Changes in SOM associated with tree mortality and secondary succession
Our results suggest that the ongoing eco-physiological (tree senescence) and ecological (secondary succession) changes occurring aboveground were hardly reflected in the quantity and the composition of SOM (Tables 1 and 2; Tables S1 and S2). Both the quality and the quantity of SOM were, apparently, not affected by tree defoliation, subsequent mortality and species replacement. As tree health worsens and mortality eventually occurs, and after 10 years of net negative C balances (due to no direct inputs of above and belowground litter and decomposition of existing SOM), we expected to observe a net increase in the relative abundance of less palatable organic compounds, such as polyphenols (aromaticity index), lipids, and other aliphatic compounds (recalcitrance index). As the input of more labile substrates as well as cellulose and hemicelluloses would decrease and eventually stop under dead canopies, we also expected to find a significant increase in the aliphacity index (A/A-O index) under dead trees. For the same reasons, we also expected to find a substantial decrease in SOM under dead trees, as is generally observed when trees are harvested (Nave et al. 2010). It seems, however, very unlikely that the important changes in soil C supply associated with tree mortality would not have resulted in parallel loss in quality and/or quality of soil C pools. In this context, the mixture of typically low palatability of pine litter (Currie 1999; Silver and Miya 2001) along with the inhibitory effect that summer droughts have on the decomposition of SOM (e.g., Curiel Yuste et al. 2007) could have resulted in relatively slow C losses, suggesting that changes in soil C could be occurring at a very slow pace following tree mortality. Other possibility is that forest compensates the losses of C in these gaps with fresh C. No differences in fine root biomass and fine root C:N ratio (which could be a proxy for fine root activity; Pregitzer et al. 1998; Burton et al. 2002) were found among microbiomes (Table 2) suggesting that the niche left by dead individuals has been occupied by new roots of surrounding living vegetation. While this could imply a more complete use of microhabitats and their resources below dead trees, it also may imply continuous incorporation of fresh carbon (in the form of mycorrhizal hyphae and/or fine roots) that compensates carbon losses via mineralization, that is, the forest living canopy is investing C in those gaps compensating for the losses associated with decomposition. Whatever the mechanisms involved, the steady state of C stocks in response to different episodes of tree mortality, being the last one a decade ago, indicates that soil C pools were not very sensitive to this environmental perturbation, at least in the time-frame tackled by this study.
Bacterial community structure and functioning associated with tree mortality and secondary succession
The high values of bacterial diversity (Table S3), together with the abundance distribution of taxons, closely resembled results obtained in other ecosystems with the same technique (e.g., Janssen 2006; Roesch et al. 2007; Acosta-Martínez et al. 2008; Will et al. 2010; Torres-Cortés et al. 2012). Bacteria ubiquity, dispersal capabilities, and abundance (Martiny et al. 2006, Finlay 2002) could help minimize community differences on a micro-local scale, but, as hypothesized, even at very small spatial scales of this study, we found differences in the structure and composition of the microbial communities under different canopies (Figs. 3-5). These observed differences clearly contrasted with the lack of differences in soil C pools observed under different canopies, and suggest that in this natural experiment microbial communities were more sensitive to tree die-off and succession than were soil C dynamics.
The observed differences in taxonomic composition and diversity between soils under healthy canopies (HOs vs. healthy Scots pines; see Figs. 4, 5, Table S3) could hardly be associated with litter chemistry, as observed in other studies (Binkley 1995; Strickland et al. 2009; Ushio et al. 2010), because quantity and/or quality of SOM (Tables 1 and 2) and nutrient composition of leaf litter (Table 3), did not differ substantially under those canopies. Other aromatic compounds, such as species-specific phenols (Kuiters 1990; Kraus et al. 2003), which may strongly affect the composition and structure of soil microbial communities (Kuiters 1990; Kraus et al. 2003; Strickland et al. 2009; Ushio et al. 2010), were not significantly different between canopies either (Tables S1 and S2). The observed differences may, therefore, respond to other species-specific plant-soil interactions not accounted for in this study, for example, species-specific volatile organic compounds emitted by litter and roots, which are known to affect the structure, stoichiometry and functioning of soil microbial communities (Ramirez et al. 2009; Asensio et al. 2012). The degree of specificity in the interaction between plants and microbes seems, therefore, a very complex matter that involve eco-physiological and/or ecological mechanisms far from being understood.
Tree dieback and death were associated with important differences in diversity (Table S3), structure (Fig. 3), and taxonomic composition of soil bacterial communities (Figs. 4, 5). Eco-physiological changes associated with tree defoliation were, in turn, associated with rapid shifts in the abundance of key taxons of the bacterial community (Figs. 4, 5). Indeed, the loss of health of Scots pine on this or similar stands has been associated with a significant drop in plant-tissue nonstructural forms of C (Galiano et al. 2011), which might have also affected resource allocation to different tree compartments (Sala et al. 2010; McDowell 2011) and hence to roots and associated soil microbes. Soil bacterial communities under DPs, whose death is dated about a decade ago (HereŞ et al. 2011), showed important differences in the diversity (Table S3), structure (Fig. 3), and relative abundance of key taxa (Figs. 4, 5) with respect to bacterial communities under living canopies (healthy and DFPs as well as HOs). We found a significant increase in bacterial diversity (Table S3) with respect to HPs and a clear divergence in community structure with respect to living pines (healthy and defoliated; Fig. 3). This increase in bacterial diversity (Table S3) and the occurrence of a number of new taxa not observed under living vegetation (Table S4) indicate that tree mortality may have stimulated the gap colonization by new bacterial taxa during the last decade, as seen in the early succession stages of microbial community colonization (Huston 1994; Jackson 2003).
Besides the relative increase observed in diversity, pine mortality-driven changes in key abundant taxa, for example, Actynomycetes, Solirubrobacterias, Rhizobiales, or Xantomonadales, pointed towards a convergence of the bacterial communities under dead trees and HOs (Figs. 3-5, and Figs. S2 and S3). Assuming that root-associated organisms are, to some extent, species-specific (Yeates 1999; De Deyn et al. 2003; Bever et al. 1997) our results may indicate that gaps generated by pines are being colonized not only by new bacterial taxa, but also by roots from healthy HOs and their associated microorganisms. We therefore hypothesize that the observed convergence may respond to an early colonization by new bacterial taxa and new HO roots and microorganisms associated into the space occupied by the dead pine area of influence. As explained above, our results indicates that the niche left by dead individuals has been occupied by new roots of surrounding living vegetation, which given the observed convergence of the bacterial communities may imply a more complete use by healthy HOs of microhabitats and the resources left by DPs. Once a tree is dead, the whole pool organic matter generated becomes very susceptible source of nutrients and energy for exploitation by colonizer plant roots (Sanford 1989; Ostertag 1998).
The Hegyi competition index, on the other hand, does not show that DPs were experiencing a higher demographic pressure than that experienced by healthy or DFPs (Table 2), thus confirming that the colonization of above-ground gaps by HOs was in an early stage with respect to below-ground colonization. Studies of the colonization of forest patches have often focused on above-ground processes, for example, dispersal of plant diasporas or the capacity of the colonizer to establish itself in a new habitat and exploit the resources, including soil water and nutrients (Willson 1992). The colonizer plants may, however, also “import” their own microbial community in order to facilitate their germination and eventual recruitment. Soil microbes are important regulators of plant dynamics and diversity that contribute to plant survival, establishment, and growth, for instance through the fixation and mineralization of nutrients (Landerweert et al. 2001; Van der Heijden et al. 2008). This is mere speculation at present but, nevertheless, additional studies need to be designed to better understand the below-ground mechanisms involved in plant colonization, as well as the above-ground ones.
Tree die-off and soil ecology: indirect links between taxonomy and ecology of soil bacterial communities
Punctual measurements of soil respiration taken in spring 2010 over the same soils sampled for this study suggest that during this period of the year soil respiration was slightly, but consistently higher under dead individuals with respect to living trees (Table 2). A mixed-effect model approach revealed that variables describing quantity and/or quality of SOM could not explain these differences, but the increase in soil respiration under dead individuals was further stimulated by the presence of HOs in the surroundings (J. Barba, J. Curiel Yuste, J. Martínez-Vilalta, F. Lloret, unpubl. data). While these results indicates a direct relation between drought-induced secondary succession and soil respiration, the observed differences in structure of soil bacterial community under different canopies may help supporting these conclusions. Indeed, we found anomalously high representation of bacteria from the Bacteroidetes and β-Proteobacteria taxa under DPs, where soil respiration was higher (Fig. 3). These are two taxons commonly associated with opportunistic r-strategic bacteria typical of environments rich in labile organic matter (Simon et al. 2002, Fazi et al. 2005; Fierer et al. 2007). This finding points to a more copiotrophic, richer in organic matter, environment under DPs with respect to healthy vegetation, which on the other hand contradicts the results obtained with 13C CP MAS NMR spectra (Table 1 and Table S1). The higher soil metabolic activity (soil respiration) and the differences in soil microbial ecology (diversity and taxonomic composition) under similar nutritional conditions (same quality, same quantity, and same stoichiometry) points to a more efficient use of resources in the “successional hot-spots” (areas under DPs nearby colonizer HOs) of the forest. Although the lack of replication prevented us from drawing conclusions, our observations adds to a growing number of studies claiming a role of the microbial community in soil C emissions and, in general, in soil C processes (e.g., Balser and Firestone 2005; Waldrop and Firestone 2006a,b; Balser and Wixon 2009; Curiel Yuste et al. 2011; Mackelprang et al. 2011) . As at least half of the CO2 emitted by terrestrial ecosystems is produced by microbial-mediated decomposition of SOM, understanding climate change-related shifts in the composition and diversity of microbes may be crucial to our understanding of future CO2 emissions, particularly from the Mediterranean basin.