1
Non-random patterns of functional redundancy revealed in ground beetle communities facing
2
an extreme flood event
3
4
Michael Gerisch*
5
6
* Department of Conservation Biology, Helmholtz Centre for Environmental Research – UFZ,
7
Permoserstr. 15, 04318 Leipzig, Germany
8
E-mail: [email protected]
9
10
11
Keywords: assemblages, biodiversity recovery, carabids, ecosystem functioning, species
12
turnover, stability, stochastic events, traits
1
13
Abstract
14
15
Theory predicts that species performing similar roles for ecological processes or functions can
16
compensate for the loss of others and that this functional redundancy may promote resilience.
17
However, there is no clear evidence for this mechanism, because functional redundancy has
18
been observed to be low in many ecosystems. By using a severe flood event this study tests
19
whether functional redundancy exists in floodplain ground beetle communities, how it is
20
controlled, and how it connects to post-flood resilience.
21
Ground beetles were sampled in floodplain grassland of the Elbe River, Germany. Functional
22
redundancy was estimated as the proportion of species in a community having neutral effects
23
on functional diversity. Null models were used to determine whether functional redundancy is
24
higher or lower than expected and mixed effects modelling was applied to estimate the
25
relationships between functional redundancy, flood disturbance, and sampling season.
26
It was found that highly redundant ground beetle communities experienced fewer losses in
27
functional diversity caused by an extreme flood than less redundant communities. Functional
28
redundancy was lowest immediately after the flood, but quickly increased with time since
29
flooding. It was significantly higher in spring than in autumn seasons, and significantly higher
30
in habitats exposed to frequent flooding. Null models confirmed that these patterns were to a
31
high degree non-random.
32
The results indicate that functional redundancy plays an important role for stabilizing ground
33
beetles under regular, predictable flooding. However, given sizeable differences in functional
34
diversity before and after the extreme flood, this effect may be lower than expected during
35
extreme events. Other regulating forces, such as stochastic colonization processes and habitat
36
templets play more important roles directly after extreme disturbances. I therefore assume that
37
a temporal hierarchy of mechanisms, including FR, controls the functional diversity of ground
38
beetles in riparian habitats.
39
2
40
Introduction
41
42
Extreme events, such as catastrophic floods, droughts, or fires, have serious implications for
43
ecosystems (Easterling et al. 2000; Jentsch, Kreyling & Beierkuhnlein 2007). They can cause
44
sudden breakdowns of species numbers and populations (Thibault & Brown 2008), change
45
competitive interactions between organisms (Jentsch & Beierkuhnlein 2003), or even shift
46
ecotone boundaries (Allen & Breshears 1998). Hence, one of the primary effects of
47
unexpected weather extremes is that they abruptly and persistently change the performance of
48
ecological processes, such as biomass production (Ciais et al. 2005), or properties realized by
49
single species or communities, for example resistance to invasive species (Sorte, Fuller &
50
Bracken 2010). There is also evidence that ecological complexity, i.e. the richness of
51
organisms and functional traits, enhances the resilience and resistance of biodiversity to
52
extraordinary weather extremes. That is, spatial and temporal hierarchies of response
53
mechanisms are assumed to maintain the ecological functioning of the entire system (White et
54
al. 2000; Jentsch et al. 2011). However, general mechanisms are still unclear and questions
55
regarding how biodiversity affects the functioning of ecological systems are increasingly
56
stimulating ecological debate (Díaz & Cabido 2001; Naeem & Wright 2003).
57
58
A central theme within these discussions is still basic in nature and encompasses the
59
understanding of the relationships between the number of species present in a community and
60
the performance of ecological processes or properties (i.e. functioning). Functional
61
redundancy (FR) is one of many potential concepts used to predict the effects of species
62
richness on ecosystem functioning, particularly after disturbances. FR is based on the
63
principle that some species perform similar functional roles in ecosystems, and might
64
therefore be substitutable with little impact on ecosystem functions, e.g. biomass productivity
65
or nutrient fluxes, or community properties, such as resilience following disturbances (Walker
66
1992; Lawton & Brown 1993; Rosenfeld 2002). In fact, FR and specific aspects of resilience
67
are closely connected. Some authors consider FR an equivalent to the functional resilience of
68
ecological systems (Naeem, Loreau & Inchausti 2002; Petchey & Gaston 2009; Konopka
69
2009; Dalerum et al. 2010). Functional resilience is assumed to explain different parts of
70
community reorganization from purely taxonomic approaches, which usually estimate
71
compositional dissimilarities to some reference conditions, i.e. beta-diversity (Moretti et al.
3
72
2009; Petchey & Gaston 2009; Dalerum et al. 2010). From a functional perspective, species
73
communities are considered to be highly resilient if many species can be lost without
74
changing the range, dispersion, and relative abundance of functional traits, i.e. functional
75
diversity (Walker 1992; Díaz et al. 2007; Dalerum et al. 2010).
76
In the past few years, the number of studies dealing with resilience-redundancy relationships
77
has increased for various different ecological systems (Micheli & Halpern 2005; Petchey et al.
78
2007; Bêche & Statzner 2009; Sasaki et al. 2009; Bihn, Gebauer & Brandl 2010; Joner et al.
79
2011; Guillemot et al. 2011). They highlight that FR differs considerably among habitats,
80
taxonomic groups, and functional units, but also in response to various types of environmental
81
stressors and disturbance agents. Nevertheless, a common finding of most of the studies is the
82
low degree of FR detected (but see e.g. Villéger et al. 2010), which means that either FR
83
cannot be measured properly, or that its role for maintaining community functioning, and
84
especially resilience, is smaller than expected. This weak support for FR endorses the views
85
of ecologists who discuss the concept as rather controversial, and many suggest that no two
86
species can be exactly equivalent, which would obviously contradict classical niche theory
87
and stable coexistence, and downplay the role of similar species for stability (Loreau 2004;
88
Resetarits & Chalcraft 2007). Yachi & Loreau (1999) highlighted the differences among
89
species and suggested that the contribution of some species to ecosystem processes may
90
change over time, depending on the type of disturbance and the traits of the species. These
91
differences, rather than redundancy sensu stricto, should insure ecosystems against declines in
92
their functioning caused by disturbances. Other theories predict that the functioning of
93
ecosystems is mainly maintained by keystone species or that species impacts on ecosystem
94
functioning is context dependent, and therefore unpredictable (Naeem, Loreau & Inchausti
95
2002). Despite the large body of theoretical groundwork, it has not been possible to draw a
96
general picture from previous works dealing with either of these conceptual frameworks
97
because studies that systematically test such theories are still lacking. By estimating FR, the
98
present study aims to gain a better understanding of the role that functional diversity and
99
species richness, and thus FR, has for the stability of ecological systems in dynamic
100
environments.
101
102
In the summer of 2002, unpredictable severe precipitation led to the most severe flooding ever
103
recorded along the river Elbe in Germany. This flood was extreme in terms of its height,
104
duration, and seasonal and spatial occurrence (Schiermeier 2003). In a previous study,
4
105
Gerisch et al. (2012) showed a rapid recovery of ground beetle species richness and diversity
106
after being massively reduced by this extreme summer flood. The present study builds upon
107
this work and aims to obtain a better understanding of the mechanisms enabling biodiversity
108
to recover from such extreme floods, and what role is played by functionally redundant
109
species in community re-organization. The main aims are to determine whether FR is present
110
in species communities inhabiting frequently flooded habitats, to elaborate whether
111
functionally redundant species can provide “insurance” against functional diversity losses
112
caused by flooding, and to identify potential drivers of FR. This study focuses on ground
113
beetles in floodplain grasslands because they are one of the most abundant semiterrestrial
114
macroinvertebrate groups in Central Europe. They are also known to respond quickly and
115
differentially to habitat disturbances (Ribera et al. 2001; Niemelä & Kotze 2009) and, beyond
116
that, functional traits and ecological preferences are widely known for the species.
117
The primary hypothesis to be tested in this study was that, due to the high environmental
118
stochasticity present in floodplains, FR of ground beetle communities should be higher than
119
would be expected by chance. This should support the concept of ecological insurance and
120
confirm the second expectation that FR is buffering communities from functional diversity
121
losses, despite them suffering from severe species losses caused by an extreme flood.
122
Nevertheless, species communities assemble randomly after stochastic events due to priority
123
effects (Chase 2007) or random ecological drift (Hubbell 2006), which implies high trait
124
variation immediately after the flood among the species. I therefore expected ground beetle
125
communities recorded immediately after an extreme flood to be characterized by low FR
126
levels, and a quick increase of FR over time when ecological drift is replaced by other
127
regulating forces. I also hypothesized that FR is strongly related to the phenology of the
128
species and to regular flood disturbance. The latter was assumed based on previous work
129
showing that communities are more species rich but less functionally diverse in highly flood
130
disturbed habitats, suggesting both low FR and strong trait filtering (Gerisch et al. 2012a).
5
131
Material and methods
132
133
Study sites & data sampling
134
With a length of approx. 1,100 km, the Elbe River is one of the largest rivers in Germany and
135
covers a catchment area of about 150,000 km2 ranging from the German-Czech border to the
136
North Sea near Cuxhaven. Its hydrological regime is close to its natural state and discharge is
137
characterized by high water levels in winter and spring, but low flow in summer (Scholten et
138
al. 2005).
139
Ground beetles were sampled at two large study sites in seasonally flooded grassland habitats
140
of the UNESCO Biosphere Reserve “Elbe River Landscape” in Central Germany. The main
141
study site is located near the village of Steckby (51.913°N, 11.977°E) and a secondary site
142
was established approximately 25 km upstream, near the village of Wörlitz (51.857°N,
143
12.384°E). Forty-eight sampling plots were installed, with 36 plots at the main study site
144
“Steckby,” and 12 plots at the site “Wörlitz” (Fig. 1). Following a stratified randomized
145
sampling design, each site was divided into 3 strata in terms of terrain morphology and
146
vegetation type. The sampling plots were then randomly located within each of the strata,
147
which represent different habitat types: (1) wet grasslands, representing frequently flooded
148
oxbow channels, (2) moist grasslands, representing habitats in intermediate conditions, and (3)
149
medium dry grassland, representing elevated, rarely flooded habitats. The distance between
150
the plots ranged between 30 and several hundred meters. All sampling plots were flooded for
151
several weeks in August 2002, but differ considerably in hydrological conditions during
152
normal years. See Henle et al. (2006) for a detailed description of the study design and
153
Gerisch et al. (2012b) for a hydrological description of the different habitat types.
154
155
Figure 1
156
157
On each plot, 5 pitfall traps were installed and filled with a 7% solution of acetic acid and a
158
detergent to reduce surface tension. Using RTK differential GPS, the traps were placed at
159
exactly the same location in each sampling period. The traps were retrieved biweekly from
160
May to June (spring period) and from September to October (autumn period) between 1998
161
and 1999 (pre-flood period), and between 2002 and 2005 (post-flood period). Sampling in the
6
162
flood year 2002 was carried out only in autumn, as soon as the floodwater had receded.
163
Owing to accidental loss of some traps through wild boar and flooding, species abundances
164
were standardized by the number of functioning trap-days. All adult ground beetles were
165
identified to the species level and all recorded species of a study plot sampled in a particular
166
season were regarded as a community.
167
168
Species traits assumed to be important for quick re-colonization of ground beetles after flood
169
disturbance were collected from standard identification keys and ground beetle compendia
170
(functional effect traits, see Table 1 and Table SI1). Other traits were considered which are
171
not necessarily related to disturbance, but which illustrate important survival and response
172
strategies to environmental variability and enable species with similar effect traits to exploit
173
different ecological niches (functional response traits, see Table 1 and Table SI1). Effect and
174
response traits were not weighted a priori for their importance, to avoid a subjective
175
overemphasis of certain traits.
176
177
Table 1
178
179
Measuring functional diversity and functional redundancy
180
Functional diversity of the communities was estimated by means of the functional dispersion
181
index, which is the mean distance of individual species from the centroid of all species in
182
multivariate trait space (Laliberté & Legendre 2010). Functional dispersion is an abundance
183
weighted index of functional diversity and reflects how strongly species are spread throughout
184
this multidimensional space. This index requires the calculation of pairwise dissimilarities
185
between all species of a community based on their functional traits. Given the different scale
186
units of the trait data, dissimilarities were calculated using the Gower distance, which can
187
handle both continuous and categorical data. Large functional dispersion values reflect a large
188
distance between the species in the trait space, meaning that several species possess traits that
189
differ from the multivariate average.
190
The main interest of this study was to estimate functional redundancy, i.e. the proportion of
191
species that did not contribute to an increase in the functional diversity of a community. In
192
this study the proportion instead of the number of functionally redundant species was used to
7
193
account for the close relationships between species richness and the number of redundant
194
species (Pearson correlation, r=0.89, p < 0.001) and minimize the effect species richness has
195
on FR patterns.
196
Species were considered redundant to functional diversity if their losses would not result in a
197
decrease of functional diversity. As this number depends strongly on which species assemble,
198
an average functional diversity value was calculated for different species combinations. For
199
this, 2 to i species subsets were created for each community, where 2 is the minimum number
200
of species needed to calculate functional diversity, and i is the total number of species
201
recorded. Each subset contained 1,000 different species combinations randomly assembled
202
from the species pool of the community, and for each subset functional diversity was
203
calculated. The average functional diversity of each subset was then calculated as the median
204
of the 1,000 different values. FR was estimated as the proportion of species that have neutral
205
effects on the functional diversity of a community. This was done by identifying analytically
206
the position along the x-axis from where the average functional diversity values did not
207
increase further in a graph with the number of species drawn on the x-axis and functional
208
diversity on the y-axis (see Fig. 2 for the conceptual approach to calculating FR in this study).
209
Pearson correlations were calculated for each season between FR and the total number of
210
species, Shannon diversity, and functional diversity in order to identify potential
211
interrelationships between these variables
212
213
To test the hypothesis that FR can buffer communities from functional diversity loss despite
214
them suffering from species loss caused by an extreme flood, the change in functional
215
diversity before and directly after the flood was related to FR. For each community, the mean
216
functional diversity of the first two post-flood seasons was subtracted from the mean
217
functional diversity of all available pre-flood sampling seasons. Pearson correlation was then
218
applied to relate this difference with the average FR of all pre-flood seasons. Because species
219
sampling was interrupted between spring 2000 and summer 2002, the available pre-flood data
220
were pooled to average potential changes in functional diversity during this time.
221
Nevertheless, this period was hydrologically very similar compared to the years 1998 and
222
1999 (Gerisch et al. 2012b), suggesting similar patterns of species and functional diversity
223
also in the non-sampled seasons.
224
225
226
8
227
Figure 2
228
Null models to test for non-random FR patterns
229
Following a null model approach, it was tested whether the observed FR differed from the FR
230
resulting from random species assembly. Null models produce the community patterns that
231
can be expected when particular ecological mechanisms do not operate and are therefore
232
suitable for detecting environmental impacts on ecological properties (Gotelli & Graves 1996).
233
The underlying null hypothesis assumes that species occurrences are not constrained by
234
external driving factors and, hence, that species assemble randomly. To test the alternative
235
hypothesis (that the observed FR values differ from a random distribution), 999 artificial
236
communities were generated for each sampling plot. Permutation was carried out by
237
randomly re-assigning species from the total species pool (which is the full set of species
238
recorded during the study) to the sampling plots. Total species richness and abundance was
239
held constant for each plot, but individual numbers were distributed randomly among the
240
species to break ties between species and trait abundances. FR of the artificial communities
241
was calculated for each of the 48 plots as described above. To estimate whether the observed
242
FR is higher or lower than a random observation, the probability (P) that a simulated FR value
243
of the null distribution takes a value of the observed FR value or smaller was calculated as
244
follows:
245
𝑃=
𝑛𝑢𝑚𝑏𝑒𝑟(𝐹𝑅𝑟𝑎𝑛 < 𝐹𝑅𝑜𝑏𝑠 ) +
𝑛𝑢𝑚𝑏𝑒𝑟(𝐹𝑅𝑟𝑎𝑛 = 𝐹𝑅𝑜𝑏𝑠 )
)
2
𝑛𝑟𝑎𝑛
246
Where FRran is the FR value of a randomized community, FRobs is the FR value of the
247
observed community, and nran represents the total number of randomized communities.
248
The probability values were transformed to effect sizes varying between -1 and +1, whereby
249
values close to 0 indicate a random FR observation and values approaching -1 or +1 represent
250
observed FR smaller or higher than expected, respectively. A Wilcoxon test was performed
251
on the effect sizes of all sampling plots to determine whether the FR in a season was
252
significantly different from 0 (i.e., nonrandom).
253
254
9
255
Mixed effects modeling
256
Owing to the hierarchical character of the study design, mixed effects modeling was applied
257
to quantify the response of FR to the extreme flood and to test how FR is governed by other
258
potential variables, such as sampling season (i.e., species phenology) and regular flood
259
disturbance. The latter was expressed as a synthetic index based on the following
260
environmental variables measured for each plot and each season: mean groundwater depth
261
(cm), duration of inundation (weeks), maximum flood height (cm), variation coefficient of
262
groundwater depth, and altitude of the sampling plot (m.a.s.l.). The variables were centered
263
and standardized to zero mean and unit variance and processed by a principal component
264
analysis (PCA). The first axis of the PCA explained 89.4 % of the data and was strongly
265
related to hydrological variation and altitude of the plot, which was used as a proxy for flood
266
disturbance.
267
Given the discontinuous pre-flood data, the temporal changes of FR were modeled as a
268
function of time elapsed since the extreme flood, without including any pre-flood data in the
269
models. The model was fitted with FR as response variable and with time after the flood event
270
(in months), sampling season, and regular flood disturbance as fixed effects, respectively.
271
Sampling plots were treated as random effect, because they were surveyed repeatedly over
272
subsequent periods. Model residuals were considerably temporally autocorrelated, tending to
273
be higher within the same seasons (e.g. spring vs. spring periods) than between seasons (e.g.
274
spring vs. autumn periods). To account for this, an auto-regressive moving average process of
275
second order was added, which improved the AIC of the model significantly.
276
Heteroscedasticity of within-group errors was modelled using a constant variance function.
277
The model did not account for spatial autocorrelation, because another study on exactly the
278
same sampling plots revealed only little spatial dependence of ground beetle diversity
279
(Gerisch 2011) and including further correlation structures would have made the model overly
280
complex.
281
282
All analyses were carried out using R (R Development Core Team 2013) and the packages
283
FD (Laliberté & Shipley 2011), nlme (Pinheiro et al. 2012), and vegan (Oksanen et al. 2012).
284
285
10
286
Results
287
288
A total of 153 ground beetle species were sampled during the 11 sampling periods. Five
289
species (Poecilus versicolor, Agonum emarginatum, Nebria brevicollis , Carabus granulatus,
290
Pterostichus melanarius) made up 49.7% (n = 47,985) of the total density and there were 19
291
species caught with only 1 specimen. See Table SI1 in Supplementary Information for a
292
complete species list including information on the traits for each species. Figure 3 gives an
293
overview on the mean species richness, functional diversity, and functional redundancy of the
294
ground beetle communities for each sampling season (see Table SI2 in Supplementary
295
Information for details).
296
297
Figure 3
298
299
A main hypothesis of this study was that FR of floodplain ground beetles should be higher
300
than would be expected by chance. Figure 4A and Table SI2 display the outcomes of the null
301
models, showing that the degree of overall randomness is relatively large, with mean values
302
varying between -0.1 and +0.3. The results also show a seasonal dependency of the effect
303
sizes: While FR was frequently non-random in spring it tended to be random in all autumn
304
seasons. The highest numbers of random observations were recorded before the extreme flood
305
event and immediately following the flood in autumn 2002 and autumn 2003, whereas the
306
number of random observations decreased with time elapsed since the extreme flood.
307
It was also expected that environmental stochasticity is a main cause for a higher than
308
expected ground beetle FR. Figure 4B outlines the results of the null models along the main
309
axes of hydrological disturbance. The axes represent the orthogonal principal components
310
revealed by a PCA of the environmental variables, explaining 89.4% (axis 1) and 5.9% (axis
311
2), respectively. The results show that, next to the high degrees of randomness, FR is
312
considerably higher than expected in areas with high groundwater levels (i.e., low
313
groundwater depth) and high flood duration. On the contrary, on rarely flood disturbed sites
314
with high groundwater depth, FR is random or even lower than would be expected by chance.
315
11
316
Figure 4
317
318
There was a high, significant correlation between the total number of species and FR (Pearson
319
correlation r=0.77, p<0.001). Correlation was lower between FR and Shannon Diversity
320
(r=0.54, p<0.001) and there was no correlation between FR and the functional diversity of a
321
community (r=0.07, p=0.08). Significantly lower differences in ground beetle functional
322
diversity were found in communities that exhibited a higher FR before the extreme flood
323
event (r=-0.4, p=0.009).
324
325
Mixed effects modeling revealed that FR was significantly positively related to flood
326
disturbance, and also to the time elapsed since the extreme flood event. FR was found to be
327
significantly higher in spring seasons compared to autumn (Table 3).
328
329
Table 2
12
330
Discussion
331
332
Floodplain ground beetle communities exhibited a sizeable proportion of species that did not
333
contribute to functional diversity. This functional redundancy (FR) was, however, to a high
334
degree randomly distributed among the communities, and its magnitude was often lower than
335
expected. There are different perceptions on the extent of FR in various ecological systems
336
and for different taxa, but in many studies it was observed to be lower than expected. For
337
example, Petchey et al. (2007) found no redundancy in British bird communities and there are
338
several studies that detected low or no redundancy in coral reef assemblages (e.g. (Bellwood,
339
Hoey & Choat 2003; Micheli & Halpern 2005; Guillemot et al. 2011). Flynn et al. (2009)
340
highlighted that declines in functional diversity were related to declines in species diversity,
341
suggesting low functional redundancy in multiple taxonomic groups to land use effects.
342
Sasaki et al. (2009) and (Laliberté et al. 2010) found differential patterns for plants, as they
343
reported on increasing FR with decreasing grazing and land use intensity, respectively. But
344
there are also studies reporting on high FR levels in other taxonomic groups and ecosystems,
345
and which support the links between FR and resilience (e.g., Duffy et al. 2001; Moretti et al.
346
2009; Joner et al. 2011; Pillar et al. 2013).
347
The present study stands between the supportive and the contradictory opinions on the
348
importance of FR for community resilience and stability. The often lower than expected FR in
349
this work contradict the initial hypothesis that FR should generally be higher than expected.
350
Possible reasons for this outcome are short-term species replacements, which are typical for
351
invertebrates in dynamic riparian habitats. Floodplain species are exposed to a variety of
352
disturbance agents, such as different flooding or mowing regimes, and even extreme
353
environmental events. Such complex interactions among multiple stressors, but also their
354
spatial heterogeneity and temporal lags, control species and functional diversity (Statzner &
355
Bêche 2010; Tockner et al. 2010), and hence also FR. Loreau (2004) predicts that only neutral
356
coexistence, i.e. the co-occurrence of functionally similar species, allows for FR. But neutral
357
coexistence is unlikely to occur in highly dynamic floodplain habitats, which are
358
characterized by random extinctions and colonization processes in complex spatiotemporal
359
trajectories. For example, flooding eliminates species not adapted to the hydrological
360
conditions in a nonrandom manner, because extinction is known to be clumped in trait space
361
(Petchey et al. 2007). But although the remaining species are well adapted to flooding, they
362
still vary in particular traits, and many of them occur in different microhabitats and appear at
13
363
different times. This high temporal and spatial niche differentiation among species, however,
364
contradicts FR. Interestingly, this effect was less pronounced in the most dynamic habitats,
365
such as margins of frequently flooded oxbows, where FR was significantly higher than
366
expected. I suggest that above certain environmental thresholds the role of environmental
367
filtering and trait convergence increases. Under such conditions, coexistence among
368
functionally similar species seems to be possible, but only during very short timescales. This
369
is supported by predictions that species may play important roles for FR or functional
370
diversity only seasonally, at irregular intervals, or under extreme environmental conditions
371
(Tilman & Downing 1994). Seasonal effects were also revealed by Duffy et al. (2001), who
372
showed that phenologies of species may alter the functional composition of communities, and
373
also their FR. This might also be the case in this study, as the degree of FR was significantly
374
related to the seasonal occurrence of the species.
375
Another argument for the low FR identified is that the artificial communities used here to
376
estimate the randomness of FR lack several species that potentially occur in the study area,
377
but which were not recorded during the study period. The consequence is that the observed
378
FR is, in many cases, similar to what one would expect, because the artificial communities are
379
quite similar to the observed ones (Chase et al. 2011). Moreover, the observed communities
380
might contain too few species or trait variation among the species was too little to detect
381
nonrandom FR patterns more frequently (Petchey et al. 2007).
382
383
A key result of this study is that FR can buffer riparian ground beetles from functional
384
diversity losses during regular flood dynamics, but, given the sizeable differences in
385
functional diversity before and after the extreme flood, this effect was found to be lower than
386
expected in the course of extreme events. Under such conditions, many biological
387
mechanisms maintaining the stability against environmental pressure are overruled because
388
the tolerance to resist disturbances is exceeded for most of the species (Diez et al. 2012).
389
Severe disturbance cause random species losses and this is also a reason for the low FR
390
observed in many post-flood communities, especially immediately following the extreme
391
event. From previous studies it is known that ground beetles can either hibernate during
392
critical seasons or quickly evade disturbed areas as a result of their high mobility
393
(Rothenbücher & Schaefer 2006; Bates, Sadler & Fowles 2006; Lambeets et al. 2008). The
394
present study provides evidence that the flood-surviving species used both strategies, and that
395
they differed in their ecological requirements, their morphology, and in their life-history
396
characteristics. The combination of random colonization by few species that are functionally
14
397
different implies little likelihood that functionally similar species would meet at a random
398
locality, causing relatively low FR immediately after the flood. I recognize that this might be
399
taxa dependent and that more sessile organism groups, such as plants, could show contrary
400
response signals (Ilg et al. 2008).
401
But there are also indications that FR is a primary mechanism for stabilizing the functional
402
diversity of ground beetles under regular disturbances. For example, the quick post-flood
403
increase, the significant relationship to flood disturbance and season, and the fewer functional
404
diversity losses in redundant communities demonstrate that FR insures biodiversity
405
functioning as predicted by Yachi & Loreau (1999). With time elapsing since the extreme
406
flood, stochastic colonization processes were reduced and local regulating forces, such as
407
periodic flood disturbance, subsequently controlled community assembly. This shift from
408
random to directed colonization has led to an increase of functionally similar species,
409
particularly in the spring season when floods normally occur, because the changing habitat
410
templet has shaped the trait requirements, and hence sorted the species with inappropriate
411
functional attributes (Lambeets et al. 2008; Gerisch et al. 2012a). The results therefore
412
underpin the prediction that a temporal hierarchy of functional relationships between species
413
and their environment control community resilience after extreme events (Jentsch et al. 2011).
414
415
One should, however, note that overall community resilience, not only functional resilience as
416
studied here, depends on different controlling agents under different environmental conditions
417
or successional stages of the habitat. It becomes obvious from this study that the capacity of
418
ground beetles to re-establish networks of ecological processes and properties depends not
419
only on FR and the intensity of disturbances, but also on priority effects and, of course, on
420
time. Similar findings were also reported for post-disturbance succession of zooplankton in
421
ponds (Jenkins & Buikema 1998). However, little is known about the roles that particular
422
species or variables have for community functioning, and much of the work (including this
423
study) relates only limited sets of biological traits to processes or properties that were
424
considered to be important for ecological functioning. Therefore, further effort needs to be put
425
into basic questions of ecology. These include identifying the traits that are relevant for
426
specific properties such as resilience, improving the transfer of theoretical background into
427
empirical studies, and finding appropriate ways to generalize among the diversity of taxa,
428
ecosystems, and stress agents.
429
430
15
431
Acknowledgements
432
This research was supported by the HABEX project (RA/V1658), funded by the German
433
Federal Institute of Hydrology, and by the BIOFRESH project
434
(http://www.freshwaterbiodiversity.eu, contract no. 226874), funded by the EU under the 7th
435
framework program. I am indebted to Kurt Jax, Frank Dziock, and Owen Petchey for support
436
in shaping the conceptual body of the paper. Two anonymous reviewers gave very helpful
437
comments that improved the manuscript considerably. Species identification was done by
438
Arno Schanowski.
439
16
440
References
441
442
443
444
Allen, C.D. & Breshears, D.D. (1998) Drought-induced shift of a forest-woodland ecotone:
rapid landscape response to climate variation. Proceedings of the National Academy of
Sciences of the United States of America, 95, 14839–42.
445
446
Bates, A.J., Sadler, J.P. & Fowles, A.P. (2006) Condition-dependent dispersal of a patchily
distributed riparian ground beetle in response to disturbance. Oecologia, 150, 50–60.
447
448
Bêche, L. a. & Statzner, B. (2009) Richness gradients of stream invertebrates across the USA:
taxonomy- and trait-based approaches. Biodiversity and Conservation, 18, 3909–3930.
449
450
451
Bellwood, D.R., Hoey, A.S. & Choat, J.H. (2003) Limited functional redundancy in high
diversity systems: resilience and ecosystem function on coral reefs. Ecology Letters, 6,
281–285.
452
453
Bihn, J.H., Gebauer, G. & Brandl, R. (2010) Loss of functional diversity of ant assemblages
in secondary tropical forests. Ecology, 91, 782–92.
454
455
456
Chase, J.M. (2007) Drought mediates the importance of stochastic community assembly.
Proceedings of the National Academy of Sciences of the United States of America, 104,
17430–4.
457
458
459
Chase, J.M., Kraft, N.J.B., Smith, K.G., Vellend, M. & Inouye, B.D. (2011) Using null
models to disentangle variation in community dissimilarity from variation in α-diversity.
Ecosphere, 2, art24.
460
461
462
463
464
465
466
Ciais, P., Reichstein, M., Viovy, N., Granier, a, Ogée, J., Allard, V., Aubinet, M., Buchmann,
N., Bernhofer, C., Carrara, a, Chevallier, F., De Noblet, N., Friend, a D., Friedlingstein,
P., Grünwald, T., Heinesch, B., Keronen, P., Knohl, a, Krinner, G., Loustau, D., Manca,
G., Matteucci, G., Miglietta, F., Ourcival, J.M., Papale, D., Pilegaard, K., Rambal, S.,
Seufert, G., Soussana, J.F., Sanz, M.J., Schulze, E.D., Vesala, T. & Valentini, R. (2005)
Europe-wide reduction in primary productivity caused by the heat and drought in 2003.
Nature, 437, 529–33.
467
468
469
Dalerum, F., Cameron, E.Z., Kunkel, K. & Somers, M.J. (2010) Interactive effects of species
richness and species traits on functional diversity and redundancy. Theoretical Ecology,
129–139.
470
471
Díaz, S. & Cabido, M. (2001) Vive la différence : plant functional diversity matters to
ecosystem processes. Trends in Ecology & Evolution, 16, 646–655.
472
473
474
475
Díaz, S., Lavorel, S., de Bello, F., Quétier, F., Grigulis, K. & Robson, T.M. (2007)
Incorporating plant functional diversity effects in ecosystem service assessments.
Proceedings of the National Academy of Sciences of the United States of America, 104,
20684–9.
476
477
478
479
Diez, J.M., D’Antonio, C.M., Dukes, J.S., Grosholz, E.D., Olden, J.D., Sorte, C.J.,
Blumenthal, D.M., Bradley, B.A., Early, R., Ibáñez, I., Jones, S.J., Lawler, J.J. & Miller,
L.P. (2012) Will extreme climatic events facilitate biological invasions? Frontiers in
Ecology and the Environment, 10, 249–257.
480
481
482
Duffy, J.E., Macdonald, K.S., Rhode, J.M. & Parker, J.D. (2001) Grazer diversity, functional
redundancy, and productivity in seasgrass beds: an experimental test. Ecology, 82, 2417–
2434.
483
484
485
Easterling, D.R., Meehl, G.A., Parmesan, C., Changnon, S.A., Karl, T.R. & Mearns, L.O.
(2000) Climate extremes: observations, modeling, and impacts. Science, 289, 2068–
2074.
17
486
487
488
Flynn, D.F.B., Gogol-Prokurat, M., Nogeire, T., Molinari, N., Richers, B.T., Lin, B.B.,
Simpson, N., Mayfield, M.M. & DeClerck, F. (2009) Loss of functional diversity under
land use intensification across multiple taxa. Ecology letters, 12, 22–33.
489
490
Gerisch, M. (2011) Habitat disturbance and hydrological parameters determine the body size
and reproductive strategy of alluvial ground beetles. ZooKeys, 370, 353–70.
491
492
493
Gerisch, M., Agostinelli, V., Henle, K. & Dziock, F. (2012a) More species, but all do the
same: contrasting effects of flood disturbance on ground beetle functional and species
diversity. Oikos, 121, 508–515.
494
495
496
497
Gerisch, M., Dziock, F., Schanowski, A., Ilg, C. & Henle, K. (2012b) Community resilience
following extreme disturbances: the response of ground beetles to a severere summer
flood in a Central European lowland stream. River Research and Applications, 92, 81–
92.
498
Gotelli, N.J. & Graves, G.R. (1996) Null Models in Ecology. Washington, D.C.
499
500
501
Guillemot, N., Kulbicki, M., Chabanet, P. & Vigliola, L. (2011) Functional Redundancy
Patterns Reveal Non-Random Assembly Rules in a Species-Rich Marine Assemblage (ed
RKF Unsworth). PLoS ONE, 6, e26735.
502
503
504
505
Henle, K., Dziock, F., Foeckler, F., Volker, K., Hüsing, V., Hettrich, A., Rink, M., Stab, S. &
Scholz, M. (2006) Study Design for Assessing Species Environment Relationships and
Developing Indicator Systems for Ecological Changesin Floodplains – The Approach of
the RIVA Project. International Review of Hydrobiology, 91, 292–313.
506
507
Hubbell, S.P. (2006) Neutral theory and the evolution of ecological equivalence. Ecology, 87,
1387–98.
508
509
510
Ilg, C., Dziock, F., Foeckler, F., Follner, K., Gerisch, M., Glaeser, J., Rink, A., Schanowski,
A., Scholz, M., Deichner, O. & Henle, K. (2008) Long-term reactions of plants and
macroinvertebrates to extreme floods in floodplain grasslands. Ecology, 89, 2392–8.
511
512
Jenkins, D. & Buikema, A. (1998) Do similar communities develop in similar sites? A test
with zooplankton structure and function. Ecological Monographs, 68, 421–443.
513
514
Jentsch, A. & Beierkuhnlein, C. (2003) Global climate change and local disturbance regimes
as interacting drivers for shifting altitudinal vegetation patterns. Erdkunde, 57, 216–231.
515
516
Jentsch, A., Kreyling, J. & Beierkuhnlein, C. (2007) A new generation of climate-change
experiments: events, not trends. Frontiers in Ecology and the Environment, 5, 365–374.
517
518
519
520
521
Jentsch, A., Kreyling, J., Elmer, M., Gellesch, E., Glaser, B., Grant, K., Hein, R., Lara, M.,
Mirzae, H., Nadler, S.E., Nagy, L., Otieno, D., Pritsch, K., Rascher, U., Schädler, M.,
Schloter, M., Singh, B.K., Stadler, J., Walter, J., Wellstein, C., Wöllecke, J. &
Beierkuhnlein, C. (2011) Climate extremes initiate ecosystem-regulating functions while
maintaining productivity. Journal of Ecology, 99, 689–702.
522
523
Joner, F., Specht, G., Müller, S.C. & Pillar, V.D. (2011) Functional redundancy in a clipping
experiment on grassland plant communities. Oikos, 120, 1420–1426.
524
Konopka, A. (2009) What is microbial community ecology? The ISME journal, 3, 1223–30.
525
526
Laliberté, E. & Legendre, P. (2010) A distance-based framework for measuring functional
diversity from multiple traits. Ecology, 91, 299–305.
527
528
Laliberté, E. & Shipley, B. (2011) FD: measuring functional diversity from multiple traits,
and other tools for functional ecology. R package version 1.0-11.
529
530
Laliberté, E., Wells, J. a, Declerck, F., Metcalfe, D.J., Catterall, C.P., Queiroz, C., Aubin, I.,
Bonser, S.P., Ding, Y., Fraterrigo, J.M., McNamara, S., Morgan, J.W., Merlos, D.S.,
18
531
532
Vesk, P. a & Mayfield, M.M. (2010) Land-use intensification reduces functional
redundancy and response diversity in plant communities. Ecology letters, 13, 76–86.
533
534
535
Lambeets, K., Vandegehuchte, M.L., Maelfait, J. & Bonte, D. (2008) Understanding the
impact of flooding on trait-displacements and shifts in assemblage structure of predatory
arthropods on river banks. Journal of Animal Ecology, 1162–1174.
536
537
Lawton, J.H. & Brown, K.V. (1993) Redundancy in Ecosystems. Biodiversity and ecosystem
function (eds E.D. Schulze & H.A. Mooney), pp. 255–270. Springer, New York.
538
Loreau, M. (2004) Does functional redundancy exist? Oikos, 104, 606–611.
539
540
Micheli, F. & Halpern, B.S. (2005) Low functional redundancy in coastal marine
assemblages. Ecology Letters, 8, 391–400.
541
542
543
Moretti, M., de Bello, F., Roberts, S.P.M. & Potts, S.G. (2009) Taxonomical vs. functional
responses of bee communities to fire in two contrasting climatic regions. The Journal of
animal ecology, 78, 98–108.
544
545
546
547
Naeem, S., Loreau, M. & Inchausti, P. (2002) Biodiversity and ecosystem functioning: the
emergence of a synthetic ecological framework. Biodiversity and Ecosystem
Functioning. Synthesis and Perspectives (eds M. Loreau, S. Naeem & P. Inchausti), pp.
3–17. Oxford University Press, Oxford.
548
549
Naeem, S. & Wright, J.P. (2003) Disentangling biodiversity effects on ecosystem functioning:
deriving solutions to a seemingly insurmountable problem. Ecology Letters, 6, 567–579.
550
551
Niemelä, J. & Kotze, D.J. (2009) Carabid beetle assemblages along urban to rural gradients:
A review. Landscape and Urban Planning, 92, 65–71.
552
553
554
Oksanen, J., Guillaume Blanchet, F., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B.,
Simpson, G.L., Solymos, P., Stevens, H.H. & Wagner, H. (2012) vegan: Community
Ecology Package. R package version 2.0-1.
555
556
Petchey, O.L., Evans, K.L., Fishburn, I.S. & Gaston, K.J. (2007) Low functional diversity and
no redundancy in British avian assemblages. The Journal of animal ecology, 76, 977–85.
557
558
Petchey, O.L. & Gaston, K.J. (2009) Effects on ecosystem resilience of biodiversity,
extinctions, and the structure of regional species pools. Theoretical Ecology, 2, 177–187.
559
560
561
Pillar, V.D., Blanco, C.C., Müller, S.C., Sosinski, E.E., Joner, F. & Duarte, L.D.S. (2013)
Functional redundancy and stability in plant communities (ed F de Bello). Journal of
Vegetation Science, 24, 963–974.
562
563
Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. & R Development Core Team. (2012) nlme:
Linear and Nonlinear Mixed Effects Models. R package version 3.1-102.
564
R Development Core Team. (2013) R: A language and environment for statistical computing.
565
566
567
Resetarits, W.J. & Chalcraft, D.R. (2007) Functional diversity within a morphologically
conservative genus of predators: implications for functional equivalence and redundancy
in ecological communities. Functional Ecology, 21, 793–804.
568
569
Ribera, I., Dolédec, S., Downie, I.S. & Foster, G.N. (2001) Effect of land disturbance and
stress on species traits of ground beetle assemblages. Ecology, 82, 1112–1129.
570
571
Rosenfeld, J.S. (2002) Functional redundancy in ecology and conservation. Oikos, 98, 156–
162.
572
573
Rothenbücher, J. & Schaefer, M. (2006) Submersion tolerance in floodplain arthropod
communities. Basic and Applied Ecology, 7, 398–408.
19
574
575
576
Sasaki, T., Okubo, S., Okayasu, T., Jamsran, U., Ohkuro, T. & Takeuchi, K. (2009) Twophase functional redundancy in plant communities along a grazing gradient in
Mongolian rangelands. Ecology, 90, 2598–608.
577
Schiermeier, Q. (2003) Analysis pours cold water on flood theory. Nature, 425, 111.
578
579
580
581
Scholten, M., Anlauf, A., Büchele, B., Faulhaber, P., Henle, K., Kofalk, S., Leyer, I.,
Meyerhoff, J., Purps, J., Rast, G. & Scholz, M. (2005) The Elbe River in Germany present state, conflicts, and perspectives of rehabilitation. Archiv für Hydrobiologie,
Supplement, 579–602.
582
583
Sorte, C.J.B., Fuller, A. & Bracken, M.E.S. (2010) Impacts of a simulated heat wave on
composition of a marine community. Oikos, 119, 1909–1918.
584
585
Statzner, B. & Bêche, L. (2010) Can biological invertebrate traits resolve effects of multiple
stressors on running water ecosystems? Freshwater Biology, 55, 80–119.
586
587
588
Thibault, K.M. & Brown, J.H. (2008) Impact of an extreme climatic event on community
assembly. Proceedings of the National Academy of Sciences of the United States of
America, 105, 3410–5.
589
590
Tilman, D. & Downing, J.A. (1994) Biodiversity and stability in grasslands. Nature, 367,
363–365.
591
592
Tockner, K., Pusch, M., Borchardt, D. & Lorang, M.S. (2010) Multiple stressors in coupled
river-floodplain ecosystems. Freshwater Biology, 55, 135–151.
593
594
595
596
Villéger, S., Ramos Miranda, J., Flores Hernández, D. & Mouillot, D. (2010) Contrasting
changes in taxonomic vs. functional diversity of tropical fish communities after habitat
degradation. Ecological applications : a publication of the Ecological Society of
America, 20, 1512–22.
597
598
Walker, B.H. (1992) Biodiversity and Ecological Redundancy. Conservation Biology, 6, 18–
23.
599
600
601
White, T.A., Campbell, B.D., Kemp, P.D. & Hunt, C.L. (2000) Sensitivity of three grassland
communities to simulated extreme temperature and rainfall events. Global Change
Biology, 6, 671–684.
602
603
604
Yachi, S. & Loreau, M. (1999) Biodiversity and ecosystem productivity in a fluctuating
environment: the insurance hypothesis. Proceedings of the National Academy of
Sciences of the United States of America, 96, 1463–1468.
605
20
606
Tables
607
608
Table 1. Species traits used to calculate functional diversity and redundancy.
Type of
functional
trait
Trait
Trait
categories
Relevance/ecological meaning
No. of
species
Effect traits
Wing
morphology
Macropterous
High dispersal capacity, i.e. high
recolonization potential
105
Brachypterous
Low dispersal capacity, i.e. low
recolonization potential
11
Dimorphic
Intermediate dispersal capacity
37
Adult
Is correlated with time of
activity, but seems to the better
predictor. Species emerge as
adults in early spring and can
therefore quickly recolonize
flooded areas.
113
Larvae
Species emerge as adults in late
summer or autumn.
22
Both
Species can reproduce either in
spring or autumn. Possibly
dependent on habitat or
geographical locality.
18
Body size
Continuous
One of the most important traits
shaping species physiology and
life-history.
153
Feeding mode
Carnivorous
Phytophagous
Polyphagous
Species can co-occur in the
same habitat, but differ in their
feeding strategies.
93
23
36
Daily activity
Diurnal
Nocturnal
Both
Species can co-occur in the
same habitat in the same season,
but have separated temporal
niches.
69
75
9
Mode of
overwintering
Response
traits
21
Light
preference
Unshaded
Partly shaded
Mainly shaded
Species can co-occur in the
same habitat, but have different
preferences for micro habitats.
92
37
24
Humidity
preference
Hygrophilus
Mesophilous
Xerophious
Important preference trait in
wetlands for niche separation.
96
21
36
Habitat
specialization
Eurytopic
Stenotopic
Generalist species are often the
first arrivals after disturbances.
106
47
609
610
22
611
612
613
614
615
Table 2. Relationships between functional redundancy (FR) and sampling season, regular
flood disturbance, and time since the extreme flood. Results are based on a linear mixed
effects model with FR as response variable and sampling season, flood disturbance, and time
since the extreme flood as fixed effects. Sampling plots were used as random effects. E =
parameter estimate, SE = standard error, DF = degrees of freedom, t = t-value, p = p-value.
Explanatory variables
E
SE
DF
t
p
Intercept
0.029
0.017
278
1.778
0.076
Season
0.182
0.016
278
11.008
< 0.001
0.161
0.022
46
7.162
< 0.001
0.004
0
278
7.966
< 0.001
(scale)
(categorical, spring & autumn)
Flood disturbance
(continuous)
Time after the extreme flood
(categorical, months)
616
617
23
618
Figure legends
619
620
Figure 1. Study area and location of the study sites in Germany (artwork: Wilfried Rohloff,
621
Berlin).
622
623
Figure 2. Conceptual approach to calculate functional redundancy (FR) in this study. The
624
basis is a species-by-site matrix with abundances within cells and another matrix containing
625
the traits of each species. Initially, the total number of species per plot (i) was calculated and
626
then 2 to i species were randomly combined into a species subset and a corresponding trait
627
subset. This step was repeated 1,000 times, resulting in 1,000 random species–trait
628
combinations for each subset size. For each pair of species and trait subsets and for each
629
replication step, functional diversity was calculated and stored in a matrix (rows = number of
630
replication, columns = subset size). For all subset sizes, the average functional diversity
631
among the replications was then calculated and FR estimated as the proportion of species that
632
did not increase the average functional diversity of a community.
633
634
Figure 3. Change of species richness, functional diversity, and functional redundancy (FR)
635
during the study period. Each boxplot contains the values of 48 sampling plots. The gap
636
between month -34 and month +1 displays the period where sampling was interrupted (spring
637
2000-summer 2002, not scale-corrected due to visual reasons).
638
639
Figure 4. Graphical representation of the null model results to identify whether the observed
640
functional redundancy (FR) differs from random expectations. Randomness of FR is
641
expressed by means of effect sizes, with values close to 0 (light green) indicate random
642
observations, while values close to -1 (red) and +1 (blue) indicate considerably lower or
643
higher than expected FR, respectively. (A) Distribution of effect sizes during spring and
644
autumn seasons for each sampling year. Polygons of each beanplot represent a kernel density
645
function of the distribution, each white horizontal line represents a data point, black
646
horizontal lines represent the means of the effect sizes in each season; (B) Distribution of
647
effect sizes along two main gradients of flood disturbance (PCA1, PCA2), revealed by a PCA
24
648
based on hydrological parameters. A loess-smoother was used to interpolate between each
649
combination of PCA1 and PCA2 scores.
25