Assemblage composition of fungal wood-decay species has a major influence on how climate and wood quality modify decomposition

Abstract

The interactions among saprotrophic fungal species, as well as their interactions with environmental factors, may have a major influence on wood decay and carbon release in ecosystems. We studied the effect that decomposer diversity (species richness and assemblage composition) has on wood decomposition when the climatic variables and substrate quality vary simultaneously. We used two temperatures (16 and 21°C) and two humidity levels (70% and 90%) with two wood qualities (wood from managed and old-growth forests) of Pinus sylvestris. In a 9-month experiment, the effects of fungal diversity were tested using four wood-decaying fungi (Antrodia xantha, Dichomitus squalens, Fomitopsis pinicola and Gloeophyllum protractum) at assemblage levels of one, two and four species. Wood quality and assemblage composition affected the influence of climatic factors on decomposition rates. Fungal assemblage composition was found to be more important than fungal species richness, indicating that species-specific fungal traits are of paramount importance in driving decomposition. We conclude that models containing fungal wood-decay species (and wood-based carbon) need to take into account species-specific and assemblage composition-specific properties to improve predictive capacity in regard to decomposition-related carbon dynamics.

climate change, fungal assemblage composition, fungal diversity, interaction, species specificity, species richness, wood decomposition, wood quality

INTRODUCTION

The effects of species diversity on the functioning and stability of vital ecosystem processes have been gaining increasing attention in recent times (Chapin et al. 2000; Hooper et al.2005). Biodiversity has two prominent aspects: species richness and composition (Cardinale, Nelson and Palmer 2000; Hooper et al.2005). Given the ongoing decline of global biodiversity, it is essential that the consequences of diversity loss of wood-decaying species (Siitonen 2001) for ecosystem functioning processes such as carbon and nutrient cycling are understood and quantified (Gessner et al.2010; Valentín et al.2014).

Coarse woody debris is an integral above-ground carbon pool in boreal ecosystems (Harmon and Hua 1991; Siitonen 2001; Cornelissen et al.2012; van der Wal, Ottosson and de Boer 2015), and saproxylic fungi are its main decomposition agents (Boddy 2001; Boddy, Frankland and van West 2008; Stokland, Siitonen and Jonsson 2012). It is well established that the three most important drivers of coarse wood decomposition are climate variables (Gange et al.2007; Venugopal et al.2016), substrate quality (Edman, Möller and Ericson 2006; Venugopal et al.2016) and the decomposer community composition (Boddy 2001; Boddy, Frankland and van West 2008). However, it is a major challenge to disentangle the relative contributions of these different drivers and their underlying dynamics (Cornelissen et al.2012).

Fungal species dynamics and interactions during decomposition, as well as the response of the fungal decomposers to environmental variations, are poorly understood (van der Wal et al.2013; Wang et al.2015). In particular, elucidating the effects of decayer diversity on decomposition is a major challenge as decomposition is not performed by a uniform group of decomposers but by a wide array of interacting organisms. Because of this complexity, decomposer diversity (both in richness and composition) has often been neglected in predictive ecosystem models, despite its plausible significance for maintaining ecosystem resilience to climate change (Allison and Martiny 2008; van der Wal, Ottosson and de Boer 2015). Furthermore, many earlier studies of ecosystem functioning (Lawton and Brown 1994; Andrén, Bengtsson and Clarholm 1995; Naeem 1998) assumed redundancy among decomposers, albeit mostly without clear evidence (Loreau et al.2002). Most wood decay models, (Radtke et al.2009; Vávřová, Penttilä and Laiho 2009; Zell, Kändler and Hanewinkel 2009) only use factors such as wood properties and climate variables to model decomposition and carbon storage. Ecosystem models often disregard the effects of the composition of decayer assemblages on ecosystem functioning under the presumption that the microbial composition is resistant, resilient and/or functionally redundant. These models may be unreliable when the rates of microbe-driven ecosystem processes under disturbance are predicted (Allison and Martiny 2008).

Species richness, which is influenced by factors such as facilitation, complementary resource use and antagonistic mechanisms, is considered to be a major determinant of ecosystem processes (Cardinale et al.2011), such as decomposition and nutrient cycling (Gessner et al.2010). However, the positive relationship between microbial diversity and decomposition and nutrient cycling (via the additive or synergistic activities of fungal species) is a popular yet much debated issue in studies related to biodiversity and ecosystem functioning (Loreau et al.2001). Although species complementary effects have been proved to increase decomposition rates in soil communities at low species levels (up to 10), the relationship generally saturates at higher species richness (Setälä and McLean 2004; Gessner et al.2010). In addition, antagonistic interactions between saprotrophic fungi are usually strong and may counteract the positive diversity effects on decay rates at all stages of wood decomposition (Woodward and Boddy 2008).

Fungal decomposition is known to be prominently affected by metabolic properties of the fungal species, by wood type and by abiotic factors such as humidity and temperature (van der Wal et al.2013; Venugopal et al.2016; Zuo et al.2016), and these can also affect species interactions in decomposer communities (Castro et al.2010; Edman and Fällström 2013; van der Wal, Ottosson and de Boer 2015). Species-specific differences in wood decay (Boddy 2001), as well as the differential response of species to climate and wood type (Venugopal et al.2016), suggest that a wider understanding is required to identify the role each species may have in the intra- and interspecific interaction processes and in the consequent decomposition (Mikola, Bardgett and Hedlund 2002; Venugopal et al.2016). The direct and interactive effects of fungal decomposer communities and their key drivers should be understood when assessing, for example, the carbon sequestration capacity of forests under climate change (van der Wal, Ottosson and de Boer 2015).

This study aims to investigate the effect of the climatic variables temperature and humidity, and wood type on fungal decomposition of coarse woody debris in boreal forests, using species of wood-decay fungi at multiple species richness levels. In our previous study, we found species-specific responses to climate variables in monocultures (Venugopal et al.2016). In this study, we expand our earlier findings and look at the effects that wood-decay species richness and assemblage composition may have on decomposition using the following hypotheses.

• Higher species richness in saprophytic fungi is generally assumed to enhance the decomposition process (when compared with monocultures) owing to increased resource partitioning and facilitative interactions among species. Hence, we postulate an increased overall decay rate at higher decomposer species richness levels.

• The effect of climatic factors, such as temperature and humidity, and wood type on decomposer activity is species specific in pure fungal monocultures. Hence, we postulate that decomposition by multi-species assemblages is also modified by climate and wood quality factors in assemblage-specific ways.

Thus, in short, we hypothesize that decomposer communities at multiple richness levels will exhibit differential wood decay rates that will vary with climatic conditions and wood quality. This would indicate that global carbon and decomposition models would need to consider differential sensitivities for different fungal decomposers and substrates.

MATERIALS AND METHODS

Overview

To test the above hypotheses, we used experimentally composed fungal assemblages that consisted of one, two or four species that were allowed to decompose wood of two different qualities at four temperature–humidity levels. Four wood-decaying basidiomycete fungi (Antrodia xantha, Dichomitus squalens, Fomitopsis pinicola and Gloeophyllum protractum) were used. Wood type was represented by fast-grown and slow-grown Scots pine (Pinus sylvestris L.) wood from managed and natural forest, respectively. Temperature and humidity levels were manipulated using climate chambers (2 × 2 factorial experiment), to represent current and predicted climate values (Fig. 1). The experiment was run for 9 months.

Figure 1.

Schematic outline of the experimental set-up. The number of jars with wood samples inoculated with fungal cultures under each treatment that is used in the final analysis from each chamber is indicated by the number in brackets (n). Each chamber includes the same subunits as shown for the first chamber. Ax, A. xantha; Ds, D. squalens; Fp, F. pinicola; Gp, G. protractum.

Wood sample extraction and preparation

Wooden disks, each approximately 10 cm in height, were sawn off at 15 ± 10 cm above ground level with a chainsaw from randomly selected freshly harvested Scots pine stumps after final felling in managed stands in North Karelia, eastern Finland during late winter 2012. In total, 120 disks were extracted from the Scots pine stumps that belonged to the diameter class 60–80 cm. In the old-growth forest, 108 disks (of the same height) were extracted from felled kelo (slowly grown and died decorticated pine trees that host many specialist fungal species; Niemelä, Wallenius and Kotiranta 2002) tree logs. The disks were processed and dried according to procedures described by Venugopal et al. (2016). The kelo samples used in the study had >15 rings per centimeter—indicating slow growth rate—whereas the average ring count of the Scots pine wood samples from the managed stands varied between six and eight per centimeter. Four and five sets of wood samples per disk were extracted from each of the Scots pine (managed) and kelo disks, respectively. The sapwood, outer heartwood bordering the sapwood and the first 20 annual rings (outwards from the pith, considered as juvenile wood) were excluded from the disk using a band saw (Scheppach hbs 500, Germany). The sample sets were extracted approximately equidistant from each other from the middle heartwood of each disk using a circular sawbench (Scheppach ts 2500, Germany). Each set, which consisted of two wood pieces, each of size 1 × 1 × 1.5 cm, was taken from the same 10 cm vertical stump column in a way that the rings overlaid each other in both pieces. This was done to keep both pieces as similar as possible. This amounted to 480 sets comprising 960 wood pieces of wood from managed stands. The wood pieces from the slowly grown kelo disks were extracted and prepared in a similar manner (and dimensions) to those from the managed forest stand disks. In total, 1080 pieces from kelo wood were prepared. We randomly selected 432 wood sets out of each wood type for the experiment. The dry weight (after an oven-dry test) of the wood pieces ranged from 0.5 to 1.4 g.

Fungal cultures

We selected four saproxylic polypore fungi species (Aphyllophorales, basidiomycetes) for the experiments: Antrodia xantha (Fr.) Ryvarden (a common species; brown rot decayer), Fomitopsis pinicola (Sw.:Fr.) P. Karst. (very common; brown rot), Gloeophyllum protractum (Fr.) Imazeki (rare; brown rot), and Dichomitus squalens (P. Karst.) D. A. Reid (rare; white rot). These species were selected based on their traits such as specificity to conifers, rarity (in a Fennoscandian context) and type of rot. Based on our preliminary culture preparations, all these species have a roughly similar radial growth rate (expressed as days to reach 2 cm radius around the point of inoculation) under in vitro condition. Each fungal species was represented by four different geographical strains, used as replicates, that were collected from Scots pine trunks from different parts of Fennoscandia.

The growth media and pure cultures of the four fungal species were prepared as described in detail in Venugopal et al. (2016). We isolated fungal plugs from the peripheral growth zone of the pure fungal cultures grown on Petri dishes; four plugs for species richness level 1 (one fungal species, hereafter referred to as S1), two plugs each for species richness level 2 (two species, S2) and one each for species richness level 4 (four species, S4) from their respective fungal assemblage composition. All the four strains of each species were mixed in the monospecific set-ups. Although all fungal strain combinations corresponding to the multi-species combinations were established, only 16 were randomly selected per assemblage composition for the final set-up. This was done to equally distribute the variance caused by different fungal strains. The fungal plugs were extracted using sterilized glass straws. A total of 1728 0.35-liter glass jars were prepared, each with 150 ml water agar (1.5% agar + 0.5% glucose). We inoculated each jar with four 0.5 cm diameter fungal inoculums (placed equidistant from each other) according to the factorial combination arrangements and based on the species richness levels (Fig. 1). The inoculations were done 10–16 days prior to the addition of the wood pieces in order to allow establishment of each fungal species on the water agar. For the jars with pure culture (one species or S1), all the four inoculums were from identical species; the one with two species (S2) was inoculated with two strains per species; and the one with four fungal species (S4) was inoculated with one strain per species. The fungal mycelia were allowed to grow for 2 weeks in the glass jars before the gamma sterilized wood pieces (Venugopal et al., 2016) were added. The wood-inoculated cultures were then randomly distributed to the climate chambers according to the experimental design (Fig. 1).

Climate chamber set-up

Four climate chambers (Conviron PGW-model growth chambers, Manitoba, Canada), located in the University of Eastern Finland, Joensuu, were set up using the factorial combination treatments. The treatments simulated the current and predicted climate in terms of temperature and air humidity. We used a standard temperature of 16°C and relative humidity (RH) of 70% for the ‘current climate’ values based on the average weather values for June–September in the region where the wood samples were obtained (Finnish Meteorological Institute 1981–2010). We based the ‘predicted climate’ values on the Air Pollution and Climate Secretariat (Olsson 2009) estimations of an average 3–5°C rise in temperature and a 15–20% increase in RH in the boreal zone over the next century. Based on these estimations, a high temperature of 21°C and RH of 90% were used for the predicted future climate scenario in this study.

The moisture content (MC, %) from one of the wood pieces/set was measured and from this we thus assumed that all the wood pieces in the set had a similar % MC. The moisture value was obtained from % MC = ((initial weight – oven dry weight) × 100)/oven dry weight)) after drying samples at 103 ± 2°C for 24 h. The wood sample was then placed in sterile milli-Q water (100 ml) for 5 h and transferred onto 4 × 4 cm² sterile plastic wire meshes (Venugopal et al.2016) that were placed on the pre-inoculated fungal cultures. The chambers were surface sterilized using 90% ethyl alcohol before the onset of the experiment. The loosely closed jars (to allow moisture and gas exchange with the chamber) containing the wood piece on the fungal inoculum were then transferred into the climate chambers (Fig. 1).

The wood samples were wetted and frequently checked for contamination as described in Venugopal et al. (2016). The treatment factors were controlled independently for each chamber and were addressed with careful randomization of the materials and treatments pertaining to the experimental units.

After the 9-month experiment, all the wood pieces with successful fungal establishment (visually determined) were used for further preparation. This resulted in a total of 959 decayed wood pieces after elimination of samples that were discarded on grounds of contamination as well as no visible mycelia. Climate chamber I (16°C and 70%) had 238 decayed samples out of which S1 were 78, S2 were 111 and S4 were 49. Climate chamber II (16°C and 90%) had 238 decayed samples out of which S1 were 82, S2 were 109 and S4 were 47. Climate chamber III (21°C and 70%) had a total of 230 decayed samples out of which 76 were S1, 101 were S2 and 53 were S4. Climate chamber IV (16°C and 70%) had 238 decayed samples out of which S1 were 88, S2 were 119 and S4 were 46. The number of decayed samples as per the fungal assemblage composition in each climate chamber is listed in Table 1.

Table 1.

Number of decayed wood samples analyzed under different temperature–humidity treatments in the climate chamber under different fungal assemblage compositions.

	Climate chamber
Fungal assemblage composition	I	II	III	IV
A. xantha	18	27	19	27
D. squalens	20	18	24	18
F. pinicola	31	32	28	35
G. protractum	9	5	5	8
A. xantha–D. squalens	28	30	27	30
A. xantha–G. protractum	14	16	12	15
F. pinicola–D. squalens	31	23	27	34
G. protractum–D. squalens	38	40	35	40
A. xantha–D. squalens–F. pinicola–G. protractum	49	47	53	46

The samples were cleaned by carefully brushing off the fungal surface mycelia under running water and were oven dried as described earlier to estimate the final dry mass of each wood piece. Mass-loss (decay rate) (%) was calculated as [(dry weight of the initial wood piece – dry weight of the wood piece after decay) × (100/dry weight of the initial wood piece)].

Statistical analyses

The independent and interactive effects of temperature, humidity, wood quality and fungal species richness on fungal mediated wood mass loss (‘decay rate’) were analyzed with a general linear model (GLM) univariate ANOVA in SPSS Statistics for Windows, Version 21.0 (IBM Corp., Armonk, NY, USA). For the assessment of effects due to fungal assemblage composition, we used all the possible combinations at S1 and S4, and four combinations at S2. All explanatory variables were treated as fixed factors. A least significant difference (LSD)-type post hoc comparison test was used to determine the significance of pairwise differences between decay rates within different fungal species richness levels.

The normality of the data was assessed using the Shapiro–Wilk test and the homogeneity of variances with Levene's test. The data deviated slightly from the assumptions of normality, but ANOVAs are considered to tolerate violations of normality rather well (Zar 1999).

RESULTS

The decay rates (expressed as mass-loss %) exhibited several significant responses to temperature, humidity, wood type, species richness levels and fungal assemblage composition changes (Table 2). In general, the largest effects (i.e. largest MS values in Table 2) were observed in wood type, fungal species richness and composition of assemblages, although several other factors were also statistically significant (see MS and P values in Table 2). Overall, decay increased with an increase in temperature (P = 0.009; Table 2; Fig. 2a) and humidity levels (P = 0.031; Table 2; Fig. 2b). In addition, the wood samples from the managed forest decayed faster than the old-growth forest kelo samples (P < 0.0001; Table 2; Fig. 2c).

Figure 2.

The effects of (a) temperature, (b) humidity, (c) wood type, (d) fungal species richness levels and (e) fungal assemblage composition on wood decay rates (mean ± 1 SE). Note different scales of the y-axes. In (d–e), bars with the same letter are not significantly different according to LSD post hoc test at P < 0.05. In (a–c), all differences are statistically significant at P < 0.05. The four different fungal species are represented as follows: Ax, A. xantha; Ds, D. squalens; Fp, F. pinicola; Gp, G. protractum.

Table 2.

Effects of temperature, humidity and wood type on decay rates. Asterisks indicate the statistical significance of the findings: *P < 0.05; **P < 0.01. MS, mean square value; F, F-test value.

Factor	MS	F	P
Temperature	571.40	6.79	0.009**
Humidity	393.22	4.67	0.031*
Wood type	1614.28	19.18	0.000**
Fungal diversity	1372.40	13.76	0.000**
Fungal assemblage composition	1158.93	13.53	0.000**
Temperature × Humidity	1134.24	11.37	0.001**
Fungal diversity × Wood type	788.291	7.90	0.000**
Temperature × Wood type	4.87	0.05	0.825
Humidity × Wood type	488.253	4.90	0.027*
Fungal diversity × Temperature	569.39	5.71	0.003**
Fungal diversity × Humidity	23.33	0.23	0.791
Fungal assemblage composition × Wood type	803.746	9.547	0.000**
Fungal assemblage composition × Temperature	146.50	1.74	0.109
Fungal assemblage composition × Humidity	79.64	0.95	0.461

The overall decay rate remained constant from S1 to S2 (P = 0.482; Table 2; Fig. 2d) but decreased significantly at S4 (P < 0.001; Table 2; Fig. 2d). Notably, the decay rates also varied between many of the different fungal assemblage compositions (P < 0.001; Table 2; Fig. 2e). For example, decay rates exhibited by single species A. xantha, D. squalens and F. pinicola differed from each other, while the decay rates by A. xantha and G. protractum were similar to each other and also to some of the fungal compositions at S2 (Fp–Ds; Ax–Gp) and S4 (Fig. 2e).

An increase in temperature appeared to increase decay rates, except for S4 where the decay rate almost halved (Fig. 3b). Of the wood types, kelo decayed more slowly at all species richness levels when compared with the wood samples from the managed forests, S2 being most sensitive to the change in wood type (Fig. 3a). For both wood types, the S4 assemblage showed the lowest decay rate (Figs 2d, 3a). Although humidity did not appear to have any significant interaction with the species richness, the decay rate seemed to be slightly accelerated by increase in moisture content, S4 exhibiting the least decay at most humidity levels (Fig. 3c). It was also interesting that, albeit non-significant, wood from the managed forest decayed faster under increased humidity levels whereas kelo was only slightly affected (Fig. 3d). The decay rate at the lower humidity level was found to be sensitive to warming (Fig. 3d), whereas at the higher humidity level this was not observed (Fig. 3e). The response of the fungal assemblage compositions to the different wood types, humidity and temperature was quite varied and species- and combination-specific differences were evident (Fig. 3f–h). Most of the fungal combinations, with the exception of F. pinicola and G. protractum, resulted in more rapid decay in the wood from the managed forests than the kelo forests (Fig. 3f).

Figure 3.

Two-factor interaction effects on decay mass-loss (% of dry weight). (a) Fungal species richness × Wood type; (b) Fungal species richness × Temperature; (c) Fungal species richness × Humidity; (d) Wood type × Humidity; (e) Temperature × Humidity; (e) Wood type × Fungal assemblage composition; (f) Temperature × Fungal assemblage composition; (g) Humidity × Fungal assemblage composition. Note different scales of the y-axes. The four different fungal species are represented as follows: Ax, A. xantha; Ds, D. squalens; Fp, F. pinicola; Gp, G. protractum.

DISCUSSION

We found that the decay rates were significantly affected by all the tested environmental factors (temperature, humidity and wood type). Significant effect of wood type and interactions of the given environmental variables on fungal wood decomposition (in monocultures) were previously observed in our earlier study (Venugopal et al.2016). We can further verify that fungal species richness and assemblage composition also affect decomposition rates. Finally, we were able to show that there are indeed interactive effects of fungal assemblage composition and environmental factors on wood decay rates. However, it was notable that at higher species richness, the decomposition rates decreased with warming (Fig. 3b), thereby contradicting our first hypothesis. Despite the lower overall decay rate at higher species richness (S4) (Fig. 2d), specific species or their combinations (at S1, S2 and S4) showed either lower or higher decay rates (Fig. 2e). This indicates the possibility that the effect of fungal species combinations could overrule the overall effect of species richness. The specific traits of fungal species could be responsible for the underlying mechanisms of this phenomenon. Thus, decay patterns may be species specific or assemblage specific, with no clear relationship to assemblage richness.

Although we were not able to track the contribution of each species within the assemblage, there are a few obvious differences on their decay rates. The faster growing wood-decay species (at S1; F. pinicola and D. squalens) produced more rapid decay when growing as single species cultures rather than together, which possibly indicates that production of defensive metabolites affects the decay processes. The mass loss in the four-species combinations was less and more similar to the slower growing fungi than that of the faster decaying species. Similar observations of competition inhibiting decay have been previously observed (Hiscox et al.2015; van der Wal, Ottosson and de Boer 2015). Negative relationships between fungal diversity and decomposition rates using wood in natural high-diversity systems have been published also by Yang et al. (2016). Unlike the combination of faster growing decay species, the decay rates of the two slower growing species, A. xantha and G. protractum, remained constant even when paired. This might indicate the possibility of functional redundancy among combinations of slower growing decay species. Functional redundancy in fungal decomposers where dissimilar fungal communities exhibited functional similarity when subjected to similar environmental changes was previsouly reported by Haynes et al. (2015). It was also seen that when a slower decaying species was paired with a faster decaying species, the decay rate increased in comparison with the rate exhibited by the slow decayer on its own, and in one case even when compared with the fast decayer. This might indicate certain complementarity when faster and slower decaying species are combined. Our findings are in general agreement with the ‘jack-of-all-trades’ effect proposed by van der Plas et al. (2016) where the function levels in multi-species communities are expected to be intermediate and never as extremely low or high as in their corresponding monocultures. This is not in accordance with our first hypothesis where higher decay rates at higher species richness at all times were anticipated.

Temperature is known to enhance biological activities (Brown et al.2004; Bronson et al.2008; Allison and Treseder 2011). Crockatt and Bebber (2015) have previously stated that wood decomposition generally increases with a moderate increase in temperature and humidity. This phenomenon was clearly seen in the overall decay results in our experiment. The results corroborate our previous findings (Venugopal et al.2016), as well as those from Dang et al. (2009), which showed species-specific responses to temperature changes. Although overall humidity was found to be significant, we failed to observe any significant interactions between humidity and fungal diversity (species richness and assemblage compositions). Suttle, Thomsen and Power (2007) reported that species interactions override the direct response to climate change over extended time scales. Hence, we can conclude that humidity could be a limiting factor only at low levels.

The differential decay rates exhibited at different species richness levels and assemblage compositions in relation to wood types in our results were in line with the findings by Rajala et al. (2012) who suggested that substrate quality influences the fungal interactions in wood. Species-specific response to wood quality has also been previously demonstrated (Edman, Möller and Ericson 2006; Venugopal et al.2016). Kelo, the slow-grown substrate, was found to decay at slower rates with most of the assemblage types in comparison to its faster grown managed-wood counterpart. Kelo samples were found to have narrower rings compared with the wood from managed stands and this could indicate higher wood density in the former (Raiskila et al.2006). This, in turn, might be attributed to the unique growth and death patterns of kelo trees (Niemelä, Wallenius and Kotiranta 2002), which might have contributed to their higher decay resistance.

Overall, our results indicate that decomposition was found to be relatively more sensitive to fungal species richness, assemblage composition and wood type changes when compared with climate variables. This supported our second hypothesis that predicted an assemblage-specific response to the environmental variables. This observation was also shared by Bradford et al. (2014) who had earlier pointed out that climate may not be the predominant factor in organic matter decomposition. Our findings specify that the interaction within the species richness levels and assemblage composition were predominantly modified by the differential sensitivity of the interacting fungi to the environmental parameters: wood type, temperature and humidity. In other words, the traits of some fungal species overruled the species number effect and its interaction with the environmental factors. Our results, however, should be interpreted with caution as the study was performed under laboratory conditions using a limited number of species.

To summarize, our results clearly indicate the need to better understand several species- and environment- or wood quality-specific factors that modify decomposition of wood. Also, the models that analyze wood and carbon dynamics in forest ecosystems should address the sensitivity of the models to the characteristics of decomposer communities. This is needed to improve the predictive capacity of carbon models that often have either disregarded decomposition communities or assumed the same sensitivity for different saproxylic species in regard to decomposition and carbon cycling when environmental conditions change.

Acknowledgments

This study was funded by Maj and Tor Nessling Foundation (project 2014530, grant to author JK). We thank Jarmo Pennala and Juha Vattulainen at the School of Forest Sciences, University of Eastern Finland, for their invaluable assistance in the field work and sample preparation. We fondly remember Maini Mononen, School of Forest Sciences, University of Eastern Finland for her timely guidance and assistance in the laboratory. We extend our gratitude to Metsähallitus, Finland and its field staff for their valuable help in providing the samples.

Conflict of interest.None declared.