INTRODUCTION
For decades, ecologists have realized that global change has caused rapid losses in biological diversity, which may in turn led to reductions in important ecosystem functions and services (Duffy 2009; Isbell et al. 2011; Hooper et al. 2012; Liang et al. 2016; Huang et al. 2018). The relationship between biodiversity and ecosystem functions (BEF), e.g. productivity or biomass, and their stability have thus become an important part of ecological research since the 1990s (Wardle et al. 2000; Ruijven & Berendse 2005; Tilman et al. 2014; van der Plas 2019). Previous theoretical and experimental studies have shown that diversity has a significant effect on ecosystem functions and stability (Naeem & Li 1997; Loreau 1998; Cadotte et al. 2012; Loreau & de Mazancourt 2013; Pennekamp et al. 2018). While there were a number of BEF studies, however, many of them were conducted in algae, wetland or grassland ecosystems (Tilman et al. 2006; Isbellet al. 2009; de Mazancourt et al. 2013; Morin et al. 2014; Craven et al. 2018; Pennekamp et al. 2018). In natural forests, whether biodiversity is important for forest biomass and productivity is still hotly debated (Fotis et al. 2018; Ouyang et al. 2019). Moreover, knowledge on the biodiversity effect on forest ecosystem stability is still limited (Mazzochiniet al. 2019), especially in tropical forests which cover only 7 ~ 10% of the Earth’s land surface but account for 55% of carbon pool and ~70% of carbon sink in global forests, and hold 96% of the world’s estimated 45,000 tree species (Panet al. 2011; Poorter et al. 2015).
Two hypotheses were commonly used to explain the effect of biodiversity on ecosystem functions. The complementarity effect hypothesis assumes that difference in niches among coexisting species lead to higher resource utilization efficiency, thereby, increasing diversity enhances ecosystem productivity through niche differentiation and facilitation (Tilman et al. 1997). Meanwhile, the sampling effect hypothesis proposes that higher diversity increases community productivity through an increased chance of possessing highly productive species (Hooperet al. 2005). As for the positive influence of diversity on ecosystem stability, the insurance hypothesis suggests that the asynchrony of different species responding to environmental changes ensures that ecosystem functions have a lower variability when disturbance occurs, and consequently enhances ecosystem stability (Naeem & Li 1997; Mori et al. 2013). At the same time, the change of interactions among species with time can also enhance ecosystem stability, which reflects the niche complementary on the temporal scale (Morin et al. 2014; del Río et al. 2017; Mazzochiniet al. 2019). The niche complementary of coexisting species on temporal and spatial scales ultimately leads to lower variability of ecosystem functions in higher-diversity communities.
Earlier BEF studies generally examined the effect of taxonomic diversity (TD) on ecosystem functions, and species richness is still the most widely used metric till now (Liang et al. 2016; Huang et al. 2018). However, the necessity of including other diversity dimensions and components is increasingly recognized (Craven et al. 2018; van der Plas 2019). In addition to TD, functional diversity (FD) and phylogenetic diversity (PD) are also critical diversity dimensions, which are useful to better address the processes responsible for spatial and temporal dynamics of species co-occurrence (Jarzyna & Jetz 2016). Some studies have shown that functional and/or phylogenetic diversity are stronger predictors than taxonomic diversity for ecosystem functions (Craven et al. 2018; Hao et al. 2018; Mazzochiniet al. 2019; Staples et al. 2019). Meanwhile, diversity still includes three components, namely richness, evenness and divergence (Mason et al. 2005; Helmus et al. 2007; Mouchetet al. 2010), with increasing studies found that evenness and divergence may also be important for ecosystem functions and stability (Zhang et al. 2012; Potter & Woodall 2014; Shirima et al.2015). However, previous BEF studies have seldom included these diversity components and dimensions simultaneously (Flynn et al.2011; Zhang et al. 2012; Craven et al. 2018), it still remains unclear about the relative importance of the richness, evenness and divergence of species, functional and phylogenetic diversity in natural forests. It also remains unclear whether their relative roles differ between distinctive ecosystem functions, e.g. forest biomass and productivity (stock and rate, respectively; see Schmid et al. 2009), and differ between ecosystem functions and their temporal stability.
Experiments that manipulate diversity generally report a strong effect of diversity on ecosystem functions (Isbell et al. 2017; Isbellet al. 2018; van der Plas 2019). However, whether these conclusions obtained from artificial communities can be extrapolated to natural communities has long been controversial, especially for natural forests with far more complex community structure (Poorter et al.2017; van der Sande et al. 2017; Satdichanh et al. 2019). Increasing studies have shown that though diversity had a significant effect on biomass or productivity, the effect can be very weak compared to environmental gradients, species (functional) identity and stand structure (Ma et al. 2010; Wu et al. 2015a; van der Sandeet al. 2018; Ouyang et al. 2019; Staples et al.2019). For instance, two recent studies have illustrated the much stronger role of stand structure attributes over diversity in shaping forest biomass and productivity patterns, respectively in temperate (Fotis et al. 2018) and subtropical forests (Ouyang et al. 2019). While these findings still need further testing, they also raised an interesting question which has not been examined before: what are the relative roles of diversity vs. stand factors in affecting the temporal stability of forest biomass and productivity? Based on the above-mentioned hypotheses on ecosystem stability (e.g. the insurance hypothesis), and evidence from manipulation experiments (del Ríoet al. 2017; Schnabel et al. 2019), it can be predicted that diversity metrics are critical for the stability of forest productivity. However, we also can not reject the possibility that stand factors may be more important than diversity in natural forest, similar as what was found for biomass and productivity per se (Fotiset al. 2018; Ouyang et al. 2019).
Meanwhile, the relationship between diversity and ecosystem functions may change with grain size (Chisholm et al. 2013; Thompsonet al. 2018; Luo et al. 2019). Chisholm et al. (2013) found that the relationship of species richness with productivity and biomass is the strongest at a 400m2 plot size. However, later studies suggest that the relationship is the strongest at about 1000 m2 (Thompson et al. 2018), or even larger grain size (Luo et al. 2019; Mazzochini et al.2019). Thus, the scale dependence of BEF relationships still need further examination. It is also notable that how the biodiversity effect on ecosystem stability changes with grain size has much less been examined compared with forest biomass and productivity (Wang & Loreau 2014), despite it is an urgent question for sustainable forest ecosystem services in the context of rapid global climate change (Mazzochini et al. 2019).
In summary, here we aimed to examine three inter-related questions as follows: 1) In tropical forests, what are the relative effects of diversity vs. stand factors on biomass/productivity, as well as ecosystem stability? 2) How does the relative importance of diversity and stand factors change with grain size? 3) Which of the three diversity components (richness, evenness and divergence) and three dimensions (taxonomic, phylogenetic and functional diversity) are more important for ecosystem functions and stability?
To examine these questions, we used a dataset from 10 permanent plots (50 * 50 m each) in the tropical rainforests of Xishuangbanna (southwest China), which were revisited annually from 2004 to 2010. We also split each 0.25 ha plot into subplots with three grain sizes (400, 800 and 1200 m2) to examine the scale-dependence of diversity effect on stand biomass, productivity and their stability. Based on previous studies, we tested several predictions for our questions as follows. 1) Stand factors, instead of diversity, are the main drivers of forest biomass and productivity (Gough et al. 2019; Ouyanget al. 2019). By contrast, diversity may be more important for the temporal stability of productivity and biomass (Delsol et al. 2018). 2) The complementarity and asynchrony between coexisting species is hypothesized to increase with lager grain size (Bond & Chase 2002; Delsol et al. 2018), and thus it is predicted that the effect of diversity on ecosystem functions and stability should increase with plot size. 3) Functional and phylogenetic diversity are stronger predictors for ecosystem functions and stability than taxonomic diversity (Craven et al. 2018; Haoet al. 2018), while richness, evenness and divergence may affect ecosystem functions and stability together (Potter & Woodall 2014).