Contents lists available atScienceDirect
Biological Conservation
journal homepage:www.elsevier.com/locate/biocon
Saproxylic biodiversity and decomposition rate decrease with small-scale
isolation of tree hollows
Laia Mestre
a,⁎, Nicklas Jansson
b, Thomas Ranius
aaDepartment of Ecology, Swedish University of Agricultural Sciences, Box 7044, 750 07 Uppsala, Sweden bIFM Biology, Conservation Ecology Group, Linköping University, 581 83 Linköping, Sweden
A R T I C L E I N F O Keywords:
Biodiversity-ecosystem functioning Coleoptera
Forests
Trophic rank hypothesis Habitat amount hypothesis Mass-ratio hypothesis
A B S T R A C T
Biodiversity is fundamental for ecosystem functioning, but little is known about how function responds to biodiversity loss following habitat disturbance in natural systems. Due to the global decay of veteran trees, many associated saproxylic (i.e. deadwood-dependent) insects are considered threatened. Nevertheless, the role of habitat spatial configuration on saproxylic insect biodiversity and dead wood decomposition is poorly under-stood. We performed a six-year landscape-scale colonization experiment on saproxylic beetles inhabiting hollow oaks, using boxesfilled with wood mould as standardized habitat patches. We placed boxes either on a hollow tree or on another tree 61–324 m from the hollows, thereby creating two habitat isolation levels. We quantified wood mould decay and biodiversity in the boxes, measuring species richness, total abundances and community-weighted mean of body mass (CWM) as an index of community functional composition. Isolation had a persistent negative effect on primary consumer biodiversity, but it only impaired decay at the beginning of the experiment. All effects were independent of landscape-level (500-m radius) habitat amount surrounding the boxes. Wood mould decay was mediated by CWM of primary consumers. Therefore function was driven by the body masses of the dominant primary consumer species but not by species numbers (richness) or individual numbers (abun-dance). Our experiment shows that small-scale habitat isolation leads to biodiversity loss and reduced function and indicates that habitats created by conservation efforts will be used by more saproxylic species if located within sites with a high density of veteran trees.
1. Introduction
Research into the link between biodiversity and ecosystem func-tioning (BEF) shows that the occurrence and abundance of a few common species can be more important than total species richness, which includes many rare species (Dangles and Malmqvist, 2004; Winfree et al., 2015). Biodiversity experiments that assemble random species mixtures of a single trophic level differing in richness and composition have been a common tool to research BEF (Balvanera et al., 2006; Tilman et al., 2014). While these experiments have in-creased our mechanistic understanding of BEF, their ability to predict the effects of biodiversity loss on real ecosystems is constrained because natural communities are highly diverse, comprise multiple trophic le-vels and undergo non-random changes in species composition brought about by different causes (De Laender et al., 2016; Srivastava and Vellend, 2005). Moreover, species richness is a poor predictor of eco-system functioning when species in the community differ in traits that are relevant to the function considered (“functional effect traits”)
(Violle et al., 2007). Thus, functional trait approaches are needed for further insights into the BEF relationship (Cadotte et al., 2011;Petchey et al., 2004).
Conceptual frameworks for the negative impacts of habitat frag-mentation on biodiversity are provided by island biogeography and metapopulation theory, which predict that increasing patch isolation reduces colonization rates, species persistence and species richness (Hanski, 1998; McArthur and Wilson, 1967). The trophic rank hy-pothesis predicts that these detrimental isolation effects will be com-pounded up the food chain (Holt, 1996), resulting in a disproportionate loss of predators and the disruption of species interactions in highly fragmented landscapes (Ewers and Didham, 2006). Therefore, an ex-cellent way to understand how non-random biodiversity loss affects ecosystem functioning is to study the BEF relationship as affected by manipulating the configuration of relevant habitats in a landscape context. Nevertheless, most research into landscape effects on function is correlational or merely quantifies biodiversity responses without actually measuring function (Chaplin-Kramer et al., 2011). Hence,
https://doi.org/10.1016/j.biocon.2018.09.023
Received 12 March 2018; Received in revised form 7 September 2018; Accepted 17 September 2018
⁎Corresponding author at: Institute for Environmental Sciences, University of Koblenz-Landau, Fortstraße 7, 76829 Landau, Germany. E-mail address:mestre@uni-landau.de(L. Mestre).
Available online 22 September 2018
0006-3207/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
landscape-scale experiments at the community level are vital to assess the relevance of spatial configuration in biodiversity conservation and ecosystem functioning.
Saproxylic insects, defined by their dependence on dead wood to complete their life cycle, comprise a large part of the fauna in forests (Grove, 2002; Speight, 1989). Agricultural expansion and intensive forest management remove dead wood from ecosystems and pose a serious threat to worldwide saproxylic biodiversity (Cálix et al., 2018; Seibold et al., 2015; Stokland et al., 2012). However, the effect of spatial configuration on habitat use by saproxylic insects remains un-clear, because in many studies sample sizes increase with habitat patch size or dispersal sources are poorly identified (Ranius et al., 2015; Seibold et al., 2017). Moreover, effects of spatial configuration may be hidden due to habitat history (but seeRanius et al., 2011) and the effect of isolation can be due to both the amount of habitat within the local landscape and to inter-patch distances, which are difficult to disen-tangle in correlational studies (Fahrig, 2013). Therefore, there is a need for experiments manipulating inter-patch distances that control for habitat amount, quality and history.
The decomposition of organic matter plays a key role in carbon storage and nutrient cycling (van der Wal et al., 2013). While the link between biodiversity and organic matter decomposition in freshwater and soil systems is well known (Gessner et al., 2010;Hättenschwiler et al., 2005), the relationship between dead wood decomposition and biodiversity is largely unexplored. Saproxylic insects may facilitate this process by fragmenting and partially digesting litter, increasing its availability to microbial decomposers (Ulyshen, 2016). However, the importance of their contribution is controversial: caging dead wood to exclude saproxylic insects decreases decomposition rates, but caging itself confounds the results by altering microclimate conditions and thus changing the decomposer fauna (Ulyshen and Wagner, 2013).
We performed a landscape-level experiment to determine the im-pact of small-scale isolation on saproxylic beetle biodiversity and dead wood decomposition in hollow oaks. Tree hollows occur mainly in veteran trees, which are globally declining (Lindenmayer et al., 2014). Hollows form when oaks are over 200 years old and become progres-sively larger andfilled with wood mould – a mixture of decaying wood, litter and fungal hyphae that is occupied by saproxylic insects (Siitonen and Ranius, 2015). Wooden boxes with artificial wood mould are an effective surrogate for tree hollows (Jansson et al., 2009) and provide an ideal study system to manipulate spatial habitat configuration in-dependently of habitat amount. Specifically, we created two levels of habitat isolation on the basis of the distance between a box and a natural hollow by placing boxes either on a hollow oak or on another tree away from the natural hollows. Over six years we monitored wood mould decay and saproxylic biodiversity. We considered two trophic levels: primary consumers (consuming fungi, organic matter or various wood substrates; see“Sampling methods”) and predators (eating other saproxylic insects). Apart from measuring species richness and total abundances per box, we also calculated the community-weighted mean of body mass (CWM) as an index of community functional composition (Gagic et al., 2015;Lavorel et al., 2008), because the body mass of an organism is related to energy demands, potentially linking this trait to dead wood decomposition. The community-weighted mean of a trait is the mean value of the trait weighted by the abundance of each species. Therefore, its value is determined by the values of the most abundant species (Laliberté and Legendre, 2010) and serves as a test of the mass-ratio hypothesis, which postulates that function is essentially driven by the functional traits of the dominant species in the community (Grime, 1998).
We predicted that (1) small-scale habitat isolation will decrease saproxylic beetle biodiversity due to longer distances between boxes and source hollows. Moreover, (2) effects will be stronger on predators, according to the trophic rank hypothesis. In addition, lower saproxylic biodiversity will lead to (3) lower wood mould decay rates due to a positive biodiversity–decomposition relationship driven by primary
consumers. We also expected that (4) CWM of body mass will better predict decomposition than species richness or total abundance, in support of the mass-ratio hypothesis.
2. Material and methods 2.1. Study sites and habitat
We selected eight oak pastures (Quercus robur L.) in the county of Östergötland (southeastern Sweden) that are at a low elevation (< 200 m a.s.l) and have a humid continental climate. Historically, these pastures were semi-open lands managed by grazing or hay-making, and most of them are still grazed by cattle (See Appendix A: Fig. A.1). Today the landscapes are dominated by managed coniferous forests and arablefields.
We used 0.70 × 0.30 × 0.30 m oak-wood boxes to mimic the en-vironment in the natural hollows in oaks and attached them to the trunk of oaks or other deciduous trees with a metallic band at a height of 2.4–3.9 m. The boxes had an 80 mm-diameter circular orifice on one of the long sides and looked like large bird nest boxes. To let some rain water in, the boxes had a cross milled on the roof and four 8 mm-dia-meter holes drilled in the corners. To retain moisture, we covered the bottom inside each box with a plastic sheet forming a 50 mm-high container.
Wefilled the boxes to 70% with 45 L of artificial wood mould, consisting of a mixture of sawdust and dead oak leaves. We also added 1 L of lucernflour and 5 L of water, to create a humid environment rich in nutrients (Carlsson et al., 2016;Jansson et al., 2009), which is im-portant for the development of beetle larvae (Landvik et al., 2016). Sawdust came from fresh wood of Quercus robur and we inoculated it with a block of agar (30 mm3) with actively growing mycelia of the polypore Laetiporus sulphureus, an important fungus creating hard wood decay in the trunks of old oaks. We cultured the fungus on HAGEM-agar in 90 mm-diameter Petri dishes, kept in darkness at 20 ± 2 °C for one month before inoculating the artificial wood mould in each box. 2.2. Experimental design
In 2009 we installed 10 wood mould boxes per landscape (total = 80) and left them until 2014. In each landscape we selected a core area with hollow oaks (defined as those with at least one entrance hole over 10 cm in diameter). At each site, we assigned each box to either of two habitat isolation treatments according to their distance to hollows. We placed connected boxes (N = 5) one each onfive selected hollow oaks in the core area and thus their distance from a hollow oak was 0. We placed isolated boxes (N = 5) each on a tree with no hollows located 61–324 m away from the core area (mean distance = 175 m) in a direction without hollow trees and at least 68–253 m from the closest hollow oak (mean minimum distance = 136 m; see Appendix A: Fig. A.1). In oak pastures, the occurrence of saproxylic beetles in hollows responds to the density of surrounding hollow oaks at scales ranging from 50 to 5000 m, with a median among species of around 500 m (Bergman et al., 2012: 500 m; Ranius et al., 2010: 400–900 m de-pending on which metrics were used). Thus, we deem the isolation distances that we achieved to be moderate. Our experimental design kept the amount of new habitat (wood mould boxes) constant within a site while having two different levels of isolation distances. Never-theless, sites had different numbers of hollow oaks (10–42), and therefore varying numbers of dispersal sources for saproxylic beetles potentially colonizing the boxes. Based on previous studies (Bergman et al., 2012;Ranius et al., 2010), we considered 500 m to be a relevant spatial scale for the overall saproxylic beetle community in our study system, and so we calculated the number of non-experimental hollow oaks within a 500-m radius for each box as a measure of colonization sources (“amount of habitat in the local landscape” sensuFahrig (2013)) available around them. We found that connected boxes were
surrounded, on average, by slightly more hollow oaks than isolated boxes (mean ± SE) (F1,75= 4.51, p = 0.037; connected = 28.6 ± 1.8 hollows; isolated = 22.5 ± 1.3 hollows).
2.3. Sampling methods
Boxes were open for colonization throughout the experiment (2009–2014). In 2010, 2012 and 2014, we put a pitfall trap in each box to sample adult saproxylic beetles, which had either migrated into the boxes or emerged from larvae that had developed inside. The traps operated for one week three times each year (May, June, July). They consisted of a plastic jar (70 mm-diameter opening and 70 mm height) placed with their opening level with the wood mould surface in the box. Wefilled pitfalls to 70% with a preservative liquid (50:50 propylene glycol:water) and added a few drops of liquid detergent to reduce the surface tension. In 2014, pitfall traps in three boxes were lost and we could not use them for analysis. In 2012 and 2014 we measured changes in the level of wood mould (in cm) inside the boxes to use wood mould decrease as a proxy for decomposition. Adjacent to each of the four sides of the boxes, we measured differences between the ori-ginal and the current height to calculate an average decrease. By using level of wood mould as a proxy for wood mould decay, we were able to conduct repeated measurements over the years without disturbing the beetle community, and therefore to reveal temporal patterns in biodi-versity and decomposition. Data about wood mould was lost for four boxes during the experiment.
2.4. Species functional traits
Saproxylic beetle species (Dahlberg and Stokland, 2004) were identified mainly by NJ and Gunnar Sjödin. We used the categories provided byKöhler (2000)to assign species to one of two trophic levels: primary consumers and predators. Primary consumers comprised my-cetophagous (feeding on wood fungi), necrophagous (dead animals), xylophagous (woody substrate), xylo-saprophagous (wood and dead organic material), saprophagous (dead organic material) and xylomy-cetophagous (fungi-infested wood) species. Predators comprised zoo-phagous species (animals) and a single xylo-zoozoo-phagous species (both wood and animals). We obtained data on species' adult body sizes (total body length) from the literature and used length–weight regression equations to estimate their body mass for CWM calculations (see Ap-pendix A: Table A.1).
2.5. Statistical analysis
We tested for isolation effects on biodiversity by analyzing species richness, total abundance and CWM of each trophic level (primary consumers, predators) in the years 2010, 2012 and 2014. Habitat iso-lation was the result of our experimental design: connected boxes, being each on a hollow oak, were a distance of 0 m from the dispersal sources, whereas isolated boxes, each on non-hollow trees, were, on average, 136 m away from hollows. For each year, we used (generalized) linear mixed models with“treatment” (connected, isolated) and “number of surrounding hollow oaks” (surrounding each box within a 500-m ra-dius) as fixed factors and “landscape” as a random intercept. We in-cluded the variable“number of surrounding hollow oaks” as a way to statistically control for differences in colonization sources (amount of habitat in local landscape) between sites (see “2.2 Experimental de-sign”). We used a Poisson error distribution when modelling count data of species richness and abundance (checked for under- and over-dispersion) and a Gaussian distribution for CWM, log10-transformed to normalize residual distributions. There was no collinearity (VIF <
1.3).
We analyzed how isolation influenced wood mould decay per box between the years when we measured wood mould level (2009–2012 and 2012–2014), and overall (2009–2014) using linear mixed models
with “treatment” (connected, isolated) and “number of surrounding hollow oaks” as fixed factors and “landscape” as a random intercept. Because wood mould level in the boxes could not decrease beyond 0, we rescaled the response variable to a proportion to make it comparable between time periods. Therefore, we considered wood mould decay as the proportion of decrease relative to the beginning (2009–2012 and 2009–2014) or to the previous level of wood mould (2012–2014) and we applied logit-transformation to fit the assumptions of a normal distribution, as recommended byWarton and Hui (2011). There was no collinearity (VIF < 1.1).
Finally, we analyzed the link between saproxylic biodiversity and wood mould decay. We ran three linear mixed models for each of the periods 2009–2012, 2012–2014 and overall (2009–2014). Each model included six continuous predictors:“species richness of primary con-sumers”, “species richness of predators”, “total abundance of primary consumers”, “total abundance of predators”, “CWM of primary con-sumers”, and “CWM of predators”, with “landscape” as a random in-tercept. There was no collinearity (VIF < 2.1). The predictor variables for each of these three models had different values: to better predict wood mould decay in each 3–5 year time period, we considered all developmental stages of beetles occurring inside the boxes. Larvae of saproxylic beetles have developmental times that can be as long as a few years (Palm, 1959). Therefore we calculated richness, abundance and CWM by including all captured adult beetles during the time period encompassed by each model: in the model for 2009–2012, we included individuals sampled in 2010 and 2012; in the 2012–2014 model, those sampled in 2012 and 2014; in the overall model (2009–2014), all in-dividuals. We then performed automated model selection based on Akaike's information criterion for small sample sizes (AICc) using the “dredge” function with a ΔAICc < 4 cut-off, indicating that the level of empirical support of the reduced model is no longer“substantial” (sensu Burnham and Anderson, 2002). We analyzed data with R 3.5.0 (R Core Team, 2018), with the packages“lme4” (Bates et al., 2015) and“car” (Fox and Weisberg, 2011) to calculate p values,“FD” to calculate CWM (Laliberté et al., 2014) and “MuMIn” for model selection (Bartoń, 2016).
3. Results
We collected 2768 saproxylic beetles, comprising 143 species from 33 families (see Appendix A: Table A.1). There were 67 predator and 76 primary consumer species. There were 85 obligate saproxylic species, which represented 29.6% of the individuals, and the rest were fa-cultative saproxylic. The most common family was Staphylinidae (39.2% of the individuals), followed by Anobiidae (16.7%) and Latridiidae (9.2%). The three most abundant (i.e. dominant) primary consumer species were Ptinus fur (Anobiidae), Cryptophagus scanicus (Cryptophagidae) and Corticaria serrata (Latridiidae), whereas the three most common predator species were Haploglossa villosula (Staphylinidae), Atheta harwoodi (Staphylinidae) and Atheta nigricornis (Staphylinidae). These six species together accounted for 42% of the total individuals in the boxes.
3.1. Habitat isolation effect on saproxylic biodiversity and wood mould decay
Across years and treatments, there was an average of 4.9 species and 12.1 saproxylic beetle individuals in each wood mould box. Isolation had a generally negative effect on species richness (2010: Wald χ2
1= 7.48, p = 0.0063; 2012: Waldχ21= 8.92, p = 0.0028; 2014: Wald χ2
1= 13.31, p = 0.00026;Fig. 1), but it varied between trophic levels: while isolation reduced primary consumer richness by half every year (55.7% lower richness in 2010: Waldχ21= 9.81, p < 0.0001; 49.3% in 2012: Wald χ2
1= 12.25, p = 0.00046; 49.8% in 2014: Wald χ2
1= 22.63, p < 0.0001), predator species numbers were 26.5% higher in isolated boxes in 2010 (Wald χ21= 4.32, p = 0.038) and were
thereafter unaffected by isolation (2012: Wald χ2
1= 0.51, p = 0.48; 2014: Waldχ21= 1.02, p = 0.31;Fig. 1). Analyzing the effects of iso-lation on abundances by trophic level uncovered analogous differences. Effects were again negative on primary consumers throughout the ex-periment (46.4% lower abundances in 2010: Wald χ21= 20.50, p < 0.0001; 55.5% in 2012: Wald χ2
1= 9.15, p = 0.0025; 69.1% in 2014: Waldχ2
1= 23.01, p < 0.0001;Fig. 1), and positive for predators only at the beginning (110% increase in 2010: Wald χ21= 6.39, p = 0.011; 2012: Wald χ2
1= 0.18, p = 0.68; 2014: Waldχ21= 0.019, p = 0.89). In contrast, isolation reduced primary consumer CWM only in 2010 (by 66%: F1,63.81= 14.89, p = 0.00027; 2012: F1,51.73= 0.12, p = 0.73; 2014: F1,60.69= 2.46, p = 0.12; Fig. 1) and increased pre-dator CWM in all years (by 39.8% in 2010: F1,71.46= 4.20, p = 0. 044; by 93.8% in 2012: F1,54.89= 10.07, p = 0. 0025; by 65% in 2014: F1,52.99= 4.43, p = 0. 04). The number of non-experimental hollow oaks surrounding each box had no influence on any biodiversity mea-sure except for a slight decrease in primary consumer richness in 2014 (β = −0.024; Wald χ2
1= 4.45, p = 0.035).
In the initial period (2009–2012), there was 20.1% less wood mould decay in isolated boxes than in connected ones (F1,72.02= 5.55, p = 0.021;Fig. 2). However, this negative effect of isolation on decay waned in 2012–2014 (F1,69.44= 0.70, p = 0.41), resulting in a lack of overall (2009–2014) isolation effects on wood mould decay (F1,73.09= 1.86, p = 0.18). The number of hollow oaks surrounding
boxes had no effect on decay (p > 0.068).
3.2. Saproxylic biodiversity effect on wood mould decay
In 2009–2012, saproxylic biodiversity promoted wood mould decay through an effect of primary consumer CWM (Fig. 3). This model with one predictor (F1,67.73= 18.93, p < 0.0001, marginal R2= 0.2) was the only one in the subset of models withΔAICc < 4 resulting from model selection (see Appendix A, Table A.2). In 2012–2014 and 2009–2014, the only selected submodel was the null model.
4. Discussion
Small-scale habitat isolation delayed wood mould decomposition and reduced saproxylic biodiversity in the boxes. Contrary to the trophic-rank hypothesis, there were strong negative effects on primary consumers but not on predators, maybe because the latter were more frequently using other habitats as well. During thefirst three years of the experiment, decomposition rate was solely linked to CWM of pri-mary consumers, indicating that function was driven by the body mass of the dominant species, as predicted by the mass-ratio hypothesis. Our six-year experiment shows that small-scale habitat isolation leads to biodiversity loss and changes in function in a natural system and em-phasizes the importance of both trait-based and experimental Fig. 1. Effect of isolation on species richness (number of species), abundances (number of individuals) and on the log-transformed community-weighted mean of body mass (CWM) of saproxylic beetles over time. We ran (generalized) linear mixed models with Poisson error distribution for richness and abundances and a Gaussian error distribution for CWM, with“landscape” as a random intercept. Codes: ALL – all saproxylic beetles, PCN – primary consumers, PRE – predators. Means ± SE are shown. *p < 0.05, **p < 0.01, ***p < 0.001.
approaches to determine biodiversity conservation strategies. 4.1. Effects of isolation on biodiversity
We found that isolated wood mould boxes had overall lower sa-proxylic biodiversity than connected ones. This study shows isolation effects on colonization directly by means of a landscape experiment, emphasizing the influence of spatial configuration of conservation ef-forts on habitat utilization by saproxylic beetles. Earlier research was based on snapshot studies of occurrence patterns (Bergman et al., 2012; Ranius et al., 2011; but seeSeibold et al., 2017) or on assessments of the dispersal ability of individuals (Chiari et al., 2013; Ranius, 2006). However, the outcome of snapshot studies may be affected by unknown correlations with other factors and the dispersal distance of individuals is only one of the several factors affecting colonization rates. It is im-portant to note that the isolation effects in our experiment can be en-tirely attributed to our manipulations of distance between boxes and hollows: all boxes provided the same amount of wood mould and we statistically controlled for the number of naturally occurring hollow oaks (dispersal sources) within a 500-m radius of each wood mould box
(i.e. amount of habitat in the local landscape sensu Fahrig (2013)). Therefore, our experiment reveals that habitat configuration per se can be a strong driver of biodiversity loss regardless of the amount of ha-bitat in the surrounding landscape (Haddad et al., 2017;Hanski, 2015). Changes in food web structure due to isolation were contrary to the predictions of the trophic rank hypothesis. Unlike another study on saproxylic beetles (Buse et al., 2016), we did notfind stronger negative effects of isolation at the highest trophic level. Instead, effects on richness, total abundances and CWM of primary consumers were al-ways negative, but were absent or even positive on predators. One possible explanation why we did not detect a bottom-up cascade of negative effects from primary consumers to predators (Holt, 1996) is that primary consumers in our system depended on dead wood and hollow habitats more than predators: 41% of primary consumer in-dividuals were obligate saproxylic (63% of species), in contrast to 19% of predators (55% of species) (low beetle numbers per box did not allow us to split trophic levels into sub-groups for analysis). Hence, our iso-lation measure (distance from boxes to hollows) may be poor at re-flecting habitat isolation as experienced by the predators. As shown by the effect on CWM, isolation discriminated against the body mass of colonizing species depending on their trophic level, a fact that deserves further investigation. Indeed, beyond trophic rank, traits like habitat specificity and dispersal ability influence responses of organisms to fragmentation (Davies et al., 2000). It is possible that predators and primary consumers differed in their dispersal abilities, but data on sa-proxylic beetles are only available for a few species (Bouget et al., 2015; Gibb et al., 2006).
There was considerable temporal variation in how biodiversity re-sponded to isolation. For primary consumers, species numbers and total abundances were lower in isolated than in connected boxes throughout the experiment, whereas negative effects on CWM only occurred at the beginning. Conversely, isolation always affected predator CWM posi-tively, although it had only a small effect on richness and abundances at the beginning. One possible explanation is that our experiment created new habitat patches (boxes). Therefore, the initial saproxylic commu-nity comprised only colonizing individuals, so effects of isolation on immigration had a critical influence on community composition in the boxes. Afterwards, other processes came into play, such as population growth, emigration, species interactions, and turnover in species com-position, whose influence on community functional composition (CWM) potentially overrode that of immigration.
4.2. Effects of habitat isolation and saproxylic beetle biodiversity on wood mould decay
The only biodiversity-ecosystem function relationship identified was the positive effect of primary consumer CWM on wood mould decay during thefirst part of the experiment (Fig. 3). It was also at the beginning of the experiment when isolation decreased both primary consumer CWM and wood mould decay. Combined, these results in-dicate that the negative effects of isolation on wood mould Fig. 2. Effect of isolation on wood mould decay, measured as the proportion of
decrease relative to the previous level of wood mould to make it comparable between different time periods (2009–2012, 2012–2014 and overall). We ran linear mixed effects models on logit-transformed proportional data, with “landscape” as a random intercept. Means ± SE are shown. *p < 0.05.
Fig. 3. Effect of community-weighted mean of body mass (CWM) of primary consumers on wood mould decay, measured as the proportion of decrease relative to the previous level of wood mould to make it comparable between different time periods (2009–2012, 2012–2014 and overall). We ran linear mixed effect models on logit-transformed proportional data, with “landscape” as a random intercept. Only the model in 2009–2012 was selected by model se-lection. We tested this model individually to show its p value and marginal R2. The grey shaded area represents the 95% confidence in-terval.
decomposition originate from decreases in CWM of the lowest trophic level of the saproxylic community, which constitutes about one half of all sampled individuals (Fig. 1). Most other studies on dead wood de-composition have assessed dede-composition rates from single dry weight measurements (Edman et al., 2006;Venugopal et al., 2017), which is a destructive method. Instead, we used a non-destructive method of measuring the decomposition rate, which might have reduced the ac-curacy of the estimate but which allowed us to measure dead wood decomposition over a period offive years without disturbing saproxylic biodiversity.
As expected from real-world ecological communities, the saproxylic community in the boxes had a skewed species-abundance distribution (Magurran and Henderson, 2003; see Appendix A: Fig. A.2), indicating that species richness is driven by the presence of many rare species, and total abundance by the presence of a few dominant ones. In contrast, CWM is driven by the body mass of dominant species, so the positive link between CWM and wood mould decay supports the mechanism predicted by the mass-ratio hypothesis (Grime, 1998). However, the positive link between primary consumer CWM and wood mould decay only held in thefirst three years of the experiment (2009–2012) and was absent in the last two years, when the decay rate was lower. This finding could reflect a change in the microbial decomposer fauna over time linked to changes in the wood mould: dead oak leaves decompose before wood sawdust, which is more nutrient limiting (Ulyshen, 2016). Our results suggest that the body mass of the dominant primary con-sumer species were at least partly responsible for ecosystem func-tioning, while the effects of mere species presence were negligible, since the persistent negative effects of isolation on richness and abun-dance of primary consumers did not have consequences for wood mould decay. In spite of mounting evidence that the preservation of ecosystem functions and services does not necessarily equate to biodi-versity conservation (Kleijn et al., 2015; Senapathi et al., 2015), our findings show higher connectivity to benefit both saproxylic biodi-versity and function. This study shows that going beyond presence-based indicators like richness and abundance and considering func-tional indices like CWM can contribute to explaining function in natural ecosystems.
5. Conclusions
Our six-year experiment demonstrates a negative impact of habitat isolation on dead wood decomposition through changes in CWM of saproxylic primary consumers. By varying the isolation levels of newly created habitat (boxes) from original dispersal sources (natural hol-lows), we avoided the biases that confound wood-caging studies (Ulyshen and Wagner, 2013) and we were able to show the contribution of saproxylic beetles to dead wood decay. Veteran trees are declining globally due to removal, increased mortality, and lack of recruitment (Lindenmayer et al., 2014), which threatens many associated species (Cálix et al., 2018). Conservation mainly involves protecting existing trees and avoiding vegetation overgrowth, but can also include pro-moting tree regeneration and the formation of important microhabitats (i.e. veteranization). Because isolation decreased saproxylic biodiversity (essentially of primary consumers), our results suggest that focusing conservation measures within well-connected veteran tree sites will result in higher habitat use by the saproxylic fauna than if efforts are distributed among veteran trees irrespective of their spatial location. A further measure can be to create stepping-stones between veteran tree sites. If so, efforts should be focused in areas outside these sites by clustering wood mould boxes where the distances between sites are shortest. Although there are large differences in the biology of sa-proxylic species and the dynamics of their habitats, those associated with veteran trees are probably less dispersal-prone, since they are adapted to a naturally stable and long-lived habitat (Siitonen and Ranius, 2015). Thus, to conserve overall saproxylic biodiversity and to stabilize ecosystem function we recommend a mixed strategy
considering many different habitats and their spatial configuration at various spatial scales.
Acknowledgments
We thank E. Hemmingsson and S. Carlsson, who helped withfield work, K. Sancak, who sorted the material, E. Ottosson who cultured Laetiporus sulphureus, and the landowners for letting us use their trees and properties. The Swedish Research Council Formas supported this study (grant number 2008-539).
Appendix A. Supplementary data
Supplementary data to this article can be found online athttps:// doi.org/10.1016/j.biocon.2018.09.023.
References
Balvanera, P., Pfisterer, A.B., Buchmann, N., He, J.S., Nakashizuka, T., Raffaelli, D., Schmid, B., 2006. Quantifying the evidence for biodiversity effects on ecosystem functioning and services. Ecol. Lett. 9, 1146–1156.
Bartoń, K., 2016. MuMIn: Multi-model Inference. R Package Version 1.15.6.https:// CRAN.R-project.org/package=MuMIn.
Bates, D., Mächler, M., Bolker, B.M., Walker, S.C., 2015. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48.
Bergman, K.-O., Jansson, N., Claesson, K., Palmer, M.W., Milberg, P., 2012. How much and at what scale? Multiscale analyses as decision support for conservation of sa-proxylic oak beetles. For. Ecol. Manag. 265, 133–141.
Bouget, C., Brin, A., Tellez, D., Archaux, F., 2015. Intraspecific variations in dispersal ability of saproxylic beetles in fragmented forest patches. Oecologia 177, 911–920. Burnham, K.P., Anderson, D.R., 2002. Model Selection and Multimodel Inference: A
Practical Information-theoretic Approach, 2nd edn. Springer, New York, UK. Buse, J., Entling, M.H., Ranius, T., Assmann, T., 2016. Response of saproxylic beetles to
small-scale habitat connectivity depends on trophic level. Landsc. Ecol. 31, 939–949. Cadotte, M.W., Carscadden, K., Mirotchnick, N., 2011. Beyond species: functional
di-versity and the maintenance of ecological processes and services. J. Appl. Ecol. 48, 1079–1087.
Cálix, M., Alexander, K.N.A., Nieto, A., Dodelin, B., Soldati, F., Telnov, D., Vazquez-Albalate, X., Aleksandrowicz, O., Audisio, P., Istrate, P., Jansson, N., Legakis, A., Liberto, A., Makris, C., Merkl, O., Mugerwa Pettersson, R., Schlaghamersky, J., Bologna, M.A., Brustel, H., Buse, J., Novák, V., Purchart, L., 2018. European Red List of Saproxylic Beetles. IUCN, Brussels, Belgium.
Carlsson, S., Bergman, K.-O., Jansson, N., Ranius, T., Milberg, P., 2016. Boxing for bio-diversity: evaluation of an artificially created decaying wood habitat. Biodivers. Conserv. 25, 393–405.
Chaplin-Kramer, R., O'Rourke, M.E., Blitzer, E.J., Kremen, C., 2011. A meta-analysis of crop pest and natural enemy response to landscape complexity. Ecol. Lett. 14, 922–932.
Chiari, S., Carpaneto, G.M., Zauli, A., Zirpoli, G.M., Audisio, P., Ranius, T., 2013. Dispersal patterns of a saproxylic beetle, Osmoderma eremita, in Mediterranean woodlands. Insect Conserv. Diver. 6, 309–318.
Dahlberg, A., Stokland, J.N., 2004. Vedlevande arters krav på substrat: sammanställning och analys av 3600 arter. Skogsstyrelsen.
Dangles, O., Malmqvist, B., 2004. Species richness-decomposition relationships depend on species dominance. Ecol. Lett. 7, 395–402.
Davies, K.F., Margules, C.R., Lawrence, J.F., 2000. Which traits of species predict popu-lation declines in experimental forest fragments? Ecology 81, 1450–1461. De Laender, F., Rohr, J.R., Ashauer, R., Baird, D.J., Berger, U., Eisenhauer, N., Grimm, V.,
Hommen, U., Maltby, L., Melian, C.J., Pomati, F., Roessink, I., Radchuk, V., Van den Brink, P.J., 2016. Reintroducing environmental change drivers in biodiversity-eco-system functioning research. Trends Ecol. Evol. 31, 905–915.
van der Wal, A., Geydan, T.D., Kuyper, T.W., de Boer, W., 2013. A thready affair: linking fungal diversity and community dynamics to terrestrial decomposition processes. FEMS Microbiol. Rev. 37, 477–494.
Edman, M., Möller, R., Ericson, L., 2006. Effects of enhanced tree growth rate on the decay capacities of three saprotrophic wood-fungi. For. Ecol. Manag. 232, 12–18. Ewers, R.M., Didham, R.K., 2006. Confounding factors in the detection of species
re-sponses to habitat fragmentation. Biol. Rev. 81, 117–142.
Fahrig, L., 2013. Rethinking patch size and isolation effects: the habitat amount hy-pothesis. J. Biogeogr. 40, 1649–1663.
Fox, J., Weisberg, S., 2011. An {R} Companion to Applied Regression, 2nd edition. Sage, Thousand Oaks, CA.
Gagic, V., Bartomeus, I., Jonsson, T., Taylor, A., Winqvist, C., Fischer, C., Slade, E.M., Steffan-Dewenter, I., Emmerson, M., Potts, S.G., Tscharntke, T., Weisser, W., Bommarco, R., 2015. Functional identity and diversity of animals predict ecosystem functioning better than species-based indices. Proc. R. Soc. B Biol. Sci. 282, 20142620.
Gessner, M.O., Swan, C.M., Dang, C.K., McKie, B.G., Bardgett, R.D., Wall, D.H., Hättenschwiler, S., 2010. Diversity meets decomposition. Trends Ecol. Evol. 25, 372–380.
Gibb, H., Hjältén, J., Ball, J.P., Pettersson, R.B., Landin, J., Alvini, O., Danell, K., 2006. Wing loading and habitat selection in forest beetles: are red-listed species poorer dispersers or more habitat-specific than common congenerics? Biol. Conserv. 132, 250–260.
Grime, J.P., 1998. Benefits of plant diversity to ecosystems: immediate filter and founder effects. J. Ecol. 86, 902–910.
Grove, S.J., 2002. Saproxylic insect ecology and the sustainable management of forests. Annu. Rev. Ecol. Syst. 33, 1–23.
Haddad, N.M., Gonzalez, A., Brudvig, L.A., Burt, M.A., Levey, D.J., Damschen, E.I., 2017. Experimental evidence does not support the Habitat Amount Hypothesis. Ecography 40, 48–55.
Hanski, I., 1998. Metapopulation dynamics. Nature 396, 41–49.
Hanski, I., 2015. Habitat fragmentation and species richness. J. Biogeogr. 42, 989–993. Hättenschwiler, S., Tiunov, A.V., Scheu, S., 2005. Biodiversity and litter decomposition in
terrestrial ecosystems. Annu. Rev. Ecol. Evol. Syst. 36, 191–218.
Holt, R.D., 1996. 29. Food webs in space: an island biogeographic perspective. In: Food Webs. Springer USA, pp. 313–323.
Jansson, N., Ranius, T., Larsson, A., Milberg, P., 2009. Boxes mimicking tree hollows can help conservation of saproxylic beetles. Biodivers. Conserv. 18, 3891–3908. Kleijn, D., Winfree, R., Bartomeus, I., Carvalheiro, L.G., Henry, M., Isaacs, R., Klein, A.-M.,
Kremen, C., M'Gonigle, L.K., Rader, R., Ricketts, T.H., Williams, N.M., Lee Adamson, N., Ascher, J.S., Baldi, A., Batary, P., Benjamin, F., Biesmeijer, J.C., Blitzer, E.J., Bommarco, R., Brand, M.R., Bretagnolle, V., Button, L., Cariveau, D.P., Chifflet, R., Colville, J.F., Danforth, B.N., Elle, E., Garratt, M.P.D., Herzog, F., Holzschuh, A., Howlett, B.G., Jauker, F., Jha, S., Knop, E., Krewenka, K.M., Le Feon, V., Mandelik, Y., May, E.A., Park, M.G., Pisanty, G., Reemer, M., Riedinger, V., Rollin, O., Rundlof, M., Sardinas, H.S., Scheper, J., Sciligo, A.R., Smith, H.G., Steffan-Dewenter, I., Thorp, R., Tscharntke, T., Verhulst, J., Viana, B.F., Vaissiere, B.E., Veldtman, R., Ward, K.L., Westphal, C., Potts, S.G., 2015. Delivery of crop pollination services is an insufficient argument for wild pollinator conservation. Nat. Commun. 6, 1–8.
Köhler, F., 2000. Totholzkäfer in Naturwaldzellen des nördlichen Rheinlands: Vergleichende Studien zur Totholzkäferfauna Deutschlands und deutschen Naturwaldforschung. Landesamt Agrarordnung NRW, Recklinghausen.
Laliberté, E., Legendre, P., 2010. A distance‐based framework for measuring functional diversity from multiple traits. Ecology 91, 299–305.
Laliberté, E., Legendre, P., Shipley, B., 2014. Measuring Functional Diversity (FD) From Multiple Traits, and Other Tools for Functional Ecology. R Package Version 1.0-12. Landvik, M., Niemelä, P., Roslin, T., 2016. Mother knows the best mould: an essential role
for non-wood dietary components in the life cycle of a saproxylic scarab beetle. Oecologia 182, 163–175.
Lavorel, S., Grigulis, K., McIntyre, S., Williams, N.S.G., Garden, D., Dorrough, J., Berman, S., Quétier, F., Thébault, A., Bonis, A., 2008. Assessing functional diversity in thefield – methodology matters!. Funct. Ecol. 22, 134–147.
Lindenmayer, D.B., Laurance, W.F., Franklin, J.F., Likens, G.E., Banks, S.C., Blanchard, W., Gibbons, P., Ikin, K., Blair, D., McBurney, L., Manning, A.D., Stein, J.A.R., 2014. New policies for old trees: averting a global crisis in a keystone ecological structure. Conserv. Lett. 7, 61–69.
Magurran, A.E., Henderson, P.A., 2003. Explaining the excess of rare species in natural species abundance distributions. Nature 422, 714–716.
McArthur, R.H., Wilson, E.O., 1967. The Theory of Island Biogeography. Princeton University Press, New Jersey, USA.
Palm, T., 1959. Die Holz- und Rindenkäfer der süd- und mittelschwedische Laubbäume. Opusc. Entomol.(supplementum XVI) (374 pp.).
Petchey, O.L., Hector, A., Gaston, K.J., 2004. How do different measures of functional diversity perform? Ecology 85, 847–857.
R Core Team, 2018. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.
Ranius, T., 2006. Measuring the dispersal of saproxylic insects: a key characteristic for their conservation. Popul. Ecol. 48, 177–188.
Ranius, T., Johansson, V., Fahrig, L., 2010. A comparison of patch connectivity measures using data on invertebrates in hollow oaks. Ecography 33, 971–978.
Ranius, T., Johansson, V., Fahrig, L., 2011. Predicting spatial occurrence of beetles and pseudoscorpions in hollow oaks in southeastern Sweden. Biodivers. Conserv. 20, 2027–2040.
Ranius, T., Johansson, V., Schroeder, M., Caruso, A., 2015. Relative importance of habitat characteristics at multiple spatial scales for wood-dependent beetles in boreal forest. Landsc. Ecol. 30, 1931–1942.
Seibold, S., Brandl, R., Buse, J., Hothorn, T., Schmidl, J., Thorn, S., Muller, J., 2015. Association of extinction risk of saproxylic beetles with ecological degradation of forests in Europe. Conserv. Biol. 29, 382–390.
Seibold, S., Bässler, C., Brandl, R., Fahrig, L., Förster, B., Heurich, M., Hothorn, T., Scheipl, F., Thorn, S., Müller, J., 2017. An experimental test of the habitat-amount hypothesis for saproxylic beetles in a forested region. Ecology 98, 1613–1622. Senapathi, D., Biesmeijer, J.C., Breeze, T.D., Kleijn, D., Potts, S.G., Carvalheiro, L.G.,
2015. Pollinator conservation—the difference between managing for pollination services and preserving pollinator diversity. Curr. Opin. Insect Sci. 12, 93–101. Siitonen, J., Ranius, T., 2015. The importance of veteran trees for saproxylic insects. In:
Kirby, K.J., W. C. (Eds.), Europe's Changing Woods and Forests: From Wildwood to Cultural Landscape. CAB international, pp. 140–153.
Speight, M.C.D., 1989. Saproxylic Invertebrates and their Conservation. Council of Europe.
Srivastava, D.S., Vellend, M., 2005. Biodiversity-ecosystem function research: is it re-levant to conservation? Annu. Rev. Ecol. Evol. Syst. 36, 267–294.
Stokland, J.N., Siitonen, J., Jonsson, B.G., 2012. Biodiversity in Dead Wood. Cambridge University Press.
Tilman, D., Isbell, F., Cowles, J.M., 2014. Biodiversity and ecosystem functioning. Annu. Rev. Ecol. Evol. Syst. 45, 471–493.
Ulyshen, M.D., 2016. Wood decomposition as influenced by invertebrates. Biol. Rev. 91, 70–85.
Ulyshen, M.D., Wagner, T.L., 2013. Quantifying arthropod contributions to wood decay. Methods Ecol. Evol. 4, 345–352.
Venugopal, P., Junninen, K., Edman, M., Kouki, J., 2017. Assemblage composition of fungal wood-decay species has a major influence on how climate and wood quality modify decomposition. FEMS Microbiol. Ecol. 93fix002.
Violle, C., Navas, M.-L., Vile, D., Kazakou, E., Fortunel, C., Hummel, I., Garnier, E., 2007. Let the concept of trait be functional!. Oikos 116, 882–892.
Warton, D.I., Hui, F.K., 2011. The arcsine is asinine: the analysis of proportions in ecology. Ecology 92, 3–10.
Winfree, R., F., J.W., Williams, N.M., Reilly, J.R., Cariveau, D.P., 2015. Abundance of common species, not species richness, drives delivery of a real-world ecosystem service. Ecol. Lett. 18, 626–635.
