Journal of Ecology 2012, 100, 619–630
doi: 10.1111/j.1365-2745.2011.01943.x
Multiple mechanisms for trait effects on litter decomposition: moving beyond home-field advantage with a new hypothesis Gre´goire T. Freschet*†, Rien Aerts and Johannes H. C. Cornelissen Department of Systems Ecology, Institute of Ecological Science, Faculty of Earth and Life Sciences, VU University, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands
Summary 1. Evidence is growing that leaf litter generally decomposes faster than expected in its environment of origin, owing to specialization of litter and topsoil decomposer communities to break down litter encountered most often. Nevertheless, this home-field advantage (HFA) in decomposition is inconsistently supported by experimental data and fails to account for situations where contrasting qualities of litter coexist within the same litter matrix. 2. In contrast to the HFA hypothesis, which expects a positive interaction between every litter species produced locally and the local decomposer communities irrespective of litter species quality, we define here an alternative substrate quality–matrix quality interaction (SMI) hypothesis that expects a continuum from positive to negative interaction between specific litters (substrates) and decomposer communities as specific litters and the ecosystem litter layer (i.e. the matrix, which drives local decomposer community composition) become increasingly dissimilar in quality. 3. To test this hypothesis, we conducted a reciprocal transplant decomposition experiment of eight leaf, six fine-stem and nine fine-root litter species from three neighbouring ecosystems of the subarctic biome: dry forest, riparian forest and forest-surrounded pond; and characterized the quality (represented by lignin content and an integrated measure of carbon ⁄ nutrient economics) of each litter species and each ecosystem litter layer. 4. We found substantial overall effects of SMI on decomposition rates of leaf (20% explained variance), stem (14%) and root (15%) litters, although this effect was lower than the single effects of litter quality and microclimate (remaining explained variance). Despite being partly inconsistent across litter species, likely due to the complexity of litter quality–decomposer community relationships, the SMI hypothesis appeared more broadly applicable than the HFA hypothesis. 5. Synthesis. We demonstrate here that plant traits, likely via their control on litter and topsoil decomposer community composition, have indirect effects on litter breakdown rates, not only at the interface between ecosystems but also within ecosystems, with likely implications for many other ecosystems world-wide. These results suggest functional variation in decomposer communities between ecosystems with respect to their efficiency to degrade litters with contrasting qualities, such as different lignolytic and detoxification activities but also contrasting efficiencies to degrade nonrecalcitrant tissues. Key-words: decomposer community, home-field advantage, leaf, litter decomposition, litter mixture, plant functional traits, plant–soil (below-ground) interactions, reciprocal transplant experiment, root, stem Introduction
*Correspondence author. Department of Forest Ecology and Management, Swedish University of Agricultural Science, 901 83 Umea˚, Sweden. E-mail:
[email protected] †Present address: Department of Forest Ecology and Management, Swedish University of Agricultural Science, 901 83 Umea˚, Sweden.
Climate, substrate quality and decomposer community composition are the three fundamental parameters driving plant litter decomposition (Swift, Heal & Anderson 1979). However, whereas climate and substrate quality might together explain at least 57% and approximately 90% of global scale variation in leaf and root litter decomposition, respectively (Berg et al.
2011 The Authors. Journal of Ecology 2011 British Ecological Society
620 G. T. Freschet, R. Aerts & J. H. C. Cornelissen 1993; Aerts 1997; Silver & Miya 2001; Cornwell et al. 2008), the factors contributing to the remaining variance and their relative contribution are yet uncertain. In particular, surprisingly little is known on the interactions between litter decomposition drivers (but see Meentemeyer 1978; Aerts 1997; Cornelissen et al. 2007 for litter quality–climate interaction; Wall et al. 2008 for climate–decomposer interactions). Globally, there is strong overall coordination of the dynamic chemical stoichiometries of litter and their decomposers (Parton et al. 2007; Manzoni et al. 2010). At the same time, deviations from such coordination of litter chemistry and decomposer activity, interpreted as local-scale interactions between substrate and decomposer community, are widespread (Bardgett & Wardle 2010). For instance, in several empirical studies, litter substrates decomposed more rapidly than expected when incubated in their own environment than away (Bocock et al. 1960; Hunt et al. 1988; de Toledo Castanho & de Oliveira 2008; Vivanco & Austin 2008), which has been referred to as the home-field advantage hypothesis (HFA; Hunt et al. 1988; Gholz et al. 2000). This could be explained by the specialization of decomposer communities to their particular substrate type (Ayres et al. 2009a; Strickland et al. 2009a), although microclimate may also take part in the interaction (e.g. de Toledo Castanho & de Oliveira 2008; Vivanco & Austin 2008). Owing to the ongoing competition for litter resources among decomposers, a selective pressure is likely to favour the most efficient decomposers in processing litter resources and therefore in degrading their litter matrix (Ayres et al. 2009b). Several studies have demonstrated such persistent differences in soil decomposer community composition functions, as evidenced by their dissimilar capacities to decompose litter of contrasting quality, both for microbes (Strickland et al. 2009a,b; Keiser et al. 2011) and for soil fauna (e.g. Scheu et al. 2003), with strong effects on litter decomposition rates (Hobbie et al. 2006; Laganie`re, Pare´ & Bradley 2010). However, such selective pressure should come on top of potentially large fluctuations in decomposer community composition induced by changes in climatic factors, discrete external input events and other environmental factors that occur at multiple time scales (Bardgett & Wardle 2010). Finally, some authors have hypothesized that functions associated with breakdown of recalcitrant material (e.g. lignolytic activity, degradation of secondary compounds) could be at least partly lost by decomposer communities historically living on non-recalcitrant materials (Keiser et al. 2011; Milcu & Manning 2011). However, while the theory underlying the HFA hypothesis is progressively taking shape, evidence for HFA effects is strongly context dependent (Prescott et al. 2000; Ayres, Dromph & Bardgett 2006; Chapman & Koch 2007). Studies that uncovered positive interactions between litter substrates (i.e. litters from one single species) and their soil or litter matrix of origin (i.e. the soil or multi-specific litter layer of the ecosystem where the litter species comes from) generally involved reciprocal litter transplant between highly contrasting ecosystems (Bocock et al. 1960; Hunt et al. 1988; de Toledo Castanho & de Oliveira 2008), highly contrasting litter qualities (Hunt et al. 1988; Gholz et al. 2000; Strickland et al. 2009a,b;
Jacob et al. 2010) and most often took place in (eco-)systems dominated by single plant species (Hunt et al. 1988; Vivanco & Austin 2008; Ayres et al. 2009a). In contrast, studies that featured litter substrates of rather similar quality (Ayres, Dromph & Bardgett 2006) or ecosystems dominated by multiple plant species (Prescott et al. 2000; Chapman & Koch 2007) were generally unsupportive of the HFA theory. This context dependence of the HFA effect has so far hampered its firm incorporation into ecological theory.
BEYOND THE HOME-FIELD ADVANTAGE
Here we propose that the HFA hypothesis represents only one aspect of a bigger framework accounting for ubiquitous effects of substrate quality–decomposer community interactions. The HFA hypothesis expects positive interactions when the site of litter incubation and the site of litter origin match. Thereby, it implicitly assumes that in a multispecies ecosystem with a composite litter matrix, the local decomposer community should be adapted to decompose the litter of many species present in this ecosystem, irrespective of their quality (HFA hypothesis in Fig. 1, which assumes a positive interaction for all different litter types at their own site). While most soils are indeed likely to contain a range of decomposers, which together are able to decompose a wide range of substrates (i.e. represent most functional groups of decomposers), differences in breakdown efficiency should exist between soil communities owing to the varying relative proportion of each functional group of decomposers. A decomposer community is relatively unlikely to be, for instance, both fungal and bacterial dominated, or in other words to degrade highly efficiently high-and low-quality litters (Wardle et al. 2004). In this respect, the HFA hypothesis is unable to account for the tremendous range of litter quality and decomposability co-occurring at small spatial scales in most ecosystems (e.g. Santiago 2007; Freschet, Aerts & Cornelissen 2011a). Alternatively, we follow in this context the mass ratio hypothesis (Grime 1998; Garnier et al. 2004) to propose that decomposer community composition, under influence of resource history (Keiser et al. 2011), should reflect the average quality of the litter matrix (likely driven by the most dominant species; Strickland et al. 2009b). In this context, we expect a continuum from positive to negative interaction between specific litters (substrates) and decomposer communities as specific litters and the ecosystem litter layer (i.e. the matrix, which drives local decomposer community composition) become increasingly dissimilar in quality. This substrate quality–matrix quality interaction (SMI) hypothesis implies for instance that low-quality substrates (e.g. pine needles) will decompose faster than expected when incubated in the decomposition matrix of poor quality (e.g. dry birch forest: left environment in Fig. 1), but in contrast will decompose slightly slower than expected in intermediatequality matrices (e.g. wet birch forest, central environment in Fig. 1) and substantially, slower in matrices composed of high-quality litters (forest-surrounded pond, right environment in Fig. 1). Within this framework, HFA effects can (although not necessarily) be interpreted as situations where
2011 The Authors. Journal of Ecology 2011 British Ecological Society, Journal of Ecology, 100, 619–630
Substrate–matrix quality interactions in decay 621
Fig. 1. Schematic diagram of the interactions in decomposition expected by the substrate–matrix quality interaction (SMI) hypothesis, the home-field advantage (HFA) hypothesis and the null (H0) hypothesis of no interaction. Symbols + and ) indicate expectations of positive and negative interactions, that is faster and slower decomposition rates than expected by the null hypothesis, respectively. Briefly, the SMI hypothesis expects interactions between ‘Litter substrate quality’ (e.g. leaf litter of low , intermediate and high quality) and ‘Incubation site litter matrix quality’ (as represented by the mass ratio of each leaf litter type contributing to the litter matrix quality) and the HFA hypothesis between ‘Substrate site of origin’ and ‘Incubation site’. It is noted that the predictions of SMI are more differentiated at ecosystem scale and differ in most cases from that of HFA.
the quality of a given litter type is highly similar to the overall litter matrix quality in its own site. To test our theory comprehensively and compare its relevance to this of other current hypotheses (HFA, H0), we set up a realistic reciprocal litter transplant experiment across three neighbouring ecosystems of a subarctic forest, namely dry birch forest, riparian birch forest and forest-surrounded pond. This way, we intended to bridge two rarely interacting research areas, that is terrestrial and freshwater decomposition studies (Wagener, Oswood & Schimel 1998; Grimm et al. 2003). We simultaneously transplanted leaf, fine-stem and fine-root litter species. These different litter types included substantial variation in litter quality and litter decomposability. To characterize litter quality for decomposer communities, we used two of the best local single predictors of litter decomposability, previously identified in a common-garden decomposition study of many leaf, fine-stem and fine-root litters (Freschet, Aerts & Cornelissen 2011a): (i) an integrated measure of plant carbon and nutrient economics (Freschet et al. 2010a) and (ii) litter lignin content. We also used several measures of secondary chemistry. In addition, we measured annual leaf litter mass inputs of each ecosystem to estimate the average ecosystem leaf litter matrix quality. We first tested whether litter traits controlling decomposition were consistent across aquatic and terrestrial systems before including the aquatic data set and testing the SMI hypothesis on both leaf and nonleaf substrates. The SMI effect was tested as the interaction between substrate quality and matrix quality in a general linear model (GLM) model of litter decomposition rate predictions. The relevance of this SMI interaction was assessed by comparison with the major local determinants of litter decomposition: substrate quality and matrix quality. Additionally, we extracted the nonexplained variance of a GLM model of litter decomposition rates using only species and site as predictor variables and confronted it, species by species, to SMI, HFA and H0 (no interaction) expectations. The relevance of each hypothesis
was discussed with regard to the number of litter species and types that differed (analysis of variance) in decomposition rates, as predicted by the hypothesis, between environments.
Materials and methods LITTER TRANSPLANT EXPERIMENT
The plant species were sampled around the Abisko Research Station, North Sweden (6821¢N, 1849¢E), at low altitude (350–400 m a.s.l.), below the tree line. During the 1999–2008 decade, this area had a mean annual rainfall of 352 mm and mean January and July temperatures of )9.7 and 12.3 C, respectively (meteorological data, Abisko Research Station). The forested area, which was the focus of this study, features strongly organic Podsol soils and covers most of the landscape below 700 m a.s.l. except for occasional treeless mires, fens and bogs. The three most distinct ecosystem types within the selected forested sites were dry birch forest with ericaceous understorey, riparian birch forest with herbaceous and shrubby understorey and forested freshwater systems (ponds and streams). Seven sampling sites (c. 20 m transects) each including all three ecosystem types were used to collect freshly senesced litters (‘substrates’) of a total of eight leaf, six fine-stem (
