toı
et al. 1999). The links between weather and vegetation
phenology, however, are complex, involving a number
of climatic variables and depending on location, while
the link between primary productivity and NDVI is
direct and has been shown to be linear, robust,
consistent, and strong in temperate areas (Pettorelli et
al. 2005b). We considered three species of mountain
ungulates: bighorn sheep Ovis canadensis, Alpine ibex
Capra ibex, and mountain goat Oreamnos americanus in
four populations. These species have broadly similar
habitat use, social organization, foraging behavior,
sexual size dimorphism, fertility, and body mass
(Festa-Bianchet 1988a, Festa-Bianchet et al. 1997, Toı
et al. 1999, Coˆ te
´and Festa-Bianchet 2003). Because
juveniles are the age class most likely to be affected by
both extrinsic and density-dependent processes in
ungulates (Gaillard et al. 2000), we focused on the
relationships between vegetation phenology and juvenile
growth and survival.
Early vegetation onsets positively affect the perfor-
mance of ungulates inhabiting highly seasonal environ-
ments (Giacometti et al. 2002, Pettorelli et al. 2005c). We
consequently expected (H
, hypothesis 1) a positive
effect of early vegetation onsets on juvenile growth and
survival. Rapid changes in plant productivity during
green-up should be associated with a reduced period of
access to high-quality forage, either through rapid
vegetation growth or reduced spatial heterogeneity in
the timing of the vegetation onset. We therefore
expected that (H
, hypothesis 2) rapid changes in NDVI
during green-up would be negatively related to juvenile
survival and growth. Finally, we expected that late
vegetation onsets or rapid changes in plant productivity
would have stronger effects under harsh environmental
conditions such as at high population density (H
MATERIALS AND METHODS
Study areas and sample collection
Ram Mountain (528N, 1158W; elevation, 1700–2200
m), Alberta, Canada, is an isolated mountain ;30 km
east of the main Canadian Rockies. The area used by
sheep (;38 km
) is characterized by alpine and
subalpine habitat. Each year, sheep are trapped from
the end of May to late September and weighed to within
125 g with a spring scale. All ewes have been marked
since 1976, and .80%of the lambs were caught in most
years. Here we considered the body mass of 332 lambs
weighed between 1982 (when NDVI measurements ﬁrst
became available) and 2004. We also calculated the
proportion of marked lambs in September (n¼634) that
were reobserved as yearlings the following spring (end of
May) to estimate ﬁrst-year overwinter survival (see
Festa-Bianchet et al. 1997 for more details on ﬁeld
The Sheep River bighorn sheep population (508N,
1148W) is also in Alberta, 160 km south of Ram
Mountain. In winter, the population uses a low-
elevation range (1450–1700 m) in the eastern slopes of
NATHALIE PETTORELLI ET AL.382 Ecology, Vol. 88, No. 2
the Rocky Mountains. In spring, ewes migrate to higher
elevation (1800–2250 m) ;12–15 km west of the winter
range (Festa-Bianchet 1988a). The winter range (14 km
is characterized by grassy meadows interspersed with
aspen (Populus tremuloides) copses, while the summer
range (;50 km
) consists of alpine and subalpine
habitat. Since 1981, .95%of the sheep have been
marked (Festa-Bianchet 1988a, b). Each autumn, lambs
aged 4–6 months are immobilized with a dart gun and
their chest girth is measured. Chest girth is highly
correlated with lamb mass (r¼0.90, Pelletier et al. 2005).
We considered the mean chest girth adjusted for capture
date, per year and per sex, of 428 lambs born between
1982 and 2004. To estimate lamb overwinter survival, we
considered the proportion of lambs marked in Septem-
ber–November (n¼464) that were reobserved as
yearlings the following spring (end of April–May).
Caw Ridge is located in the Rocky Mountains of
west-central Alberta (548N, 1198W). Mountain goats
use ;28 km
of alpine tundra and open subalpine forests
from 1750 to 2185 m (Coˆ te
´and Festa-Bianchet 2001a).
Since 1989, the population has ﬂuctuated from 76 to 147
individuals. We considered the sex-speciﬁc average mass
of 137 kids weighed between late May and early October
in 1989–2004. No kids were weighed in 1998–2000. We
used the proportion of kids in September (n¼349) that
were reobserved as yearlings the following spring (end of
May–June) to estimate kid overwinter survival, from
1989 to 2004.
The Gran Paradiso National Park (GPNP hereafter)
in northwestern Italy (458N, 78E) is composed entirely
of mountainous terrain. Alpine pastures, moraines,
cliffs, glaciers, and rock account for 59%of its 720
. Ibex use elevations ranging from ;800 m to
beyond the upper limit of vegetation at ;3200 m. Yearly
autumn counts are conducted in the entire park by ;30
park wardens over two consecutive days in September,
when the number of kids, yearlings, and adult males and
females are determined (Jacobson et al. 2004). We used
the proportion of kids that were seen as yearlings the
following autumn to estimate ﬁrst-year overwinter
survival in 1982–2004 (number of yearlings in year
t/number of kids in year t1). Because censuses are
done in September, juvenile survival in the GPNP was
from 4 to 16 months of age. Estimates of survival based
on population counts are subject to biases, and their
quality is far lower than estimates based on marked
individuals (Gaillard et al. 2000). However, those biases
should not vary from year to year, making the
comparison of vegetation phenology and survival
estimates possible in GPNP.
Alberta.—Data on snowfall (in centimeters), precipi-
tation in water equivalent (in millimeters) and average
temperature (in degrees Celsius) in April and May were
obtained from the Environment Canada meteorological
stations near the study sites (Grande Cache [1255 m] for
Caw Ridge, Nordegg [1326 m] for Ram Mountain, High
River [1219 m] for Sheep River). Those data were
available from 1988 to 2004 for Caw Ridge, and from
1982 to 2004 for Sheep River and Ram Mountain.
GPNP.—Data on snow depth (in centimeters), rain
(in millimeters), and average temperature (in degrees
Celsius) in April and May were obtained from the Serru`
meteorological Station (Azienda Elettrica Municipal-
izzata Torino) located inside the GPNP at an elevation
of 2240 m (1982–2004). The average temperature was
deﬁned as (average maximal temperature þaverage
Data collected by the National Oceanic and Atmo-
spheric Administration satellites and processed by the
GIMMS group (Tucker et al. 2005) are available. From
these, NDVI values have been produced from visible
and near-infrared reﬂectance measurements (NDVI ¼
[NIR VIS]/[NIR þVIS], where NIR is the near
infrared light reﬂected by the vegetation, and VIS is the
visible light reﬂected by the vegetation). We used the
best available corrected NDVI time series for the
number of years considered. The spatial scale of
resolution (pixel size) for that series is 64 km
NDVI value is available on a bimonthly basis, from
July 1981 to now (Pettorelli et al. 2005b). Bimonthly
NDVI values are based on 15-d temporal composites
(maximum value compositing) to reduce cloud contam-
ination problems (Pettorelli et al. 2005b). We used
NDVI averages from one pixel for Ram Mountain,
eight pixels for Sheep River, ﬁve pixels for Caw Ridge,
and 11 pixels for the GPNP. For Ram Mountain, we
thus used the minimum possible scale, i.e., 64 km
the other study areas, we used the total spatial range of
the populations, covering areas likely much larger than
those actually used by the study populations. We used
the minimum number of pixels possible considering the
shape of the study areas and their overlap with NDVI
pixels. We thus traded spatial resolution for data
quality, assuming that the annual phenological signal
captured at the scale of the GIMMS data would be
correlated with vegetation productivity in the different
study sites. We could not, however, test this assump-
tion, because there are no data on the phenology of
forage available for our study sites.
We used NDVI measurements around the mean date
of green-up to distinguish early from late annual onsets
of vegetation growth. The timing of the onset of
vegetation growth was thus indexed using the sum of
the two bimonthly NDVI values in May (vegetation
growth typically starts in May in all study sites; see Fig.
1a), an index that correlates with vegetation biomass
(integrated NDVI [INDVI], Pettorelli et al. 2005b). The
rate of change in plant productivity during green-up,
when indexed using NDVI, is deﬁned as the rate of
increase between two ﬁxed dates (generally between the
estimated date when vegetation starts growing and the
February 2007 383PLANT PHENOLOGY AND ALPINE UNGULATES
estimated date when vegetation biomass reaches a
plateau; Pettorelli et al. 2005b). Because for all sites
vegetation growth reaches a plateau in July (Fig. 1a),
we considered the slope between early May and early
July as an index of the rate of vegetation changes
during green-up. This last index, however, does not
capture any deviation from a linear increase in NDVI
between those two dates and smooths the rate of change
during green-up. For example, a linear and a logarith-
mic increase between the two dates would provide the
same slope. We therefore also indexed the rate of
vegetation changes during green-up as the maximal
slope between any two consecutive bimonthly NDVI
values from early May to early July (Reed et al. 1994,
Kaduk and Heinmann 1996, Fig. 1b). Higher maximal
increases indicate faster changes in vegetation growth
and higher deviations from a linear increase in NDVI
FIG. 1. (a) Average normalized difference vegetation index (NDVI) values (1982–2004) in four study areas: Ram Mountain
(Alberta), Sheep River (Alberta), Caw Ridge (Alberta), and Gran Paradiso National Park (GPNP, Italy). (b) Illustration of the
concept of the maximal increase in NDVI. We present two years (1984 and 1997) in the Gran Paradiso National Park exhibiting
contrasted vegetation dynamics. The x-axis represents each NDVI picture available (two per month, for a total of 24 pictures per
year), ranging from 1 to 24 and starting on 1 January. The period considered in the analyses extended from early May to early July.
In both years (1984 and 1997), the maximal increase in NDVI occurred between 1–15 May and 15–30 May; however the rate of
increase between these two periods in 1997 was double that in 1984.
NATHALIE PETTORELLI ET AL.384 Ecology, Vol. 88, No. 2
We expected climatic conditions in April/May to
determine the average and maximal rate of changes in
primary productivity during green-up. Precipitation was
transformed as ‘‘log(precipitation þ4)’’ to stabilize the
variance (Pettorelli et al. 2005a). Because there were no a
priori reasons to believe that different relationships
between climate and phenology should be expected
between Ram Mountain, Sheep River, and Caw Ridge,
climatic data for Alberta were pooled and study site was
considered as a factor. Because temperature and
snowfall in April were correlated in Alberta (R
0.11, slope ¼0.05 60.02, mean 6SE, P¼0.01), we
considered the residuals of this regression as an index of
snowfall in April when analyzing the relationships
between phenological measures and weather data.
As lamb chest girth and body mass increased during
the period of capture, both parameters were adjusted
(girth at Sheep River to 20 November; body mass at
Ram Mountain to 15 September; Festa-Bianchet et al.
1997, Pelletier et al. 2005) before determining the yearly
sex-speciﬁc means. At Caw Ridge, kid body mass was
adjusted to 30 July (the mean date of capture for the 137
kids considered) using the slope of the linear regression
between mass and date without distinction for sex (Coˆ te
and Festa-Bianchet 2001a), before determining the
yearly sex-speciﬁc means. Mass and chest girth were
ln-transformed to stabilize the variance (Sokal and
Rohlf 1995). Using linear models weighted for yearly
sample size, we explored the linear and quadratic
relationships between NDVI measures and the chest
girth or body mass of juveniles.
The annual proportion of juveniles that survived the
winter was arcsine square-root transformed (Sokal and
Rohlf 1995). First-year winter survival is not sex-biased
(Festa-Bianchet et al. 1997, Coˆ te
2001a). Lamb survival at Sheep River and Ram
Mountain was affected by cougar (Puma concolor)
predation episodes (Sheep River: one from 1993 to
1995, and one from 1999 to 2004; Ram Mountain: one
from 1997 to 2003; Festa-Bianchet et al. 2006) and a
pneumonia epizootic in Sheep River (1985–1986; Festa-
Bianchet 1988a). To account for such effects, we used a
dummy variable (0/1, with 1 coding for cougar
predation or the occurrence of the epizootic). Using
linear models on arcsine square-root transformed
proportions, we then explored the linear and quadratic
relationships between NDVI measures and survival.
Because density-dependent responses were expected at
Ram Mountain and the GPNP (Portier et al. 1998,
Jacobson et al. 2004), we took into account population
density while modeling the effect of NDVI on body mass
and ﬁrst-year survival. Density at Ram Mountain was
indexed as the ln-transformed average body mass of
yearling females in June, an index of resource availabil-
ity (Coltman et al. 2003, Festa-Bianchet et al. 2004).
Because the yearly maximal increase in NDVI at time t
correlated with this index of density at time t(R
slope ¼0.63 60.27, P¼0.03), we used the residuals of
this regression to account for changes in density.
Density in the GPNP was indexed using yearly total
population counts (Jacobson et al. 2004). There was no
correlation between this index of density and the yearly
maximum increase in NDVI at the GPNP.
In three out of the four populations, several factors
have been described to inﬂuence juvenile survival and
growth such as birth date and age or condition of the
mother (Festa-Bianchet 1988b, Festa-Bianchet et al.
2000, Coˆ te
´and Festa-Bianchet 2001a, b, Gendreau et al.
2005). Here we used yearly averages because we were
interested in the impact of annual variations in
phenology on annual average performance. We also
wanted to explore whether similar patterns could be
observed in Canada and in Europe, in the GPNP, where
individual data are not available.
Model selection was performed using Akaike’s
Information Criterion corrected for small sample sizes
; Burnham and Anderson 1998). All continuous
variables were standardized when checking for interac-
tions between them. Temporal autocorrelation among
residuals was checked, and was not signiﬁcant for all
comparisons. All statistical analyses were performed in
the statistical package R (available online).
Phenology and local climate
Indicators of warm and wet springs inﬂuenced
vegetation phenology as assessed by NDVI in our study
sites. In Alberta (Appendix: Table A1), warmer springs
tended to be associated with rapid changes in plant
productivity: temperatures in April (slope ¼0.006 6
0.003, P¼0.08) and precipitation in May (slope ¼0.03
60.02, P¼0.06) tended to affect positively maximal
increases in NDVI. On the other hand, low snowfall in
April (slope ¼0.11 60.06, P¼0.05) and high
precipitation in May (slope ¼0.14 60.04, P¼0.002)
favored high INDVI values in May. Neither tempera-
ture in April nor snowfall in April or precipitation in
May was signiﬁcantly related to the average slope of
NDVI from early May to early July (all P.0.13;
Appendix: Table A1). In the GPNP (Appendix: Table
A2), both high temperatures in May and low snow depth
in April appeared to favor both the INDVI in May
(suggesting early onset of vegetation growth) and the
maximum increase in NDVI (suggesting rapid changes
in plant productivity), but these relationships were
nonsigniﬁcant (all P.0.09).
Vegetation dynamics and early performance
Contrary to our ﬁrst hypothesis (H
), high INDVI in
May did not inﬂuence the growth of bighorn lambs at
Sheep River or mountain goat kids at Caw Ridge (Table
1; Appendix: Table A3; all P.0.77). At Ram
February 2007 385PLANT PHENOLOGY AND ALPINE UNGULATES
Mountain, however, INDVI in May tended to negative-
ly affect lamb mass in mid-September (slope ¼0.19 6
0.11, P¼0.08, Table 1; Appendix: Table A3). We found
no effect in any study area of INDVI in May on ﬁrst-
year overwinter survival (Table 2; Appendix: Table A4;
As expected (H
), rapid changes in plant productivity
during green-up negatively inﬂuenced the growth of
lambs and kids of both sexes at Ram Mountain, Sheep
River, and Caw Ridge (Table 1, Fig. 2a–c; Appendix:
Table A3). At Sheep River, the yearly average chest
girth differed by up to 2.9 cm for males (3.6%of average
male chest girth) between years with low and high
maximal increases in NDVI, while the yearly average
mass of lambs at Ram Mountain varied by as much as
3.06 kg for males (;11.2%of the average). At Caw
Ridge, yearly average kid mass differed by up to 2.5 kg
for males (15.5%of the average) between years with low
and high maximal increases in NDVI. Similar results
were obtained for female lambs and kids.
The negative effect of the maximal increase in NDVI
on lamb growth was followed by a similar negative effect
on ﬁrst-year overwinter survival at Ram Mountain and
TABLE 1. Parameter estimates from linear models weighted for
sample sizes for the mean chest girth of bighorn lambs at
Sheep River and the body mass of bighorn lambs at Ram
Mountain and mountain goat kids at Caw Ridge.
Parameter LSM SE TP
A) Sheep River
Intercept 4.43 0.02 317.76 ,0.001
(Females males) 0.04 0.007 5.41 ,0.001
Max. inc. 0.37 0.10 3.50 0.001
B) Ram Mountain
Intercept 3.41 0.04 82.43 ,0.001
(Females males) 0.09 0.02 3.85 ,0.001
Res(BMY) 0.79 0.16 4.81 ,0.001
Max. inc. 0.48 0.17 2.86 0.006
C) Caw Ridge
Intercept 3.03 0.13 23.76 ,0.001
(Females males) 0.07 0.04 1.67 0.11
Max. inc. 1.70 0.82 2.08 0.05
Notes: Model selection procedures are presented in the
Appendix (Table A3). ‘‘Res(BMY)’’ is the residual from the
linear relationship between the maximum increase in NDVI and
the average body mass of yearling females, indexing resource
availability; ‘‘Max. inc.’’ is the maximum increase in NDVI
TABLE 2. Parameter estimates for overwinter survival of
juvenile ungulates at (a) Ram Mountain, (b) Sheep River,
(c) Caw Ridge, and (d) the GPNP, Italy.
Parameter LSM SE TP
A) Ram Mountain
Intercept 1.22 0.14 8.45 ,0.001
Res(BMY) 1.56 0.53 2.94 0.008
Max. inc. 1.64 0.65 2.51 0.02
B) Sheep River
Intercept 1.31 0.17 7.67 ,0.001
Predation 0.27 0.08 3.29 0.004
Max. inc. 2.62 1.36 1.92 0.07
C) Caw Ridge
Intercept 1.07 0.04 26.95 ,0.001
D) Gran Paradiso National Park
Intercept 1.07 0.10 10.63 ,0.001
Max. inc. 2.34 0.73 3.19 0.004
Notes: Model selection procedures are presented in the
Appendix (Table A4). ‘‘Predation’’ is a dummy variable (0,
absence of heavy predation or pneumonia epizootic; 1, heavy
predation or pneumonia); ‘‘Res(BMY)’’ is the residual from the
linear relationship linking the maximum increase in NDVI and
the average body mass of yearling females, indexing resource
availability; ‘‘Max. inc.’’ is the maximum increase in NDVI
FIG. 2. (a) Log-transformed mean body mass (measured in
kg) adjusted for density of male and female bighorn lambs at
Ram Mountain (Alberta) according to the maximal increase in
NDVI. (b) Mean chest girth for male and female bighorn lambs
according to the maximal increase in NDVI at Sheep River
(Alberta). (c) Log-transformed mean body mass (measured in
kg) of male and female mountain goat kids at Caw Ridge
(Alberta) according to the maximal increase in NDVI.
NATHALIE PETTORELLI ET AL.386 Ecology, Vol. 88, No. 2
the GPNP (Table 2, Fig. 3; Appendix: Table A4). The
annual maximal increase in NDVI explained 18%of the
variability in the average ﬁrst-year overwinter survival
of lambs at Ram Mountain and 33%for kids in the
GPNP. At Sheep River, the effect was in the same
direction and approached signiﬁcance (slope ¼2.62 6
1.36, P¼0.07). At Caw Ridge, however, the maximal
increase in NDVI was not related to kid survival (Table
2; Appendix, Table A4). The slope in NDVI between
early May and early July did not affect juvenile growth
or survival in any populations (all P.0.10).
Although population density negatively affected
growth and survival of lambs at Ram Mountain (Tables
1 and 2), contrary to (H
), density did not interact
signiﬁcantly with INDVI in May, the NDVI slope
between early May and early July, or the maximal
increase in NDVI in affecting body mass or survival of
juveniles at Ram Mountain or in the GPNP (Appendix:
Tables A3 and A4).
As expected, we found a negative effect of rapid
changes in plant productivity during green-up on
juvenile growth in three alpine ungulate populations.
Reduced growth presumably led to a negative effect of
rapid changes in plant productivity on juvenile survival
(Gaillard et al. 1997). However, we found no positive
effect on juvenile growth and survival of either early
vegetation onset as indexed by INDVI in May or
negative effect of steep vegetation onset as indexed by
the slope between NDVI in early May and early July. In
populations where density dependence was previously
reported (Portier et al. 1998, Jacobson et al. 2004), we
did not ﬁnd any interaction between INDVI in May or
high maximal increase in NDVI and population density
in determining growth or overwinter juvenile survival.
The reported patterns were coherent among the four
populations, three species, and two continents consid-
The duration of the vegetation growing period when
herbivores can access high-quality forage appears
mainly constrained by spring weather (Pettorelli et al.
2005a). Topographic variability in alpine habitats, on
both meso- and micro-scales, and associated differences
in snowmelt, can result in swards of different phenolog-
ical stages in close proximity, generating spatially
heterogeneous vegetation (Kudo 1991). Warm temper-
atures in spring may reduce this spatial heterogeneity if
they generate rapid snowmelt over the landscape,
reducing the period during which herbivores can access
high-quality forage. Warm temperatures and high
moisture favor rapid plant growth (Deﬁla 1991), which
also shortens the period of high forage quality (Hay and
Heide 1984). In Alberta, we found a positive association
between warm springs and high values of the maximal
increase in NDVI, which allowed us to establish a link
between warm springs and rapid changes in plant
productivity during green-up. Snowfall in April was
positively correlated with the maximum increase in
NDVI but had negative effects on INDVI in May.
Because the timing of snowmelt is the main determinant
of vegetation onset in mountainous environments (Kudo
1991), the negative relationship between INDVI in May
and April snowfall was expected. Water from abundant
snowfall in April and heavy precipitation in May,
associated with warm temperatures in April/May, may
lead to a vegetation bloom and fast changes in plant
FIG. 3. Winter survival of juveniles according to the maximal increase in NDVI at Ram Mountain (bighorn sheep, Alberta),
Sheep River (bighorn sheep, Alberta), Caw Ridge (mountain goats, Alberta), and the GPNP (ibex, Italy). Survival was adjusted for
density at Ram Mountain and for predation or pneumonia epizootic at Sheep River.
February 2007 387PLANT PHENOLOGY AND ALPINE UNGULATES
Nutritional requirements of animals vary with their
physiological state. For ungulate females, nutritional
demand peaks in late gestation and during lactation
(Clutton-Brock et al. 1989). The daily energetic require-
ment may increase by 150%during peak lactation
compared to maintenance (Loudon 1985). In highly
seasonal environments such as alpine habitats, spring
forage conditions can thus have profound effects on the
energy balance during late gestation and lactation.
Much emphasis has been placed on the importance for
herbivores inhabiting seasonal environments to match
vegetation green-up and birth period to access the
longest possible vegetation growing period (Bunnell
1982, Rutberg 1987). Females with access to high-
quality forage during lactation could provide greater
maternal care than females on a low nutritional plane
and reach sufﬁcient body condition in autumn to
conceive again. Offspring should beneﬁt from greater
maternal care by maximizing growth and overwinter
survival. Indeed, late-born offspring are more likely to
die during winter since they are smaller in autumn
(Festa-Bianchet 1988b). Early vegetation onset mea-
sured by INDVI in May, however, did not affect
juvenile growth or survival in our study.
Contrary to the timing of vegetation onset, little
attention has been paid to the role of the rate of change
in plant productivity during green-up and the possible
variation in the duration of the period of access to high-
quality forage in determining performance of herbivores
(but see Langvatn et al. 1996, Mysterud et al. 2001,
Pettorelli et al. 2005a). Here, we underlined the
relevance of indexing the rate of changes in plant
productivity during green-up using the maximum rate of
change in NDVI, and highlighted the sensitivity of
alpine ungulates to the shape of the vegetation
phenology curve. The only exception was in mountain
goats, where overwinter survival was not related to
vegetation dynamics indexed by NDVI. A previous
study at Caw Ridge reported that fecal crude protein
content in June, an index of forage quality, affected
positively kid mass, but not kid survival (Coˆ te
Festa-Bianchet 2001a). The absence of an effect of fecal
crude protein on survival could be due to an effect on
growth that was not sufﬁcient to lead to a higher
juvenile mortality (Gaillard et al. 1997), consistent with
the results presented here.
The slope between NDVI values in early May and
early July, which describes the steepness of the entire
green-up period, did not correlate with spring weather
or juvenile performance. The rate of changes in plant
productivity during green-up in a particular year would
always be smoothed when considering the slope between
two ﬁxed dates. Therefore, the slope between NDVI
values in early May and early July was less likely than
the maximal increase in NDVI to measure rapid changes
in vegetation productivity inﬂuencing the period of
access to high-quality forage by herbivores.
Surprisingly, we did not ﬁnd any interaction between
density and the maximal increase in NDVI in determin-
ing early growth or survival of alpine ungulates in sites
where density dependence was previously reported. The
interaction of density dependence and climate affects the
early performance of ungulates (e.g., Gaillard et al.
1997, Portier et al. 1998). The absence of interaction in
our study could result from density variations that were
insufﬁcient to exacerbate the effects of rapid changes in
plant productivity during the green-up on early perfor-
In all study sites, NDVI was assessed at a scale that
was generally higher than the areas used by animals. We
believe that the phenological signal captured at such
large scale is likely to be coherent with vegetation
phenology in the study sites, but it is possible that NDVI
data at a smaller spatial resolution could enhance the
signal we already captured.
Rapid changes in plant productivity during green-up
decreased juvenile performance in all study sites. We
suggest that rapid changes in NDVI during green-up
reduce the period of access to high-quality forage. This
could occur through a faster growth rate of plants, a
reduction in the spatial heterogeneity of snowmelt, or a
change in the plant community accessible to animals. To
test such hypotheses will require independent data on
the spatiotemporal availability and phenology of forage
over many years in the four study sites, data that are not
currently available. However, data on feeding sites of
female mountain goats during two years indicate that
vegetation quality, as measured by protein content of
plants in June, is high when the maximum increase in
NDVI is low (2003, proteins ¼20.1 60.5, maximum
NDVI increase ¼0.118) and low when the maximum
increase in NDVI is high (2002, proteins ¼17.3 60.6,
maximum NDVI increase ¼0.172; comparison of
protein content between years, F
Further work is required to explain the interannual
variability in the maximal increase in NDVI and to
understand the mechanisms by which rapid changes in
plant productivity (as reﬂected by NDVI) during the
green-up affect early performance. Considering its
importance in determining growth and survival in all
four populations measured, those studies are critically
Since we were unable to highlight any consistent effect
of INDVI in May, our results point toward the greater
inﬂuence of a measure of the average duration of the
period of access to high-quality forage such as maximal
increase in NDVI, than a measure of the average timing
of vegetation onset (INDVI in May), in determining
growth and survival of juvenile alpine ungulates.
Previous work had shown that in Norway warmer
winters could lead to higher snowfalls and delayed
vegetation onsets at high altitudes, affecting negatively
reindeer performance (Pettorelli et al. 2005c). Our
results highlight another mechanism that could become
more frequent with global climate change (Lapp et al.
NATHALIE PETTORELLI ET AL.388 Ecology, Vol. 88, No. 2
2005): warmer springs could negatively affect alpine
ungulates through a shorter period of access to high-
quality forage. In two of our study areas, the annual
maximal increase in NDVI appeared to increase over
time (Appendix: Fig. A2), possibly reﬂecting a warming
trend. Finally, our study illustrates how satellite-based
information on vegetation can be useful in investigating
the coupling between vegetation and herbivore perfor-
mance, particularly in highly seasonal environments
where phenological signals are strong.
We are grateful to the many people who helped with ﬁeld
work. We thank J. Hogg, K. Smith, and J. Jorgenson for their
pivotal contribution to ﬁeld research in the Alberta sites, and S.
Hamel for data on plant protein content from Caw Ridge.
Special thanks to J.-M. Gaillard, A. Mysterud, T. Coulson, A.
Provenzale, and J. Huot for ideas, comments, and suggestions
on previous drafts of this work. Research at Caw Ridge, Ram
Mountain, and Sheep River was ﬁnanced by the Alberta Fish
and Wildlife Division, the Natural Sciences and Engineering
Research Council of Canada, the Rocky Mountain Goat
Foundation, the Alberta Conservation Association, the Alberta
Sport, Recreation, Parks and Wildlife Foundation, the Fonds
´becois pour la formation de Chercheurs et l’Aide a
Recherche, the Universite
´de Sherbrooke, and Universite
We are thankful to A. E. M. Torino for providing the weather
data from the Serru` Meteorological Station.
Albon, S. D., and R. Langvatn. 1992. Plant phenology and the
beneﬁts of migration in a temperate ungulate. Oikos 65:502–
Beniston, M., and D. G. Fox. 1996. Impacts of climate change
on mountain regions. Pages 191–213 in R. T. Watson, M. C.
Zinyowera, and R. H. Moss, editors. Climate change 1995:
impacts, adaptations and mitigation of climate change.
Contribution of working group II to the Second Assessment
Report of the International Panel on Climate Change.
Cambridge University Press, New York, New York, USA.
Berteaux, D., D. Re
´ale, A. G. McAdam, and S. Boutin. 2004.
Keeping pace with fast climate change: Can arctic life count
on evolution? Integrative and Comparative Biology 44:140–
Bunnell, F. L. 1982. The lambing period of mountain sheep:
synthesis, hypotheses and tests. Canadian Journal of Zoology
Burnham, K. P., and D. R. Anderson. 1998. Model selection
and inference: a practical information-theoretic approach.
Springer Verlag, Berlin, Germany.
Chmielewski, F. M., and T. Ro
¨tzer. 2002. Annual and spatial
variability of the beginning of growing season in Europe in
relation to air temperature changes. Climate Research 19:
Clutton-Brock, T. H., S. D. Albon, and F. E. Guinness. 1989.
Fitness costs of gestation and lactation in wild mammals.
Coltman, D. W., P. O’Donoghue, J. T. Jorgenson, J. T. Hogg,
C. Strobeck, and M. Festa-Bianchet. 2003. Undesirable
evolutionary consequences of trophy hunting. Nature 426:
´, S. D., and M. Festa-Bianchet. 2001a. Birthdate, mass and
survival in mountain goat kids: effects of maternal charac-
teristics and forage quality. Oecologia 127:230–238.
´, S. D., and M. Festa-Bianchet. 2001b. Reproductive
success in female mountain goats: the inﬂuence of maternal
age and social rank. Animal Behaviour 62:173–181.
´, S. D., and M. Festa-Bianchet. 2003. Mountain goat.
Pages 1061–1075 in G. A. Feldhamer, B. C. Thompson, and
J. A. Chapman, editors. Wild mammals of North America:
biology, management and conservation. Second edition.
John Hopkins University Press, Baltimore, Maryland, USA.
Crawley, M. J. 1983. Herbivory: the dynamics of animal–plant
interactions. Blackwell, Oxford, UK.
Crick, H. Q. P., and T. H. Sparks. 1999. Climate change related
to egg-laying trends. Nature 399:423–424.
Deﬁla, C. 1991. Pﬂanzenpha
¨nologie der Schweitz. Publication
of the Swiss Meteorological Institute. Verlag Schweizerische
Meteorologische Anstalt Zu
Diaz, H. F., and R. S. Bradley. 1997. Temperature variations
during the last century at high elevation sites. Climate
Festa-Bianchet, M. 1988a. Seasonal range selection in bighorn
sheep: conﬂicts between forage quality, forage quantity, and
predator avoidance. Oecologia 75:580–586.
Festa-Bianchet, M. 1988b. Birthdate and survival in bighorn
lambs (Ovis canadensis). Journal of Zoology 214:653–661.
Festa-Bianchet, M., D. W. Coltman, L. Turelli, and J. T.
Jorgenson. 2004. Relative allocation to horn and body
growth in bighorn rams varies with resource availability.
Behavioral Ecology 15:305–312.
Festa-Bianchet, M., T. Coulson, J.-M. Gaillard, J. T. Hogg,
and F. Pelletier. 2006. Stochastic predation events and
population persistence in bighorn sheep. Proceedings of the
Royal Society B: Biological Sciences 273:1537–1543.
Festa-Bianchet, M., J. T. Jorgenson, C. Be
´, C. Portier, and
W. D. Wishart. 1997. Body mass and survival of bighorn
sheep. Canadian Journal of Zoology 75:1372–1379.
Festa-Bianchet, M., J. T. Jorgenson, and D. Re
´ale. 2000. Early
development, adult mass and reproductive success in bighorn
sheep. Behavioral Ecology 11:633–639.
Fryxell, J. M. 1991. Forage quality and aggregation by large
herbivores. American Naturalist 138:478–498.
Gaillard, J.-M., J. M. Boutin, D. Delorme, G. Van Laere, P.
Duncan, and J. D. Lebreton. 1997. Early survival in roe deer:
causes and consequences of cohort variation in two
contrasted populations. Oecologia 112:502–513.
Gaillard, J.-M., M. Festa Bianchet, N. G. Yoccoz, A. Loison,
and C. Toigo. 2000. Temporal variation in ﬁtness compo-
nents and population dynamics of large herbivores. Annual
Review of Ecology and Systematics 31:367–393.
Gendreau, Y., S. D. Co
´, and M. Festa-Bianchet. 2005.
Maternal effects on post-weaning physical and social
development in juvenile mountain goats (Oreamnos ameri-
canus). Behavioral Ecology and Sociobiology 58:237–246.
Giacometti, M., R. Willing, and C. Deﬁla. 2002. Ambient
temperature in spring affects horn growth in male alpine
ibexes. Journal of Mammalogy 83:245–251.
Grifﬁth, B., D. Douglas, N. E. Walsh, D. D. Young, T. R.
McCabe, D. E. Russell, R. G. White, R. D. Cameron, and
K. R. Whitten. 2002. The Porcupine caribou herd. Pages 8–
37 in D. C. Douglas, P. E. Reynolds, and E. B. Rhode,
editors. Arctic refuge coastal plain terrestrial wildlife research
summaries. U.S. Geological Survey, Biological Resources
Division,Biological Science Report USGS/BRD BR-2002-
Hay, R. K. M., and O. M. Heide. 1984. The response of high
latitude Norwegian grass cultivars to long photoperiods and
cool temperatures. Pages 46–50 in H. Riley and A. O.
Skjelvag, editors. The impact of climate on grass production
and quality. Proceedings of the 10th General Meeting of the
European Grassland Federation, AS, Norway.
Hughes, L. 2000. Biological consequences of global warming: is
the signal already apparent? Trends in Ecology and
Hulme, P. E. 2005. Adapting to climate change: is there scope
for ecological management in the face of a global threat?
Journal of Applied Ecology 42:784–794.
February 2007 389PLANT PHENOLOGY AND ALPINE UNGULATES
Inouye, D. W., B. Barr, K. B. Armitage, and B. D. Inouye.
2000. Climate change is affecting altitudinal migrants and
hibernating species. Proceedings of the National Academy of
Sciences (USA) 97:1630–1633.
Jacobson, A. R., A. Provenzale, A. von Hardenberg, B.
Bassano, and M. Festa-Bianchet. 2004. Climate forcing and
density dependence in a mountain ungulate population.
Kaduk, J., and M. Heinmann. 1996. A prognostic phenology
model for global terrestrial carbon cycle models. Climate
Kudo, G. 1991. Effects of snow-free period on the phenology of
alpine plants inhabiting snow patches. Arctic and Alpine
Langvatn, R., S. D. Albon, T. Burkey, and T. H. Clutton-
Brock. 1996. Climate, plant phenology and variation in age
at ﬁrst reproduction in a temperate herbivore. Journal of
Animal Ecology 65:653–670.
Lapp, S., J. Byrne, I. Townshend, and S. Kienzle. 2005. Climate
warming impacts on snowpack accumulation in an alpine
watershed. International Journal of Climate 25:521–536.
Laycock, W. A., and D. A. Price. 1970. Factors inﬂuencing
forage quality, environmental inﬂuences on nutritional value
of forage plants. USDA Forest Service Miscellaneous
Loudon, A. S. I. 1985. Lactation and neonatal survival of
mammals. Symposia of the Zoological Society of London 54:
Mysterud, A., R. Langvatn, N. G. Yoccoz, and N. C. Stenseth.
2001. Plant phenology, migration and geographic variation
in body weight of a large herbivore: the effect of a variable
topography. Journal of Animal Ecology 70:915–923.
Oechel, W. C., T. Callaghan, T. Gilmanov, J. I. Holten, D.
Maxwell, U. Molau, and S. Sveinbjornsson. 1997. Global
change and Arctic terrestrial ecosystems. Springer Verlag,
New York, New York, USA.
Pelletier, F., K. A. Page, T. Ostiguy, and M. M. Festa-Bianchet.
2005. Fecal counts of lungworm larvae and reproductive
effort in bighorn sheep. Oikos 110:473–480.
Pettorelli, N., A. Mysterud, N. G. Yoccoz, R. Langvatn, and
N. C. Stenseth. 2005a. Importance of climatological down-
scaling and plant phenology for red deer in heterogeneous
landscapes. Proceedings of the Royal Society B: Biological
Pettorelli, N., J. O. Vik, A. Mysterud, J.-M. Gaillard, C. J.
Tucker, and N. C. Stenseth. 2005b. Using the satellite-derived
normalized difference vegetation index (NDVI) to assess
ecological effects of environmental change. Trends in
Ecology and Evolution 20:503–510.
Pettorelli, N., R. Weladji, Ø. Holand, A. Mysterud, H. Breie,
and N. C. Stenseth. 2005c. The relative role of winter and
spring conditions: linking climate and landscape-scale plant
phenology to alpine reindeer body mass. Biology Letters 1:
Portier, C., M. Festa-Bianchet, J. M. Gaillard, J. T. Jorgenson,
and N. G. Yoccoz. 1998. Effects of density and weather on
survival of bighorn sheep lambs. Journal of Zoology 245:
Post, E., and N. C. Stenseth. 1999. Climate change, plant
phenology, and northern ungulates. Ecology 80:1322–1339.
Reed, B. C., J. F. Brown, D. Vanderzee, T. R. Loveland, J. W.
Merchant, and D. O. Ohlen. 1994. Measuring phenological
variability from satellite imagery. Journal of Vegetation
Rutberg, A. T. 1987. Adaptive hypotheses of birth synchrony in
ruminants: an interspeciﬁc test. American Naturalist 130:
Schwartz, M. D., and B. E. Reiter. 2000. Changes in North
American spring. International Journal of Climatology 20:
Sokal, R. R., and F. J. Rohlf. 1995. Biometry. Third edition.
W.H. Freeman, New York, New York, USA.
Thomas, D. W., J. Blondel, P. Perret, M. M. Lambrechts, and
J. R. Speakman. 2001. Energetic and ﬁtness costs of
mismatching resource supply and demand in seasonally
breeding birds. Science 291:2598–2600.
¨go, C., J.-M. Gaillard, and J. Michallet. 1999. Cohort
affects growth of males but not females in alpine ibex.
Journal of Mammalogy 80:1021–1027.
Tucker, C. J., J. E. Pinzon, M. E. Brown, D. A. Slayback, E. W.
Pak, R. Mahoney, E. F. Vermote, and N. El Saleous. 2005.
An extended AVHRR 8-km NDVI data set compatible with
MODIS and SPOT vegetation NDVI data. International
Journal of Remote Sensing 26:4485–4498.
Visser, M. E., C. Both, and M. M. Lambrechts. 2004. Global
climate change leads to mistimed avian reproduction.
Advances in Ecological Research 35:89–110.
Walther, G. R., E. Post, P. Convey, A. Menzel, C. Parmesan,
T. J. C. Beebee, J. M. Fromentin, O. Hoegh-Guldberg, and
F. Bairlein. 2002. Ecological responses to recent climate
change. Nature 416:389–395.
White, R. G. 1983. Foraging patterns and their multiplier
effects on productivity of northern ungulates. Oikos 40:377–
Wilmshurst, J. F., J. M. Fryxell, and R. J. Hudson. 1995.
Forage quality and patch choice by wapiti. Behavioral
Tables showing climatic factors inﬂuencing the INDVI in May, the maximum NDVI increase, and the average slope of NDVI
between early May and early July in three study sites in Alberta; climatic factors inﬂuencing the INDVI in May, the maximum
NDVI increase, and the average slope in NDVI between early May and early July, and correlation coefﬁcients between climatic
variables for the GPNP (Italy); and model selection procedures. Also included are ﬁgures showing interannual variations in NDVI
during the study periods and maximum increase in NDVI from 1982 to 2004 in four study sites in Alberta and Italy (Ecological
NATHALIE PETTORELLI ET AL.390 Ecology, Vol. 88, No. 2