Dynamic feedbacks among tree functional traits, termite populations and deadwood turnover

Biodiversity is declining dramatically as a result of habitat alteration, unsustainable harvesting and climate change, which may affect ecosystem processes functions (IPBES Report, 2019). Much research on terrestrial plant biodiversity has focused how the loss or change species functional groups affects primary productivity, nutrient cycling carbon fluxes (Handa et al., 2014; Hättenschwiler 2005; Hooper & Vitousek, 1998), for example, through niche complementarity plants microbes. Within this biodiversity–ecosystem function debate, emphasis shifted from effects taxonomic diversity different coarse to continuous variation in traits at level Violle 2007). However, work hardly addressed services via trophic interactions (but see Bascompte 2003; Scherber 2010). Those are often hard unravel, because, one hand, ‘top-down control’ by consumers detritivores, their predators, flows material food webs (Srivastava 2009); such might have particularly strong impact if keystone predators involved (Jones 1997). On other quality availability plant-based resources population dynamics ‘bottom-up (Haddad 2009; Variation traits, species' resource complementarity, regulate basal consequently higher levels (Thakur Eisenhauer, 2015). natural ecosystems, supplied over time, consequence interspecific phenology consumer web changes dead matter quantity detritivore web. Green (plant) can be structured environmental time-scales, seasonal, inter-annual decadal (McMeans 2015; Voigt 2003). no previous studies temporal variance across could lead brown (detritivore) webs. Such occur, instance, deadwood tree with initial trait values that decompose rates. To our knowledge, theoretical empirical hypothesis. Here we test concept an intriguing system warm-climate forests featuring feedback among dynamic derived them, termites decomposition (Figure 1). Termites invertebrate decomposer globally (Cornwell Ulyshen, 2016; Wood Sands, 1978), been found contribute, about half turnover southern China (Liu At bottom end web, quality, determined several wood palatability termites. Previous shown prefer consume softer Takamura, 2001). Therefore, composition favourable termite consumption should help populations thrive, thereby also supporting predatory vertebrates pangolins affected each given decay trajectory, may, turn, forest time. How feedbacks between rates these time remains unknown. Based theory above field observations subtropical broad-leaved China, conceptual model 1) proposes that, after logging typhoon impact, set functionally woody (ranging widely quality) will provide major floor. Termites, being selective consumers, first high-quality species, low density, dry lignin content, high cellulose content. associated resource-acquisitive growth strategy along ‘wood economics spectrum’ (WES; Freschet 2012; Pietsch Zuo 2018). Over mostly eaten (mainly termites), predict initially medium-quality becomes more palatable owing microbial activity decomposition. Thus, (i.e. intermediate WES values) now consumed preferentially This pattern would imply case broad range qualities WES, is, diversity, there stable provisioning sufficient maintain In sustain same mechanism changing We put framework experimental comparing 34 abundance contributions two evergreen sites Eastern China: without Chinese Manis pentadactyla. contrast was check whether relationships robust top-down control termite-specialized predator. sites, carried out identical 18-month experiment exclusion. conducted study Zhejiang Province, East (a) Tiantong National Forest Park (TT) (29°52?N, 121°39?E), monsoon climate. The dominant vegetation forest. (b) Putuo island (PT) (29°97?N, 121°38?E), Zhoushan archipelago 6.5 km mainland, marine supports dominated Cyclobalanopsis glauca mixed evergreen–deciduous forests. pentadactyla critically endangered (IUCN, 2016) pangolin evidence observed TT during years. contrast, PT thriving because conservation efforts isolation mainland. both Coptotermes formosanus, Odontotermes Reticulitermes chinensis, speratus) predominant macroinvertebrates S5) decomposers (Yi 2006), they many Asian (see e.g. Griffiths All belong pangolins' preferred (Wu 1999, 2005). Table 1 further data abiotic biotic features Yan 2006). From October November 2017, selected 32 important representative PT, 2 endemic (Pittosporum tobira, Eurya emarginata). included 7 trees 11 shrubs, deciduous 4 conifer monocot (bamboo; full names S1). ‘shrubs’ short stature, single-stem understorey multi-stemmed shrubs exposed habitats. total, 147 individual healthy adult trees/shrubs were selected, 19 × 3 individuals 15 6 shrubs. larger number shrub individuals, shorter needed obtain 5 cm diameter segments three replications below). chain-sawed those collected samples 20-cm long stem sections ± 0.5 so size Adjacent end, 2-cm thick disk sawn analyses traits. used living standardizing initial, undegraded phase all samples, avoiding complications prior pathogen attack having altered providing direct link (Cornelissen 2012). Moreover, studied typhoon-induced wind-throw well very common, large fraction still alive when downed. therefore consider going into process. 25 litter bags differing mesh assess contribution decomposition: exclusion treatment, enclosed stainless steel 0.05-mm access nylon 4-mm mesh. survey period, strongly group macro-invertebrates Isoptera-termites S5). Hymenoptera had site, remaining fauna taxa, Coleoptera, Araneae, Isopoda, Geophilomorpha, Lumbricida, Lepidoptera larva contributed less than 5% macro-invertebrate abundance. based observations, few entered bags. Acarina, Collembola Diptera meso-invertebrates while Tubificida, Corrodentia, Hemiptera larva, Pseudoscorpionida, Thysanoptera Symphyla <5% Although did possibly not exclude smaller mesofauna, expect negligible, taxa do feed itself but substrates frass fungi. Each fresh sample weighed immediately cutting below measurement water content), sealed bag. 1,224 litter-bag produced, treatments (plots) incubation harvest times. December established 20 m 30 replicate plots TT, respectively. Plots within site broadly similar slope, altitude, soil type, substrate structure 10 apart minimize interference them. started temperatures. pinned onto floor respective subplots randomly, per treatment subplot, harvest. distances approximately m. Litter harvested months (July 2018), 12 (December 2018) 18 After collection plots, transported them laboratory where carefully removed extraneous materials days. tray tall enough edges crawl out. cut small pieces cleaned mud sand (brought termites) brush carefully. then counted oven-dried 75°C constant mass (dry mass). For considered relevant decomposition, subsample stored plastic bag field, kept cool until processing. hr, bark removed, volume using Archimedes' principle displacement (Williamson Wiemann, subsamples dried 72 hr determine mass. Initial density (g/cm3) calculated volume. (fresh ? mass)/fresh mass, calculate samples. ground mill passed 0.15-mm sieve. Thereafter, 0.2 g digested concentrated H2SO4 N P concentrations infrared spectrophotometer (Smartchem 200, Alliance, France). Lignin content acidolysis-titration method, anthrone–sulphuric acid colorimetry (Poorter Villar, Bark possible influence (cf. Kahl 2017) promotion Ulyshen measured companion focus study. quantify situ, sampled July 2018 2019 adopting adjusting transects method Eggleton, 2000). Since area m, took entire plot belt transect, surveyed ten diagonal. section, meter (with grid), counts grid cells scaled up examined leaf total plot. See counting mounds plot, simultaneously, nearby statistical performed R language version 3.5.1. main axes principal component analysis (PCA) (r package ‘vegan’, ‘rda’ function). axis (PC1), accounting 55.9% related contents nutrients (nitrogen phosphorus), lignin, density. PC1 scores represent position positions subsequent analyses. second (PC2) 17.1% PCA community-level variability PT. accounted 50.3% 40.9% variance, respectively, economic (specific area, nitrogen, phosphorus mean area) community abundance-weighted (CWM) specific compare differences identity Student's t tests. derive CWM community-weighted multiplied its relative community. r ‘lme4’ applied linear effect construction. Prior analysis, logit-transformed (Griffiths 2019; Warton Hui, 2011). performance models assessed marginal conditional squares, indicate explained fixed (; marginal) whole conditional), explanatory power independent variable period cumulative suggested Nakagawa Schielzeth (2013). calculation, proportion ratio ) response cumulative). Second, regression and/or nonlinear (cumulative) % (for treatments) separately periods. Also, regressions relationship (period-specific cumulative) (in samples) Again, sites. analyses, ‘ggplot2’ visualization, stat smooth ggplot2, 6–12 12–18 Comparably, binomial site. glm (generalized models) months) Gaussian months). data, Levene's examine homogeneity Shapiro–Wilk normality. Comparisons support view play role 2), only tropical Section Specifically, resulted average 26.6% (range: 1.5%–84.6%) 21.1% 1.18%–68.8%) species. 14.7% 1.1%–60.7%) 12.7% 1.0%–45.7%) S2a,b). difference generally decreased 2, Figure showed interaction significantly (Table 2). joint 42% 43% ), Termite 8%, 2% 20% loss, Harvest alone significant jointly 11% loss. 13%, 10% = 0.419 0.420 0.426 0.427 0.459 0.460 0.372 0.373 46% 37% respectively 2%, 3% 16% 5%, 25% Multiple notably N, P, together PC1, interpreted here Methods values), decomposed faster conservative species; S1, positive score periods 0–6 non-significant ones months), cumulatively months, Our revealed accelerated rate overall clear acceleration 3), hypothesized year (periods 0–6, positively linearly 2a–d), indicating (high dense structure). Thereby, termites' amplified decomposability, it increased slope. partly activity, acquisitive considerably depleted months. termites, modulated trajectory WES. Now, relation access, humpback least termite-rich 2e, corresponding (equation 2e). apparent, steep latter 2d,f) 2b). suggests negative These patterns period-specific confirmed presence/absence Together, 1), deviating caused convergence 2g,h), indicated lack consistent 3) S3) significant, exponential increase 3a,b). peak occurred slightly medium 3c,d). By moved towards 3e), centre range, matching height width range) 3f). whole, peaks process, itself. experimentally unravel animal matter. findings added new dimension ‘afterlife effects’ pinpointing ‘brown’ chain, consequences populations. While, involvement decomposers, predicted spectrum (WES); participated actively ranking terms interesting parallel recent modulating rankings spectrum, (Guo study, modulation due single outbreak detritivorous moth larvae, decomposing invertebrates present considerable forests, changed progressed. Below, discuss ecosystems. next steps regulating without-termites means side composition. se chemical (positively) (Maillard Zak 2011) (negatively) (Austin Ballare, 2010), captured suite biochemical belonging (resource) (Freschet correlations 2c,d). most likely explanation rapid stage (via leaching decomposition) deadwood, labile carbohydrates reduced low-quality recalcitrant, reducing qualities. addition, winter temperature leads weakening (e.g. fungi), obvious tendency Consistent hypothesis, supported sustenance complementary varied capture throughout structural analysed (Zuo 2020). As Oberle al. (2019) demonstrated, soft wood, denser lost slowly quickly advancing decay, thus key vessel diameter) diverse experienced Indeed, located contents, some that. once positioned middle become palatable. explain why reached maximum quality. maintained resources. pattern. specific, decomposability quality), increasing slope (the richest termites). mid-WES) Interestingly, (or turnover) closely attributed becoming increasingly infiltrated degraded fungi (Stokland 2012); import surroundings process goes (authors' unpublished data). decayed rather undecomposed again (Waller 1987). preliminary lower 2001), expected (like study) degradation especially concurrently. weaker What worth discussing If intense shape details, ecological floristic dissimilarities geographically distant 1, Figures S4 stronger disturbance regime, organic matter, (CWM pangolins. another frequent severe disturbance, composition, findings, revealing abundances potential implications right side). prediction else equal conditions regional pool available, wide-ranging qualities, supply allowing establish. occasional disturbances typhoon-prone events simultaneously bring down depending potentially knock-on top termite-feeding designed predictions explicitly, merit so, advance understanding needs detritivores predators. add real-world complexity, involve ants. Some ants known staple (Buczkowski Bennett, 2007), compete vertebrate feeders pangolins, themselves controlled vertebrates, indirectly questions investigation. It noted recorded identify (only overall, Methods). Future pay much effort investigate associates doing, benefit mechanistic functions. Another issue future extend approach experiment, relatively class, standard logs classes. (more recalcitrant) heartwood sapwood thicker bark, compared class. experiments enrich variation, evidence, beyond chain here. Terrestrial integrated multiple chains better understand generality here, need leaf, root) beetles tend temperate ecosystems; either directly feeding facilitating similar-sized animals invade tunnelling, feeding, nesting intended microbe cultivation (Ulyshen, 2014). wide variety millipedes; although far (much) affecting in-depth Importantly, stages, seasons), interact function. harvests tracking (Schwarzmüller 2015) insights suppression influenced availability. conclude, turnover. suggest provision In-depth indirect authors like thank Minshan Xu, Liting Zheng, Xiangyu Liu, Umar Aftab Abbasi, Tian Su, Wujian Xiong, Qiming Liang, Xiaotong Zhu, Liang Li, Dong He Liangyan Wang assistance preparing laboratory. State Key Program Natural Science Foundation (grant no. 32030068). declare competing interests. E.-R.Y., J.H.C.C. C.G. conceptualized experiment; C.G., B.T. H.C. data; wrote manuscript; E.-R.Y. revised improved manuscript. substantially revisions. peer review history article available https://publons.com/publon/10.1111/1365-2745.13604. Data Dryad Digital Repository https://doi.org/10.5061/dryad.wpzgmsbm6 2021). Please note: publisher responsible functionality any information authors. Any queries (other missing content) directed author article.