Journal of Ecology 2014, 102, 28–35
doi: 10.1111/1365-2745.12182
The phenology–substrate-match hypothesis explains
decomposition rates of evergreen and deciduous oak
leaves
Ian S. Pearse1,2*, Richard C. Cobb3 and Richard Karban1
1
Department of Entomology, UC Davis, One Shields Ave, Davis, CA 95616, USA; 2Cornell Lab of Ornithology, 159
Sapsucker Woods, Ithaca, NY 14850, USA; and 3Department of Plant Pathology, UC Davis, One Shields Ave, Davis,
CA 95616, USA
Summary
1. There is substantial evidence that the rate of litter decomposition is affected by the match
between the litter substrate and the soil matrix (decomposer community). We introduce and test the
phenology–substrate-match hypothesis, which predicts that both litter composition and soil matrix
will change over the course of the year and that a lagged match between litter type and soil matrix
will result in an optimal decomposition environment.
2. We conducted a decomposition experiment in a Mediterranean mixed deciduous–evergreen oak
savanna in California. We initiated litter decomposition of both a deciduous oak (whose leaves fall
in autumn) and an evergreen oak (whose leaves fall in spring) in both autumn and spring.
3. Consistent with the phenology–substrate-match hypothesis, we found that decomposition of
deciduous oak litter was accelerated compared to evergreen oak litter when decomposition was initiated in spring, while evergreen litter was accelerated compared to deciduous litter when decomposition was initiated in autumn.
4. We also found a small effect of microsite on leaf decomposition, where both evergreen and
deciduous oak leaves decomposed faster under the canopy of a conspecific.
5. Synthesis. Our study extends theory of litter quality and the decomposer community into a temporal context, which may be an important source of variation in decomposition rates when species
with different litterfall phenologies co-occur.
Key-words: climate change, decomposer community, home–field advantage, leaf longevity, litter
decomposition, plant–soil (below-ground) interactions, Quercus, substrate quality matrix quality
interaction hypothesis
Introduction
Litter decomposition accounts for the energetic fate of the
majority of biomass in temperate forests and other terrestrial
systems (Cebrian 1999), making decomposition an even more
important factor in the cycling of carbon and nutrients than
the actions of herbivores and higher trophic levels. Decomposition at a global scale is largely a function of substrate quality (plant chemistry) and climate (Meentemeyer 1978; Berg
et al. 1993). However, there is substantial variation in decomposition rates that cannot be described by substrate or climate
alone, but is critical to understanding and predicting the
dynamics of this process at smaller spatial scales (Hunt et al.
1988; Freschet, Aerts & Cornelissen 2012). Ecosystems are
heterogeneous, and a nuanced understanding of function is
needed to forecast how a critical process such as decomposi*Correspondence author. E-mail: [email protected]
tion will change with shifts in composition of floral and faunal communities as well as climate (Eviner & Chapin 2003;
O’Neill et al. 2003).
Functionally, decomposition as a process is mediated by
the actions of a diverse guild of decomposer organisms and
their surrounding abiotic environments. Several hypotheses
seek to explain differences in rates of decomposition based on
differences in decomposer communities and their interactions
with particular litter substrates. Most prominently, the home–
field advantage (HFA) hypothesis suggests that decomposer
communities are locally adapted and will more efficiently
decompose litter from a local source (Hunt et al. 1988; Gholz
et al. 2000; Ayres et al. 2009). For example, the decomposition of litter from a broad-leaf deciduous species was accelerated in deciduous forests as compared to conifer-dominated
forests, and conifer litter decomposition was accelerated in
conifer-dominated forests (Gholz et al. 2000). However, performance of the HFA is variable and has been shown to bet-
© 2013 The Authors. Journal of Ecology © 2013 British Ecological Society
Decomposition rates of oak leaves 29
ter explain site-specific rates of litter decomposition when
comparing between species with low-quality litter (recalcitrant
litter) than species with high-quality litter (Hunt et al. 1988;
Ayres et al. 2009). Similarly, recent studies that compared
site-specific rates of litter decomposition within a particular
site found little evidence for the HFA (Ayres, Dromph &
Bardgett 2006; Chapman & Koch 2007; Freschet, Aerts &
Cornelissen 2012). The observation that the HFA is context
dependent has led to the substrate quality–matrix quality
interaction hypothesis (SMI), a generalization of the HFA
hypothesis (Freschet, Aerts & Cornelissen 2012). The SMI
states that the match between a substrate (litter) and the surrounding matrix (i.e. the decomposer community and environment) will affect the rate of litter decomposition. The SMI
hypothesis recognizes that the match between substrate and
matrix is determined by litter quality and the decomposer
community. Unlike the HFA, it recognizes that there may be
better matches between substrate and matrix than the home
site. As theory and field study of litter decomposition have
improved, it is clear that the relationship between litter chemistry and the biotic and abiotic environment are important
drivers of this critical ecosystem process.
Phenology of leaf senescence is a poorly studied, but
potentially important, driver of litter decomposition. Across
biomes, leaves may senesce synchronously in autumn (deciduous), continuously (continuous evergreen) or synchronously
at the time of new leaf expansion (vernal evergreen) (Axelrod
1966; Reich 1995). Even within temperate deciduous forests,
which are characterized by a synchronous autumnal leaf drop,
there is substantial interspecific variation in the exact timing
of leaf drop, and thus timing of peak litter input (Augspurger,
Cheeseman & Salk 2005). Rates of litter decomposition also
vary with the seasonality of abiotic factors such as temperature, rainfall or nutrient inputs (Berg et al. 1993; Rayment &
Jarvis 2000) as well as biotic factors such as the phenology
of decomposer organisms (Verhoef & Brussaard 1990;
Benstead & Huryn 2011), root activity (Bader & Cheng
2007) or the activity of catabolic enzymes (Criquet et al.
2000). Given that the decomposition environment and the litter substrate both vary throughout the year, it is likely that
the match between litter type and season-specific soil matrix
will affect decomposition rates analogously to the site-specific
expectations of litter decomposition presented in the HFA and
SMI hypotheses. Here, we introduce and test the phenology–
substrate-match hypothesis, a temporal extension of the SMI
hypothesis, to explain season-specific decomposition of two
Mediterranean oak species.
THE PHENOLOGY–SUBSTRATE-MATCH HYPOTHESIS
While hypotheses that use the ‘match’ between litter substrate
and soil matrix were initially conceived to explain variation
in litter decomposition between sites, a similar concept could
be used to explain litter decomposition initiated at different
times. The interaction between the litter substrate and soil
matrix explains variation in litter decomposition as suggested
by the SMI hypothesis (Freschet, Aerts & Cornelissen 2012).
However, there is likely a lag between initial litterfall and
optimal decomposition matrix because populations of decomposers increase over time when litter substrate is available
and abiotic conditions are appropriate. Many lagged soil–
substrate interactions have been documented including
increased laboratory-measured decomposition rates following
inoculation with litter-specific microbial communities (Keiser
et al. 2011), delayed peaks between gross primary productivity and soil respiration in mixed evergreen (Vargas et al.
2010) and oak forests (DeForest et al. 2006), and increased
abundance of litter-layer nematodes and microarthropods with
advancement of decay in deciduous broad-leaved, evergreen
broad-leaved and evergreen needle-leaved species (Reynolds,
Crossley & Hunter 2003; Garcia-Pausas, Casals & Romanya
2004). These patterns may be even more pronounced when
species with very different phenologies occur in close proximity because of temporal dependence of the substrate–matrix
interaction. In this case, we expect each litter type would
decompose most rapidly when introduced after (not during)
the peak litter input because decomposer populations will
have been given time to build-up on an abundant substrate.
This may be a common scenario in many Mediterranean and
other ecosystems where species vary dramatically in their leaf
phenologies, and evergreen leaves are often of poorer quality
than deciduous leaves (Klemmedson 1992; Aerts 1995;
Chapman, Schweitzer & Whitham 2006). While the phenology–substrate-match hypothesis is a type of substrate–matrix
interaction, for simplicity, we refer to the SMI hypothesis
below with reference to geographic differences in matrix,
while we refer to the phenology–substrate-match hypothesis
with reference to seasonal differences in matrix.
EXPERIMENTAL TESTS OF THE SMI HYPOTHESIS AND
THE PHENOLOGY–SUBSTRATE-MATCH HYPOTHESIS
We tested the phenology–substrate-match hypothesis and SMI
hypothesis by measuring litter decomposition of Quercus
wislizeni (vernal evergreen) and Q. douglasii (deciduous) at a
site where these species co-occur in close proximity. In this
system, peak litterfall of Q. douglasii occurs in autumn, and
the litter layer is dominated by partially decomposed
Q. douglasii litter through early spring (Fig. 1). Peak litterfall
of Q. wislizeni occurs in spring, and the litter layer is dominated by Q. wislizeni litter through early autumn (Fig. 1). We
initiated decomposition of each litter in the autumn and spring
under the canopies of Q. douglasii and Q. wislizeni. We use
this factorial design to address four research questions: (i)
Does the season of initiation of decomposition affect the
overall rate of decomposition? (ii) Are there inherent differences in the rate of decomposition of each substrate independent of phenology? (iii) Is the rate of decomposition of each
substrate accelerated when decomposition is initiated at a time
when the litter layer is dominated by its own partially decomposed litter (phenology–substrate interaction hypothesis)? (iv)
Is the rate of decomposition of each substrate accelerated
when decomposition takes place at a home site, that is, under
the canopy of a conspecific (SMI/HFA hypothesis).
© 2013 The Authors. Journal of Ecology © 2013 British Ecological Society, Journal of Ecology, 102, 28–35
30 I. S. Pearse, R. C. Cobb & R. Karban
Fig. 1. Leaf and litter phenology in a Californian mixed oak forest.
Leaves from deciduous oaks such as Quercus douglasii senesce and
drop in autumn and begin to decompose in the litter layer over the
following winter. Leaves from evergreen oaks such as Quercus wislizeni drop (mostly) in early spring and form the litter layer over the
following spring and summer.
Materials and methods
STUDY SITE AND CLIMATE
We conducted our study at Quail Ridge Biological Reserve (UTM
Zone: 10; Easting: 574326; Northing: 4260037) near Winters, California. Our study site was a mixed oak forest, whose overstorey was
dominated by the white oak Quercus douglasii and the red oak
Q. wislizeni. Both species are native to the west coast of North America where they are restricted to low- to mid-elevation ecosystems
where they can co-occur or respectively dominate the overstorey.
Q. douglasii senesces and drops most of its leaves in autumn, and its
litter dominates the litter layer until the following spring (Fig. 1).
Quercus wislizeni, like many evergreen oaks (Pearse & Karban
2013), drops most of its leaves in early spring, as the new flush of
leaves is developing, and its litter dominates the litter layer until the
following autumn (Fig. 1). This type of vernal evergreen litterfall is
distinct from continuous leaf senescence in other Mediterranean oaks
(e.g. Q. ilex) or tropical plants where senescence is more evenly distributed over the year (e.g. Aide 1993). Ground cover and understorey
plants at our study site are characterized by the woody shrubs manzanita (Arctostaphylos sp.) and poison oak (Toxicodendron diversilobum), non-native grasses and mixed-species litter. Grass cover was
greater under deciduous Q. douglasii canopies (microsites), and shrub
cover was greater under Q. wislizeni canopies.
We used publicly accessible daily weather information collected
two kilometres from our study site to estimate daily actual evapotranspiration (AET) for the duration of the study (NOAA Markeley
Cove station). AET was estimated for a 5-year span including the
3 years prior to deployment of our litter bags. We calculated potential
evapotranspiration with the Penman–Monteith equation and modified
the Thornthwaite water-balance AET model according to Pastor &
Post (1984). Soil field capacity was parameterized to 13.97 cm3 in
accordance with characteristics of the Bressa–Dibble complex soils
that dominate our study site. We report the cumulative precipitation
and AET throughout our two experimental time periods.
FIELD METHODOLOGY AND EXPERIMENTAL DESIGN
In spring 2010, we collected recently senesced Q. wislizeni litter from
our study site. These leaves were air-dried to a constant mass and
then stored at room temperature until the following autumn. In the
following autumn (2010), this procedure was repeated for Q. douglasii litter. We filled 20 9 20 cm litter bags made of 1-mm2 plastic
mesh with either 3 g of Q. douglasii (n = 120) or Q. wislizeni
(n = 120) litter. Subsamples of initial litter (three 3-g samples per
species) were dried at 60 °C to a constant mass (~48 h) to correct for
initial moisture content (5.3% and 4.8% for Q. douglasii and Q. wislizeni respectively). Initial litter chemistry was assayed on 5 L samples of each species. We assessed lignin content (ash-free ADF) using
a concentrated sulphuric acid extraction of acid detergent fibre and
total carbon and nitrogen with a LECOâ TruSpec CN Analyzer
(LECO, St. Joseph, Michigan, USA). All chemical analyses were conducted by the UC Davis Analytical Laboratory.
We designed a three-way, full factorial experiment to test the interactive effects of litter species (Q. douglasii or Q. wislizeni), time of
initiation of decomposition (either autumn or spring) and microsite
(under the canopy of either Q. douglasii or Q. wislizeni) on rates of
litter decomposition. We identified sites under 20 individual trees
located at least 10 m apart and within a 100 m of all other trees, 10
Q. douglasii and 10 Q. wislizeni (microsite); trees were selected such
that litterbags could be deployed under the canopy of an individual
species. In December 2010, we deployed three litter bags of each oak
species under each tree. In April 2011, we deployed another three litter
bags of each oak species under the same trees. The bags were
retrieved according to the following schedule: fall-deployed bags were
collected in April 2011, June 2011 and November 2011; springdeployed bags were collected in June 2011, November 2011 and April
2012. At each collection, we manually sorted contaminant material
from litter and dried the sorted litter at 60 °C to a constant mass
(within 48 h) to determine mass loss by decomposition.
DATA ANALYSIS
Decomposition constants, k, were calculated according to Adair,
Hobbie & Hobbie (2010): ln[X(t)] = ln[X(0)] –kt = 0 kt, where X
(t) is the proportion of mass remaining at time t (days) given the
initial mass, and X(0) is the proportion remaining at t = 0 (X
(0) = 1). This formulation of exponential litter decay reduces bias
in estimates of k when fit with nonlinear regression procedures
(Adair, Hobbie & Hobbie 2010). We analysed variation in k converted to annual values with a mixed model parameterized with site
(tree) as a random factor and season of initiation (spring/fall), litter
species (Q. douglasii or Q. wislizeni) and microsite (under
Q. douglasii or Q. wislizeni) as fixed factors. We analysed this
three-way factorial design in order to distinguish between main
effects of species, phenology of litterfall and microsite, as well as
interactive effects that correspond to the phenology–substrate-match
hypothesis and the SMI/HFA hypothesis. Specifically, significant
season 9 species interactions are indicative of phenology–substratematch while significant microsite 9 litter interactions with more
rapid decomposition under a respective species canopy were considered evidence of SMI/HFA effects. Our initial model included all
possible interactions between fixed effects. We removed the threeway interaction from the final model because it was not significant
and did not correspond to a specific initial hypothesis. This did not
qualitatively affect any of the model results. Because the two intermediate sampling periods were unevenly spaced between phenological treatments, we conducted an initial analysis using only the final
sampling period (mass lost after 12 months) as a response. Results
of this analysis were qualitatively identical to analyses of k values,
and therefore, we report k values based on the entire data set
throughout the manuscript. For each model, assumptions of normal
error distribution and homoscedastic variance were evaluated with a
goodness-of-fit of the residuals to a normal error distribution and
visual examination of heteroscedasticity; data transformations were
© 2013 The Authors. Journal of Ecology © 2013 British Ecological Society, Journal of Ecology, 102, 28–35
Decomposition rates of oak leaves 31
not necessary to meet these assumptions. The coefficient of determination (r2) of each factor was calculated by a comparison of loglikelihood values between models differing only in the presence of
that factor (Kramer 2005). Effect sizes were calculated by comparing least-square means of treatments. All graphs display actual values as opposed to least-square estimates. All statistics were
calculated in R using package nlme (R Core Development Team
2008; Pinheiro et al. 2009).
Results
OVERALL SPECIES DIFFERENCES
Initial litter chemistry differed significantly between our two
oak species. Quercus wislizeni litter (vernal evergreen) had a
lower nitrogen concentration, higher C:N ratio and higher lignin concentration relative to Q. douglasii (deciduous;
Table 1). Lignin:N ratios were also lower in Q. douglasii relative to Q. wislizeni. As a whole, these foliar chemistry data
are consistent with studies in closely related Quercus species
(Pearse 2011; Pearse & Hipp 2012). Following lignin-based
models of decomposition, the results suggest that Q. wislizeni
litter is likely of lower chemical quality in regards to decomposability than Q. douglasii litter (e.g. Meentemeyer 1978;
Aerts 1997). However, in contrast to our initial expectations,
the species factor in our mixed model was not a significant
source of variation in rates of decomposition (F1,55 = 0.01,
P = 0.91, Table 2). Without regard to the season in which litter decomposition was initiated, Q. douglasii and Q. wislizeni
decomposition was not significantly different despite significant differences in initial litter chemistry.
in the spring (Fig. 2; Table 2). Fall-initiated decomposition
resulted in greater overall amounts and rates of mass loss relative to spring-initiated decomposition for both species. These
patterns occurred despite equal temporal duration of decomposition and the same seasonal environments, but with a different order of seasons over a full year (Table 2 and Fig. 3).
Fall-initiated decomposition began rapidly over the peak period of precipitation during the winter season, slowed during
low-precipitation spring and slowed to an even greater degree
during the summer drought typical of a Mediterranean climate
(Figs 2 and 3). In contrast, spring-initiated decomposition
began slowly and remained constant even during the wet winter months at the end of the season (Fig. 2).
The climate at the site was typical of Mediterranean habitats (Fig. 3). Precipitation and actual evapotranspiration
(AET), two important climatic factors that drive rates of
decomposition (Meentemeyer 1978; Berg et al. 1993; Aerts
1997; Yahdjian, Sala & Austin 2006), were both low in the
summer and autumn months and were higher in winter and
spring (Fig. 3). Differences between rates of rainfall between
winter 2011 and winter 2012 resulted in lower cumulative
precipitation for litter deployed in spring 2011 compared to
litter deployed in autumn 2010. However, estimates of overall
cumulative AET were similar for both treatments (Fig. 3).
PHENOLOGY–SUBSTRATE INTERACTIONS
(PHENOLOGY–SUBSTRATE-MATCH HYPOTHESIS)
We tested whether the decomposition rate for a particular species of litter was higher when decomposition was initiated at
a time of year when the litter layer, not necessarily litterfall,
is dominated by the respective species’ litter (i.e. the phenol-
PHENOLOGICAL TRENDS
We observed twofold greater decomposition rates (k) when
litter was distributed in the fall compared to litter distributed
Table 1. Initial litter chemistry of Quercus douglasii, a deciduous
oak, and Quercus wislizeni, a co-occurring evergreen oak (n = 5)
Species
Q. douglasii
Q. wislizeni
Leaf
phenology
Deciduous
Evergreen
P
N (%)
C:N
0.866 0.36
0.721 0.019
<0.001
Lignin (%)
51.2 1.9
65.5 1.5
<0.001
9.38 0.97
11.84 0.34
<0.001
Table 2. GLM results describing the interactive effects of season, litter species and microsite (canopy) on leaf litter decomposition
(decomposition constant – k). Tree was included in the model as a
random effect
Factor
Microsite
Species
Season
Species*Season
Species*Microsite
Season*Microsite
d.f.
F-value
1,18
1,55
1,55
1,55
1,55
1,55
0.521
0.014
651.021
48.698
6.279
8.281
r2
0.01
0.00
0.82
0.38
0.08
0.10
P
0.479
0.908
<0.0001
<0.0001
0.015
0.006
Fig. 2. Proportion of mass lost from evergreen (Quercus wislizeni –
solid squares) and deciduous (Quercus douglasii – empty circles) oak
leaves over time. A set of litter bags for each species was distributed
during spring (solid lines) and fall (dashed lines) into the litter of a
mixed oak woodland. Data are means (n = 20) with one standard error.
© 2013 The Authors. Journal of Ecology © 2013 British Ecological Society, Journal of Ecology, 102, 28–35
32 I. S. Pearse, R. C. Cobb & R. Karban
Fig. 4. Box–whisker plots of phenological effects on exponential
decomposition constants (k) for blue oak and interior live oak senescent litter. Litter bags were distributed during the fall and spring for
each species. The thick line represents the median (n = 20) value.
Fig. 3. Cumulative actual evapotranspiration (AET) and precipitation
during autumn and spring distributed litter bags. Incubation periods
overlapped during the last 8 months of autumn and the first 8 months
of spring are identical, thus data from this time period are identical.
ogy–substrate-match hypothesis). In support of the phenology–substrate-match hypothesis, we found that the rate of
decomposition was accelerated when litter was distributed
several months following peak litterfall as compared to when
litter was distributed at the time of peak litterfall
(F1,55 = 48.70, P < 0.0001, Fig. 4). Specifically, Q. douglasii
litter decomposed faster than Q. wislizeni litter when decomposition was initiated in the spring (F1,18 = 24.99,
P = 0.0001), when the litter layer was dominated by partially
decomposed Q. douglasii leaves (Fig. 2). Quercus wislizeni
litter decomposed faster than Q. douglasii litter, when decomposition was initiated in the fall (F1,18 = 37.15, P < 0.0001;
Figs 2 and 4), when the litter layer was dominated by partially decomposed Q. wislizeni leaves.
INTERACTIONS WITH MICROSITE (SMI HYPOTHESIS)
The overall rate of decomposition did not vary between microsites, that is, whether litterbags were placed under Q. douglasii
or Q. wislizeni (Table 2). However, in support of the SMI
hypothesis, we found that there was an effect of microsite that
depended on the species identity of the substrate (Table 2,
Fig. 5). Specifically, Quercus douglasii litter decomposed 7%
faster when it was under its own canopy than when it was under
the canopy of Q. wislizeni (F1,18 = 3.61, P = 0.07). Quercus
wislizeni decomposed at a similar rate under its own canopy as
when under the canopy of Q. douglasii (F1,18 = 1.07, P = 0.32).
We also found that the effect of microsite depended on the season of initiation of litter decomposition (F1,55 = 8.28,
P = 0.006, Table 2). Specifically, for both species, litter decomposition that was initiated in the spring progressed 10% faster,
when that litter was underneath the canopy of Q. douglasii,
whose leaves dominate the spring litter layer (F1,18 = 10.05,
Fig. 5. Box–whisker plots of least-square means of the effect of
microsite on litter decomposition of fall-deciduous Quercus douglasii
and vernal evergreen Quercus wislizeni. Least-squares means were
calculated with our mixed model integrating seasonal, species and
microsite effects (Table 2). The thick line represents the median
(n = 20) value.
P = 0.005). In contrast, microsite did not affect decomposition
of litter distributed in autumn (F1,18 = 0.68, P = 0.42).
Discussion
SUPPORT FOR THE PHENOLOGY–SUBSTRATE-MATCH
HYPOTHESIS
We found strong support for the phenology–substrate-match
hypothesis in a Mediterranean mixed evergreen–deciduous
oak woodland. The phenology–substrate-match hypothesis is
an extension of the matrix quality–substrate quality (SMI)
hypothesis (Freschet, Aerts & Cornelissen 2012), and it predicts that rates of decomposition reflect phenological dynamics of both decomposer communities and periods of peak
litterfall. Specifically, it predicts that there will be a substrate–
© 2013 The Authors. Journal of Ecology © 2013 British Ecological Society, Journal of Ecology, 102, 28–35
Decomposition rates of oak leaves 33
matrix interaction between litter and seasonal soil microbial
or faunal decomposer communities, where the optimal conditions for decomposition occur after the peak litterfall of a particular species. These types of interactions may be common
in plant communities where co-occurring species reach peak
litterfall at times of year that are also marked by substantially
different climate, such as the Mediterranean conditions of our
study site. In our field test of this hypothesis, seasonal effects
were sufficient to overwhelm differences in litter quality
(Fig. 2), typically a principal driver of decomposition rates
across species (Meentemeyer 1978; c.f. Cobb 2010; Freschet,
Aerts & Cornelissen 2012). In both fall and in spring, the litter that decomposed the fastest was from the oak species that
senesced several months prior to the initiation of the decomposition experiment (Fig. 2). Based on litter chemistry alone,
decomposition rates of Q. douglasii should have been more
rapid than Q. wislizeni (Table 1; Meentemeyer 1978), but
phenological differences had greater effects than variation in
litter chemistry.
Our experimental support for the phenology–substratematch hypothesis was consistent with trends observed in other
studies (Criquet et al. 2000; Steffen et al. 2007; Strickland
et al. 2009; Keiser et al. 2011) although ours, and many
others, did not quantify our hypothesized mechanistic driver of
a shifting decomposer community. In general, the decomposition of evergreen and deciduous leaves differs across many
ecosystems, likely due to differences in chemical composition
of these litters (Schlesinger & Hasey 1981; Aerts 1995). While
decomposer communities of evergreen and deciduous species
are poorly described, some specialization of decomposer communities to each litter type is likely. In our case, the decomposer communities of Q. douglasii and Q. wislizeni litter are
poorly known. More generally, many specialist taxa are
known to be important components of decomposer communities in oak savannas; saprophytic fungi in the genus Marasmius are highly species-specific in their litter preference and
decomposition efficiency, and they are some of the first fungal
colonists of oak leaves (Steffen et al. 2007). A range of other
endophytic fungi have been shown to be effective in decomposing soluble and structural leaf compounds (Koide, Osono
& Takeda 2005; Osono 2006), and the host range of these
organisms includes many apparently host-specific ascomycetes
(Arnold & Lutzoni 2007), suggesting that effects caused by
phenology–substrate-matches may occur in many ecosystems.
Fungal phenology and enzymatic activity are strongly linked
with climate (Kauserud et al. 2012) and seasonal dynamics
(Criquet et al. 2000), suggesting that seasonal changes in
fungal colonization and enzymatic activity may be important
drivers of decomposition rates within ecosystems.
One of the most striking features of our study was that the
relative differences in decomposition rates between
Q. wislizenii and Q. douglasii were maintained after
12 months of incubation for both seasons (Fig. 2). This pattern is suggestive of seasonal differences in assembly history
of the decomposer community, and assembly history has been
shown to influence rates of wood decomposition when antagonistic fungi are the first colonizers (Fukami et al. 2010). The
next step in demonstrating the phenology–substrate-match
hypothesis will be to show that species-specific decomposer
communities change over the season, increase following peak
litter fall and that mechanistic factors such as specific enzymatic activities are associated with seasonal variation in
decomposition rates among species. Questions of decomposer
community structure and function are becoming increasingly
tractable with modern molecular techniques (Zak, Blackwood
& Waldrop 2006), and recent studies of mixed litter decomposition have demonstrated sensitivity of microbial communities to the diversity and identity of litter types in mixture
(Chapman & Newman 2010; Chapman et al. 2013). While
our proposed mechanistic link between litterfall phenology
and seasonality of decomposer communities requires further
evaluation, the phenology–substrate-match hypothesis implies
that integrating the temporal dynamics of litterfall may be
necessary to understand the non-additive interactions
addressed in field studies of mixed litter decomposition
(Chapman & Newman 2010; Cobb 2010; Chapman et al.
2013). Most studies of mixed litter decomposition have
focused on species with approximately synchronous patterns
of litterfall or implicitly assume that phenological differences
among species are not important. However, ecosystems dominated by species with pronounced phenological differences in
litterfall are certain to produce litter mixes of different species
at different levels of decomposition.
In what situations is the phenology–substrate-match hypothesis likely to be an important driver of decomposition rates?
The phenology–substrate-match hypothesis requires variable litter qualities, litter phenologies and interannual climate that
would favour season-specific optimization of the decomposer
communities. Each of these requirements is met in many Mediterranean habitats where temperature generally does not limit
decomposition (Hart, Firestone & Paul 1992; Garcia-Pausas,
Casals & Romanya 2004), and overstorey species with vernal,
autumnal or continuous patterns of leaf senescence can occur
in close proximity (Dallman 1998; Allen-Diaz, Standiford &
Jackson 2007). In particular, vernal evergreen species often cooccur with fall-deciduous and drought-deciduous species in
Mediterranean climates (Allen-Diaz, Standiford & Jackson
2007). Seasonal lags in optimal decomposition rates may also
occur in tropical habitats with even greater diversity of overstorey species and patterns of leaf senescence. In temperate deciduous forests characterized by a pulse fall leaf drop, some
species such as Aesculus glabra drop their leaves months
before peak autumn leaf drop (Augspurger, Cheeseman & Salk
2005), while other ‘marcescent’ species such as Fagus sylvatica retain a substantial proportion of dead leaves in the canopy
until the following spring (Otto & Nilsson 1981). One expectation of the phenology–substrate-match hypothesis, as well as
the SMI hypothesis, is that natural mismatches between litter
type and decomposer communities should not occur in forests
with one dominant litter type. Regardless of biome or ecosystem type, our hypothesis predicts that decomposition rates
could be slower when two species with phenologically separated decomposers co-occur compared to stands where any one
species dominates the litterfall.
© 2013 The Authors. Journal of Ecology © 2013 British Ecological Society, Journal of Ecology, 102, 28–35
34 I. S. Pearse, R. C. Cobb & R. Karban
EFFECTS OF PHENOLOGY ON ALL DECOMPOSING
SUBSTRATES
The season in which decomposition was initiated had a large
effect on the decomposition rates of all substrates at our site
(Fig. 2 and Table 2). Both precipitation and AET varied considerably between seasons at our site (Fig. 3), and each set of litter
bags experienced these seasons in a different order. There are two
potential explanations, then, for accelerated rates of fall-initiated
decomposition. First, differences in aspects of climate that limit
the rate of decomposition varied between the seasonal treatments.
We found support for this in precipitation, which affects decomposition in arid climates (Yahdjian, Sala & Austin 2006), and
was greater in the fall-initiated treatment, due to a wet 2011
winter (Fig. 3). Secondly, the order of climatic events could
affect the decomposition of litter due to climate having different
effects on early vs. late decomposition processes. While our study
cannot address this hypothesis, it has received little support from
experiments investigating the effects of timing of climatic events
on litter decomposition in other systems (O’Neill et al. 2003).
SMI AND HFA HYPOTHESES OVER MICROSITES
We found a small, but significant substrate-matrix interaction
that was consistent with the home–field advantage (HFA) and
substrate–matrix interaction (SMI) hypotheses (Gholz et al.
2000; Freschet, Aerts & Cornelissen 2012) in that Q. douglasii and Q. wislizeni leaves decomposed faster under the canopy of a conspecific than under the canopy of the other
species (Fig. 5). The HFA hypothesis received mixed support
from studies that dealt with either litter of very different quality or in very different or geographically separate decomposition environments (Gholz et al. 2000; Ayres et al. 2009;
Makkonen et al. 2012). The HFA received little support from
studies that dealt with similar litters at nearby sites (Ayres,
Dromph & Bardgett 2006; Freschet, Aerts & Cornelissen
2012). In our study, we found marginal support for an HFA,
despite a close spatial proximity and taxonomic similarity of
litter sources. However, differences in decomposition rates,
due to phenology of deployment and due to phenology–substrate interactions, were far greater than the effects expected
by the HFA hypothesis (Fig. 4). We suggest that seasonal
specificity in litter decomposition rates may reflect a more
general substrate-matrix match at a local scale than spatial
specificity in litter decomposition because the seasonal
changes are consistent. In contrast, microsite specificity may
be swamped by transport of litter from another location or
obscured by non-additive interactions of mixed-species litter
decomposition (e.g. Cobb 2010; Chapman et al. 2013). In
addition, there was a potential relationship between these two
hypotheses, as the importance of microsite on decomposition
rates varied marginally with season (Table 2).
Conclusion
Our study suggests climate-driven phenological changes could
have a significant impact on litter decomposition in addition
to the direct impacts of altered climate. Globally, plant phenology is rapidly changing, mostly towards earlier leaf-out
and later leaf-drop dates (Gordo & Sanz 2009; Polgar & Primack 2011). Likewise, in Europe, the fruiting phenology of
fungal decomposers has shifted towards later dates over the
past 30 years (Kauserud et al. 2012). This suggests that phenological ‘match’ between fungal community and type of litter input has the potential to shift with a changing climate.
However, our study suggests that this shift could be either in
the direction of increased or decreased decomposition rates,
as the optimal decomposition environment is lagged relative
to the period of peak litterfall (Fig. 2).
We showed that the phenological match between litter fall
and the soil matrix affects decomposition rates. There is a broad
consensus that to understand decomposition at a fine scale a better understanding of litter quality and decomposer community
interactions are needed (Prescott 2002; Fukami et al. 2010;
Freschet, Aerts & Cornelissen 2012). Our study suggests that it
is also necessary to explore both the phenology of litter drop
and the phenology of decomposer communities. Such phenological dynamics and influences are likely to be of greater
importance when several dominant plant species have very different timing of leaf senescence, but occur in close enough
proximity that underlying soil decomposer communities and
substrates are influenced by multiple species. These types of
interactions may be common in Mediterranean-type environments with autumnal, vernal and drought-deciduous species.
However, many other ecosystems are notable for distinct wet or
dry seasons and corresponding differences in leaf senescence
that could lead to seasonal effects.
Acknowledgements
We would like to thank Virginia ‘Shorty’ Boucher and Shane Waddell for
access to Quail Ridge Nature Preserve. Sandra Ferguson and Jill Baty helped
collect leaves for the study. Greg Crutsinger, Dave Rizzo, Cassandra Swett, Jill
Baty and Amanda Hodson provided valuable discussion on this project. We
received funding from USDA FS 12-CS-11052007-026 and USDA NC-7.
Data accessibility
Decomposition data available from the Dryad Digital Repository: doi:10.5061/
dryad.nj383 (Pearse, Cobb & Karban 2013).
References
Adair, E.C., Hobbie, S.E. & Hobbie, R.K. (2010) Single-pool exponential
decomposition models: potential pitfalls in their use in ecological studies.
Ecology, 91, 1225–1236.
Aerts, R. (1995) The advantages of being evergreen. Trends in Ecology &
Evolution, 10, 402–407.
Aerts, R. (1997) Climate, leaf litter chemistry and leaf litter decomposition in
terrestrial ecosystems: a triangular relationship. Oikos, 79, 439–449.
Aide, T.M. (1993) Patterns of leaf development and herbivory in a tropical
understorey community. Ecology, 74, 455–466.
Allen-Diaz, B., Standiford, R. & Jackson, R.D. (2007) Oak woodlands and forests. Terrestrial Vegetation of California (eds M.G. Barbour, T. Keeler-Wolf &
A.A. Shoenherr), pp. 313–338. University of California Press, Berkeley, CA.
Arnold, A.E. & Lutzoni, F. (2007) Diversity and host range of foliar fungal
endophytes: are tropical leaves biodiversity hotspots? Ecology, 88, 541–549.
Augspurger, C.K., Cheeseman, J.M. & Salk, C.F. (2005) Light gains and physiological capacity of understorey woody plants during phenological avoidance
of canopy shade. Functional Ecology, 19, 537–546.
© 2013 The Authors. Journal of Ecology © 2013 British Ecological Society, Journal of Ecology, 102, 28–35
Decomposition rates of oak leaves 35
Axelrod, D.I. (1966) Origin of deciduous and evergreen habits in temperate forests. Evolution, 20, 1–15.
Ayres, E., Dromph, K.M. & Bardgett, R.D. (2006) Do plant species encourage
soil biota that specialise in the rapid decomposition of their litter? Soil Biology & Biochemistry, 38, 183–186.
Ayres, E., Steltzer, H., Simmons, B.L., Simpson, R.T., Steinweg, J.M., Wallenstein, M.D., Mellor, N., Parton, W.J., Moore, J.C. & Wall, D.H. (2009)
Home-field advantage accelerates leaf litter decomposition in forests. Soil
Biology & Biochemistry, 41, 606–610.
Bader, N.E. & Cheng, W.X. (2007) Rhizosphere priming effect of Populus fremontii obscures the temperature sensitivity of soil organic carbon respiration.
Soil Biology & Biochemistry, 39, 600–606.
Benstead, J.P. & Huryn, A.D. (2011) Extreme seasonality of litter breakdown
in an arctic spring-fed stream is driven by shredder phenology, not temperature. Freshwater Biology, 56, 2034–2044.
Berg, B., Berg, M.P., Bottner, P., Box, E., Breymeyer, A., Deanta, R.C. et al.
(1993) Litter mass-loss rates in pine forests of Europe and Eastern United States some relationships with climate and litter quality. Biogeochemistry, 20, 127–159.
Cebrian, J. (1999) Patterns in the fate of production in plant communities. The
American Naturalist, 154, 449–468.
Chapman, S.K. & Koch, G.W. (2007) What type of diversity yields synergy
during mixed litter decomposition in a natural forest ecosystem? Plant and
Soil, 299, 153–162.
Chapman, S.K. & Newman, G.S. (2010) Biodiversity at the plant-soil interface:
microbial abundance and community structure respond to litter mixing. Oecologia, 162, 763–769.
Chapman, S.K., Schweitzer, J.A. & Whitham, T.G. (2006) Herbivory differentially alters plant litter dynamics of evergreen and deciduous trees. Oikos,
114, 566–574.
Chapman, S.K., Newman, G.S., Hart, S.C., Schweitzer, J.A. & Koch, G.W.
(2013) Leaf litter mixtures alter microbial community development: mechanisms for non-additive effects in litter decomposition. PLoS ONE, 8, 9.
Cobb, R.C. (2010) Species shift drives decomposition rates following invasion
by hemlock woolly adelgid. Oikos, 119, 1291–1298.
Criquet, S., Farnet, A.M., Tagger, S. & Le Petit, J. (2000) Annual variations of
phenoloxidase activities in an evergreen oak litter: influence of certain biotic
and abiotic factors. Soil Biology & Biochemistry, 32, 1505–1513.
Dallman, P.R. (1998) Plant Life in the World’s Mediterranean Climates. University of California Press, Berkeley, CA.
DeForest, J.L., Noormets, A., McNulty, S.G., Sun, G., Tenney, G. & Chen, J.Q.
(2006) Phenophases alter the soil respiration-temperature relationship in an
oak-dominated forest. International Journal of Biometeorology, 51, 135–144.
Eviner, V.T. & Chapin, F.S. (2003) Functional matrix: a conceptual framework
for predicting multiple plant effects on ecosystem processes. Annual Review
of Ecology Evolution and Systematics, 34, 455–485.
Freschet, G.T., Aerts, R. & Cornelissen, J.H.C. (2012) Multiple mechanisms
for trait effects on litter decomposition: moving beyond home-field advantage
with a new hypothesis. Journal of Ecology, 100, 619–630.
Fukami, T., Dickie, I.A., Wilkie, J.P., Paulus, B.C., Park, D., Roberts, A.,
Buchanan, P.K. & Allen, R.B. (2010) Assembly history dictates ecosystem functioning: evidence from wood decomposer communities. Ecology Letters, 13,
675–684.
Garcia-Pausas, J., Casals, P. & Romanya, J. (2004) Litter decomposition and
faunal activity in Mediterranean forest soils: effects of N content and the
moss layer. Soil Biology & Biochemistry, 36, 989–997.
Gholz, H.L., Wedin, D.A., Smitherman, S.M., Harmon, M.E. & Parton, W.J.
(2000) Long-term dynamics of pine and hardwood litter in contrasting environments: toward a global model of decomposition. Global Change Biology,
6, 751–765.
Gordo, O. & Sanz, J.J. (2009) Long-term temporal changes of plant
phenology in the Western Mediterranean. Global Change Biology, 15,
1930–1948.
Hart, S.C., Firestone, M.K. & Paul, E.A. (1992) Decomposition and nutrient
dynamics of Ponderosa pine needles in a Mediterranean-type climate. Canadian Journal of Forest Research, 22, 306–314.
Hunt, H.W., Ingham, E.R., Coleman, D.C., Elliott, E.T. & Reid, C.P.P. (1988)
Nitrogen limitation of production and decomposition in prairie, mountain
meadow, and pine forest. Ecology, 69, 1009–1016.
Kauserud, H., Heegaard, E., Buntgen, U., Halvorsen, R., Egli, S., Senn-Irlet, B.
et al. (2012) Warming-induced shift in European mushroom fruiting phenology. Proceedings of the National Academy of Sciences of the United States
of America, 109, 14488–14493.
Keiser, A.D., Strickland, M.S., Fierer, N. & Bradford, M.A. (2011) The effect
of resource history on the functioning of soil microbial communities is maintained across time. Biogeosciences, 8, 1477–1486.
Klemmedson, J.O. (1992) Decomposition and nutrient release from mixtures of
Gambel oak and Ponderosa pine leaf litter. Forest Ecology and Management,
47, 349–361.
Koide, K., Osono, T. & Takeda, H. (2005) Colonization and lignin decomposition
of Camellia japonica leaf litter by endophytic fungi. Mycoscience, 46, 280–286.
Kramer, M. (2005) R2 Statistics For Mixed Models. Proceedings of the Conference on Applied Statistics in Agriculture, pp. 148–160.
Makkonen, M., Berg, M.P., Handa, I.T., Hattenschwiler, S., van Ruijven, J.,
van Bodegom, P.M. & Aerts, R. (2012) Highly consistent effects of plant litter identity and functional traits on decomposition across a latitudinal gradient. Ecology Letters, 15, 1033–1041.
Meentemeyer, V. (1978) Macroclimate and lignin control of litter decomposition rates. Ecology, 59, 465–472.
O’Neill, E.G., Johnson, D.W., Ledford, J. & Todd, D.E. (2003) Acute seasonal
drought does not permanently alter mass loss and nitrogen dynamics during
decomposition of red maple (Acer rubrum L.) litter. Global Change Biology,
9, 117–123.
Osono, T. (2006) Role of phyllosphere fungi of forest trees in the development
of decomposer fungal communities and decomposition processes of leaf litter. Canadian Journal of Microbiology, 52, 701–716.
Otto, C. & Nilsson, L.M. (1981) Why do beech and oak trees retain leaves
until spring. Oikos, 37, 387–390.
Pastor, J. & Post, W.M. (1984) Calculating Thornthwaite and Mather actual
evapotranspiration using an approximating function. Canadian Journal of
Forest Research, 14, 466–467.
Pearse, I.S. (2011) Leaf defensive traits in oaks and their role in both preference and performance of a polyphagous herbivore, Orgyia vetusta. Ecological Entomology, 36, 635–642.
Pearse, I.S., Cobb, R.C. & Karban, R. (2013) Data from: The phenology substrate match hypothesis explains decomposition rates of evergreen and deciduous oak leaves. Dryad Digital Repository, doi:10.5061/dryad.nj383.
Pearse, I.S. & Hipp, A.L. (2012) Global patterns of leaf defenses in oak
species. Evolution, 66, 2272–2286.
Pearse, I.S. & Karban, R. (2013) Leaf drop affects herbivory in oaks. Oecologia, 173, 925–932.
Pinheiro, J., Bates, D., DebRoy, S. & Sarkar, D. & The R Core Development
Team (2009) nlme: Linear and Nonlinear Mixed Effects Models. R package
version 3.1-93. The R Core Development Team, Vienna.
Polgar, C.A. & Primack, R.B. (2011) Leaf-out phenology of temperate woody
plants: from trees to ecosystems. New Phytologist, 191, 926–941.
Prescott, C.E. (2002) The influence of the forest canopy on nutrient cycling.
Tree Physiology, 22, 1193–1200.
R Core Development Team (2008) R v. 2.10.1. The R Foundation, Vienna.
Rayment, M.B. & Jarvis, P.G. (2000) Temporal and spatial variation of soil
CO2 efflux in a Canadian boreal forest. Soil Biology & Biochemistry, 32,
35–45.
Reich, P.B. (1995) Phenology of tropical forests - patterns, causes, and consequences. Canadian Journal of Botany, 73, 164–174.
Reynolds, B.C., Crossley, D.A. & Hunter, M.D. (2003) Response of soil invertebrates to forest canopy inputs along a productivity gradient. Pedobiologia,
47, 127–139.
Schlesinger, W.H. & Hasey, M.M. (1981) Decomposition of chaparral shrub
foliage - losses of organic and inorganic constituents from deciduous and
evergreen leaves. Ecology, 62, 762–774.
Steffen, K.T., Cajthaml, T., Snajdr, J. & Baldrian, P. (2007) Differential degradation of oak (Quercus petraea) leaf litter by litter-decomposing basidiomycetes. Research in Microbiology, 158, 447–455.
Strickland, M.S., Lauber, C., Fierer, N. & Bradford, M.A. (2009) Testing the
functional significance of microbial community composition. Ecology, 90,
441–451.
Vargas, R., Baldocchi, D.D., Allen, M.F., Bahn, M., Black, T.A., Collins, S.L.,
Yuste, J.C., Hirano, T., Jassal, R.S., Pumpanen, J. & Tang, J.W. (2010)
Looking deeper into the soil: biophysical controls and seasonal lags of soil
CO2 production and efflux. Ecological Applications, 20, 1569–1582.
Verhoef, H.A. & Brussaard, L. (1990) Decomposition and nitrogen mineralization in natural and agroecosystems - the contribution of soil animals. Biogeochemistry, 11, 175–211.
Yahdjian, L., Sala, O. & Austin, A.T. (2006) Differential controls of water
input on litter decomposition and nitrogen dynamics in the Patagonian
steppe. Ecosystems, 9, 128–141.
Zak, D.R., Blackwood, C.B. & Waldrop, M.P. (2006) A molecular dawn for
biogeochemistry. Trends in Ecology & Evolution, 21, 288–295.
Received 20 September 2013; accepted 25 October 2013
Handling Editor: Rien Aerts
© 2013 The Authors. Journal of Ecology © 2013 British Ecological Society, Journal of Ecology, 102, 28–35