Field Exclusion of Large Soil Predators Impacts Lower Trophic Levels, And

DR NEREIDA MELGUIZO-RUIZ (Orcid ID : 0000-0001-5153-5281) DR AMY T AUSTIN (Orcid ID : 0000-0002-7468-5861)

Article type : Research Article

Handling Editor: Dr Natalie Clay

Field exclusion of large soil predators impacts lower trophic levels, and decreases leaf-litter decomposition in dry forests Nereida Melguizo-Ruiz 1,2,3, Gerardo Jiménez-Navarro 1,3, Eva De Mas 1, Joaquina Pato2, Stefan Scheu 4,5, Amy T. Austin 6, David H. Wise 7 and Jordi Moya-Laraño 1

1 Estación Experimental de Zonas Áridas, Functional and Evolutionary Ecology, Consejo Superior de Investigaciones Científicas (CSIC), La Cañada de San Urbano, Almería, Spain.

2 Research Unit of Biodiversity (UO/CSIC/PA), Oviedo University, Mieres, Spain.

3 CIBIO/InBio Research Center in Biodiversity and Genetic Resources, 7000-651, Évora, Portugal.

4 J.F. Blumenbach Institute of Zoology and Anthropology, Animal Ecology, University of Göttingen, Göttingen, Germany.

5 Centre of Biodiversity and Sustainable Land Use, University of Göttingen, Göttingen, Germany.

6 Instituto de Investigaciones Fisiológicas y Ecológicas Vinculadas a la Agricultura (IFEVA) and Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Facultad de Agronomía, Universidad de Buenos Aires, Buenos Aires, Argentina.

This article has been accepted for publication and undergone full peer review but has not been through the copyediting,Accepted Article typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/1365-2656.13101

Corresponding author: Nereida Melguizo-Ruiz, CIBIO-InBIO, Research Center in Biodiversity and Genetic Resources, Casa Cordovil, Rua Dom Augusto Eduardo Nunes 7000-651, Evora, Portugal. E- mail: [email protected]. Accepted Article

1. Shifts in densities of apex predators may indirectly affect fundamental ecosystem processes, such as decomposition, by altering patterns of cascading effects propagating through lower trophic levels. These top-down effects may interact with anthropogenic impacts, such as climate change, in largely unknown ways. 2. We investigated how changes in densities of large predatory arthropods in forest leaf-litter communities altered lower trophic levels and litter decomposition. We conducted our experiment in soil communities that had experienced different levels of long-term average precipitation. We hypothesized that altering abundances of apex predators would have stronger effects on soil communities inhabiting dry forests, due to lower secondary productivity and greater resource overexploitation by lower trophic levels compared to wet forests. 3. We experimentally manipulated abundances of the largest arthropod predators (apex predators) in field mesocosms replicated in the leaf-litter community of Iberian beech forests that differed in long-term mean annual precipitation by 25% (three dry forests with MAP < 1250 mm and four wet forests with MAP > 1400 mm). After one year, we assessed abundances of soil fauna in lower trophic levels and indirect impacts on leaf-litter decomposition using litter of understory hazel, Corylus avellana. 4. Reducing densities of large predators had a consistently negative effect on final abundances of the different trophic groups and several taxa within each group. Moreover, large predatory arthropods strongly impacted litter decomposition, and their effect interacted with the long-term annual rainfall experienced by the soil community. In the dry forests, a 50% reduction in the densities of apex predators was associated with a 50% reduction in decomposition. In wet forests, the same reduction in densities of apex soil predators did not alter the rate of litter decomposition. 5. Our results suggest that predators may facilitate lower trophic levels by indirectly reducing competition and resource overexploitation, cascading effects that may be more pronounced in drier forests where conditions have selected for greater competitive ability and more rapid resource utilization. These findings thus provide insights into the functioning of soil invertebrate communities and their role in decomposition, as well as potential consequences of soil community responses to climate change. Accepted Article

This article is protected by copyright. All rights reserved KEYWORDS Apex predators; body size; climate change; field mesocosms; leaf-litter decomposition; precipitation; soil food webs; top-down control Accepted Article

This article is protected by copyright. All rights reserved Changes in the densities of large apex consumers (top predators) may strongly affect ecosystem structure and dynamics (Duffy, 2003; Schmitz, Hawlena, & Trussell, 2010). These effects may interact with other anthropogenic impacts such as climate change, alterations in land use, and habitat loss; therefore, understanding the role of top predators in food webs and their interactions with environmental drivers may be central to predicting the consequences of biodiversity loss in ecosystem functioning.

Trophic cascades, which occur when consumer impacts propagate through lower trophic levels, may lead to strong ecosystem-level effects (see Pace et al., 1999 for a review). Predators can indirectly affect ecosystem processes such as primary productivity (Schmitz, Hambäck, & Beckerman, 2000; Halaj & Wise, 2001) and litter decomposition (Kajak, 1995; Lawrence & Wise, 2004; Koltz, Classen, & Wright, 2018). These indirect effects of predators can be mediated by changes in population densities of lower trophic levels, i.e. density-mediated indirect interactions (DMII); or by changes in specific behaviours of lower levels, i.e. behavioural or trait-mediated indirect interactions (TMII) (see Werner & Peacor 2003 for a review). The pattern of results to date is complex, as the sign of the predator-induced cascades can vary from negative to neutral to positive, suggesting that factors such as abiotic conditions may strongly influence the nature of the trophic cascade (Chase, 1996; Lawrence & Wise, 2004; Lensing & Wise, 2006; Wu et al., 2011; Kardol et al., 2016; McCluney & Sabo, 2016).

Furthermore, not all effects may propagate to lower trophic levels or affect ecosystem processes in all systems. There are compensatory mechanisms that can dampen, minimize and even suppress cascades, such as trophic-level omnivory or the regulation of other trophic levels not predicted by cascading trophic interactions (Pace et al., 1999). In particular, intraguild predation (Polis et al., 1989) can complicate the influence of top-down patterns on lower trophic levels (Finke & Denno, 2005), since generalist predators have the capacity not only to directly reduce prey densities, but also to indirectly increase the abundance of lower trophic levels by preying on other natural enemies (e.g. Snyder & Wise 2001). Accepted Article

This article is protected by copyright. All rights reserved Manipulative field experiments to uncover the nature of trophic cascades have removed or added one or more taxa (e.g. Lawrence & Wise, 2004), or have manipulated abundances of predators with different functional attributes or foraging strategies (such as web-building and hunting spiders and ants, Sanders et al., 2011). To our knowledge no study has altered the density of an entire size- defined trophic functional group. Because predators are usually larger than their prey (Schneider & Brose, 2013), the top (apex) predators are often the largest in the system (Woodward et al., 2005). Thus, manipulating abundances of the largest predators from a natural food web could reveal the cascading indirect effects of such apex predators on decomposition rates, through changes in abundances of lower trophic levels. As apex predators are more sensitive to climate change, they tend to be the first organisms lost in food webs. This is probably due to their relatively higher metabolic requirements and smaller population sizes; or owing to bottom–up effects, i.e. top predators may be the most sensitive to the lowered productivity in the basal trophic level (Petchey et al., 1999; Voigt et al., 2003; Gilman et al., 2010). Therefore, it becomes urgent to understand how changes in densities of apex predators could interact with climate to impact other trophic levels and associated ecosystem processes. This is particularly important in soil ecosystems and their upper layer, the leaf-litter food web, because we know little about how biotic interactions from top predators to basal resources (Fig. 1) interact with abiotic factors to determine the rate of leaf-litter decomposition (but see Lensing & Wise, 2006).

Climate-change predictions focus on both increases in global temperature and changes in rainfall patterns (Stocker et al., 2013), with an expected decrease in precipitation in several parts of the earth. In soil ecosystems, water availability resulting from short-term precipitation patterns can indirectly impact how trophic cascades affect decomposition. Drier soil conditions during droughts appear to strongly affect fungal growth (A'Bear, Jones, & Boddy, 2014), which in turn may alter how trophic cascades affect rates of litter decomposition through the activity of fungivores (Lensing & Wise, 2006). For instance, intermediate levels of fungal grazing by springtails can stimulate fungal growth and promote decomposition, while overgrazing by springtails, in the absence of predators, may lessen fungal populations, resulting in reduced rates of decomposition (Lawrence & Wise, 2004), with effects being more pronounced in drier soils. Moreover, soil moisture may influence invertebrate activity and abundance, increasing predator-prey encounters as fauna aggregate in water-rich areas Accepted Article

This article is protected by copyright. All rights reserved (Melguizo-Ruiz et al., 2012; Verdeny-Vilalta, 2013; Verdeny-Vilalta & Moya-Laraño, 2014; Melguizo-Ruiz, Jiménez-Navarro, & Moya-Laraño, 2016) and predators seek prey as sources of water (McCluney & Sabo, 2009). Thus, short-term changes in water availability can alter the direction of trophic cascades. In addition, terrestrial ecosystems that vary in long-term precipitation patterns may exhibit distinct microbial, plant and animal communities, well-adapted to the average climatic conditions, and therefore may respond differently to changes in predator densities and resulting trophic cascades.

To examine the joint cascading effects of altered densities of apex predators and long-term precipitation patterns on lower trophic levels and leaf-litter decomposition, we experimentally reduced, and also increased, abundances of the largest arthropod predators from the litter layer in field mesocosms located in three dry (Mean Annual Precipitation, MAP < 1250 mm) and four wet (MAP > 1400 mm) Iberian beech forests (Fagus sylvatica L.). These rainfall estimates were obtained from rainfall time series of at least 20 years (Ninyerola, Roure & Pons, 2005), which suggest that these forests have differed in rainfall patterns for at least several seasons. Performing field experiments in contrasting climatic regions combines the strengths of correlational and experimental approaches while mitigating their respective limitations (Stewart et al., 2013).

Short and long-term water availability is linked to primary productivity (Loustau, Hungate, & Drake, 2001), rates of fungal growth (A'Bear, Jones, & Boddy, 2014), predator-prey interactions (Verdeny-Vilalta & Moya-Laraño, 2014; McCluney & Sabo, 2016) and the composition of plant and animal communities. In addition, higher MAP is linked to higher secondary productivity across Iberian beech forests (Melguizo-Ruiz et al., 2012). In dry forests, equilibrium densities may be lower and populations may have experienced lower resource availability for many generations than in wet forests; if so, a sudden reduction of large predators might lead to relatively faster resource overexploitation from lower trophic levels and, consequently, to slower decomposition rates in the dry forests. This would occur because resource scarcity would prompt the evolution of higher competitive abilities in dry forests (Carroll et al., 2007; Rodrigues et al., 2016), which could entail faster resource depletion when populations of lower trophic levels are released from top predators. Hence, we predicted that i) reduced numbers of large predators will cause changes in densities in lower trophic levels (through both DMII and TMII, see Werner & Peacor, 2003), producing effects that cascade Accepted Article

This article is protected by copyright. All rights reserved down to leaf-litter decomposition; ii) the cascading effects of removing predators will have a stronger impact on decomposition in dry than in wet forests. This will occur because in dry forests top-down control will more strongly affect populations at lower trophic levels as these populations will more readily overexploit their resources. An implicit prediction stemming from the latter is that fauna should be more abundant in wet forests. Finally, we expected that iii) increased abundance of large predators should be attenuated in the mid-term, owing to density-dependent compensatory mechanisms such as intraguild predation, which should result in weakened differences between increased and control densities of large predators in terms of effects on lower trophic levels and decomposition rates. Alternatively, higher abundances of large predators could result in overexploitation of lower trophic levels and to a decrease of decomposition rates.

ALS AND METHODS

2.1 | Study area

We selected eight beech forests (F. sylvatica) across the Cantabrian Mountains in northwest Spain that differed in mean annual precipitation -MAP (see Appendix S1 and Appendix S2: Table S1 for details). MAP ranged between 973 and 1232 mm/year at four of the sites -'dry forests'-, and between 1432 and 1510 mm/year at the other four sites -'wet forests' (Ninyerola, Pons, & Roure, 2005).

2.2 | Experimental design, faunal manipulations and sampling

From May to July 2012, ten experimental plots arranged across two blocks (ca. 100 m apart from each other) were established in each of the eight forests. Each block consisted of five 50 x 50-cm plots that were fenced with metal flashing buried 10 cm into the soil to minimize horizontal migrations of animals. To prevent macrofauna from recolonizing or leaving the plots after our manipulations, a 1.2 mm-mesh fiberglass net was located on top and another at the bottom (see Appendix S1 and Appendix S3: Fig. S1). In total 80 experimental plots were established across the eight beech forests. Due to unanticipated events during the experiment, including some replicates being destroyed by bears, we ended up with 62 plots from seven forests for the analysis of final responses (see Appendix S1 for details). The average MAP for the three dry forests was 1095 mm compared with 1472 mm for Accepted Article

This article is protected by copyright. All rights reserved the four wet forests (extracted from the digital atlas of the Iberian Peninsula (Ninyerola, Pons, & Roure, 2005), with average precipitation values based on at least twenty years measured since 1950). Thus, dry forests were ~ 25% drier than wet forests, a difference that is slightly above the predictions of rainfall reduction for the end of the 21st century in southern Europe (Stocker et al., 2013). Even though it did rain substantially more during the year of our experiment, these differences were maintained, as between 2012 and 2013 MAP for the three dry forests was 1333 mm, compared with 1843 mm for the four wet forests (values obtained by applying an interpolation algorithm (Ninyerola, Pons, & Roure, 2007) to data from weather stations across the Iberian Peninsula).

We conducted a factorial randomized-block experiment. Within each block, two plots were randomly assigned to a 'Predator removal' treatment (0X), where all large predators were removed (see details below); another two plots were assigned to a 'Predator addition' treatment (2X), where the density of large predators was doubled, and one 'Control' plot (1X), which included natural densities of large predators (see Appendix S1 for details on predator manipulations). Predator manipulations based on natural densities present at each site are more realistic than imposing an arbitrary set of different densities across all sites, since each site may have a characteristic equilibrium density of top predators. Thus, the “Predator addition” treatment can be interpreted as a doubling of the equilibrium density reflective of local conditions at the time of the manipulation.

Before installing the fenced plots, we collected five 28-L leaf-litter samples from each block (by filling, with fallen beech leaves, identical plastic containers of 28-L volume, one from each of the five locations where an experimental plot would be established. These samples were the basic units from which we sorted and counted, in the laboratory, all the meso- and macrofauna. To minimize damage to the animals, we did not sift the litter, but instead carefully scattered all litter in a tray and closely examined every leaf to separate the large predators from the rest of the fauna. All specimens were identified to the lowest taxonomic level possible (Appendix S1), and their length was measured to the nearest 0.5 mm. Large (apex) predators were kept in separate vials until their reassignment to the experimental treatments, while the leaf litter and remaining animals were released into their respective 28-L containers. Within each block all the large predators were randomly assigned to treatments and released into the appropriate container, such that 2/5 were placed in each of the two 'Predator addition' containers (double density – 2X), 1/5 in the control container (natural density – Accepted Article

This article is protected by copyright. All rights reserved 1X), and none in the two 'Predator removal' containers (0X). Thus, 2X plots had two times more apex predators than 1X plots, being the natural densities block-specific. The litter and fauna were then introduced into the corresponding fenced plots.

In autumn of 2012, the experimental manipulations were reinforced by removing all the large predators found in the 'Predator removal' plots and randomly transferring them to each of the two 'Predator addition' plots (Appendix S1). At the end of the experiment (April to June 2013), all animals found in the plots were counted, measured and identified to the same taxonomic level as when setting up the experiment.

2.3 | Assignment of the fauna to size-based trophic groups

Individuals were assigned to five size-based trophic groups (Fig. 1) as follows:

—Large predators: web-building and wandering spiders, lithobiomorph and geophilomorph centipedes, carabid and staphylinid beetles, and predatory larvae larger than the manipulation criteria (Appendix S1).

—Intermediate predators: these were the same taxa as the manipulated predators but with a size below the threshold. Most were likely offspring of the manipulated predators.

—Small predators: predators, other than offspring of those manipulated, that were smaller than the manipulation threshold even at adult stages. This group included pseudoscorpions, small opilionids, predatory mites (Prostigmata and Mesostigmata) and campodeid Diplura.

—Microbivores and fungivores: springtails (Collembola), enchytraeid worms, symphylids, polyxenid millipedes and oribatid mites —these groups mostly graze on bacteria and fungi.

—Detritivores: millipedes (julids, glomerids and polydesmids), isopods and earthworms.

During the experiment 36,079 invertebrates were counted and measured. Due to loss of replicates, 34,744 individuals were included in the analyses: 16,289 in 2012 and 18,455 in 2013 (Appendix S2: Table S2).

2.4 | Decomposition rates Accepted Article

This article is protected by copyright. All rights reserved Three 10 x 10-cm litterbags each containing 1 g of hazel (Corylus avellana) leaf litter were placed in each plot (see details in Appendix S1). We used hazel leaves instead of beech leaves, as hazel trees occur in, or near, all our forests as understory trees, and its leaves decompose at a much faster rate than beech leaves (Sanpera-Calbet, Lecerf, & Chauvet, 2009), thus ensuring sufficient decomposition during the course of the experiment to enable assessment of the experimental manipulations on decomposition rate.

One bag was collected after 6 months, and after 12 months the remaining two bags were collected to increase the precision of the estimate. The bags were dried at 60 ºC for 24 h and the remaining litter weighed to the nearest 0.01 g.

We estimated the decomposition constant (k) for each plot by first fitting the following equation (Olson, 1963) through the three collected bags:

–kt Mt = M0 e with Mt the mass of litter at time t, M0 the initial mass of litter, k the decomposition constant and t the amount of time elapsed since the initial measurement. We fitted a least-squares line to the model [– log (Mt / M0) = kt] and then used the fitted k to characterize the rate of decomposition in each plot.

2.5 | Statistical analyses

All analyses were performed using the statistical software R 3.4.4 (R Core Team, 2018).

Abundances of trophic and taxonomic groups were analyzed with generalized linear mixed- effects models (GLMM) using the "glmer.nb" function in the "lme4" package of R (Bates et al., 2015) and a negative binomial distribution. We also tested GLMM models with a Poisson distribution, but the overdispersion was higher than that for the negative binomial model. 'Rainfall' (dry vs. wet), 'Treatment' (0X, 1X, 2X) and the logarithm of the 'Initial abundance' (faunal abundances in 2012; continuous variable to control for initial conditions) were included as fixed factors, while 'Site' and 'Block' were added as random factors to control for pseudoreplication. We systematically tested two models, one additive and one including the Rainfall x Treatment interaction, to examine whether effects of manipulating densities of the top predators differed between rainfall levels. Since the Accepted Article

This article is protected by copyright. All rights reserved “glmm.nb” function did not converge, we forced the “nAGQ” option to 0, which usually provides a more liberal convergence criterion. This procedure was unlikely to bias our results, as a sensitivity analysis using the R functions glmmadmb(), MCMCglmm() and glmmTMB() always provided consistent outputs (not shown). Due to the statistical constraints imposed by small numbers, we only analyzed the responses of the most abundant taxonomic groups of predators and prey (Moya-Laraño & Wise, 2007). When testing multiple taxonomic or trophic groups for statistical significance (defined as p < 0.05), we accounted for increases in type I error rates when performing multiple tests by applying the false discovery rate (FDR) adjustment to correct alpha levels (Benjamini & Hochberg, 1995). Post-hoc comparisons to check for statistically significant effects between treatments were performed using Tukey tests in the "glht" function within the "multcomp" package (Hothorn, Bretz, & Westfall, 2008). To maintain power relatively high but keep Type I errors low, such post-hoc comparisons were performed for each trophic group but not for taxa within each group. All model effects were tested with a type-III Likelihood Ratio Test using the function "Anova" within the "car" package (Fox & Weisberg, 2011).

Decomposition rates were analyzed with a mixed-effects linear model ("lmer" function of the "lme4" package) with a square-root transformation. The square root of the k value for each plot was the level of replication.

Predicted effects and 95% confidence intervals of statistically significant (p < 0.05) variables in models were plotted using the R library "effects" (Fox 2003). Raw data and the R code used to analyse the abundance of one of the trophic groups are available in Supporting Information.

3 | RESULTS

Reducing densities of large predators had a consistently negative effect on final abundances of the different trophic groups and several taxa within each group. Further, large predatory arthropods strongly impacted litter decay, but only in the dry forests. Reducing densities of large predators decreased decomposition rate by 50% in the dry sites. Finally, we found evidence that wet forests may be more productive, as both initial and final abundances of all functional groups tended to be higher in wet than dry forests (Appendix S2: Tables S2); furthermore, several taxonomic groups (predatory Accepted Article

This article is protected by copyright. All rights reserved larvae, predatory mites, symphylids and earthworms) were clearly more abundant in wet relative to dry forests (p’s < 0.0095, 0.0472, 0.0004, 0.005, respectively; Appendix S2: Table S3a).

3.1 | Effectiveness of the manipulation of large predators

Our experimental manipulations established a clear difference in the densities of apex predators (Fig. 2a; χ² = 55.7, d.f. = 2, p < 0.0001). At the end of the experiment, large predators were ca. two times more abundant in the predator-addition plots (2X and 1X) than in the predator-removal plots (Tukey contrasts, p < 0.0001; plot of effect sizes with 95% CIs in Fig. 2a). No Rainfall x Treatment interaction was detected. Densities in 2X and 1X treatments tended to converge by the end of the experiment (Tukey contrasts, p = 0.09; plot of 95% CI for 2X/1X in Fig. 2a). The following groups of apex predators differed among treatments (Appendix S2: Table S3a): large web-building spiders, wandering spiders, lithobiomorph centipedes, carabid beetles and predatory larvae.

3.2 | Faunal responses to manipulations of large predators

Intermediate predators: These predators likely were the offspring of the manipulated apex predators. Abundances differed across treatments (Fig. 2b; χ² = 14.3, d.f. = 2, p = 0.0008), and there was no Rainfall x Treatment interaction. The pattern roughly mirrored that of the large predators, with increasing abundances of intermediate predators in 1X and 2X plots, although proportional effect sizes were slightly smaller (comparison of plots of effect sizes and 95% CIs in Figs. 2a and 2b). Within the intermediate predators, the web-building spiders appear to drive the trend, as only the abundance of this taxon clearly differed across treatments (Appendix S2: Table S3a).

Small predators: This was the only trophic group that responded differently in wet and dry forests (Fig. 2c; Rainfall x Treatment χ² = 10.6, d.f. = 2, p = 0.005). Manipulating apex predators had a much stronger impact on small predators in the dry forests, where they were twice as abundant in both 2X and 1X treatments as in the 0X treatment. In contrast, in the wet forests, abundances of the 2X and 1X treatments were more similar to the 0X treatment (Fig. 2c). Within this group only campodeid Diplura showed a clear interaction with MAP, with a negative effect of large-predator removal on Diplura in dry sites (Appendix S2: Table S3b). Abundances of pseudoscorpionids, Accepted Article

This article is protected by copyright. All rights reserved opilionids and predatory mites were also negatively affected by large-predator removal, although effects were similar for wet and dry forests (Appendix S2: Table S3a).

Microbivores and fungivores: This trophic group was ca. 1.5x more abundant in both 2X and 1X treatments than in 0X plots (Fig. 2d; χ² = 31.4, d.f. = 2, p < 0.0001), and there was no Rainfall x Treatment interaction. Within this functional category, springtails, oribatid mites and polyxenid millipedes responded positively to higher densities of apex predators (Appendix S2: Table S3a). Detritivores: Abundances of detritivores also responded positively to increased densities of apex predators (Fig. 2e; χ² = 15.3, d.f. = 2, p = 0.0005), but only to the 2X treatment, where densities were ca. 1.5x higher than in 0X plots (Tukey contrasts, p < 0.001; Fig. 2e). Numbers of julid millipedes and isopods displayed the clearest responses (Appendix S2: Table S3a). There was no Rainfall x Treatment interaction.

3.3 | Ecosystem-level impact: Cascading effects of large predators on litter decomposition

Lower abundances of apex predators decelerated litter decomposition in dry forests, but had no impact in the wet sites (Fig. 3; Rainfall x Treatment χ²= 6.3, d.f. = 2, p = 0.043; Appendix S2: Table S4). In the dry forests, decomposition rate was at least 1.5x higher in the two predator-addition treatments than in the 0X plots, while there was no indication of an effect in the wet forests (post-hoc tests and overlap of proportional effect-size CI’s with effect = 1; Fig. 3).

Our experiment, replicated in dry and wet beech forests, was designed to test the hypothesis that the amount of rainfall experienced by the litter arthropod community over the long-term would alter the indirect impact of apex predators on decomposition through changes in lower trophic levels. We postulated that in dry forests, resource overexploitation from lower trophic levels would be more likely, and thus the cascading effects of removing predators could have a stronger impact on decomposition. The results support this hypothesis, as the cascade initiated by removing apex predators altered decomposition only in the dry forests. A 50% reduction in the densities of apex Accepted Article

This article is protected by copyright. All rights reserved predators in the dry forests caused a 50% reduction in the rate of litter decomposition; in contrast, decreased densities of apex predators did not impact decomposition in the wet forests. Below we discuss the implications of this finding, but first we address an initially perplexing aspect of the data: decreased densities of large predators had a consistently negative effect on the abundances of lower trophic levels within the food web; thus, such changes in abundances did not always alternate as expected from “classical” theories of trophic cascades (i.e. DMII, Werner & Peacor, 2003). Instead, at the end of the experiment all faunal groups were more abundant in the plots with higher densities of apex predators.

acts of apex predators on lower trophic levels Our manipulation of densities of large (apex) predators was successful, as their abundances were lower in 0X than 1X and 2X plots at the end of the experiment. On the other hand, the tendency for densities of apex predators to have converged by the end of the experiment (Fig. 2a) suggests that increasing their densities in the 2X treatment may have led to density-dependent effects as communities converged on their carrying capacity, likely from increased rates of intraguild predation and cannibalism. The fact that overexploitation of resources on lower trophic levels did not occur after doubling predator densities (i.e. numbers of small and intermediate predators did not change in 2X compared to 1X plots), suggests that resource competition among top predators was not severe, with predation and cannibalism being stronger. We observed clearly cascading effects of altered predator densities from the upper to the lower levels of the food web, which integrate the net outcomes of all direct and indirect biotic interactions during the experiment. The removal of large predators caused abundances of all the other trophic groups to decrease. Most responses are readily explainable. Differences between treatments in densities of intermediate predators, which are young stages of apex predators, reflected numbers of apex predators; however, the density differences were proportionally less (Fig. 2a vs. Fig. 2b), probably because of lower intraguild predation in the 0X and 1X plots compared to the 2X treatment. Meanwhile, abundances of the lowest trophic levels (microbivores, fungivores and detritivores) increased in response to higher densities of the larger predators (adult apex predators and their offspring), in agreement with a density-mediated indirect interaction, which would be due to reduced Accepted Article

This article is protected by copyright. All rights reserved predation pressure by lower-level predators on their prey. Among decomposers, smaller animals, such as springtails, oribatid mites and polyxenids, appear to have been most strongly affected by lower densities of large predators; densities of larger decomposers, such as enchytraeids and earthworms, did not respond to the apex predator manipulation. As small, lower-trophic level predators may primarily feed on smaller decomposers (rather than on large macroinvertebrates), we expected that the small predators would have been less, not more, abundant in plots with higher numbers of apex predators. This initially perplexing pattern has at least two possible non-mutually exclusive and parsimonious explanations, which are fully detailed in Fig. 4. First, non-consumptive effects (Werner & Peacor, 2003; Schmitz, Krivan, & Ovadia, 2004; Madji et al., 2014) likely contributed to part of the trophic cascade pattern in this system. For example, small predators may have increased their foraging activity when apex predators were removed, leading to decreased densities of their prey (e.g. Gordon et al., 2015). Secondly, we sampled faunal densities (other than apex predators) only at the end of the experiment, which may have caused us to miss density-dependent indirect interactions (DDII) affecting small predators during the course of the study. In other words, our sampling schedule likely failed to fully uncover the dynamics of the system. An early increase in numbers of small predators in the 0X plots may have initially led to decreased abundances of fungivores and detritivores, leading to a subsequent decline in the final numbers of small predators in the 0X treatment (Fig. 4). As predicted, in dry forests, this effect was likely stronger, causing resource overexploitation on fungivores and detritivores that later lead to the detected lower densities of their (small) predators. Without further research, it is not possible to disentangle all potential interactions involving the intermediate trophic levels; nevertheless, it is clear that large predators appear to facilitate higher densities of lower trophic levels, as reducing numbers of the apex predators caused densities of all lower functional trophic groups to decline, which then depressed decomposition rate – but solely in dry forests.

teractive effects of precipitation and densities of apex predators Our results suggest that beyond the direct effects of mean annual precipitation (MAP) affecting water availability and soil moisture, long-term patterns of rainfall can influence the indirect effect of large predators on decomposition via various mechanisms, including potential differences in the Accepted Article

This article is protected by copyright. All rights reserved composition of soil communities as well as local adaptation of soil fauna to specific biotic and abiotic conditions. The interactive effect on decomposition of manipulating the density of apex predators in dry and wet forests did not exactly parallel the patterns found for decomposers, probably because such patterns resulted from multiple intricate interactions that occurred among soil organisms during the one-year experiment. Field experiments conducted to date reveal that the strength of the trophic cascade affecting decomposition is variable and complex, and sometimes is related to differences between sites in precipitation and/or soil moisture. Lawrence and Wise (2000) discovered that natural densities of spiders retarded decomposition during a year experiencing normal rainfall in a deciduous forest in Kentucky, eastern USA. In contrast, spiders enhanced decomposition during a drought year, possibly, they speculated, by preventing Collembola and other fungivores from overgrazing drought- stressed fungi (Lawrence & Wise, 2004). In the same forest a few years later, spiders accelerated litter decomposition under experimentally imposed low rainfall, but not under high rainfall, and only on the site with moister soil (Lensing & Wise, 2006). In these last two studies, when drought appeared to have been a factor, lowered decomposition rates likely were caused by fungivore densities increasing in response to lowered predation by spiders. In contrast, other experiments have shown that arthropod predators, including spiders, can retard decomposition by limiting densities of fungivores (Kajak & Jakubczyk, 1977; Kajak, 1997; Lawrence & Wise, 2000; Liu et al., 2014). This mixture of negative and positive effects of predators on decomposition rates can be attributed to differences in moisture stress, length of the experiment, and differences in microbial communities and composition of the fauna between dry and wet forests, as well as differences in how predators were manipulated. In this study, we examined how changes in densities of large apex consumers could indirectly affect litter decay in dry- and wet-adapted soil communities. Because we compared top-down effects between forests that differed in long-term MAP by 25%, a difference that is close to predicted changes in rainfall for the end of the century (Stocker et al., 2013), our findings are directly relevant to understanding the implications of climate change for both trophic interactions and indirect effects on decomposition. Our study design contrasts with that of research in which rainfall was experimentally manipulated over much shorter time periods (e.g. Lensing & Wise, 2006; Wise & Lensing, 2019; 2 and 3 growing seasons, respectively). Accepted Article

This article is protected by copyright. All rights reserved We suggest that long-term differences in secondary productivity between dry and wet forests explain our finding that apex predators positively impacted decomposition only in dry forests. Terrestrial productivity is positively correlated with rainfall (Loustau, Hungate, & Drake, 2001), which in a system such as ours would be expressed as more litterfall on the forest floor (Jackson et al., 2001), as well as higher soil moisture and fungal growth in the litter. Higher amounts of litterfall, together with higher moisture and more fungi should translate into higher faunal biomass. Consistent with higher secondary productivity in wetter forests, we found higher densities of several groups (predatory larvae, predatory mites, symphylids and earthworms) in the wet forests we studied, a pattern found in other studies (Lindberg, Bengtsson, & Persson, 2002; Melguizo-Ruiz et al., 2012). We argue that in drier forests, small and intermediate predators may have evolved higher competitive abilities and, as a consequence, may have over-exploited fungivores and detritivores after experimental release from apex predators, leading to a decrease of fauna across trophic levels and a subsequent decrease in the rate of decomposition. Thus, in the dry forests, the indirect effects of apex predators on lower-level predators translated into greater impacts on the lowest levels, with strong consequences on the rate at which the leaf-litter resource decomposed. Studies similar to ours have conceptually treated predators as a single trophic level, making the manipulated system an abstract chain of four trophic levels: predators, decomposers, fungi and litter. Ours is an exception because we specifically targeted apex predators in the food web. Separating predators into size categories has rarely been done in mesocosm experiments (Schneider, Scheu, & Brose, 2012). In our case, instead of lumping all predators together by taxon, we selectively manipulated densities of the larger individuals across all predatory taxa, leaving densities of smaller predators undisturbed at the onset of the experiment. This approach allowed us to test the impact of the highest predatory trophic level on lower levels, including lower-level predators in addition to their prey.

Effects of climate change on decomposition could be multiplicative, as impacts of the soil fauna on decomposition, which can be substantial, depend on climate (Wall et al., 2008; García- Palacios et al., 2013). In addition, larger-bodied animals – i.e. apex predators and higher trophic levels in general – face the highest extinction risk from global change (Duffy, 2003; Voigt et al., 2003; Lotze et al., 2006; Byrnes, Reynolds, & Stachowicz, 2007), especially under climatic disturbances Accepted Article

This article is protected by copyright. All rights reserved (Gilman et al., 2010). Therefore, changes in long-term precipitation patterns due to climate change may alter the cascading effects of apex predators both by reducing their densities and by changing the impact of the altered cascade on decomposition.

Researchers have stressed the importance of large predators in structuring ecological communities (Terborgh et al., 2001; Duffy, 2003; Estes et al., 2011), as well as cascading effects of their disappearance in marine, terrestrial and freshwater ecosystems (Estes et al., 2011). However, few experimental studies have documented the effects of adding or removing top predators on ecosystem processes (Ford & Goheen, 2015). Experimental manipulation of large vertebrates in natural ecosystems is difficult because of their long generation times and expansive spatial ranges. Our field experiment clearly reveals that large predatory arthropods have strong indirect effects on lower trophic levels and leaf-litter decomposition, and that these cascading effects vary between dry and wet beech forests, with direct implications for ecosystem responses to future climate change. Indeed, dry- adapted communities may be more severely affected by the loss of apex predators. To disentangle the mechanisms behind the observed complex top-down effects on lower trophic levels and litter decay, future studies should incorporate monitoring of the short-term dynamics of the soil system through intermediate sampling). In addition, directly manipulating water availability in dry and wet forests would distinguish effects caused by local adaptation to low productivity from those resulting from sudden, short-term changes in precipitation.

ACKNOWLEDGEMENTS

We acknowledge J. M. Herrera and D. Ruiz Lupión for helping in the field, E. Palop and M. Morán for giving a hand whenever it was needed, and all the people of the Research Unit of Biodiversity (UO/CSIC/PA), that made our work easier. B. Bolker and J. Fox provided important insights on statistical analysis and M. Ninyerola extracted MAP values for the year of the study. M. A. Arnedo suggested using C. avellana leaves for decomposition. We also want to thank E. Montoya, A. González and the forest rangers of the Asturian environmental administration for administrative and logistic help. Finally we are grateful to two anonymous reviewers for their helpful comments and suggestions which have greatly improved the manuscript. This study has been funded by Spanish Accepted Article

This article is protected by copyright. All rights reserved Ministry of Science and Innovation grants CGL2010-18602, CGL2015-66192-R and Andalusian grant P12-RNM-1521-EEZA to J.M.L.; the European Regional Development Fund, Agencia Nacional de la Promoción de Ciencia y Tecnología (PICT 2016-1780), Argentina. to A.T.A.; and FPI

fellowship (BES-2011-043505) to N.M.R. This work has been conducted under permit 2011/059163 of the Asturias Government.

AUTHORS’ CON

NMR, DHW, ATA, SS and JML devised the study, NMR, GJN, EDM, JP and JML performed the field experiment and collected the data, NMR, JML and DHW performed statistical analyses, NMR designed the drawings, NMR and GJN constructed the tables and figures, NMR, DHW and JML wrote the first draft of the manuscript. All authors contributed substantially to revisions, and gave final approval for publication.

DATA A Data are available at https://digital.csic.es/handle/10261/183996

REFERENCES

A' Bear, A. D., Jones, T. H., & Boddy, L. (2014). Potential impacts of climate change on interactions among saprotrophic cord-forming fungal mycelia and grazing soil invertebrates. Fungal Ecology, 10, 34-43. https://doi.org/10.1016/j.funeco.2013.01.009

Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1-48. https://doi.org/10.18637/jss.v067.i01

Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B (Methodological), 57(1), 289–300. Accepted Article

This article is protected by copyright. All rights reserved Byrnes, J. E., Reynolds, P. L., & Stachowicz, J. J. (2007). Invasions and extinctions reshape coastal marine food webs. PLoS ONE, 2(3), 1–7. https://doi.org/10.1371/journal.pone.0000295

Carroll, S. P., Hendry, A. P., Reznick, D. N., & Fox, C. W. (2007). Evolution on ecological time- scales. Functional Ecology, 21, 387-393.

Chase, J. M. (1996). Abiotic controls of trophic cascades in a simple grassland food chain. Oikos, 77, 495–506. DOI: 10.2307/3545939

Duffy, J. E. (2003). Biodiversity loss, trophic skew and ecosystem functioning. Ecology Letters, 6, 680–687. https://doi.org/10.1046/j.1461-0248.2003.00494.x

Estes, J. A., Terborgh, J., Brashares, J. S., Power, M. E., Berger, J., Bond, W. J., … Wardle, D. A. (2011). Trophic downgrading of Planet Earth. Science, 333, 301–306. DOI: 10.1126/science.1205106

Finke, D. L., & Denno, R. F. (2005). Predator diversity and the functioning of ecosystems: the role of intraguild predation in dampening trophic cascades. Ecology letters, 8, 1299-1306. https://doi.org/10.1111/j.1461-0248.2005.00832.x

Ford, A. T., & Goheen, J. R. (2015). Trophic cascades by large carnivores: a case for strong inference and mechanism. Trends in Ecology & Evolution, 30(12), 725–735. https://doi.org/10.1016/j.tree.2015.09.012

Fox, J. (2003). Effect displays in R for Generalised Linear Models. Journal of Statistical Software, 8(15), 1-27.

Fox, J., & Weisberg, S. (2011). Multivariate linear models in R. An R Companion to Applied Regression. Los Angeles, Thousand Oaks, 2nd Ed. Thousand Oaks, CA, Sage.

García-Palacios, P., Maestre, F. T., Kattge, J., & Wall, D. H. (2013). Climate and litter quality differently modulate the effects of soil fauna on litter decomposition across biomes. Ecology Letters, 16(8), 1045–1053. https://doi.org/10.1111/ele.12137 Accepted Article

This article is protected by copyright. All rights reserved Gilman, S. E., Urban, M. C., Tewksbury, J., Gilchrist, G. W., & Holt, R. D. (2010). A framework for community interactions under climate change. Trends in Ecology & Evolution, 25(6), 325- 331. https://doi.org/10.1016/j.tree.2010.03.002

Gordon, C. E., Feit, A., Grüber, J., & Letnic, M. (2015). Mesopredator suppression by an apex predator alleviates the risk of predation perceived by small prey. Proceedings of the Royal Society of London. Series B, Biological Sciences, 282(1802), 20142870. https://doi.org/10.1098/rspb.2014.2870

Halaj, J., & Wise, D. H. (2001). Terrestrial trophic cascades: how much do they trickle? The American Naturalist, 157, 262-281. https://doi.org/10.1086/319190

Hothorn, T., Bretz, F., & Westfall, P. (2008). Simultaneous inference in General Parametric Models. Biometrical Journal, 50(3), 346–363. https://doi.org/10.1002/bimj.200810425

Jackson, R. B., Lechowicz, M. J., Li, X., & Mooney, H. A. (2001). Phenology, growth, and allocation in global terrestrial productivity. In J. Roy, B. Saugier, & H. A. Mooney (Eds.), Terrestrial global productivity (pp. 61-82). San Diego, California: Academic Press.

Kajak, A. (1995). The role of soil predators in decomposition processes. European Journal of Entomology, 92, 573–580.

Kajak, A. (1997). Effects of epigeic macroarthropods on grass litter decomposition in mown meadow. Agriculture, Ecosystems & Environment, 64(1), 53–63. https://doi.org/10.1016/S0167- 8809(96)01125-5

Kajak, A., & Jakubczyk, H. (1977). Experimental studies on predation in the soil-litter interface. Ecological Bulletins, 25, 493–496.

Kardol, P., Spitzer, C. M., Gundale, M. J., Nilsson, M. C., & Wardle, D. A. (2016). Trophic cascades in the bryosphere: the impact of global change factors on top—down control of cyanobacterial

N2-fixation. Ecology Letters, 19(8), 967-976. https://doi.org/10.1111/ele.12635 Accepted Article

This article is protected by copyright. All rights reserved Koltz, A. M., Classen, A. T., & Wright, J. P. (2018). Warming reverses top-down effects of predators on belowground ecosystem function in Arctic tundra. Proceedings of the National Academy of Sciences of the Unite States of America, 115(32), E7541-E7549. https://doi.org/10.1073/pnas.1808754115

Lawrence, K. L., & Wise, D. H. (2000). Spider predation on forest-floor Collembola and evidence for indirect effects on decomposition. Pedobiologia, 44(1), 33–39. https://doi.org/10.1078/S0031- 4056(04)70026-8

Lawrence, K. L., & Wise, D. H. (2004). Unexpected indirect effect of spiders on the rate of litter disappearance in a deciduous forest. Pedobiologia, 48(2), 149–157. https://doi.org/10.1016/j.pedobi.2003.11.001

Lensing, J. R., & Wise, D. H. (2006). Predicted climate change alters the indirect effect of predators on an ecosystem process. Proceedings of the National Academy of Sciences of the Unite States of America, 103(42), 15502–15505. https://doi.org/10.1073/pnas.0607064103

Lindberg, N., Bengtsson, J., & Persson, T. (2002). Effects of experimental irrigation and drought on the composition and diversity of soil fauna in a coniferous stand. Journal of Applied Ecology, 39(6), 924–936. https://doi.org/10.1046/j.1365-2664.2002.00769.x

Liu, S., Chen, J., He, X., Hu, J., & Yang, X. (2014). Trophic cascade of a web-building spider decreases litter decomposition in a tropical forest floor. European Journal of Soil Biology, 65, 79–86. https://doi.org/10.1016/j.ejsobi.2014.10.004

Loustau, D., Hungate, B., & Drake, B. G. (2001). Water, nitrogen, rising atmospheric CO2, and terrestrial productivity. In J. Roy, B. Saugier, & H. A. Mooney (Eds.), Terrestrial global productivity (pp. 123-167). San Diego, California: Academic Press.

Lotze, H. K., Lenihan, H. S., Bourque, B. J., Bradbury, R. H., Cooke, R. G., Kay, M. C., … Jackson, J. B. C. (2006). Depletion, degradation, and recovery potential of estuaries and coastal seas. Science, 312(5781), 1806–1809. DOI: 10.1126/science.1128035 Accepted Article

This article is protected by copyright. All rights reserved Majdi, N., Boiché, A., Traunspurger, W., & Lecerf, A. (2014). Predator effects on a detritus‐based food web are primarily mediated by non‐trophic interactions. Journal of Animal Ecology, 83(4), 953-962. https://doi.org/10.1111/1365-2656.12189

McCluney, K. E., & Sabo, J. L. (2009). Water availability directly determines per capita consumption at two trophic levels. Ecology, 90(6), 1463–1469. https://doi.org/10.1890/08-1626.1

McCluney, K. E., & Sabo, J. L. (2016). Animal water balance drives top-down effects in a riparian forest—implications for terrestrial trophic cascades. Proceedings of the Royal Society of London. Series B, Biological Sciences, 283(1836), 20160881. https://doi.org/10.1098/rspb.2016.0881

Melguizo-Ruiz, N., Jiménez-Navarro, G., & Moya-Laraño, J. (2016). Beech cupules as keystone structures for soil fauna. PeerJ, 4, e2562. https://doi.org/10.7717/peerj.2562

Melguizo-Ruiz, N., Verdeny-Vilalta, O., Arnedo, M. A., & Moya-Laraño, J. (2012). Potential drivers of spatial structure of leaf-litter food webs in south-western European beech forests. Pedobiologia, 55(6), 311–319. https://doi.org/10.1016/j.pedobi.2012.06.003

Moya-Laraño, J., & Wise, D. H. (2007). Direct and indirect effects of ants on a forest-floor food web. Ecology, 88(6), 1454–1465. https://doi.org/10.1890/05-1474.

Ninyerola, M., Pons, X., & Roure, J. M. (2005). Atlas climático digital de la Península Ibérica. Metodología y aplicaciones en bioclimatología y geobotánica. Universidad Autónoma de Barcelona, Bellaterra.

Ninyerola, M., Pons, X., & Roure, J. M. (2007). Monthly precipitation mapping of the Iberian Peninsula using spatial interpolation tools implemented in a Geographic Information System. Theoretical and Applied Climatology, 89(3-4), 195-209. DOI 10.1007/s00704-006-0264-2.

Olson, J. S. (1963). Energy storage and the balance of producers and decomposers in ecological systems. Ecology, 44(2), 322-331. https://doi.org/10.2307/1932179 Accepted Article

This article is protected by copyright. All rights reserved Pace, M. L., Cole, J. J., Carpenter, S. R., & Kitchell, J. F. (1999). Trophic cascades revealed in diverse ecosystems. Trends in Ecology & Evolution, 14, 483–488. https://doi.org/10.1016/S0169-5347(99)01723-1

Petchey, O. L., McPhearson, P. T., Casey, T. M., & Morin, P. J. (1999). Environmental warming alters food-web structure and ecosystem function. Nature, 402(6757), 69–72. https://doi.org/10.1038/47023

Polis, G. A., Myers, C. A., & Holt, R. D. (1989). The ecology and evolution of intraguild predation: potential competitors that eat each other. Annual Review of Ecology and Systematics, 20(1), 297–330. https://doi.org/10.1146/annurev.es.20.110189.001501

R Core Team (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available at: https://www.R-project.org/. Last accessed 09 January 2019.

Rodrigues, L. R., Duncan, A. B., Clemente, S. H., Moya-Laraño, J., & Magalhães, S. (2016). Integrating competition for food, hosts, or mates via experimental evolution. Trends in ecology & evolution, 31, 158-170.

Sanders, D., Schaefer, M., Platner, C., & Griffiths, G. J. K. (2011). Intraguild interactions among generalist predator functional groups drive impact on herbivore and decomposer prey. Oikos, 120(3), 418–426. https://doi.org/10.1111/j.1600-0706.2010.18924.x

Sanpera-Calbet, I. S. I. S., Lecerf, A., & Chauvet, A. (2009). Leaf diversity influences in‐stream litter decomposition through effects on shredders. Freshwater Biology, 54(8), 1671-1682. https://doi.org/10.1111/j.1365-2427.2009.02216.x

Schmitz, O. J., Hambäck, P. A., & Beckerman, A. P. (2000). Trophic cascades in terrestrial systems: a review of the effects of carnivore removals on plants. The American Naturalist, 155, 141–153. https://doi.org/10.1086/303311 Accepted Article

This article is protected by copyright. All rights reserved Schmitz, O. J., Hawlena, D., & Trussell, G. C. (2010). Predator control of ecosystem nutrient dynamics. Ecology Letters, 13, 1199–1209. https://doi.org/10.1111/j.1461-0248.2010.01511.x

Schmitz, O. J., Krivan, V., & Ovadia, O. (2004). Trophic cascades: the primacy of trait‐mediated indirect interactions. Ecology Letters, 7(2), 153-163. https://doi.org/10.1111/j.1461- 0248.2003.00560.x

Schneider, F. D., & Brose, U. (2013). Beyond diversity: how nested predator effects control ecosystem functions. Journal of Animal Ecology, 82(1), 64–71. https://doi.org/10.1111/1365- 2656.12010

Schneider, F. D., Scheu, S., & Brose, U. (2012). Body mass constraints on feeding rates determine the consequences of predator loss. Ecology Letters, 15(5), 436-443. https://doi.org/10.1111/j.1461-0248.2012.01750.x

Snyder, W. E., & Wise, D. H. (2001). Contrasting Trophic Cascades Generated By a Community of Generalist Predators. Ecology, 82(6), 1571–1583.

Stewart, R. I. A., Dossena, M., Bohan, D. A., Jeppesen, E., Kordas, R. L., Ledger, M. E., ... Suttle, B. (2013). Mesocosm experiments as a tool for ecological climate‐change research. In G. Woodward & E.J. O'Gorman (Eds), Advances in Ecological Research, (1st edn), (pp. 71–181). London, UK: Elsevier Ltd.

Stocker, T. F., Qin, D., Plattner, G. -K., Tignor, M., Allen, S. K., Boschung, J., ... Midgley, P. M. (2013). IPCC, 2013: climate change 2013: the physical science basis. Contribution of working group I to the fifth assessment report of the intergovernmental panel on climate change. Intergovernmental Panel on Climate Change, AR5, 1535.

Terborgh, J., Lopez, L., Nuñez, P., Rao, M., Shahabuddin, G., Orihuela, G., ... Lambert, T. D. (2001). Ecological meltdown in predator-free forest fragments. Science, 294(5548), 1923–1926. DOI: 10.1126/science.1064397 Accepted Article

This article is protected by copyright. All rights reserved Verdeny-Vilalta, O. (2013). Algunas consecuencias ecológicas y evolutivas del movimiento animal para las interacciones bióticas. PhD Thesis. Universidad de Granada.

Verdeny-Vilalta, O., & Moya-Laraño, J. (2014). Seeking water while avoiding predators: Moisture gradients can affect predator-prey interactions. Animal Behaviour, 90, 101–108. https://doi.org/10.1016/j.anbehav.2014.01.027

Voigt, W., Perner, J., Davis, A. J., Eggers, T., Schumacher, J., Bährmann, R., ... Sander F. W. (2003). Trophic levels are differentially sensitive to climate. Ecology, 84(9), 2444–2453. https://doi.org/10.1890/02-0266

Wall, D. H., Bradford, M. A., St. John, M. G., Trofymow, J. A., Behan-Pelletier, V., Bignell, D. E., … Zou, X. (2008). Global decomposition experiment shows soil animal impacts on decomposition are climate-dependent. Global Change Biology, 14(11), 2661–2677. https://doi.org/10.1111/j.1365-2486.2008.01672.x

Werner, E. E., & Peacor, S.D. (2003). A review of trait‐mediated indirect interactions in ecological communities. Ecology, 84(5), 1083-1100. https://doi.org/10.1890/0012- 9658(2003)084[1083:AROTII]2.0.CO;2

Wise, D. H. & Lensing, J. R. (2019). Impacts of rainfall extremes predicted by climate-change models on major trophic groups in the leaf-litter arthropod community. Journal of Animal Ecology, In press.

Woodward, G., Ebenman, B., Emmerson, M., Montoya, J. M., Olesen, J. M., Valido, A., & Warren, P. H. (2005). Body size in ecological networks. Trends in Ecology & Evolution, 20(7), 402– 409. https://doi.org/10.1016/j.tree.2005.04.005

Wu, X., Duffy, J. E., Reich, P. B., & Sun, S. (2011). A brown-world cascade in the dung decomposer food web of an alpine meadow: effects of predator interactions and warming. Ecological Monographs, 81(2), 313–328. https://doi.org/10.1890/10-0808.1 Accepted Article

This article is protected by copyright. All rights reserved A typical leaf-litter food web in a beech forest from northwest Spain. Large (apex) predators feed on each other (intraguild predation - IGP) and on lower trophic levels, which includes the offspring of large predators (intermediate predators), small predators and decomposers (detritivores, microbivores and fungivores), which in turn feed on a mixture of litter, fungi and bacteria. The sudden decrease in abundances of large predators (either from extinction or from experimental removal) may cause overexploitation of decomposers by intermediate and small predators, eventually leading to a decrease in the densities of smaller predators from IGP and strong competition for the now more-scarce resources. Since in dry beech forests, secondary productivity is lower and feeding resources scarcer, populations in lower trophic levels may have evolved stronger competitive abilities. We therefore predicted that resource overexploitation would be stronger among intermediate and smaller predators in dry forests, leading to stronger decreases in their densities (in response to fewer apex predators) as well as stronger declines in densities of the decomposer trophic level, resulting in a greater reduction in decomposition rate in dry forests. Accepted Article

This article is protected by copyright. All rights reserved Abundances (no. per plot) of the five trophic groups at the end of the experiment — (a) large, (b) intermediate, and (c) small predators; (d) microbivores and fungivores; and (e) detritivores — across the experimental treatments (0X - Predator removal treatment, in white; 1X - Control, in light green; 2X - Predator addition treatment, in dark green). Note that only significant Rainfall x Treatment interactions are shown and that, in the absence of an interaction, pooled effects of the experimental treatments are presented. Abundances are the model-predicted abundances and confidence intervals (estimates and 95% CIs) of significant variables (library "effects" - Fox 2003). Predicted abundances with the same letter do not differ (p > 0.05) according to post-hoc contrasts, whereas abundances with different letters differ at the p = 0.05 level. The second component of the figure summarizes these relationships more succinctly by showing effect size (with 95% CI) expressed as the ratio of the predicted abundances of two treatments (1X/0X, 2X/0X, and 2X/1X). The dashed line represents the absence of an effect. Accepted Article

This article is protected by copyright. All rights reserved Higher densities of apex predators accelerated decomposition in the dry forests but had no impact in wet forests. Decomposition rates (√k, with k in years-1) across the experimental treatments (0X - Predator removal treatment, in white; 1X - Control, in light green; 2X - Predator addition treatment, in dark green). Decomposition rates (√k) are model-predicted decomposition rates and confidence intervals (estimates and 95% CIs) (library “effects” - Fox 2003). Letters over model- predicted rates, and proportional effect sizes with 95% CIs, have the same meanings as in Fig. 1. Accepted Article

This article is protected by copyright. All rights reserved Outline of the hypothesized mechanisms leading to the observed decline in abundances and litter decomposition from reduction of large predators in 0X plots. In our study, cascading effects clearly influenced decomposition, although at the end of the experiment all functional groups were more abundant in the 2X plots. Black solid arrows indicate direct positive and negative effects; grey solid arrows indicate effects that will be relevant and more important at later times (e.g., when prey shared by large and intermediate predators become more scarce); the black dashed arrow indicates impact on an ecosystem process (litter decomposition). Large “X”s signify a decline in abundance (either caused directly by manipulation, or resulting from increased predation pressure or reduced food supply), and asterisks indicate significant effects in dry forests. At time T1 predators were removed from the 0X plots, which at time T2 immediately reduced predation pressure on smaller and intermediate predators, resulting in higher densities and increased activity of these predators compared to 1X and 2X treatments (T2). These increases in turn led to stronger predation pressure upon decomposers (microbivores, fungivores and detritivores), reducing their densities at T3. The reduction in decomposer numbers caused lower rates of decomposition in 0X compared to 1X and 2X plots (dashed arrow). In addition, fewer decomposers, the food of small and intermediate predators, eventually caused abundances of these predators to decrease in 0X plots. The decline in intermediate predators was also due to reduced recruitment resulting from removal of adults (the apex predators), represented by the grey arrow at T2. Effects on decomposition and small predators were only present or stronger, respectively, in dry forests, as indicated by the asterisks (supported by the Treatment x Rainfall interactions exhibited by these groups – see text). This hypothesis assumes that the direct effect of large predators on decomposers is negligible. Our sampling schedule was limited to detecting effects at T3 and T4 (supported by our statistical analyses). Accepted Article