DISCUSSION
To our knowledge, this is the first study to experimentally unravel cascading effects of tree functional diversity, through trophic chains, on a biogeochemical process through time. Our concept and findings have added a new dimension to the “afterlife effects” of functional traits by pinpointing the cascading effects of variation in deadwood quality among tree species through a “brown” food chain, with important consequences for both forest carbon turnover and animal populations. While, without the involvement of invertebrate decomposers wood decomposition rates were predicted by initial values along the WES, when termites participated in the decomposition process they actively altered the species’ ranking in terms of the WES through time and moved in abundance and activity from initially high WES species to initially medium WES species. These findings have an interesting parallel with a recent study that found a modulating effect of invertebrate consumption on leaf mass loss rankings along an axis of initial litter quality, i.e., the leaf economics spectrum, over time (Guo et al. 2019). However, in that study the modulation was due to a single outbreak by detritivorous moth larvae, while in this study the decomposing invertebrates (i.e., termites) were present in considerable abundance through time in the studied forests, but changed their relative abundance and contribution to decomposition among tree species as decomposition progressed. As a consequence, the temporally stable overall consumption of deadwood by termites should help to sustain deadwood turnover in a forest of high functional diversity. In turn, negative feedback by pangolins (Fig. 1) should contribute to net carbon sequestration via controlling the termite populations, but we could not carry out manipulation of pangolins to quantify this feedback.
Interactive effects of top-down and bottom-up forces in regulating food web structure have been shown (Faithfull et al. 2011; Leibold et al. 1997). However, in our system the positive bottom-up control of termite populations by deadwood seemed to strongly outweigh the negative top-down control by pangolins, as the termite population was still larger in PT with than in TT without pangolins. To overcome predation pressure, prey communities may show compensatory population growth via faster regeneration (Crawley 1992). This can provide an explanation why the PT termite population did not decline with a significant increase in predator population (evidenced by pangolin burrow density). In a functionally diverse forest, with the wood resource complementarity as shown in our study, the basal resource constraint is likely relatively weak through time, allowing prey (termites) to reach high population density in spite of predator (pangolin) control. If hunting and trafficking of pangolins ceased completely (e.g., in response to pangolins’ suspected role in transmission of the 2020 Covid-19 virus to humans), and pangolin populations were allowed to regain their natural population size across the entire region, the contributions of top-down versus bottom-up control of termite populations and their effect on wood decomposition and the forest carbon budget would be easier to study.
New experiments are needed to enrich our conceptual model with empirical evidence, beyond the specific food chain studied here. On the one hand, other termite predators in the world (e.g., other pangolin species; aardvarks in Africa; some armadillo species in the Americas) are also in dramatic population decline. We need to know whether, besides halting their persecution by people, improving their habitat by promoting plant functional diversity and, thereby, their staple food will help to recover or sustain the populations of these species in general. Some ants are also known to have termites as a staple foods (Buczkowski & Bennett 2007), but whether or how they compete for termites with vertebrate termite feeders such as pangolins, or are themselves eaten and thereby controlled top-down by these vertebrates, and thereby indirectly affect wood decomposition, are intriguing questions for further investigation.
Moreover, terrestrial decomposition is integrated in multiple food chains in the food web. Thus, to better understand the generality of the positive deadwood turnover feedback studied here, we need to identify the interactions between other wood (or leaf, root) decomposing invertebrates or predators of other detritivores and plant functional trait diversity. Although termites are the dominant species in warm-climate wood decomposition, other invertebrates such as beetles tend to control the animal contribution to decomposition in temperate ecosystems; either directly through feeding or by facilitating other decomposers such as fungi or similar-sized animals to invade the deadwood by tunneling, feeding, nesting or intended microbe cultivation (Ulysen 2016; Zuo et al. 2014). In our experiment we found a wide variety of deadwood beetles and millipedes; although their abundance is far less than that of termites, how (much) they affect wood decomposition, directly or by affecting other decomposers such as fungi and termites, needs in-depth study. Importantly, we need to study how these different decomposers, through time (e.g., decay stages, seasons), interact with each other and with the trait diversity of their basal resource via their decomposing function. Multiple harvests and tracking animal population dynamics (Schwarzmüller et al. 2015) in such experiments can provide further insights into time effects of decomposers, and their suppression by predators, as influenced by resource availability.
Taken together, our findings strongly point to high wood functional diversity helping to provision sufficient food resource in terms of deadwood quantity and quality for termites to maintain their population through time. In turn, this larger and stable termite population seemed to have increased the density of pangolins as indicated by the four-fold increase in density of their burrows. Thus, our results strongly suggest that woody plant functional (trait) diversity can support a (threatened) mammal species at the top of the food chain through the detritivore subsystem. We have to make the caveat that the causality of wood functional diversity supporting the pangolin population in our experiment cannot be proven and would need further, perhaps manipulative study. The fact that we added deadwood in our research plots may by itself have stimulated both termites and pangolins and, in theory, year-to-year changes in environmental conditions could also have interfered with termite and pangolin populations. Still, our results strongly suggest that, if only initially high, medium or low quality wood had been added in the experiment, the termite populations would not have increased as much as they did with the broader resource economic spectrum (also compared to the initial forest community) incubated in our forest plots over the time span of the experiment. And neither would the number of pangolin burrows in the PT site have quadrupled. Moreover, although indirect evidence, the (living) tree community in PT, with the larger overall termite density, had a greater wood trait diversity than in TT (Table 1, unweighted WES range, community-weighted WES variance), which likely translates into a wider overall range in deadwood quality in PT. Thus, our results are consistent with the hypothesis that woody plant functional (trait) diversity drives an intriguing interplay of dead wood quality, its decomposition and termites through time, thereby even supporting a threatened mammal at the top of the food chain. These findings should invite further studies, in other green or brown food webs, of how plant functional diversity can cascade dynamically through different trophic levels all the way to the top of food chains.