ArticlePDF Available

Abstract and Figures

Many perennial plants display masting, i.e., fruiting with strong interannual variations, irregular and synchronized between trees within the population. Here, we tested the hypothesis that the early flower phenology in temperate oak species promotes stochasticity into their fruiting dynamics, which could play a major role in tree reproductive success. From large field monitoring network, we compared the pollen phenology between temperate and Mediterranean oak species. Then, focusing on temperate oak species, we explored the influence of the weather around the time of bud‐burst and flowering on seed production, and simulated with a mechanistic model the consequences an evolutionary shifting of flower phenology would have on fruiting dynamics. Temperate oak species release pollen earlier in the season than Mediterranean oak species. Such early flowering in temperate oak species results in pollen being often released during unfavorable weather conditions and resulting in frequent reproductive failure. If pollen release was delayed due to natural selection, fruiting dynamics would exhibit much reduced stochastic variation. We propose that early flower phenology might be adaptive by making mast‐seeding years rare and unpredictable, which would greatly help controlling the dynamics of seed consumers.
Content may be subject to copyright.
Flower phenology as a disruptor of the fruiting dynamics in
temperate oak species
Eliane Schermer
, Marie-Claude Bel-Venner
, Jean-Michel Gaillard
,Stephane Dray
, Vincent
, Iris Le Ronce
, Gilles Oliver
, Isabelle Chuine
, Sylvain Delzon
and Samuel Venner
Laboratoire de Biometrie et Biologie
Evolutive, UMR 5558, Universite de Lyon, Universite Lyon 1, CNRS, F-69622, Villeurbanne, France;
Departement recherche, developpement et
innovation, Office National des For^ets, F-77300, Fontainebleau, France;
Centre d’
Ecologie Fonctionnelle et
Evolutive, UMR 5175, CNRS, F-34293, Montpellier, France;
Reseau National de
Surveillance Aerobiologique, F-69690, Brussieu, France;
UMR 1202, BIOGECO, Institut National de la Recherche Agronomique, F-33612, Cestas, France;
UMR 1202, Biodiversite, des
genes aux communautes, Universite de Bordeaux, F-33615, Pessac, France
Author for correspondence:
Samuel Venner
Tel: +33 4 72 43 29 02
Received: 30 July 2019
Accepted: 19 September 2019
New Phytologist (2019)
doi: 10.1111/nph.16224
Key words: masting, oak species, pollen
phenology, resource budget model,
stochastic mast seeding.
Many perennial plants display masting, that is, fruiting with strong interannual variations,
irregular and synchronized between trees within the population. Here, we tested the hypothe-
sis that the early flower phenology in temperate oak species promotes stochasticity into their
fruiting dynamics, which could play a major role in tree reproductive success.
From a large field monitoring network, we compared the pollen phenology between tem-
perate and Mediterranean oak species. Then, focusing on temperate oak species, we explored
the influence of the weather around the time of budburst and flowering on seed production,
and simulated with a mechanistic model the consequences that an evolutionary shifting of
flower phenology would have on fruiting dynamics.
Temperate oak species release pollen earlier in the season than do Mediterranean oak
species. Such early flowering in temperate oak species results in pollen often being released
during unfavorable weather conditions and frequently results in reproductive failure. If pollen
release were delayed as a result of natural selection, fruiting dynamics would exhibit much
reduced stochastic variation.
We propose that early flower phenology might be adaptive by making mast-seeding years
rare and unpredictable, which would greatly help in controlling the dynamics of seed con-
Reproduction in many perennial and wind-pollinated plant
species is characterized by masting, that is, synchronized and
highly variable amounts of seed production over the years within
a population (Janzen, 1976; Kelly & Sork, 2002). Masting is
known to impact the demography and evolution of seed con-
sumers strongly (Yang et al., 2010; Venner et al., 2011; Gamelon
et al., 2013; Pelisson et al., 2013; Rey et al., 2015; Bogdziewicz
et al., 2016), with cascading effects on forest biodiversity dynam-
ics together with major economical and societal issues (e.g. forest
regeneration, disease propagation) (Crawley, 2000; Ostfeld &
Keesing, 2000; Frey et al., 2007; Bogdziewicz & Szymkowiak,
2016). Despite the increasing number of studies addressing the
issue of masting and its consequences for ecosystem functioning
and service provisioning, its proximate causes remain difficult to
disentangle, mainly because of the diversity of candidate mecha-
nisms possibly interacting and the strong stochasticity (in the
sense of unpredictability for observers or seed consumers) in the
fruiting dynamics (Crone & Rapp, 2014; Pearse et al., 2016;
Vacchiano et al., 2018).
Fruiting of mast-seeding species, besides fluctuating strongly
and synchronously over the years, is characterized by negative
temporal autocorrelation (Sork et al., 1993; Herrera et al., 1998;
Koenig & Knops, 2000; Koenig et al., 2003). Such autocorrela-
tion is classically interpreted as resulting from the resource deple-
tion of the trees following mast-seeding years, which prevents
them from producing flowers and seeds the following year (i.e.
resource depletion hypothesis; Monks & Kelly, 2006; Barringer
et al., 2013; Crone et al., 2009, but see Kelly et al., 2013). Conse-
quently, the fruiting dynamics are potentially extremely asym-
metrical, with lean-seeding years consistently occurring after a
mast-seeding year (a deterministic component of masting as a
result of reserve depletion of trees) while mast-seeding years may
not systematically follow one lean-seeding year. This irregularity
in the occurrence of mast-seeding years (hereafter called the
stochastic component of masting corresponding to the fluctua-
tions not explained by the negative temporal autocorrelation)
Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
New Phytologist (2019) 1
would depend on weather conditions possibly affecting resource
acquisition (Smail et al., 2011), flower development, pollination
and fertilization of female flowers (Cecich & Sullivan, 1999;
Koenig et al., 2015; Pearse et al., 2015; Sabit et al., 2016;
Bogdziewicz et al., 2017a; Schermer et al., 2019) or even fruit
ripening (Richardson et al., 2005; Perez-Ramos et al., 2015;
Chang-Yang et al., 2016; Buechling et al., 2016). Furthermore,
such a weather effect can act as an ‘environmental veto’ by nearly
completely preventing flower or seed development at the tree
population scale (Feret et al., 1982; Cecich & Sullivan 1999;
Bogdziewicz et al., 2018a, 2019).
A promising avenue to understanding both the proximate and
evolutionary causes of masting is to identify the key tree life-his-
tory traits governing the stochastic component of masting.
Flower phenology would be a serious candidate as its evolution-
ary change (i.e. the timing of flowering at the population level
which may shift independently of any climate change) could the-
oretically act on masting in two complementary ways: by modify-
ing the likelihood of late frost events at the vulnerable flowering
stage (Garcıa-Mozo et al., 2001; Augspurger, 2009), which can
act as environmental veto and may strongly impede fruit set
(Feret et al., 1982; Cecich & Sullivan 1999; Bogdziewicz et al.,
2018a); and by partly setting the weather conditions influencing
pollen maturation and release which would play a key role in pol-
len limitation and then in fruiting success. In this sense, high
spring temperature has recently been shown to favor phenological
synchronization between trees, by reducing the flowering period,
which in turn would increase pollination success and promote
mast-seeding years (Koenig et al., 2008, 2012, 2015; Bogdziewicz
et al., 2017a). High spring temperature also increases the annual
amount of airborne pollen that can be mobilized for reproduc-
tion. Schermer et al. (2019), after analyzing the interannual varia-
tion of both airborne pollen amount and its temporal
distribution, suggested that pollen limitation in European tem-
perate oaks would rely more on the annual amount of airborne
pollen than on tree synchrony.
Based on these findings, the aim of our study was to test the
hypothesis that flower phenology is a key trait driving the
stochastic component of masting in two temperate oak species
(Quercus petraea and Quercus robur) by keeping mast-seeding
years rare and unpredictable. We thus examined the conse-
quences of an evolutionary shift in the flower phenology on their
masting. For this purpose, we combined empirical and theoreti-
cal approaches. First, we tested the hypothesis that pollen phenol-
ogy is early in these two temperate oak species in comparison to
Mediterranean oak species (Quercus ilex and Quercus pubescens)
and we examined the consequences of phenological differences
between the two groups on the sensitivity of annual amounts of
airborne pollen to spring weather conditions. Second, focusing
exclusively on the two temperate oak species for which we have
an extensive network of fruiting monitoring, we explored the
consequences of an evolutionary change in flower phenology on
fruiting dynamics. For this last point, we proceeded in two steps:
first, we determined the key weather conditions surrounding the
timing of budburst that should affect fruiting success (late frost
and/or weather conditions affecting pollen release and diffusion);
and then we built a mechanistic model (resource budget model
(RBM); Isagi et al., 1997; Satake & Iwasa, 2000, 2002) and we
simulated fruiting dynamics according to several evolutionary
flower phenological strategies (i.e. earlier or later phenology than
currently observed).
Materials and Methods
Study species
We focused on the four most abundant oak monoecious
species in France: Quercus robur L., Q. petraea Liebl. L.,
Q. ilex L. and Q. pubescens Willd. Q. robur and Q. petraea are
present from southern Scandinavia to Spain and western Rus-
sia in Europe. Q. ilex is the most dominant tree species in the
central and western parts of the Mediterranean basin. Q.
pubescens has an intermediate distribution, co-occurring with
Q. robur and petraea in central Europe and with Q. ilex in
southern Europe. Q. robur and Q. petraea co-occur all over
France except along the Mediterranean basin where they are
replaced by Q. pubescens up to 1200 m, and Q. ilex, especially
at lower elevations (Badeau et al., 2017). The phenologies of
the four species show some differences. Budburst occurs
between late April and early May for Q. robur, Q. petraea and
Q. pubescens (Badeau et al., 2017), and between April and
May for Q. ilex, depending on the latitude (Garcia-Mozo
et al., 2007; Ogaya & Penuelas 2004; Misson et al., 2011;
Fernandez-Martinez et al., 2012). The four species can have
either vegetative buds with leaves only, mixed buds with male
flowers, female flowers and leaves or reproductive buds with
male flowers only. Male flowers are mature 2 wk after bud
flush and 2 wk before female flowers. Leaves have reached
c. 75% of their final size when female flowers become recep-
tive (Badeau et al., 2017). In all four species, fertilization
occurs at the end of June or early July (Pesson & Louveaux,
Phenology, pollen and fruiting data
The airborne amount of oak pollen was recorded daily using
Hirst traps (Hirst, 1952) at 43 sites in France during a 22 yr sur-
vey (19942015; Reseau National de Surveillance Aerobi-
ologique; see Fig. 1a for a map; see Supporting Information
Table S1 for the pollen-sampling site characteristics). As the oak
species was not recorded in the pollen dataset, we relied on the
national forest inventory (Institut Geographique National,
France; see the forest stand dataset providing the forest cover rate
of each species) to determine within a 50 km radius at each pol-
len-sampling site the covering surface of each oak species. We
split the pollen dataset into two sub-datasets, one called ‘temper-
ate’, including sites where >80% of oaks are temperate oak
species (Q. robur and/or Q. petraea, 35 sites), and the other one
called ‘Mediterranean’, including sites where >80% of oaks are
Mediterranean oak species (Q. ilex and/or Q. pubescens, eight
sites). At each site and each year, the total amount of airborne
pollen was computed and divided by the percentage of the surface
New Phytologist (2019) Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
covered by oak trees so as to account for disparities in forest den-
sity between the sites; this corrected amount of pollen (hereafter
airborne pollen amount) will be used in all subsequent analyses.
We used data on the budburst date and the fruiting dynamics
of temperate oak species from the ONF-RENECOFOR network
(Ulrich, 1995) covering 30 sites for 14 yr (19942007) (see
Fig. S1 for a map and Table S2 for the GPS coordinates). Among
the 30 sites, 19 are dominated by Q. petraea, nine by Q. robur
and two of them are mixed oak forests (see Table S2). These sites
are all different from the pollen-sampling sites. Acorn production
was estimated yearly at each site on a fixed 1 acre (0.405 ha) sur-
face where 10 nonneighboring mature trees were each equipped
with one 0.5 m
raised litter-fall trap; the mature acorns collected
were counted exhaustively and summed for the 10 trees. The
budburst date was estimated at each site and year as the earliest
date at which the first 10% of trees had 2050% of their buds
open (phenological stage BBCH 9; Meier et al., 2009).
Meteorological data and their use
On the basis of the daily weather data extracted from the
SAFRAN spatially explicit database (8 98 km mesh size grid)
(Durand et al., 1993), we calculated for each of the 43 pollen and
30 acorn sampling sites the mean daily temperature (°C) and the
cumulative rainfall (mm) during different periods in spring to
test the effect of weather conditions on the amount of airborne
pollen in both temperate and Mediterranean oak species and on
fruit production in temperate oak species.
At each of the 30 acorn-sampling sites of the ONF-
RENECOFOR network, we also computed the minimum daily
temperature (to check for the possibility of late frost acting as an
environmental veto; see later for a discussion of threshold detec-
tion). Following Lebourgeois et al. (2008), we first modeled the
budburst date available at each of the 30 acorn-sampling sites as a
linear function of the mean March temperature recorded every
year at these sites (see Table S3; Fig. S2). Using this negative rela-
tionship, we then inferred the budburst date each year at each of
the 35 pollen-sampling temperate oak sites. This allowed us to
test if the weather conditions around the budburst date (e.g. the
occurrence of late frost within 30 d before, or the mean tempera-
ture 1 month afterwards) were linked to both amount of airborne
pollen and fruit production. Focusing on these identified key
weather conditions around budburst date and using the meteoro-
logical data retrieved at each site since 1959, we carried out fur-
ther simulations using the RBM (see RBM modeling section) to
explore the effect of a shift in flower phenology in temperate oak
species on fruiting dynamics.
Data analysis
We compared the ‘temperate’ and ‘Mediterranean’ oak popula-
tions for their pollen phenology. We analyzed the differences in
the median date of pollen release (i.e. the day by which 50% of
the annual airborne pollen has already been released) using Stu-
dent’s t-test.
We analyzed the sensitivity of airborne pollen amount to vari-
ous spring weather variables separately for the ‘temperate’ and
‘Mediterranean’ sites as follows. First, for various spring periods,
we performed a principal component analysis (PCA) on mean
temperature and cumulative rainfall, and used the first principal
component (PC1) as a synthetic weather variable reflecting both
temperature and rainfall of each spring periods (see Table S4).
Second we performed generalized linear mixed models (GLMMs
with Gaussian family and identity link) with log-transformed air-
borne pollen amount as the dependent variable and the PC1 vari-
able depending on the spring period considered (see Table S4)
(a) (b)
Fig. 1 Comparison of the pollen phenology between the temperate and Mediterranean oak forests. (a) Spatial distribution of the 43 pollen-sampling sites.
Temperate oak forests are defined to include 80% or more Quercus petraea and/or Quercus robur (35 sites; see orange circles) and Mediterranean oak
forests include 80% or more Quercus ilex and/or Quercus pubescens (eight sites; green triangles) relative to the whole oak forest area comprised within a
50 km radius around each pollen-sampling site. The GPS coordinates and the forest cover rate of each oak species of all pollen-sampling sites are indicated
in Supporting Information Table S1. (b) Cumulative frequency distribution of the median date of oak pollen release for the ‘temperate’ (orange line) and
‘Mediterranean’ (green line) oak forests. The median date was calculated each year (from 1994 to 2015) at each site as the day by which 50% of the
annual pollen amount has already been released. Dates are in Julian days, that is, the number of days elapsed since 1 January (day 1) of each year.
Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
New Phytologist (2019)
Phytologist Research 3
and the lag-1 yr autocorrelation of airborne pollen amount as
covariates, considering the factors ‘site’ and ‘year’ as random
effects. We selected the most parsimonious GLMM separately
for the two forest types using the Akaike information criterion
(see results in Table S5).
To assess the importance of pollen phenology on masting in
temperate oak species, we tested the sensitivity of fruit produc-
tion to weather at different time periods around the budburst
date, from the date on which male flowers become particularly
sensitive to frost right up to pollen release. To ensure the robust-
ness of the results reported, we split the whole acorn dataset into
two mirror sub-datasets, each one comprising full time series of
15 acorn-sampling sites evenly distributed over similar altitude,
longitude and latitude gradients (see Fig. S1). We conducted an
exploratory approach on a first sub-dataset to identify candidate
periods and their weather conditions (the minimum temperature
threshold below which late frost may act as an environmental
veto on fruiting). We fitted negative binomial GLMMs with log
link to estimate fruit production with the lag-1 yr autocorrelation
of fruit production, the mean temperature over 30 d after bud-
burst date and the occurrence of frost during several periods
around budburst date as binary factor (i.e. considering frost
whenever minimum daily temperature falls below a threshold
value tested) as covariates, considering the factors ‘site’ and ‘year’
as random effects to increase the probability of identifying candi-
date periods and minimum temperature threshold (see
Table S6). On the second sub-dataset, we tested whether the
weather variables previously identified were still detected by fit-
ting a negative binomial generalized linear model (GLM) with
the factors ‘site’ and ‘year’ as fixed effects (see Table S7).
All statistical analyses were performed with the R free software
environment (v.3.4.3, We performed
the PCA using the ‘dudi.pca’ function in the ADE4 package (Dray
& Dufour, 2007), and fitted the multiple additive GLMMs using
the ‘lmer’ function in the LME4 package (Bates et al., 2015).
RBM modeling
We built an RBM, that is, an individually based, spatially explicit
model accounting for the individual strategies for allocating
resources into reproduction, and allowing us to simulate individ-
ual flowering and fruiting dynamics within a population. We
modified a previously published RBM (Venner et al., 2016;
Schermer et al., 2019) to incorporate the effect of pollen phenol-
ogy on masting in temperate oaks. According to the former
RBM, interannual variations of fruit production may partly result
from interannual variations of airborne pollen available for repro-
duction that depend on both the amount of pollen produced by
trees and spring temperature at the time of pollen release (see
Schermer et al., 2019). Whereas in the former RBM (Schermer
et al., 2019), the timing of pollen release was set to April, irre-
spective of the year and the site, the biological realism of the
RBM presented here was improved by integrating the identified
key weather conditions around the actual timing of budburst (i.e.
pollen phenology; see Methods S1 for further details) and we
used this enhanced RBM to study the impact of a theoretical
evolutionarily shifted pollen phenology (i.e. budburst date) on
the fruiting dynamics of temperate oaks. Notably we explored
the impact on fruiting dynamics of a fixed 15 d shift in the bud-
burst date, either advanced or delayed, depending on the model,
and a 15 d delayed budburst date corresponding to the actual
pollen phenology of Mediterranean oaks (this study) and to that
of ash trees (Fraxinus spp.) in the temperate region (Vitasse et al.,
2009). We further examined a 30 d delayed pollen phenology, as
observed for beech (Fagus spp.) in the temperate region (see
Vitasse et al., 2009).
The four classical mathematical descriptors for masting are:
the individual coefficient of variation of fruiting intensity (CVi)
describing the individual between-year variability in seed pro-
duction; the degree of synchrony among trees within the popu-
lation in their fruiting interannual dynamics (classically the
mean of pairwise correlation between crop size of individuals
within the population); the population coefficient of variation
(CVp) describing the fruiting temporal variation at the popula-
tion level; and the negative temporal autocorrelation (often at
1 yr time lag) of seed production (classically ACF1; Koenig
et al., 2003; Herrera, 1998; Kelly & Sork, 2002; Buonaccorsi
et al., 2003). None of these descriptors, however, is able to
describe the asymmetry in the fruiting dynamics (see the Intro-
Here, to analyze the results of the simulations, we character-
ized the intensity of the fluctuations with the CVp parameter.
Complementarily, to quantify the deterministic and stochastic
components of fruiting dynamics and the impact of flower phe-
nology on these components, we analyzed temporal autocorrela-
tion at the population scale using the standard statistic:
ðxtxtþ1Þwith xt¼ztz
where Tis the length of the fruiting series and z,z
and x
spond, respectively, to the average annual crop size of the popula-
tion, the crop size at year tand the centered crop size at year t.
This statistic corresponds to the numerator of several standard
measures of autocorrelation (Wald & Wolfowitz, 1943; Dray
et al., 2010).
For each flower phenology, we evaluated the significance of
the observed statistic (S) by comparing its value to the distribu-
tion under the null hypothesis obtained using 999 permutations
of the fruiting series. To compare the different phenology sce-
nario for their degree of stochasticity in fruiting dynamics, we
computed the ‘standardized effect size’ (SES, Gotelli & McCabe,
2002) by standardizing the observed statistics (S) by the means
and SDs estimated under the null hypothesis. Under the null
hypothesis that there is no autocorrelation in the fruiting series,
the distribution of SES should be centered on 0 with SD =1,
while SES will be all the more negative when the negative tempo-
ral autocorrelation is strong (or when the stochastic component
of the masting is weak). Under the assumption that phenology
has no effect on the stochastic component of masting, the distri-
bution of SES should be similar between the different phenologi-
cal scenarios.
New Phytologist (2019) Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
In addition, we used a dual metric reflecting the degree of
asymmetry in the fruiting dynamics as follows: the probability
for a lean-seeding year (L) to follow a mast-seeding year
(M) at the population scale; and the probability P
of having a
mast-seeding event (M) the year following a lean-seeding year
(L). In the RBM outputs, mast- and lean-seeding years are
defined for fruiting allocation >0.7 and <0.3, respectively (the
value 1 being, on average, the mean amount of resources acquired
annually by trees that can be allocated to current reproduction or
stored for future reproduction; see Methods S1). A sensitivity
analysis was performed by testing various threshold values to
define these two categories of fruiting level and the results remain
qualitatively similar (see Fig. S3).
For each set of parameters (or sites), 100 repeated simulations
of fruiting dynamics were run over 2000 yr; we then computed,
over the last 100 years, the CVp to describe masting intensity,
the Sand SES for analysis of temporal autocorrelation of masting
(i.e. its deterministic and stochastic components), and the two
frequencies P
and P
for describing masting asymmetry.
We were then able to compare various scenarios differing in their
flower phenology for the degree of asymmetry in their associated
fruiting dynamics.
We carried out additional simulations to analyze the relative
contribution of the two modeled meteorological effects (by inte-
grating only one of the two mechanisms at a time, i.e. either the
‘environmental veto’ effect related to late frosts or the effect of
weather conditions on pollen aerial diffusion) on the fruiting
dynamics under the different phenological scenarios. Finally, to
address the issue of decoupled investment made by trees into
male and female flowers in pollen limitation and fruiting dynam-
ics (see Crone & Rapp, 2014), we carried out sensitivity analyses
considering that the relative allocation of male and female flower-
ing resources could deviate from a strict equilibrium of 0.5. We
ran analyses in two complementary ways, considering: that trees
may have their own, consistent allocation ratio into male flower-
ing (defined for each tree by randomly sampling in a Gaussian
distribution with 0.5 (0.1), mean (SD)); or each tree may vary
from one year to the next in its relative allocation into male and
female flowers (defined for each tree and each year by sampling
the ratio in a Gaussian distribution with 0.5 (0.1), mean (SD)).
Data availability
Data supporting the results are available from the Dryad Digital
Repository (doi: 10.5061/dryad.p8cz8w9k3).
Pollen phenology in temperate and Mediterranean oaks
In oak species growing in the temperate region, pollen is released
mainly from the second half of April to early May, occurring ear-
lier in the season as latitude decreases (see Fig. S4). Mediter-
ranean oak species, despite being located south of the temperate
oak forests (Fig. 1a), release their pollen mainly in May (Fig. 1b),
that is, about 2 wk later on average than temperate oaks (two-
sample Student’s t-test: t=17.42, df =676, P<0.001, 95% CI:
Depending on the region (temperate or Mediterranean), pol-
len release thus occurs under contrasting weather conditions
owing to phenological differences between oak species (Fig. 2a).
In the temperate region, the annual airborne pollen amount was
positively related to April temperature, following a logistic rela-
tionship (see Table S8 for results of the GLMM selection; see
Table S9 for results of the model selection between the logistic
and linear models) in line with a recent study (Table S5; Scher-
mer et al., 2019). Conditions for pollen release seem optimal for
mean April temperature >13°C (value determined by a threshold
model; Huber, 1964; see Fig. S5), which occurred in 11% and
100% of the sites and years for temperate and Mediterranean oak
species, respectively (Fig. 2b).
Impact of a shift in the pollen phenology on fruiting
dynamics in temperate oaks
In the temperate region, the early timing of pollen release makes
reproduction sensitive to late frost (5°C or less) whenever it
occurs within 30 d before the budburst date (Fig. 3a; Tables S6,
S7), and to the mean temperature >30 d after budburst date (i.e.
spring temperature impacting airborne pollen amount; see
Fig. S6) (see also Fig. 3b; Tables S6, S7, S9).
Our RBM simulations suggest that fruiting dynamics of tem-
perate oak tree populations would be sensitive to evolutionary
shifting pollen phenology (Figs 4, 5). Although fruiting dynamics
fluctuate greatly under all phenological scenarios (Fig. 4), the
CVp would decrease under later phenology (Fig. 5a) and the neg-
ative temporal autocorrelation would become more pronounced
(Fig. 5b) and, hence, variation in the fruiting dynamics more
deterministic. A shift towards earlier pollen phenology would be
accompanied by more pronounced asymmetry of fruiting
dynamics (Fig. 5c,d): the probability that a lean-seeding year
would follow a mast-seeding year remained unchanged and high
(i.e. mainly between 0.8 and 1) irrespective of the phenology sim-
ulated (Fig. 5c). By contrast, the probability of having a mast-
seeding year after a lean-seeding one was very variable according
to the different phenological scenarios and was lowest for the ear-
liest phenologies (Fig. 5d).
By considering weather factors in isolation in the modeling,
we revealed that the veto-like effect of late frost would play only a
minor role in fruiting dynamics under current phenology,
whereas the weather conditions that influence pollen spread, as
they stand, would retain a key role (Fig. S7b). If oak phenology
was 15 d earlier, the negative effect of late frost would then
emerge (Fig. S7a). Later flower phenologies would almost sys-
tematically meet optimal conditions for reproduction, without
severe weather conditions unfavorable to flower survival or pollen
diffusion (Fig. S7c,d).
Overall, our results suggest that if temperate oak evolved
towards delayed phenology (independently of any climate
change), their fruiting dynamics would still fluctuate, but in a
much less stochastic way, and mast-seeding years should then
become more predictable (i.e. mainly driven by negative
Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
New Phytologist (2019)
Phytologist Research 5
temporal autocorrelation). These results were obtained consider-
ing that trees invested equal amounts of energy into male and
female flowers; our findings remain robust to departures from
that assumption (see Fig. S8).
Whereas the timing of pollen release is delayed with increasing
latitude within temperate oak species, our results revealed that
these species have much earlier pollen phenology (15 d differ-
ence) than Mediterranean oaks. In the four species studied, fertil-
ization consistently takes place during the same period (late June,
early July) (Pesson & Louveaux, 1984) and at a much later date
than pollination, suggesting that the phenology of temperate oak
species could theoretically be later than it actually is. In temperate
oak forests, we show that the early pollen phenology observed in
the field is often associated with weather conditions that are unfa-
vorable to pollen maturation and/or aerial diffusion, which could
explain why reproductive failure is common. Our results suggest
that such advanced pollen phenology would give trees a selective
advantage by generating a strong stochastic component in fruit-
ing dynamics, which is possibly decisive for effective control of
seed consumer demography.
The early spring pollen maturation of temperate oak species
could be seen as maladaptive owing to the suboptimal weather
conditions encountered at the time of pollen release (c. 10% of
years have mean temperature >13°C over 30 d after budburst
date (Figs 3b, S6)) and to the probability of suffering frost dam-
age at flowering (i.e. 5% of years with minimum temperature
<5°C occurring within 30 d before the budburst date). Such
early pollen phenology might lead to frequent, massive fruiting
failure and explain why the fruiting dynamics of some oak species
are very sensitive to spring weather conditions (Pearse et al.,
2014; Koenig et al., 2015; Bogdziewicz et al., 2017a; Caignard
et al., 2017; Nussbaumer et al., 2018; Schermer et al., 2019). By
contrast, Mediterranean oak pollen, because of the warmer cli-
matic conditions encountered and their delayed pollen phenol-
ogy, experience weather conditions that are usually favorable to
pollen maturation and release (Fig. 2), with very rare exposure to
late, intense frost (Fig. S9b). The evolutionary divergence in pol-
len phenology between these oak species would then sustain the
diversity of their responses to spring weather conditions and
partly explain why finding common determinants of masting is
so difficult in the genus Quercus (Sork et al., 1993).
From an evolutionary perspective, pollen phenology could be
seen as a key life-history trait that partly controls the degree of
stochasticity in fruiting dynamics in temperate oak species. Based
on our RBM, we show that contrasting yet realistic variations in
oak flower phenology (i.e. within the range of other wind-polli-
nated forest species) would all still generate large fluctuations in
fruiting (Fig. 4). However, the stochastic component of masting
was increased only when simulating earlier pollen phenology
(Fig. 5b,d), which generated conditions that are often unfavor-
able to reproduction. As proposed from theoretical work (Rees
et al., 2002), disturbance in fruiting dynamics is probably essen-
tial to efficiently control the dynamics of seed consumer popula-
tions and maximize tree fitness. Oak acorns are a pulsed resource
for various consumers that affect their population dynamics (in-
sects, Venner et al., 2011; birds, McShea, 2000; rodents, Wolff,
1996; ungulates, Gamelon et al., 2017). Among consumers,
insects specialized in this resource are probably the most
6 7 8 9 10 11 12 13 14 15 16 17 18
Mean annual airborne pollen amount
(m−3 of air)
6 7 8 9 10 11 12 13 14 15 16 17 18
Cumulative frequencies of sites
Mean temperatures (°C)
at time of pollen release (1 month)
Fig. 2 Comparison of the sensitivity of airborne pollen and mean
temperature at time of pollen release between temperate and
Mediterranean oak forests. (a) Mean annual airborne pollen amount as a
function of mean temperature at time of pollen release for both
Mediterranean (Quercus ilex and/or Quercus pubescens; green circles)
and temperate oaks (Quercus petraea and/or Quercus robur; orange
circles). Mean temperatures were computed in April for temperate oaks
and between mid-April and mid-May for Mediterranean oaks to account
for the 15 d delayed pollen phenology (see Fig. 1). Data shown are means
SE of annual airborne pollen amount ranked according to increasing
temperature and grouped by sets of 10 consecutive values to compute
mean SE (see Supporting Information Fig. S10a,b, which shows the
same relationship with ungrouped data). Shaded areas show the 95%
confidence interval of the model estimates. (b) Cumulative frequencies of
April temperatures (orange line) and temperatures between mid-April and
mid-May (green line) at each site and year for temperate and
Mediterranean oaks, respectively. The orange vertical dotted line shown in
(a) and (b) is the 13°C threshold value above which the pollen amount of
temperate oaks reaches high values, independent of mean April
temperature (see Fig. S5 for the deviance profile from the ‘threshold
model’ (Huber, 1964)).
New Phytologist (2019) Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
problematic for the following reasons: insects are able to respond
demographically very quickly to the fluctuations of the resource
(Bogdziewicz et al., 2018b); acorn consumption by insects
severely reduces seed germination success and seedling survival
(Andersson, 1992; Mu~noz et al., 2014; Leiva et al., 2018; Yi
et al., 2019); and several weevil species commonly coexist on the
same individual trees and display widely diverse life-history traits
(Venner et al., 2011; Pelisson et al., 2012, 2013; Rey et al.,
2015), making it difficult for the trees to control the dynamics of
the whole insect community. Efficient control of such insect
diversity is probably tightly linked to strong stochastic compo-
nent in the fruiting dynamics. In temperate oak species, our
results suggest that early phenology would play this pivotal role
in inducing weather conditions most often detrimental to yearly
reproduction, thus making mast-seeding years unpredictable for
seed consumers, and hence maximizing tree fitness. Our results
are therefore in line with the recent proposal by Bogdziewicz
et al. (2019) that the weather conditions causing frequent fruiting
failure are traditionally perceived as negative for plants, but
would help to maximize their lifetime reproductive success.
Under the current phenology of temperate oak species, fruiting
failure is explained much more widely by climatic conditions that
are unfavorable to pollen maturation or diffusion (i.e. the month
following the budburst date) than by the occurrence of late frost
−10 −8 −6 −4 −2 0 2 4 6
Annual acorn production (m–2 )
Minimum temperatures (°C) occurring
within 30 d before budburst date
Veto No veto
7 8 9 10111213141516
Mean annual acorn production (m−2)
Mean temperatures (°C) at
> 30 d after budburst date
Fig. 3 Sensitivity of fruiting intensity to temperature around budburst date in temperate oak forests (Quercus petraea and/or Quercus robur). (a) Annual
acorn production along the minimum temperatures occurring within 30 d before budburst. The vertical line corresponds to the 5°C threshold value under
which frost is detected causing fruiting failure (see Supporting Information Tables S6 and S7 for results). (b) Mean annual acorn production as a function of
mean temperatures at >30 d after budburst (i.e. at the time of pollen release; see Fig. S6 for a similar relationship between airborne pollen amount and
mean temperature at >30 d after budburst date) (Table S9). The mean (SE) acorn amounts shown were computed within groups of 10 consecutive
site 9year values once being ranked according to their mean temperature (see Fig. S11, which shows the same relationship with ungrouped data). Shaded
area shows the 95% confidence interval of the model estimates.
Early pollen phenology
Population crop size
CVp = 1.39
SES = −2.46
PL M = 1
PM L = 0.034
Current pollen phenology
CVp = 1.23
SES = −4.23
PL M = 1
PM L = 0.25
Late pollen phenology
Time (yr) Time (yr)
Population crop size
CVp = 1.04
SES = −6.60
PL M = 1
PM L = 0.55
Very late pollen phenology
CVp = 0.99
SES = −8.98
PL M = 1
PM L = 0.81
Fig. 4 Examples of population fruiting
dynamics simulated by our resource budget
model over 100 yr depending of pollen
phenology: (a) earlier; (b) current; (c) later;
and (d) very late. The ‘population crop size’
axis corresponds to the index of the mean
amount of resource allocated to fruiting at
the population level. For each pollen
phenology scenario, the values of the
population coefficient of variation (CVp),
autocorrelation standardized effect size (SES)
and the probabilities P
of having a lean-
seeding year (L) the year following a mast-
seeding event (M) and P
of having a
mast-seeding event (M) the year following a
lean-seeding year (L) are indicated (see
legend of Fig. 5 for details on phenology or
masting parameters).
Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
New Phytologist (2019)
Phytologist Research 7
(the month preceding the budburst date), which is relatively rare
and which would only have minor effects on fruiting dynamics
(see Figs 3a, S9b). Although very early phenology would make
fruiting dynamics even more stochastic (e.g. by reinforcing the
‘environmental veto’ effect of late frost), it would probably be too
costly in terms of fitness, either by producing too rare mast-seed-
ing years or by impairing leaf growth (see later).
Mediterranean oak species are also exposed to greatly diverse
seed consumers whose control is also expected to require stochas-
tic fruiting dynamics. Although not studied here, the fruiting
dynamics of Mediterranean oak species are probably as variable
and stochastic as those of temperate oak species (Bogdziewicz
et al., 2017b). The late flower phenology of Mediterranean oaks
seems to promote weather conditions mainly favorable to pollina-
tion (Fig. 2b). In consequence, the stochastic component of
masting is likely to be independent of weather-driven pollination
failure, instead being a result of severe drought in spring or sum-
mer, resulting in frequent and very high fruit abortion rate (Fer-
nandez-Martinez et al., 2012; Perez-Ramos et al., 2015 (for a
review); Pearse et al., 2015; Bogdziewicz et al., 2017b). In
Mediterranean oak species, the late flower phenology could be a
way of not adding noise to the already very stochastic fruiting
dynamics and, consequently, keeping the frequency of mast-seed-
ing years at a minimum threshold. Overall, the proximate mecha-
nisms of oak masting (i.e. including environmental veto as a
result of late frost or water stress, weather conditions impacting
flower maturation and pollen diffusion) would probably depend
on the species, local ecological conditions and/or local adaptation
(Koenig et al., 2016). Considering this last point, the evolution
of flower phenology could be rapid as it is tightly linked to leaf
phenology, which is itself quickly evolving (Franjicet al., 2011)
and could thus be responsible for the short-term change in the
weight of late frost and weather conditions at the time of pollen
release in masting.
Our study, in line with previous work (Koenig et al., 2015;
Bogdziewicz et al., 2017b), underlines the need to elucidate the
interdependency between fruiting strategies (i.e. the interannual
dynamics of fruiting, possibly masting) and the phenology of
perennial plants. For example, in oak species, flower maturation
is organically linked to leaf maturation because most buds are
compound buds (i.e. containing leaves and flowers); leaf and
flower phenologies are thus tightly related (Koenig et al., 2012).
The evolution of flower phenology might thus be a by-product
of, and driven by, leaf phenology that would be predominantly
selected to maximize carbon gain through photosynthesis; in this
sense, the early flower phenology of temperate oak (and the
stochasticity induced in masting) would be an exaptation. Most
likely, the phenology of temperate oaks would result from a
tradeoff between the advantage of being early to trigger stochastic
fruiting dynamics and to lengthen the canopy duration, and the
advantage of being late to avoid exposing the nascent leaves to
late frost.
0.6 0.8 1 1.2 1.4 1.6 1.8 2
Cumulative frequencies of sites
Very late
−10 −9 −8 −7 −6 −5 −4 −3 −2 −1 0
Autocorrelation SES
0 0.2 0.4 0.6 0.8 1
Cumulative frequencies of sites
0 0.2 0.4 0.6 0.8 1
Fig. 5 Impact of an evolutionary shift in
flower phenology on the fruiting dynamics of
temperate oak species. Four flower
phenologies were tested for their impact on
fruiting dynamics through simulation with
the resource budget model. The current
phenology corresponds to the phenology
recorded in the field at the 30 fruit sites. The
early flower phenology corresponds to a 15 d
advance in the budburst date, and the late
and very late pollen phenologies correspond
to a 15 d (as observed for Mediterranean oak
species; see Fig. 1b) and a 30 d lag (as
observed for beeches in some temperate
forest communities) in the budburst date,
respectively. (ad) Cumulative frequency
distribution of sites for: (a) the population
coefficient of variation of fruiting (CVp); (b)
the autocorrelation ‘standardized effect size’
(SES), which reflects both the deterministic
and stochastic components of masting; (c)
the probability P
of having a lean-seeding
year (L) after a mast-seeding year (M); and
(d) the probability P
of having a mast-
seeding event (M) the year following a lean-
seeding year (L). Together these describe the
degree of asymmetry of the masting. The
polygons display the 95% credible interval
(i.e. including 95% of the simulations).
New Phytologist (2019) Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
More generally, it would be useful to develop integrative work
on phenology and masting through comparative approaches of
the dynamics of fruiting and phenology of flowers and leaves in
wind-pollinated perennial plant species. For example, in species
with separate flower and leaf buds, pollen phenology could be
even earlier, and fruiting dynamics more stochastic, than in other,
more constrained species. From a more theoretical perspective, it
might be worth combining several models, including those con-
sidering ecophysiological traits for their impact on plant phenol-
ogy (Chuine & Beaubien, 2001), those looking at the mechanistic
traits of fruiting dynamics (Isagi et al., 1997; Abe et al., 2016; this
work) for example, the evolution of traits affecting resource allo-
cation in reproduction and those dedicated to simulating seed
consumer dynamics (Rees et al., 2002; Tachiki & Iwasa, 2013), to
link explicitly the proximate causes of masting to fitness conse-
quences and plant regeneration success. Coupling these
approaches is all the more urgent as phenology is greatly affected
by climate change in a vast number of plant species, which could
impact their fruiting dynamics, the success of regeneration and
ultimately the assembly of perennial plant species in forest ecosys-
tems and the associated ecosystem services (Cleland et al., 2007).
We are grateful to the Reseau National de Surveillance Aerobi-
ologique (RNSA), France, which provided us with their oak
pollen database, and to the European network for the monitoring
of forest ecosystems (RENECOFOR) from the Office National
des For^ets (ONF) which provided us with the oak acorn census
database. The weather conditions were extracted from the Systeme
d’Analyse Fournissant des Renseignements Adaptes a la Nivologie
(SAFRAN) spatially explicit database from the Centre National
de la Recherche Meteorologique. Simulations were performed
using the computing cluster CC LBBE/PRABI. This research was
funded by the Federation Nationale des Chasseurs (FNC), France,
and the ONF. This work was performed within the framework of
the LABEX ECOFECT (ANR-11-LABX-0048) of Universitede
Lyon, the program Investissements d’Avenir (ANR-11-IDEX-
0007) and the program ANR FOREPRO operated by the French
National Research Agency (ANR).
Author contributions
ES, SV and M-C V conceived and led the study and wrote the
paper. ES and GO assembled the pollen dataset. ES, IC, Sylvain
Delzon and VB assembled the dataset of budburst dates in tem-
perate oak trees. ES and SV analyzed the field data and per-
formed the modeling. Stephane Dray analyzed autocorrelation
within simulated fruiting dynamics. IC, Sylvain Delzon, J-MG,
VB and ILR gave fruitful comments during the research process.
All authors revised the manuscript.
Marie-Claude Bel-Venner
Isabelle Chuine
Sylvain Delzon
Stephane Dray
Jean-Michel Gaillard
Iris Le Ronce
Eliane Schermer
Samuel Venner
Abe T, Tachiki Y, Kon H, Nagasaka A, Onodera K, Minamino K, Han Q,
Satake A. 2016. Parameterisation and validation of a resource budget model for
masting using spatiotemporal flowering data of individual trees. Ecology letters
19: 11291139.
Andersson B. 1992. Autumn frost hardiness of Pinus sylvestris offspring from seed
orchard grafts of different ages. Scandinavian Journal of Forest Research 7: 367
Augspurger CK. 2009. Spring 2007 warmth and frost: phenology, damage and
refoliation in a temperate deciduous forest. Functional Ecology 23: 10311039.
Badeau V, Bonhomme M, Bonne F, Carre J, Cecchini S, Chuine I, Ducatillion
C, Jean F, Lebourgeois F. 2017. Les plantes au rythmes des saisons, 336 pages.
Biotope, Meze, France.
Bates D, Maechler M, Bolker B, Walker S, Christensen RHB, Singmann H, Dai
B, Grothendieck G, Green P.2015. Package ‘lme4’. Convergence 12. [WWW
document] URL
Barringer BC, Koenig WD, Knops JMH. 2013. Interrelationships among life-
history traits in three California oaks. Oecologia 171: 129139.
Bogdziewicz M, Fernandez-Martınez M, Bonal R, Belmonte J, Espelta JM.
2017b. The Moran effect and environmental vetoes: phenological synchrony
and drought drive seed production in a Mediterranean oak. Proceedings of the
Royal Society B-Biological Sciences 284: 20171784.
Bogdziewicz M, Marino S, Bonal R, Zwolak R, Steele MA. 2018b. Rapid
aggregative and reproductive responses of weevils to masting of North
American oaks counteract predator satiation. Ecology 99: 25752582.
Bogdziewicz M, Steele MA, Marino S, Crone EE. 2018a. Correlated seed failure
as an environmental veto to synchronize reproduction of masting plants. New
Phytologist 219:98108.
Bogdziewicz M, Szymkowiak J. 2016. Oak acorn crop and Google search volume
predict Lyme disease risk in temperate Europe. Basic and Applied Ecology 17:
Bogdziewicz M, Szymkowiak J, Kasprzyk I, Grewling Ł, Borowski Z, Borycka
K, Kantorowicz W, Myszkowska D, Piotrowicz K, Ziemianin M et al. 2017a.
Masting in wind-pollinated trees: system specific roles of weather and
pollination dynamics in driving seed production. Ecology 98: 26152625.
Bogdziewicz M, Zwolak R, Crone EE. 2016. How do vertebrates respond to
mast seeding? Oikos 125: 300307.
Bogdziewicz M,
Zywiec M, Espelta JM, Fernandez-Martinez M, Calama R,
Ledwon M, McIntire E, Crone EE. 2019. Environmental veto synchronizes
mast seeding in four constrasting tree species. American Naturalist 194: 246
Buechling A, Martin PH, Canham CD, Shepperd WD, Battaglia MA. 2016.
Climate drivers of seed production in Picea engelmannii and response to
warming temperatures in the southern Rocky Mountains. Journal of Ecology
104: 10511062.
Buonaccorsi JP, Elkinton J, Koenig W, Duncan RP, Kelly D, Sork V. 2003.
Measuring mast seeding behavior: relationships among population variation,
individual variation and synchrony. Journal of Theoretical Biology 224:107114.
Caignard T, Kremer A, Firmat C, Nicolas M, Venner S, Delzon S. 2017.
Increasing spring temperatures favor oak seed production in temperate areas.
Scientific Reports 7: 8555.
Cecich RA, Sullivan NH. 1999. Influence of weather at time of pollination on
acorn production of Quercus alba and Quercus velutina.Canadian Journal of
Forest Research 29: 18171823.
Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
New Phytologist (2019)
Phytologist Research 9
Chang-Yang C-H, Sun I-F, Tsai C-H, Lu C-L, Hsieh C-F. 2016. ENSO and
frost codetermine decade long temporal variation in flower and seed production
in a subtropical rain forest. Journal of Ecology 104:4454.
Chuine I, Beaubien EG. 2001. Phenology is a major determinant of tree species
range. Ecology Letters 4: 500510.
Cleland EE, Chuine I, Menzel A, Mooney HA, Schwartz MD. 2007. Shifting
plant phenology in response to global change. Trends in Ecology and Evolution
22: 357365.
Crawley MJ. 2000. Seeds: the ecology of regeneration in plant communities. In:
Fenner M, ed. Seed predators and plant population dynamics, vol. 2. Wallingford,
UK: CABI, 167182.
Crone EE, Miller E, Sala A. 2009. How do plants know when other plants are
flowering? Resource depletion, pollen limitation and mast-seeding in a
perennial wildflower. Ecology Letters 12: 11191126.
Crone EE, Rapp JM. 2014. Resource depletion, pollen coupling, and the ecology
of mast seeding. Annals of the New York Academy of Sciences 1322:2134.
Dray S, Dufour A-B. 2007. The ade4 package: implementing the duality diagram
for ecologists. Journal of Statistical Software 22:120.
Dray S, Royer-Carenzi M, Calenge C. 2010. The exploratory analysis of
autocorrelation in animal-movement studies. Ecological Research 25: 673681.
Durand Y, Brun E, Merindol L, Guyomarc’h G, Lesaffre B, Martin E. 1993. A
meteorological estimation of relevant parameters for snow models. Annals of
Glaciology 18:6571.
Feret PP, Kreh RE, Merkle SA, Oderwald RG. 1982. Flower abundance,
premature acorn abscission, and acorn production in Quercus alba L. Botanical
Gazette 143: 216218.
Fernandez-Martınez M, Belmonte J, Espelta JM. 2012. Masting in oaks:
disentangling the effect of flowering phenology, airborne pollen load and
drought. Acta Oecologica 43:5159.
Franjic J, Sever K, Bogdan S,
Z, Krstonosic D, Aleskovic I. 2011.
Phenological asynchronization as a restrictive factor of efficient pollination in
clonal seed orchads of pedunculate oak (Quercus robur L.). Croatian Journal of
Forest Engineering: Journal for Theory and Application of Forestry Engineering 32:
Frey BR, Ashton MS, McKenna JJ, Ellum D, Finkral A. 2007. Topographic and
temporal patterns in tree seedling establishment, growth, and survival among
masting species of southern New England mixed-deciduous forests. Forest
Ecology and Management 245:5463.
Gamelon M, Douhard M, Baubet E, Gimenez O, Brandt S, Gaillard J-M. 2013.
Fluctuating food resources influence developmental plasticity in wild boar.
Biology Letters 9: 20130419.
Gamelon M, Focardi S, Baubet E, Brandt S, Franzetti B, Ronchi F, Venner S,
Sæther B-E, Gaillard J-M. 2017. Reproductive allocation in pulsed-resource
environments: a comparative study in two populations of wild boar. Oecologia
183: 10651076.
Garcıa-Mozo H, Pablo JH, Carmen G, Gomez-Casero MT, Domınguez E.
2001. Catkin frost damage in Mediterranean cork-oak Quercus suber L. Israel
Journal of Plant Sciences 49:4247.
Garcıa-Mozo H, Gomez-Casero MT, Domınguez E, Galan C. 2007. Influence
of pollen emission and weather-related factors on variations in holm-oak
(Quercus ilex subsp. ballota) acorn production. Environmental and Experimental
Botany 61:3540.
Gotelli NJ, McCabe DJ. 2002. Species co-occurrence: a meta-analysis of JM
Diamond’s assembly rules model. Ecology 83: 20912096.
Herrera CM. 1998. Population-level estimates of interannual variability in seed
production: what do they actually tell us? Oikos 82: 612616.
Herrera CM, Jordano P, Guitian J, Traveset A. 1998. Annual variability in seed
production by woody plants and the masting concept: reassessment of
principles and relationship to pollination and seed dispersal. American
Naturalist 152: 576594.
Hirst JM. 1952. An automatic volumetric spore trap. Annals of Applied Biology
39: 257265.
Huber PJ. 1964. Robust estimation of a location parameter. Annals of
Mathematical Statistics 35:73101.
Isagi Y, Sugimura K, Sumida A, Ito H. 1997. How does masting happen and
synchronize? Journal of Theoretical Biology 187: 231239.
Janzen DH. 1976. Why bamboos wait so long to flower. Annual Review of Ecology
and Systematics 7: 347391.
Kelly D, Sork VL. 2002. Mast seeding in perennial plants: why, how, where?
Annual Review of Ecology and Systematics 33: 427447.
Kelly D, Geldenhuis A, James A, Holland EP, Planck MJ, Brockie RE, Cowan
PE, Harper GA, Lee WG, Maitland MJ et al. 2013. Of mast and mean:
differential-temperature cue makes mast seeding insensitive to climate change.
Ecology Letters 16:9098.
Koenig WD, Knops JMH. 2000. Patterns of annual seed production by northern
hemisphere trees: a global perspective. American Naturalist 155:5969.
Koenig WD, Kelly D, Sork VL, Duncan RP, Elkinton JS, Peltonen MS,
Westfall RD. 2003. Dissecting components of population-level variation in
seed production and the evolution of masting behavior. Oikos 102: 581591.
Koenig WD, Knops JMH, Carmen WJ.2008. Timing of flowering and seed
production in three California oaks. In: Merenlender A, McCreary D,
Purcell KL, eds. Proceedings of the Sixth Symposium on Oak Woodlands:
Today’s Challenges, Tomorrow’s Opportunities. Pacific Southwest Forest and
Range Experiment Station General Technical Report PSW-GTR-217, 371
Koenig WD, Funk KA, Kraft TS, Carmen WJ, Barringer BC, Knops JMH.
2012. Stabilizing selection for within-season flowering phenology confirms
pollen limitation in a wind-pollinated tree. Journal of Ecology 100: 758
Koenig WD, Knops JMH, Carmen WJ, Pearse IS. 2015. What drives masting?
The phenological synchrony hypothesis. Ecology 96: 184192.
Koenig WD, Alejano R, Carbonero MD, Fernandez-Rebollo P, Knops JM,
Mara~non T, Padilla-Diaz CM, Pearse IS, Perez-Ramos IM, Vȧzquez-PiqueJ,
Pesendorfer MB. 2016. Is the relationship between mast-seeding and weather
in oaks related to their life-history or phylogeny? Ecology 97: 26032615.
Lebourgeois F, Pierrat J-C, Perez V, Piedallu C, Cecchini S, Ulrich E. 2008.
Determinisme de la phenologie des for^ets temperees franc
ßaises: etude sur les
peuplements du reseau RENECOFOR. Revue Forestie re Franc
ßaise 3: 323343.
Leiva MJ, Perez-Romero JA, Mateos-Naranjo E. 2018. The effect of simulated
damage by weevils on Quercus ilex subsp Ballota acorns germination, seedling
growth and tolerance to experimentally induced drought. Forest Ecology and
Management 409: 740748.
McShea WJ. 2000. The influence of acorn crops on annual variation in rodent
and bird populations. Ecology 81: 228238.
Meier U, Bleiholder H, Buhr L, Feller C, Hack H, Heß M, Lancashire PD,
Schnock U, Stauß R, Van den Boom T, et al. 2009. The BBCH system to
coding the phenological growth stages of plantshistory and publications.
Journal fur Kulturpflanzen 61:4152.
Misson L, Degueldre D, Collin C, Rodriguez R, Rocheteau A, Ourcival JM,
Rambal S. 2011. Phenological responses to extreme droughts in a
Mediterranean forest. Global Change Biology 17: 10361048.
Monks A, Kelly D. 2006. Testing the resource-matching hypothesis in the mast
seeding tree Nothofagus truncata (Fagaceae). Austral Ecology 31: 366375.
Mu~noz A, Bonal R, Espelta JM. 2014. Acorn weevil interactions in a mixed-
oak forest: outcomes for larval growth and plant recruitment. Forest Ecology and
Management 322:98105.
Nussbaumer A, Waldner P, Apuhtin V, Aytar F, Benham S, Bussotti F,
Eichhorn J, Eickenscheidt N, Fabianek P, Falkenried L et al. 2018. Impact of
weather cues and resource dynamics on mast occurrence in the main forest tree
species in Europe. Forest Ecology and Management 429: 336350.
Ogaya R, Pe~nuelas J. 2004. Phenological patterns of Quercus ilex,Phillyrea
latifolia, and Arbutus unedo growing under a field experimental drought.
E coscience 11: 263270.
Ostfeld RS, Keesing F. 2000. Pulsed resources and community dynamics of
consumers in terrestrial ecosystems. Trends in Ecology & Evolution 15: 232
Pearse IS, Koenig WD, Funk KA, Pesendorfer MB. 2015. Pollen limitation and
flower abortion in a wind-pollinated, masting tree. Ecology 96: 587593.
Pearse IS, Koenig WD, Kelly D. 2016. Mechanisms of mast seeding: resources,
weather, cues, and selection. New Phytologist 212: 546562.
Pearse IS, Koenig WD, Knops JMH. 2014. Cues versus proximate drivers:
testing the mechanism behind masting behavior. Oikos 123: 179184.
New Phytologist (2019) Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
Pelisson P-F, Bel-Venner M-C, Rey B, Burgevin L, Martineau F, Fourel F,
Lecuyer C, Menu F, Venner S. 2012. Contrasted breeding strategies in four
sympatric sibling insect species: when a proovigenic and capital breeder copes
with a stochastic environment. Functional Ecology 26: 198206.
Pelisson P-F, Bernstein C, Franc
ßois D, Menu F, Venner S. 2013. Dispersal and
dormancy strategies among insect species competing for a pulsed resource.
Ecological Entomology 38: 470477.
Perez-Ramos IM, Padilla-Dıaz CM, Koenig WD, Mara~non T. 2015.
Environmental drivers of mast-seeding in Mediterranean oak species: does leaf
habit matter? Journal of Ecology 103: 691700.
Pesson P, Louveaux J. 1984. Pollinisation et productions vegetales. Paris, France:
INRA, Quae.
Rees M, Kelly D, Bjørnstad ON. 2002. Snow tussocks, chaos, and the evolution
of mast seeding. American Naturalist 160:4459.
Rey B, Pelisson P-F, Bel-Venner M-C, Voituron Y, Venner S. 2015. Revisiting
the link between breeding effort and oxidative balance through field evaluation
of two sympatric sibling insect species. Evolution 69: 815822.
Richardson SJ, Allen RB, Whitehead D, Carswell FE, Ruscoe WA, Platt KH.
2005. Climate and net carbon availability determine temporal patterns of seed
production by Nothofagus.Ecology 86: 972981.
Sabit M, Ramos JD, Alejandro GJ, Galan C. 2016. Seasonal distribution of
airborne pollen in Manila, Philippines, and the effect of meteorological factors
to its daily concentrations. Aerobiologia 32: 375383.
Satake A, Iwasa YOH. 2000. Pollen coupling of forest trees: forming
synchronized and periodic reproduction out of chaos. Journal of Theoretical
Biology 203:6384.
Satake A, Iwasa Y. 2002. The synchronized and intermittent reproduction of
forest trees is mediated by the Moran effect, only in association with pollen
coupling. Journal of Ecology 90: 830838.
E, Bel-Venner M-C, Fouchet D, Siberchicot A, Boulanger V,
Caignard T, Thibaudon M, Oliver G, Manuel N, Gaillard J-M, et al. 2019.
Pollen limitation as a main driver of fruiting dynamics in oak populations.
Ecology Letters 22:98107.
Smaill SJ, Clinton PW, Allen RB, Davis MR. 2011. Climate cues and resources
interact to determine seed production by a masting species. Journal of Ecology
99: 870877.
Sork VL, Bramble J, Sexton O. 1993. Ecology of mast-fruiting in three species of
North American deciduous oaks. Ecology 74: 528541.
Tachiki Y, Iwasa Y. 2013. Coevolution of mast seeding in trees and extended
diapause of seed predators. Journal of Theoretical Biology 339: 129139.
Ulrich E. 1995. Le reseau RENECOFOR: objectifs et realisation. Revue Forestiere
ßaise 2: 107124.
Vacchiano G, Ascoli D, Berzaghi F, Lucas-Borja ME, Caignard T, Collalti A,
Mairota P, Palaghianu C, Reyer CPO, Sanders TGM, et al. 2018.
Reproducing reproduction: How to simulate mast seeding in forest models.
Ecological Modelling 376:4053.
Venner S, Pelisson P-F, Bel-Venner M-C, Debias F, Rajon E, Menu F. 2011.
Coexistence of insect species competing for a pulsed resource: toward a unified
theory of biodiversity in fluctuating environments. PLoS ONE 6: e18039.
Venner S, Siberchicot A, Pelisson P-F, Schermer E, Bel-Venner M-C, Nicolas
M, Debias F, Miele V, Sauzet S, Boulanger V, et al. 2016. Fruiting strategies
of perennial plants: a resource budget model to couple mast seeding to
pollination efficiency and resource allocation strategies. American Naturalist
Vitasse Y, Delzon S, Dufr^ene E, Pontailler J-Y, Louvet J-M, Kremer A, Michalet
R. 2009. Leaf phenology sensitivity to temperature in European trees: Do
within-species populations exhibit similar responses? Agricultural and Forest
Meteorology 149: 735744.
Wald A, Wolfowitz J. 1943. An exact test for randomness in the non-parametric
case based on serial correlation. Annals of Mathematics and Statistics 14: 378
Wolff JO. 1996. Coexistence of white-footed mice and deer mice may be
mediated by fluctuating environmental conditions. Oecologia 108: 529533.
Yang LH, Edwards KF, Byrnes JE, Bastow JL, Wright AN, Spence KO. 2010. A
meta-analysis of resource pulse-consumer interactions. Ecological Monographs
80: 125151.
Yi X, Bartlow AW, Curtis R, Agosta SJ, Steele MA. 2019. Responses of seedling
growth and survival to post-germination cotyledon removal: An investigation
among seven oak species. Journal of Ecology 107: 18171827.
Supporting Information
Additional Supporting Information may be found online in the
Supporting Information section at the end of the article.
Fig. S1 Spatial distribution in the temperate region of the 30
acorn-sampling sites.
Fig. S2 Relationship between budburst date and March mean
daily temperature in the temperate region.
Fig. S3 Sensitivity analysis of the resource budget model outputs.
Fig. S4 Relationships between the median date of oak pollen
release and the latitude.
Fig. S5 Deviance profile from the ‘threshold model’ applied to
the logistic relationships between mean airborne pollen amount
and mean April temperature in the temperate region.
Fig. S6 Logistic relationship between the mean amount of air-
borne pollen and mean temperature at >30 d after budburst date
in the temperate region.
Fig. S7 Analysis of the relative contribution of the late frost and
the mean temperature impacting pollen aerial diffusion in the
fruiting dynamics under four different simulated pollen pheno-
logical scenarios.
Fig. S8 Sensitivity analysis of the resource budget model outputs
to the female flower allocation.
Fig. S9 Distribution of the March minimum temperatures
depending on temperate and Mediterranean regions.
Fig. S10 Relationships between airborne pollen amount and
spring temperatures depending on temperate and Mediterranean
Fig. S11 Response of both airborne pollen amount and acorn
production to temperatures at >30 d after budburst date in the
temperate region.
Method S1 Detailed description of the resource budget model.
Table S1 Characteristics of the pollen-sampling sites.
Table S2 Characteristics of the acorn-sampling sites.
Table S3 Response of budburst date to March mean daily tem-
Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
New Phytologist (2019)
Phytologist Research 11
Table S4 Proportion of variance explained by the various
synthetic weather variables reflecting both temperature and
Table S5 Response of airborne pollen amount to spring
weather variables depending on temperate and Mediterranean
Table S6 Exploring the sensitivity of fruiting intensity to weather
conditions around the budburst date.
Table S7 Testing the sensitivity of fruiting intensity to weather
conditions around the budburst date.
Table S8 Model selection between the various models explaining
the airborne pollen amount in the temperate region.
Table S9 Model selection between the logistic model and poly-
nomial regression models for predicting airborne pollen amount
or fruit production from spring weather conditions in the tem-
perate region.
Please note: Wiley Blackwell are not responsible for the content
or functionality of any Supporting Information supplied by the
authors. Any queries (other than missing material) should be
directed to the New Phytologist Central Office.
New Phytologist is an electronic (online-only) journal owned by the New Phytologist Trust, a not-for-profit organization dedicated
to the promotion of plant science, facilitating projects from symposia to free access for our Tansley reviews and Tansley insights.
Regular papers, Letters, Research reviews, Rapid reports and both Modelling/Theory and Methods papers are encouraged.
We are committed to rapid processing, from online submission through to publication ‘as ready’ via Early View – our average time
to decision is <26 days. There are no page or colour charges and a PDF version will be provided for each article.
The journal is available online at Wiley Online Library. Visit to search the articles and register for table
of contents email alerts.
If you have any questions, do get in touch with Central Office ( or, if it is more convenient,
our USA Office (
For submission instructions, subscription and all the latest information visit
New Phytologist (2019) Ó2019 The Authors
New Phytologist Ó2019 New Phytologist Trust
... Unsurprisingly, the CV of alternate-bearing crops reported here at national scales is smaller than has been observed for both masting and alternate-bearing taxa at population scales [6,35]. In general, alternate bearing leads to lower CVs than mast-seeding at longer intervals [38,50]. More importantly, we expect that in crop plants, as in wild plants, synchrony should decay with distance, and, at the present time, the scale of synchrony in crop plants is largely unknown. ...
... Environmental vetoes-external conditions that prevent seed set-have been well-supported as a driver of synchrony in masting systems [54,55]. As a recent example, Schermer et al. [50] studied frost-induced fruit losses in relationship to flower phenology (mean or Phil. Trans. ...
Cyclical fluctuations in reproductive output are widespread among perennial plants, from multi-year masting cycles in forest trees to alternate bearing in horticultural crops. In natural systems, ecological drivers such as climate and pollen limitation can result in synchrony among plants. Agricultural practices are generally assumed to outweigh ecological drivers that might synchronize alternate-bearing individuals, but this assumption has not been rigorously assessed and little is known about the role of pollen limitation as a driver of synchrony in alternate-bearing crops. We tested whether alternate-bearing perennial crops show signs of alternate bearing at a national scale and whether the magnitude of national-scale alternate bearing differs across pollination syndromes. We analysed the Food and Agriculture Organization of the United Nations time series (1961–2018) of national crop yields across the top-producing countries of 27 alternate-bearing taxa, 6 wind-pollinated and 21 insect-pollinated. Alternate bearing was common in these national data and more pronounced in wind-pollinated taxa, which exhibited a more negative lag-1 autocorrelation and a higher coefficient of variation (CV). We highlight the mutual benefits of integrating ecological theory and agricultural data for (i) advancing our understanding of perennial plant reproduction across time, space and taxa, and (ii) promoting stable farmer livelihoods and global food supply. This article is part of the theme issue ‘The ecology and evolution of synchronized seed production in plants’.
... Trade-off between growth and reproduction In Quercus, positive correlation with the period of pollen emission and spring temperature (April). Positive relation between growth and reproduction Schermer et al. (2019) Q. robur, Q. petraea Positive correlation with amount of airborne pollen and its sensibility to spring weather Schermer et al. (2020) Q. robur, Q. petraea Positive effect of spring temperature and negative effect of rainfall. Pollen release sensitive to late frost and fruit production from 1997 to 2007 using linear mixed models (LMMs) implemented via the glmmTMB package (Brooks et al. 2017). ...
... Moreover, other direct climatic effects, such as drought, are also linked to episodic fruit production (Pérez-Ramos et al. 2010;Lebourgeois et al. 2018). Schermer et al. (2019) demonstrated that pollen dynamics was the key driver of fruit production in Quercus petraea and Q. robur, and Lebourgeois et al. (2018) showed that spring temperature was important for oak fruit production for the same dataset, due to its positive effects on pollination (Caignard et al. 2017;Schermer et al. 2020). Spring temperature during year of fruit maturation is an important determinant of fruit production, but also a major determinant of leaf season (Vitasse et al. 2009b). ...
Full-text available
Phenology and fruit production are sensitive to climate. Variation of leaf phenology alters canopy duration (an indicator of growing season length), which in turn affects forest ecosystem functioning and tree productivity. However, the influence of canopy duration on tree reproduction is poorly explored. In this study, we investigated if and to what extent the canopy duration influences fruit production in three major European deciduous trees: European beech (Fagus sylvatica) and two temperate oaks (Quercus robur and Quercus petraea). We used a long-term (11 years) monitoring dataset from 50 populations. In addition to the widely reported main effect of previous summer temperature on fruit production in European beech, we detected a nonlinear relationship between current canopy duration and fruit production. For beech, intermediate canopy duration is associated with highest fruit production. In oaks, fruit production was positively correlated with spring temperature and current canopy duration. Total variance explained by our models is 45% and 13%, for beech and oaks, respectively. These results suggest contrasting effects of expected increase in canopy duration on fruit production: nonlinear in European beech and positive linear in oaks.
... In contrast, global warming under the photoperiod-sensitivity hypothesis would lead to a lower frequency of years favorable for pollination, because warmer years would generally advance flowering in some individuals and thus desynchronize it at a population level. More frequent failures of pollination will increase the stochasticity and interannual variability of seed production Schermer et al., 2020). A better understanding of the ecophysiological processes controlling flowering phenology is thus essential for improving our understanding of the responses of trees and forests to the ongoing climate change. ...
... Flowering synchrony drives pollen limitation in oaks, which in turn is believed to interact with the dynamics of plant resources in driving mast seeding Pesendorfer et al., 2016;Schermer et al., 2019). Desynchronized flowering vetoes reproduction, which forces plants to conserve resources for subsequent years, so more frequent interference leads to more stochastic and variable patterns of reproduction Schermer et al., 2020). To the extent that phenological synchrony is involved in determining variable seed production, global warming under the microclimatic hypothesis is predicted to lead to less frequent vetoes (more frequent high flowering synchrony years), thereby decreasing masting intensity . ...
Full-text available
Annually variable and synchronous seed production, or masting, is often correlated with environmental factors and in oaks involves differential pollination success that depends on phenological synchrony in flowering. The synchronization of phenology of flowering was thought to be driven by temperature during flowering (micro-climatic hypothesis). We tested an alternative, whereby phenological synchronization is driven by the timing of the onset of flowering (photoperiod-sensitivity hypothesis). This hypothesis assumes that flowering synchrony is driven by interaction between daylength and temperature, and individual variation in sensitivity to daylength as a phenological cue. We used long-term (23-26 years) records of airborne pollen in Quercus robur, Q. petraea, Q. ilex, and Q. humilis. Late pollen seasons were short, as predicted by photoperiod-sensitivity hypothesis. The onset of pollen seasons was delayed as preseason temperatures cooled over the last three decades at our Mediterranean sites, which was paralleled by shortening in pollen seasons, providing additional support for the photoperiod-sensitivity hypothesis. Global warming under the microclimatic hypothesis is predicted to lead to less frequent reproductive failures and thus decreased variability and synchrony of mast seeding. In contrast, warming under the photoperiod-sensitivity hypothesis should advance the onset of and desynchronize flowering, a pattern supported by our data. This pattern suggests that global warming will lead to more frequent vetoes and more stochastic and variable patterns of oak reproduction.
... The effects of spring frost on temperate oak species phenology have been well documented. They are associated with the early timing of flowering, potentially resulting in reproductive failure [79], since there is a high probability that the pollen release process will occur during adverse meteorological conditions. Therefore, the unfavourable climatic conditions in 2017 during pollen release likely harmed the flowering process, interrupting microsporogenesis and the elongation of the catkins, leading to their death [80]. ...
Full-text available
Due to the visible and predictable influence of climate change on species’ spatial distributions, the conservation of marginal peripheral populations has become topical in forestry research. This study aimed to assess the spring (budburst, leaf development, and flowering) and autumn (leaf senescence) phenology of sessile oak (Quercus petraea), a species widespread across European forests close to its ranges’ eastern limit. This study was performed in Romania between spring 2017 and 2020, and it included a transect with three low-altitude populations, a reference population from its inner range, and a sessile oak comparative trial. The temperature was recorded to relate changes to phenophase dynamics. We identified small variations between the reference and peripheral populations associated with climatic conditions. In the peripheral populations, budburst timing had day-of-year (DOY) values <100, suggesting that sessile oak may be more susceptible to late spring frost. Furthermore, we found spring phenophase timing to be more constant than autumn senescence. Moreover, budburst in the sessile oak comparative trial had obvious longitudinal tendencies, with an east to west delay of 0.5–1.4 days per degree. In addition, budburst timing influenced leaf development and flowering, but not the onset of leaf senescence. These findings improve our understanding of the relationship between spring and autumn phenophase dynamics and enhance conservation strategies regarding sessile oak genetic resources.
... That paper was published when I was writing my first research proposals, so it defined my research program. I also loved Schermer et al. (2019), which links the interannual weather variation, phenology, and stochasticity of masting events. That paper made a mark on my thinking and inspired some new projects, including my Tansley insight (Bogdziewicz, 2022). ...
... Moreover, given the important role of water availability in regulating plant reproductive processes (Bykova et al., 2019;Meng et al., 2016;Nam and Kim, 2020), experimental warming could increase evapotranspiration and decrease soil water availability (Quan et al., 2018;Wan, 2012, 2013), thus reducing the relative growth rate and thereby shortening the flowering season of wind-pollinated species in this study. The lengthened flowering seasons of insect-pollinated species may be advantageous for reproductive success and increase the dominance of these plants in the community Schermer et al., 2020), and vice versa for wind-pollinated species (Kudo and Ida, 2013;Maglianesi et al., 2020). The opposite response between forb and grass species to the flowering season under experimental warming in this synthesis may be attributed to the difference in phenological sensitivity and functional traits of these species in response to climate warming. ...
Climate warming potentially changes the flowering seasons of terrestrial plants, and thus species interaction, with consequently substantial impacts on ecosystem structure and function. However, the general response patterns of flowering seasons to warming and the underlying mechanisms remain poorly understood. Here, a meta-analysis of data from 26 experimental studies examining 168 species was conducted to quantify the responses of the flowering seasons of terrestrial plants to experimental warming. The results showed that experimental warming prolonged the flowering season by 2.08% across all species included in this study. In addition, flowering season responses were dependent on plant functional types, with a significant extension in herbaceous species (+2.18%) but no change in woody species. The warming impacts on the flowering season of wind-pollinated (-4.53%) and insect-pollinated species (+4.21%) were opposite. Among herbaceous species, the flowering seasons of forb (+4.47%) and specifically legume species (+15.06%) were positive, whereas grass species (-4.53%) showed negative responses to experimental warming. Moreover, experimental warming effects on the flowering season showed quadratic relationships with the latitude and the mean annual temperature but did not change with the mean annual precipitation. The responses of the flowering season to experimental warming also differed in terms of the warming magnitude. These diverse findings indicate the need for additional experimental warming experiments, especially for underrepresented plant functional groups, to better understand the mechanistic relationships between phenology and temperature under future climate warming scenarios.
... temperature), but rather on recognizable environmental patterns (e.g. [7,86,97]) that are independent of specific site conditions or climate variability, including ongoing climate change (but see [98,99]). This could generate hypotheses on the evolution of masting under disturbances that would be testable within and across different biomes. ...
The timing of seed production and release is highly relevant for successful plant reproduction. Ecological disturbances, if synchronized with reproductive effort, can increase the chances of seeds and seedlings to germinate and establish. This can be especially true under variable and synchronous seed production (masting). Several observational studies have reported worldwide evidence for co-occurrence of disturbances and seed bumper crops in forests. Here, we review the evidence for interaction between disturbances and masting in global plant communities; we highlight feedbacks between these two ecological processes and posit an evolutionary pathway leading to the selection of traits that allow trees to synchronize seed crops with disturbances. Finally, we highlight relevant questions to be tested on the functional and evolutionary relationship between disturbances and masting. This article is part of the theme issue ‘The ecology and evolution of synchronized seed production in plants’.
Full-text available
Forest ecology traditionally focuses on plant growth and survival, leaving seed production as a major demographic process lacking a framework for how it will be affected by global change. Understanding plant reproductive responses to changing climate is complicated by masting, annually variable seed production synchronized within populations. Predicting trends in masting is crucial because masting impacts seed predation and pollination enough to over‐ride simple trends in mean seed production. I gather proximate mechanisms of seed production patterns in perennial plants to identify processes through which masting may be affected by a changing environment. Predicting trends in masting will require understanding the mechanisms that cause predictable seed failure after high‐seed years, and the stochastic mechanisms that synchronize individuals in high‐seed years.
Phenology is a major component of an organism's fitness. While individual phenological events affect fitness, growing evidence suggests that the relationship between events may be equally or more important. This may explain why temperate deciduous woody plants exhibit considerable variation in the order of reproductive and vegetative events, or ower‐leaf sequences (FLSs). There is evidence to suggest that FLSs may be adaptive, with several competing hypotheses to explain their function. Here, we advance existing hypotheses with a new framework that accounts for quantitative FLS variation at multiple taxonomic scales using case studies from temperate forests. Our inquiry provides several major insights towards a better understanding of FLS variation. First, we show that support for FLS hypotheses is sensitive to how FLSs are defined, with quantitative definitions being the most useful for robust hypothesis testing. Second, we demonstrate that concurrent support for multiple hypotheses should be starting point for future FLS analyses. Finally, we highlight how adopting a quantitative, intra‐specific approach generates new avenues for evaluating fitness consequences of FLS variation and provides cascading benefits to improving predictions of how climate change will alter FLSs and thereby re‐shape plant communities and ecosystems.
Full-text available
1. Rodents regularly rely on emerged epicotyls to locate and remove cotyledons still containing valuable nutrients. However, the extent to which acorn characteristics influence tolerance to post-germination predation has received little attention. 2. Here, we investigated the impact of cotyledon removal following epicotyl emergence on seedling performance and survival of seven oak (Quercus) species. We imitated cotyledon predation at different stages of seedling establishment and development in order to detect effects on seedling height, leaf number, and tissue/component mass. 3. Seedling growth and survival were negatively affected by cotyledon loss regardless of oak species. However, these negative effects decreased as the epicotyl length at which cotyledons were removed increased. We also found that there was a threshold epicotyl length above which seedling survival and performance were relatively unaffected in white oak species compared to red oak species. 4. Following cotyledon removal, early germinating white oak (section Quercus) seedlings survived and/or grew better than the late germinating red oak (section Lobatae) seedlings. This was likely caused by a difference in dependence on cotyledon reserves, which ultimately affected the ability of seedlings to tolerate cotyledon removal. 5. Synthesis From an evolutionary perspective, this is likely to follow from the early germination in white oaks and the ability of seed consumers to locate young seedlings from the emerging epicotyls. Our study has implications for forest regeneration by suggesting additional opportunities for white oak species to establish following epicotyl emergence. Future studies should consider quantifying the rates of post-germination cotyledon loss.
Full-text available
In many perennial wind‐pollinated plants, the dynamics of seed production is commonly known to be highly fluctuating from year to year and synchronised among individuals within populations. The proximate causes of such seeding dynamics, called masting, are still poorly understood in oak species that are widespread in the northern hemisphere, and whose fruiting dynamics dramatically impacts forest regeneration and biodiversity. Combining long‐term surveys of oak airborne pollen amount and acorn production over large‐scale field networks in temperate areas, and a mechanistic modelling approach, we found that the pollen dynamics is the key driver of oak masting. Mechanisms at play involved both internal resource allocation to pollen production synchronised among trees and spring weather conditions affecting the amount of airborne pollen available for reproduction. The sensitivity of airborne pollen to weather conditions might make oak masting and its ecological consequences highly sensitive to climate change.
Full-text available
The predator satiation hypothesis posits that masting helps plants escape seed predation through starvation of predators in lean years, followed by satiation of predators in mast years. Importantly, successful satiation requires sufficiently delayed, bottom-up effects of seed availability on seed consumers. However, some seed consumers may be capable of quick aggregative and reproductive responses to masting which may jeopardize positive density-dependence of seed survival. We used a 17-year data set on seed production and insect (Curculio weevils) infestation of three North American oaks species (northern red Quercus rubra, white Q. alba, and chestnut oak Q. montana) to test predictions of the predation satiation hypothesis. Furthermore, we tested for the unlagged numerical response of Curculio to acorn production. We found that masting results in a bottom-up effect on the insect population; both through increased reproductive output and aggregation at seed-rich trees. Consequently, mast seeding in two out of three studied oaks (white and chestnut oak) did not help to escape insect seed predation, whereas in the red oak, the escape depended on the synchronization of mast crops within the population. Bottom-up effects of masting on seed consumer populations are assumed to be delayed, and therefore to have negligible effects on seed survival in mast years. Our research suggests that insect populations may be able to mount rapid reproductive and aggregative responses when seed availability increases, possibly hindering satiation effects of masting. Many insect species are able to quickly benefit from pulsed resources, making mechanisms described here potentially relevant in many other systems.
Full-text available
Mast seeding, the synchronised occurrence of large amounts of fruits and seeds at irregular intervals, is a reproductive strategy in many wind-pollinated species. Although a series of studies have investigated mast year (MY) patterns in European forest tree species at the regional scale, there are few recent evaluations at a European scale on the impact of weather variables (weather cues) and resource dynamics on mast behaviour. Thus the main objective of this study is to investigate the impact of specific weather conditions, as environmental drivers for MYs, on resources in Fagus sylvatica L., Quercus petraea (MATT.) LIEBL., Quercus robur L., Picea abies (L.) KARST. and Pinus sylvestris L. at a European level and to explore the robustness of the relationships in smaller regions within Europe. Data on seed production originating from the International Co-operative Programme on Assessment and Monitoring of Air Pollution Effects on Forests (ICP Forests) were analysed. Three beta regression models were applied to investigate the impact of seasonal weather variables on MY occurrence, as well as the influence of fruiting intensity levels in the years prior to MYs. Resource dynamics are analysed at three different spatial scales (continent, countries and ecoregions). At a European scale, important weather cues for beech MYs were a cold and wet summer two years before a MY, a dry and warm summer one year before a MY and a warm spring in the MY. For spruce, a cold and dry summer two years prior to a MY and a warm and dry summer in the year before the MY showed the strongest associations with the MY. For oak, high spring temperature in the MY was the most important weather cue. For beech and spruce, and to some extent also for oak species, the best fitting models at European scale were well reflected by those found at smaller scales. For pine, best fitting models were highly diverse concerning weather cues. Fruiting levels were high in all species two years before the MY and also high one year before the MY in the oak species and in pine. In beech, fruiting levels one year before the MY were not important and in spruce, they were inconsistent depending on the region. As a consequence, evidence of resource depletion could only be seen in some regions for spruce.
Full-text available
Variable, synchronized seed production, called masting, is a widespread reproductive strategy in plants. Resource dynamics, pollination success, and, as described here, environmental veto, are possible proximate mechanisms driving masting. We explored the environmental veto hypothesis, which assumes that reproductive synchrony is driven by external factors preventing reproduction in some years, by extending the resource budget model of masting with correlated reproductive failure. We ran this model across its parameter space to explore how these parameters interact to drive seeding dynamics. Next, we parameterized the model based on 16-years of seed production data for populations of red (Quercus rubra) and white (Q. alba) oaks. We used these empirical models to simulate seeding dynamics, and compared simulated time series to patterns observed in the field. Simulations showed that resource dynamics and reproduction failure can produce masting even in the absence of pollen coupling. In concordance, in both oaks, among-year variation in resource gain and correlated reproductive failure were necessary and sufficient to reproduce masting, whereas pollen coupling, although present, was not necessary. Reproductive failure caused by environmental veto may drive large-scale synchronization without density-dependent pollen limitation. Reproduction-inhibiting weather events are prevalent in ecosystems, making described mechanisms likely to operate in many systems.
Full-text available
Masting is the highly variable production of synchronized seed crops, and is a common reproductive strategy in plants. Weather has been long recognized as centrally involved in driving seed production in masting plants. However, the theory behind mechanisms connecting weather and seeding variation has only recently been developed, and still lacks empirical evaluation. We used 12-years long seed production data for 255 holm oaks (Quercus ilex), as well as airborne pollen and meteorological data, and tested whether masting is driven by environmental constraints: phenology synchrony and associated pollination efficiency, and drought-related acorn abscission. We found that warm springs resulted in short pollen seasons, and length of the pollen seasons was negatively related to acorn production, supporting the phenology synchrony hypothesis. Furthermore, the relationship between phenology synchrony and acorn production was modulated by spring drought, and effects of environmental vetoes on seed production were dependent on the last year environmental constraint, implying passive resource storage. Both vetoes affected large-scale synchrony in seed production. Finally, precipitation preceding acorn maturation was positively related to seed production, mitigating apparent resource depletion following high crop production in the previous year. These results provide new insights into mechanisms beyond widely reported weather and seed production correlations.
Synchronized and variable reproduction by perennial plants, called mast seeding, is a major reproductive strategy of trees. The need to accumulate sufficient resources after depletion following fruiting (resource budget), the efficiency of mass flowering for outcross pollination (pollen coupling), or the external factors preventing reproduction (environmental veto) could all synchronize masting. We used seed production data for four species (Quercus ilex, Quercus humilis, Sorbus aucuparia, and Pinus albicaulis) to parametrize resource budget models of masting. Based on species life-history characteristics, we hypothesized that pollen coupling should synchronize reproduction in S. aucuparia and P. albicaulis, while in Q. ilex and Q. humilis, environmental veto should be a major factor. Pollen coupling was stronger in S. aucuparia and P. albicaulis than in oaks, while veto was more frequent in the latter. Yet in all species, costs of reproduction were too small to impose a replenishment period. A synchronous environmental veto, in the presence of environmental stochasticity, was sufficient to produce observed variability and synchrony in reproduction. In the past, vetoes like frost events that prevent reproduction have been perceived as negative for plants. In fact, they could be selectively favored as a way to create mast seeding.
Masting is the highly variable and synchronous production of seeds by plants. Masting can have cascading effects on plant population dynamics and forest properties such as tree growth, carbon stocks, regeneration, nutrient cycling, or future species composition. However, masting has often been missing from forest models. Those few that simulate masting have done so using relatively simple empirical rules, and lack an implementation of process-based mechanisms that control such events. Here we review more than 200 published papers on mechanistic formulations of masting, and summarize how the main processes involved in masting and their related patterns can be incorporated in forest models at different degrees of complexity. Our review showed that, of all proximate causes of masting, resource acquisition, storage and allocation were the processes studied most often. Hormonal and genetic regulation of bud formation, floral induction, and anthesis were less frequently addressed. We outline the building blocks of a general process-based model of masting that can be used to improve the oversimplified functions in different types of forest models, and to implement them where missing. A complete implementation of masting in forest models should include functions for resource allocation and depletion, and for pollination, as regulated by both forest structure and weather in the years prior to seed production. When models operate at spatio-temporal scales mismatched with the main masting processes, or if calibration data are not available, simulation can be based on parameterizing masting patterns (variability, synchrony, or frequency). Also, observed masting patterns have the potential to be used as “reality checks” for more process-based forest models wishing to accurately reproduce masting as an emergent phenomenon.
The main aim of this study is to assess the effect of moderate to low level of cotyledon damage (simulated weevils infection) on holm-oak seedling growth and physiological performance under conditions of soil water stress, a recurrent constraint in Mediterranean and other seasonally dry environment. Three levels of artificial damage were applied to the acorns (no damage NoD, low damage LD, and medium damage MD), and the germination, emergence and early seedling growth under controlled conditions were studied during ca. 1.5 months. On the other hand drought effect on seedling growth, leaf gas-exchange, PSII efficiency, photosynthetic pigments and electron transport energy fluxes was analyzed in a set of older seedlings (i.e., 6.5 months old) that also derived from treated acorns and were exposed to two irrigation treatments (well water WW and water stress WS) for 1.5 months. The results showed that LD and MD acorns germinated earlier than NoD ones but final seedling emergence was lowest in the MD treatment. Cotyledons exhibited increased level of necrotic tissue as physical damage (drilling) increased while seedling biomass and size-related traits tended to decrease. Under WS conditions seedlings derived from LD and MD exhibited higher decrease in above and belowground biomass, as well as on net photosynthetic rate (AN), stomatal conductance (gs), intercellular CO2 concentration (Ci) and PSII efficiency than those derived from NoD. These differences were much less conspicuous under WW treatment where all traits reached higher values. In view of the interactive effect of drought and acorn damage it is concluded that under natural conditions acorn infection by weevils and other insects may represent a highest limitation to holm-oak seedling recruitment than previously considered. In addition this limitation may become more important in the future whether warming and drought increase.