Cascading effects associated with climate-change-induced conifer mortality in mountain temperate forests result in hot-spots of soil CO 2 emissions

Graphical Abstract (for review) This document is the Accepted Manuscript version of a Published Work that appeared in final form in: Curiel Yuste J., Flores-Rentería D., García-Angulo D., Heres A.-M., Braga C., Petritan A.-M., Petritan I.C. 2019. Cascading effects associated with climate-change-induced conifer mortality in mountain temperate forests result in hot-spots of soil CO 2 emissions. SOIL BIOLOGY & BIOCHEMISTRY. 133. 50-59. DOI (10.1016/j.soilbio.2019.02.017). © 2019 Elsevier Ltd This manuscript version is made available under the CC-BY-NC-ND 3.0 license http:// creativecommons.org/licenses/by-nc-nd/3.0/ Episodes of drought-induced conifer mortality are becoming more frequent in Europe Conifer mortality triggers secondary succession of autochthonous broadleaf species. Tree mortality triggers complex cascading effects which ultimately affects Rs Changes in fine root specific root length (SRL) correlates negatively with RH. *Highlights (for review) Cascading effects associated with climate-change-induced conifer mortality in mountain temperate forests result in hot-spots of soil CO2 emissions Curiel Yuste J. 1,2*, Flores-Rentería D.3, García-Angulo D.4A.-M.1,5, Brag C.6, Petritan A.-M.6, Petritan I.C.5 1 BC3 - Basque Centre for Climate Change, Scientific Campus of the University of the Basque Country, 48940 Leioa, Spain; 2 IKERBASQUE - Basque Foundation for Science, Maria Diaz de Haro 3, 6 solairua, 48013 Bilbao, Bizkaia, Spain 3 CONACYTCINVESTAV Unidad Saltillo, Group of Sustainability of Natural Resources and Energy. Av. Industria Metalúrgica 1062, Parque Industrial Ramos Arizpe, Ramos Arizpe, 25900, Coahuila, Mexico. 4 Museo Nacional de Ciencias Naturales (MNCN, CSIC) Serrano 115 dpdo; E-28006 Madrid, Spain. 5 -1, 6 National Institute for ResearchEroilor, Voluntari, 077190, Romania *Corresponding author: Tlf: +34 944014690; e-mail= [email protected] Keywords: Soil respiration, heterotrophic respiration, secondary succession, tree mortality, fine roots functional traits, cascading effects *Manuscript with continous line numbering Click here to view linked References Summary As a widespread phenomenon affecting terrestrial ecosystems worldwide, the extent and spatio-temporal scales at which the increasing number of reported events of climatechange-induced tree mortality could affect the ecology and carbon (C) sink capacity of terrestrial soils, remains unknown. We here study how regional-scale drought-induced tree mortality events registered after a very dry 2012 year in the Carpathians mountain range (Romania), which affected three of the most widely distributed conifer species: Scots pine, Black pine, and Silver fir, resulted in hot-spots of biogenic soil CO2 emissions (soil respiration; Rs). Four to five years after the main mortality event, Rs-related soil CO2 emissions under dead trees were, on average, 21% higher than CO2 emissions under living trees (ranging from 18 to 35%). Total (Rs) and heterotrophic (RH)-related soil CO2 emissions were strongly determined by the soil environmental alterations following tree mortality (e.g. changes in quantity and quality of soil organic matter, microclimate, pH or fine root demography). Moreover, the massive mortality event of 2012 ultimately resulted in a stronger dominant role of successional vegetation (broadleaf seedlings, shrubland and grasses) in controlling those environmental factors that either directly or indirectly affected biotic soil fluxes (Rs and RH). We, therefore, show that apart from the well-known direct effects of climate change over soil CO2 emissions, cascading effects triggered by climatechange-induced tree mortality could also exert a strong indirect impact over soil CO2 emissions, altering the magnitude and the environmental controls of Rs and hence determining ecosystem C budget and their response to climate. 1. Introduction The number of episodes of forest defoliation and mortality associated with climate change has increased substantially during the last decades (Allen et al., 2010; Carnicer et al., 2011), and it is further expected to increase even more in future decades (IPCC 2014). In this regard, understanding physiological and ecological causes of tree mortality as well as predisposition of trees to die is nowadays a hot-topic that has attracted many attention and studies (Allen et al., 2010, 2015; Anderegg et al., 2012; McDowell et al., 2015; SangüesaBarreda et al. 2015; Rogers Brendan et al., 2016; Neumann et al., 2017; Lloret and Kitzberger, 2018). There is, however, a knowledge gap on how ecosystems are actually responding to such perturbations, i.e. whether and at which extent tree mortality could affect ecosystem functioning (Anderegg, et al. 2013) and more particularly how tree mortality could affect soil respiration (Rs), which represents the total biogenic CO2 produced and emitted from soils (Vargas et al., 2010), and is the major outgoing flux of CO2 from ecosystems to the atmosphere (Curiel Yuste et al., 2005; Davidson et al., 2005; Barba et al., 2018). Trees have the capacity to modulate the belowground environment (Flores-Rentería et al., 2015, 2016) triggering cascading causal-effect relations that could result in substantial changes in the biological functioning of the soil system and in fundamental alterations of the soil nutrient and soil CO2 emissions (Flores-Rentería et al., 2018). However, data and evidences on how these alterations occur and at which extent tree mortality could affect patterns and controls of CO2 emissions from terrestrial soils are scarce. Besides altering soil abiotic conditions, tree mortality limits the supply of substrate in the form of carbohydrates (e.g. exudates) or nutrients (litter) demanded by belowground organs (fine roots), symbionts (e.g. mycorrhiza), and soil biological communities from the rhizosphere and the soil (e.g. Högberg et al., 2001; Binkley et al., 2006; Barba et al., 2016). The disruption of the C flow to belowground has been directly related to almost immediate decreases in Rs (Högberg et al., 2001; Binkley et al., 2006; Nave et al., 2011; Levy-Varon et al., 2014), due to a parallel decrease in autotrophic (fine root and mycorrhiza) and heterotrophic (respirations from microbes and soil fauna) respiration. On medium-long term, the death of trees may have critical effects over key soil biogeochemical cycling (Rodriguez et al. 2016), resulting in chronical losses of key nutrients such as nitrogen (N), with unknown consequences for the capacity of the systems to recover pre-perturbations pools and functions (García-Angulo et al. in prep). This happens because the disruption of the flow of C from plants to soils and the changes in the microclimatic conditions associated with tree mortality may prominently alter the composition, structure and functionality of soil biological communities (Curiel Yuste et al. 2012, Avila et al. 2016) resulting in irreversible losses of key functional groups that sustain important soil functions such as N fixation or mineralization of essential nutrients (Gómez Aparicio et al. 2017). Few studies have been designed, however, to investigate in depth how processes triggered by tree mortality could affect biogenic soil CO2 emissions and at which extent. In this regard, tree mortality has been associated with ecosystems reaching new equilibriums, resulting in important changes in the diversity of soil biota and soil functions (Curiel Yuste et al., 2012; Lloret et al., 2015, Avila et al., 2016), as well as in the overall biogenic emissions of CO2 from soils (Moore et al. 2013; Avila 2018). Depending on the magnitude of the tree mortality event and/or legacies from historical management, forest ecosystems are able to counteract potential negative effects of tree mortality and recover pre- perturbations functioning rates relatively fast (Nave et al., 2011; Gough et al., 2013; LevyVaron et al., 2014; Barba et al., 2016). This is because tree mortality, depending on the ecosystem's characteristics and its initial conditions, triggers a process of recolonization by seedlings of the same species (regeneration) or of other species better fitted to present conditions (secondary succession), which slowly replace the niche left by the death of the trees (e.g. Vayreda et al., 2016; Ruiz Benito et al., 2017). How these complex aboveground ecological processes could actually impact belowground functioning, subsequently affecting magnitude and controls of biotic CO2 emissions remains unknown. The objective of this study was to deepen our mechanistic understanding of the effects of large-scale tree die-off events on total soil respiration (Rs). For that, we here show a regional scale study, spanning for two consecutive years (2016, ; and 2017, 2 ), on rates of Rs in stands located in Romania), where recent extremely dry years, such as 2012, have resulted in extended mortality rates. These mortality events have mainly affected conifer tree species, especially Scots pine (Pinus sylvestris L.), Silver fir (Abies alba Mill.), and Black pine (Pinus nigra Arnold)mortality observed in 2012 and estimated to extend over large areas Kronstadt), followed a sequence of several extraordinary dry and hot years registered during the first decade of the century (http://www.meteoromania.ro/anm/?lang=ro_ro). These affected conifers are slowly replaced by autochthonous broadleaf species, especially Quercus robur, Fagus sylvatica, Fraxinus ornus, Fraxinus excelsior, Carpinus betulus, Quercus petraea or Acer campestre (Table 1). Additionally, the study collected detailed information on variables potentially sensitive to tree die-off and directly/indirectly associated with Rs, e,g, soil water content (SWC) and temperature (Tsoil), soil C and nutrients pools, soil heterotrophic respiration (RH), and fine root biomass and functional traits (e.g. specific root area, SRA and root length, SRL). We, therefore, hypothesized that abiotic/biotic changes promoted by tree mortality will trigger a cascade of causal-effect relations that could have resulted in substantial changes in the biological functioning of the soil system, subsequently altering the soil CO2 emissions. In particular, we hypothesized that in these coniferous forests, where tree mortality is giving way to forests dominated by native hardwood species, the process of secondary succession associated with mortality of conifers would be associated with profound transformations of the microclimate, biology and chemistry of the soil, which ultimately will affect the magnitude and controls of soil CO2 emissions. 2. Materials and Methods 2.1. Study sites For this study, we selected a total number of 9 conifer stands (3 for Silver fir, 3 for Scots pine, and 3 for Black pine), all of them affected by recent mortality events (i.e., following the 2012 drought) (Table 1). These forests were located in the Transylvanian side of the Eastern Romanian Carpathians Mountain range ( County). As we wanted to avoid/limit as much as possible other disturbance factors (e.g. management), forests were selected either in protected areas or in areas where management intensity has been minimal for the last decades. All stands were located on sloppy terrains (slopes ranging from 17 at to 37 º). Both pine species stands were located between 450 and 700 m a.s.l., while the Silver almost pure stands, that were artificially regenerated 100 years ago, whereas the Silver fir stands are uneven, naturally regenerated, mixed stands with Fagus sylvatica (up to 35% in the forest composition) of more than 150 years old . Soil type in the Silver fir stands is mainly Eutricambisols, while Rendzina is the main soil type of the pine stands (Table 1). Both mean annual precipitations (MAP) and mean annual temperatures (MAT) (Climate Research Unit Time Series, CRU TS3.10; via http://climexp.knmi.nl) were relatively low and not very variable among locations, ranging, respectively from 593 to 693 mm and from 3.7 to 6.6 ºC (Table 1). Our study also shows a natural understory gradient (tree saplings and seedlings) and grass cover, from sites with very scarce understory/grass cover (7%) to sites with understory/grass cover averaging up to 70% (Table 1). The drought-induced mortality rate was estimated to round 19-23% for Silver fir, 16-27% for Black pine, and 17-22% for Scots pine (Table 1). 2.2.Field measurements 2.2.1. Experimental design and tree age estimation At each of the 9 conifer stands affected by mortality, 5 pairs of standing adult dead and living trees were sampled (see below) along a transect perpendicular to the slope. We used a paired sampling design (Bigler and Bugmann 2004), in which the selected living trees had similar size (diameter at breast height, DBH), competition level, and microsite conditions with the dead ones. Trees noticeably affected by biological agents (e.g. pathogens, fungus), wind, or human influences were avoided during the sampling. The sampling of the 5 pairs was carried out in transects starting at a random point within each stand and maintaining a constant altitude, and thus similar humidity conditions, until the required number of trees Structural equation models (SEMs) were finally used to test the direct and indirect influence of the different biotic and abiotic factors on Rs. Since the sample size was relatively small (n = 88), the number of predictors included in the model was thoroughly limited, as recommended by Shipley (2002). Our models considered a complete set of hypothesis based on literature, previous exploratory analyses (Kruskal-Wallis, correlations, etc.), and our own previous experience (Flores-Rentería et al., 2016, 2018; Pérez-Izquierdo et al., 2017, see Fig. S2). The model assumed that aboveground vegetation structure (size of the sampled living and dead trees, tree competition index, and the % of understory and grass cover) would affect the abiotic (microclimate, pH) and the biotic (root demography) soil environment, as well as the nutritional status (PC1) and the organic matter content (PC2) in soils under trees. Overall, both RH and Rs, would be strongly controlled by changes in all these environmental factors which are ultimately controlled by the aboveground vegetation structure. Several SEMs were run and the best-fitted ones were finally selected according to the covariance proximity between observed and expected data (goodness-ofgeneral model we used multigroup SEM to test whether the studied factors were linked by the same causal structure in each tree status (dead or living) and to identify the paths that did not behave similarly in the two conditions (Shipley, 2002; García-Camacho et al., 2010). For this analysis, we used the same hypothetical model used for each status group separately. A constrained model in which all free parameters were forced to be equal across the two conditions was built and contrasted with field data. Since a lack of fit was detected in the fully constrained multigroup model, a series of nested models, where equality constraints were removed one at a time, were developed to detect which one would significantly improve the model (Shipley, 2002; García-Camacho et al., 2010). Differences in X2 and AIC statistics between the fully constrained model and the models freed from a constraint were used to test for differences in parameter values between the two conditions. Standardized path coefficients were estimated by using the maximum likelihood algorithm (Shipley, 2002). SEM analyses were performed using SPSS® and SPSS® AMOS 20.0 3. Results Both SWC and Tsoil varied significantly between years 1 6) warmer and wetter on average than 2 7), although no such significant trend was found for Rs (Figs. 1 and S3). No differences in Rs were found between the three conifer species, although the soils from Silver fir stands were significantly colder (lower average Tsoil) and wetter (higher average SWC) than the ones from the Black pine and Scots pine stands (Fig. S3). When looking for differences in SWC, Tsoil and Rs considering tree status (living or dead), living trees differed significantly from the dead ones only in SWC and Rs. Specifically, dead trees showed higher SWC and Rs values than the living ones (Figs. 1 and S3). Overall, Rs was consistently higher under dead trees with respect to living ones (Fig. 2, Table S1). According to our results, CO2 emissions under dead trees were on average 21% higher than under living ones. Furthermore, RH, specific root length/area (SRL/SRA), TOC, TON, and SWC were among the environmental variables that were consistently higher, although not always significantly, under dead trees comparing with the living ones (Fig. 2, Table S1). Other variables such as pH, C:N ratio, or total fine root biomass/volume (FRB and FRV, respectively) showed less sensitivity to tree mortality and/or less consistent trends across tree species, although FRB tended to be lower under dead trees (Fig. 2, Table S1). No clear differences were found between living and dead trees when the grass and understory cover, or the tree competition index were considered (Table S1). This happened because cover of understory of broadleaf seedlings and grasses was independent of the health status of the sampled trees and very dependent on site-specifics conditions. Both mean soil temperature and quantity of SOM (PC2) were, together with tree status (living and death) the variables that better explained Rs variability across sites (Fig. 3, Table S1). As expected for ecosystems generally limited by temperature, Tsoil explained a large portion of across-site variability in Rs (Fig. 3, Table S2), whereas Rs was also strongly driven by SOM quantity (PC2, Fig. 3, Table S2). No differences in RH between soil collected under dead and living trees were found (Figure 4). On the other hand, we found a strong effect of the nutritional status (PC1) and the quantity of SOM (PC2) on RH (Fig. 4, Table S3). Additionally, the best obtained model also included a significant negative effect of SRL over RH (Fig. 4, Table S3). SEMs showed the complex causal-effect cascade of processes controlling Rs and RH (Fig. 5). Specifically, this analyses highlighted how strongly forest structure (tree DBH, tree competition, understory and grass cover) influenced, directly or indirectly, the observed variability of soil abiotic (microclimate, pH, nutrient content) and biotic (SRL, FRV, RH) variables, resulting in the observed variability in Rs across sites. Both, trees (i.e. size and tree competition) and understory cover exerted a strong effect over soil microclimate (Tsoil), soil pH, and nutrients (PC1). While conifers tend to acidify soils, we here observe how an increasing cover of broadleaf understory was associated with increases in pH, which, on the other hand, was also behind the observed improvement of soil nutritional status (PC1) and SOM sequestration (PC2). The presence of grasses was also directly associated with an increase in SOM (PC2) and SRA, but a decrease in FRV. Multigroup SEM further showed a tighter control of Tsoil over Rs under dead than under living trees, as illustrated by the significantly higher ML coefficient obtained. Moreover, multigroup SEM also showed how conifers and successional vegetation exerted an opposite effect over the demography of fine roots (SRL and FRV). Specifically, living trees exerted an overall negative effect over SRL, whereas the presence of grasses was positively associated with SRL. On the other hand, the FRV was stimulated in poor soils (high PC1) and under high SOM contents (low PC2), but was also negatively correlated with the presence of grasses and the increase in broadleaf understory cover. Furthermore, we here show how, independently of the health status of the trees, increase in the SRA negatively affected RH. Also, controls of RH differed depending on the conifer health status (living or dead): living trees exerted a positive control over RH, but when tree dies RH variability was mainly controlled by nutrient status (PC1) and SOM quantity (PC2) (besides SRL). Finally, observed variability of Rs seemed to be partially explained by variability of RH under dead trees while under living conifers no relation was found between both fluxes. 4. Discussion We here reported large events of drought-induced tree mortality on stands dominated by three of the conifer tree species most widely distributed in the Carpathian mountain range (Table 1). This drought-induced mortality coincides with globally increasing reported events of conifer decline, which are being generally attributed to historical management practices (e.g. favoring/planting conifer species outside their climatic and structural optimum; Urbieta et al., 2008; Ruiz Benito et al., 2012), or to climate-change induced increases in temperature and droughts intensity and frequency that seem to affect more the conifer-like than the angiosperms-like related functional traits (Henne Paul et al., 2015; McIntyre et al., 2015; Ruiz Benito et al., 2017). Indeed, the observed increasing presence of an understory vegetation of seedlings mainly composed by native broadleaf species (e.g. Fagus sylvatica, Fraxinus excelsior, Fraxinus ornus, Acer campestre, Quercus petraea, Ulmus glabra) (Table 1) coincides with these mentioned above observations and are, therefore, in accordance with this general decline in conifer dominance, especially ubiquitously observed in European forests. Drought induced conifer mortality resulted in large increases in biogenic soil CO2 emissions (Rs), averaging a 21% increase under the dead trees comparing with the living ones, and persisting during two consecutive years (2016-2017), four to five years after the mortality event occurred in 2012. Given the observed large proportion of tree mortality observed after the 2012 drought (Table 1), this tree mortality-induced hot-spot of CO2 emissions might be responsible for decelerating the capacity of these ecosystems to recover pre-mortality levels of C sequestration (e.g. Moore et al, 2013). Our results are not in accordance with decreases in Rs observed under experimental tree girdling manipulations (e.g. Högberg et al., 2001; Binkley et al., 2006; Nave et al., 2011; Levy-Varon et al., 2014), or under natural conditions, when tree mortality events were massively caused by bark beetle attacks (e.g. Moore et al. 2013) or by infections with pathogens, e.g. Phytophtora cinammoni (Avila et al., 2016). Accordingly, literature generally shows how tree death results in an almost immediate and dramatic decrease in Rs rates associated with the decrease in the supply of newly plants-fixed carbohydrate for belowground metabolic activity, e.g. autotrophic respiration, mycorrhiza activity and rhizosphere heterotrophic respiration (Subke et al., 2004; Högberg et al., 2007). This Rs drop associated with tree mortality may last for decades in monospecific forest (e.g. Moore et al. 2013, Avila et al., 2016), while other studies have shown how depending on the level of the perturbation and the secondary successional processes, Rs may recover pre-perturbation values after several years (e.g. Levy-Varon et al., 2014; Barba et al., 2016), or even increase during favorable seasons in mixed forest (Barba et al., 2013). Hence, due to the initial physiological collapse that tree mortality produces in a system, the functional recovery of this system in general and of the Rs in particular depends on the degree of perturbation but also on secondary successional process triggered by tree mortality (Levy-Varon et al., 2014; Lloret et al., 2015; Barba et al., 2018). In these conifer forests so representative for the Carpathians' landscape, the observed hotspots of CO2 under dead trees were strongly dominated by RH (Fig. 2 and 5). We here postulate that these mortality-triggered hot-spots of RH and Rs were mostly explained by an increase in the quality and quantity of SOM which results from both the increase in senescent material and from the successional processes following tree death (Fig. 2, 4 and 5). The observed increase in topsoil SOM under dead trees (increase in TOC, Figure 2) could be attributed, at least partially, to the accumulation of senescent plant material (leaves, roots and branches) which generally accumulate under dead trees (Moore et al. 2013). However, SOM accumulation under dead conifers alone cannot explain the observed increase in RH because, as observed, RH was also very sensitive to the increase in soil nutrient availability (PC1) (Figure 4, Table S3). Besides, we here observed how shifts in the controls of Rs after the massive mortality event of 2012 ultimately resulted in a stronger dominant role of the successional vegetation (broadleaf seedlings, shrubland and grasses) over the belowground environmental factors, directly or indirectly affecting Rs and RH fluxes (Figure 5). This shift towards greater understory control over soil functions was in detriment of the former control exerted by the conifers which influenced the microclimate (SWC and Tsoil), the abiotic soil environment (pH), the nutrient quality (PC1), SOM (PC2), and the fine root demography (specifically SRA/SRL, FRV) (Figs. 2, 3, 4 and 5). These changes were further reflected in changes in the magnitude and the controls of biotic soil fluxes (Rs and RH) (Figure 2 and 5). For instance, the increase in pH and grassland cover resulting from the shift in the aboveground vegetation dominance also played a critical role in increasing SOM (PC2; Fig. 5). This is because, under conifer influence, the generally low soil pH and the low quality of the residues due to the high proportion of recalcitrant compounds, e.g. lignin and/or allelopathic molecules (e.g. Curiel Yuste et al. 2005; Fernández-Alonso et al., 2018) slow down the breakdown of the litter and its incorporation to SOM in the mineral soil. On the other hand, incorporation of litter in SOM occurs generally faster in ecosystems dominated by broadleaf species because the generally higher pH and higher quality of the produced residues stimulates bioturbation (Frouz et al., 2009). Indeed, the multigroup SEM further showed how the increase in pH associated with the increasing presence of the understory vegetation had a direct and strong positive effect over SOM quality (PC1), suggesting that on top of the increase in SOM under dead trees, the secondary successional processes triggered by conifer mortality was positively affecting the quality of the substrate. Our results, therefore, clearly indicate how the shift in vegetation dominance associated with conifer mortality had a strong impact over the quantity and quality of SOM, resulting in increased RH, which subsequently affected Rs. This dominant role of successional vegetation after tree death was also reflected in a substantial increase in the surface of absorption of the radical system (increase in SRL and SRA; Fig. 2) which corresponds to a shift towards a fine-root demography optimized to maximize nutrients acquisition (Roumet et al., 2016). This shift, associated with an increase in the presence of grasses (Fig. 5), suggest that the belowground niche left by the death of the conifers creates an opportunity to the surrounding early successional vegetation to obtain resources (nutrients and moisture) (Curiel Yuste et al., 2012; Barba et al., 2013) whose acquisition is, otherwise, subjected to strong competition, especially in nutrientpoor, low pH conifer sites as those considered in this study. Indeed, the multigroup SEM showed how the poor nutrient conditions under conifers (PC1), while alive, promoted a bigger radical system (higher FRV), but with relatively less very fine roots (suppressing SRL). The consistent increase in the specific length and surface (SRL/SRA) of fine roots under dead trees was paralleled by the observed increase in Rs (Figure 2). This was expected, given the general observed linear relation between, on one hand, SRL and the fine roots turnover rates (Silver and Miya, 2001; Hobbie et al., 2010; Roumet et al., 2016), and, on the other hand, SRL and rates of root respiration (RA) (Reich et al., 2008; Makita et al., 2012; Picon-Cochard et al., 2012). Although RA were not measured in this study, we did not observe a significant increase in SOM turnover (rates of RH per unit of soil C; data not shown) under dead trees, suggesting that it is most plausible that the increase of SRL was associated with a parallel increase in RA. Rather than stimulating RH, our models also showed a very consistent negative relation between SRL and RH (Figure 4 and 5), suggesting that besides this expected positive effect of SRL over autotrophic activity, the net effect over RH was negative. It, therefore, could be that by increasing their capacity to absorb nutrients (increase in SRL), successional vegetation competes more efficiently for the same resources with the soil heterotrophic community (negative priming, Kuzyakov 2002), resulting in the observed suppression of RH. Indeed, an increase in competition for key nutrients (e.g. N, P, K) between roots and heterotrophs could be maximal in soils when nutrients are generally limiting, thereby resulting in the suppression of RH (e.g. Schimel et al., 1989; Wang and Bakken 1997; Kuzyakov 2002). 5. Conclusions We here collected compelling evidences to support our initial hypotheses: cascading mechanisms triggered by selective tree mortality and a subsequent secondary successional process played a critical role in regulating soil functioning and soil CO2 emissions during transitional states. Specifically, we here show how conifer mortality resulted in an average increment of biogenic emissions of 21%, 4-5 years after the large mortality event of 2012, which might be further responsible for decelerating the capacity of these ecosystems to recover pre-mortality levels of C sequestration. These transitional states after tree death resulted in a stimulation of the heterotrophic activity (RH), favored by the increase in senescent material but also by changes in the soil microenvironment (e.g. climate, pH and SOM) partially controlled by successional vegetation. A shift towards a more efficient resource-acquisitive strategy of fine roots (increase in SRL), triggered by tree mortality and also associated with the increasing dominance of the successional vegetation, was also behind the observed changes in the magnitude and controls of RH and Rs. Our results, hence, call the attention on how above-belowground ecological processes triggered by tree mortality may substantially determine dynamics of key biogeochemical cycles (e.g. C and N) at local and regional scales. One of the drawbacks of this study might be the fact that the effects of tree-mortality were only evaluated during a relatively short-term (2 years), at subdecadal time scale (4-5 years after the main mortality event), and in a limited number of sites (9). Despite its limitations, this is one of the first studies evidencing the complexity of the controls over Rs in climate-change-induced tree mortality scenarios, and as such, it might serve as a base to develop further, more extended studies on this topic. In a changing world where episodes of tree mortality associated with climate change are substantially incrementing, more studies should, therefore, be designed to deepen the observed potential impacts of tree mortality and subsequent successional processes at larger temporal and spatial scales Acknowledgements The authors thanks to the staff of Forest Districts: Kronstadt, Rasnov, Codlea, Teliu, and Sacele who administrate the 9 forests considered for this study. This work was supported by the Spanish Ministry of Economy and Competitiveness (MINECO) with the projects VERONICA (CGL2013-42271-P) and the project IBERYCA (CGL2017-84723-P), and by the Romanian Ministry of Education and Scientific Research through UEFISCDI with the projects TREEMORIS (PN-II-RU-TE-2014-4-0791), NATIvE (PN-III-P1-1.1-PD-20160583), and BIOCARB (PN-III-P1-1.1-TE-2016-1508). This research was also supported by the Basque Government through the BERC 2018-2021 program, and by the Spanish Neumann, M., Mues, V., Moreno, A., Hasenauer, H., Seidl, R., 2017. Climate variability drives recent tree mortality in Europe. Global Change Biology 23, 4788-4797. Pérez-Izquierdo, L., Zabal-Aguirre, M., Flores-Rentería, D., González-Martínez, S.C., Buée, M., Rincón, A., 2017. Functional outcomes of fungal community shifts driven by tree genotype and spatial-temporal factors in Mediterranean pine forests. Environmental Microbiology 19, 1639-1652. Picon-Cochard, C., Pilon, R., Tarroux, E., Pagès, L., Robertson, J., Dawson, L., 2012. Effect of species, root branching order and season on the root traits of 13 perennial grass species. Plant and Soil 353, 47-57. Reich Peter, B., Tjoelker Mark, G., Pregitzer Kurt, S., Wright Ian, J., Oleksyn, J., Machado, J.L., 2008. Scaling of respiration to nitrogen in leaves, stems and roots of higher land plants. Ecology Letters 11, 793-801. Rogers Brendan, M., Jantz, P., Goetz Scott, J., 2016. Vulnerability of eastern US tree species to climate change. Global Change Biology 23, 3302-3320. Roumet, C., Birouste, M., Picon Cochard, C., Ghestem, M., Osman, N., Vrignon Brenas, S., Cao, K.f., Stokes, A., 2016. Root structure function relationships in 74 species: evidence of a root economics spectrum related to carbon economy. New Phytologist 210, 815-826. Ruiz Benito, P., Gómez Aparicio, L., Zavala, M.A., 2012. Large scale assessment of regeneration and diversity in Mediterranean planted pine forests along ecological gradients. Diversity and Distributions 18, 1092-1106. Ruiz Benito, P., Ratcliffe, S., Zavala, M.A., Martínez Vilalta, J., Vilà Cabrera, A., Lloret, F., Madrigal González, J., Wirth, C., Greenwood, S., Kändler, G., Lehtonen, A., Kattge, J., Dahlgren, J., Jump, A.S., 2017. Climate and successional related changes in functional composition of European forests are strongly driven by tree mortality. Global Change Biology 23, 4162-4176. Sangüesa-Barreda, G., Linares, J.C., Camarero, J.J., 2015. Reduced growth sensitivity to climate in bark-beetle infested Aleppo pines: Connecting climatic and biotic drivers of forest dieback. Forest Ecology and Management 357, 126-137. Shipley, B., 2002. Cause and correlation in biology: a user's guide to path analysis, structural equations and causal inference. Cambridge University Press, Cambridge, UK. Silver, W.L., Miya, R.K., 2001. Global patterns in root decomposition: comparisons of climate and litter quality effects. Oecologia 129, 407-419. Subke, J.-A., Hahn, V., Battipaglia, G., Linder, S., Buchmann, N., Cotrufo, M.F., 2004. Feedback interactions between needle litter decomposition and rhizosphere activity. Oecologia 139, 551-559. Tang, J., Baldocchi Dennis, D., Xu, L., 2005. Tree photosynthesis modulates soil respiration on a diurnal time scale. Global Change Biology 11, 1298-1304. Urbieta, I.R., Zavala, M.A., Marañón, T., 2008. Human and non human determinants of forest composition in southern Spain: evidence of shifts towards cork oak dominance as a result of management over the past century. Journal of Biogeography 35, 1688-1700. van der Heijden, M.G.A., Bardgett, R.D., Van Straalen, N.M., 2008. The unseen majority: soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems. Ecology Letters 11, 296-310. Vargas, R., Detto, M., Baldocchi Dennis, D., Allen Michael, F., 2010. Multiscale analysis of temporal variability of soil CO2 production as influenced by weather and vegetation. Global Change Biology 16, 1589-1605. Vayreda, J., Martínez-Vilalta, J., Vilà-Cabrera, A., 2016. El Inventario Ecológico y Forestal de Cataluña: una herramienta para la ecología funcional. Revista Ecosistemas 25, 70-79. Fig. captions Fig. 1. Evolution of soil water content (SWC), soil temperature (Tsoil) and soil respiration (Rs) as a function of time and trees' health status (living or dead). Error bars represent the standard error of the mean. Fig. 2. Changes in soil abiotic (microclimate, pH) and biotic (fine root demography, Rs, RH) variables, as well as soil total organic carbon (TOC) and total organic nitrogen (TON) values evaluated under dead trees relative to the values evaluated under living trees. Therefore, positive values (right hand side of the vertical bar) represent an increase in that particular variable under dead with respect to under living trees. Error bars represent standard error of the mean. Asterisk represent significant from zero differences (p value > 0.05; t-test). Fig. 3. Linear mixed-effects models for soil respiration (Rs). Solid lines represent modeled Rs responses under dead trees, whereas dotted lines represent modeled Rs responses under living trees. Where, Ts = soil temperature at 5 cm depth; PC2 = second dimension of the PCA, here representing SOM. Fig. 4. Linear mixed-effects models for heterotrophic respiration (RH). Solid lines represent modeled Rs responses under dead trees, whereas dotted lines represent modeled Rs responses under living trees. SRL = specific root length; PC1= first dimension of the PCA, here representing soil nutrients; PC2 = second dimension of the PCA, here representing SOM. Fig. 5. Multigroup SEM representation. Path diagrams representing hypothesized causal relationships between aboveground vegetation, biotic and abiotic variables, soil respiration (Rs) and soil heterotrophic activity (RH) under living (a) and dead (b) conifer trees. Arrows depict causal relationships: positive and negative effects are indicated by solid and dashed lines respectively, with numbers indicating standardized estimated regression weights (SRW). Arrow widths are proportional to the significance values according to the legend. Paths with non-significant coefficients are represented in gray. Coefficients in bold characters represent those causal relationships where the strength of the relation differed between soils under living (green) and under dead (orange) trees. 2= 86.81, NFI= 0.83 y RMSEA= <0.0001, df= 122, p= 0.99 Fig.1 Click here to download high resolution image Fig.2 Click here to download high resolution image Fig.3 Click here to download high resolution image Fig.4 Click here to download high resolution image Fig.5 Click here to download high resolution image