Building a bridge between biodiversity and ecosystem multifunctionality

Referring to a general ecology textbook, an ecosystem has three basic functions: energy flow, nutrient cycling, and information transfer. Thus, it can be assumed that an ecosystem can simultaneously provide multiple functions that support human well-being, which is referred to as ecosystem multifunctionality. According to a search using "ecosystem multifunctionality" as the topic in the Web of Science, the oldest paper entitled "Biodiversity and ecosystem multifunctionality" was written by Hector and Bagchi (2007). Their study quantified the influence of biodiversity on multiple ecosystem functions for the first time and ignited the curiosity about understanding ecosystem multifunctionality. Since then, species richness (i.e., the number of species in a community; taxonomic diversity) has become a popular proxy for biodiversity when investigating ecosystem multifunctionality. In general, the development of multifunctionality within an ecosystem relies on greater species richness than that required for a single function. In addition to species richness, biodiversity encompasses other main facets, such as phylogenetic (the presence of different evolutionary lineages) and functional diversity (the variety of growth forms and resource use strategies). Greater species richness within an ecosystem is more likely to be associated with higher dissimilarity between species with distinct evolutionary histories (i.e., phylogenetic diversity) and/or complementary resource-use strategies (i.e., functional diversity). Yet species richness is not necessarily positively associated with phylogenetic or functional diversity. Phylogenetic diversity depends on the species composition, while functional diversity depends on the selected traits, even without changing the species number/composition in a community. To date, several studies have investigated all three main biodiversity facets but reported that they had opposite effects on ecosystem multifunctionality (e.g., Le Bagousse-Pinguet et al., 2019). Accordingly, the community-weighted mean (CWM) trait, which reflects the dominance/selection effect, is likely to play a promising role in mediating these apparent paradoxes (Hagan et al., 2023). And ecosystem trait is formally defined as "the functional traits or quantitative characteristics of organisms (i.e., plants, animals, and microbes) at the community level expressed as the intensity (or density) normalized per unit land area" (He et al., 2019). Furthermore, plant community traits can be further developed as two-dimensional characteristics defined as "intensity or density" on unit land area (He et al., 2023). Only when an ecosystem trait is calculated as a content or ratio (e.g., N content, P content, and N:P ratio) is it equivalent to the CWM trait (He et al., 2019). Stepping forward, ecosystem traits, combining their two-dimensional characteristics, have emerged as advantageous in predicting ecosystem gross primary productivity (GPP) and net primary productivity compared to the CWM trait (He et al., 2023), and would be a better mediator than the CWM trait. However, this must be verified using additional experimental data. Given that ecosystem multifunctionality is a promising but relatively new topic in contemporary global change biology research, much debate exists regarding the number of functions (Manning et al., 2018) and current metrics (Jing et al., 2020) for quantifying ecosystem multifunctionality. Additionally, the separate and integrated effects of biodiversity, climate, and other environmental factors on ecosystem multifunctionality remain unclear. Recently, Migliavacca et al. (2021) revealed that the overall functions of carbon, water, and energy fluxes in major terrestrial ecosystems can be mainly captured by three components: maximum GPP (GPPmax), water use efficiency (WUE), and carbon use efficiency (CUE). However, light use efficiency (LUE), a widely utilized parameter for characterizing ecosystem productivity in Earth models, has not been mentioned. These components can potentially be used to clarify the underlying mechanisms regulating ecosystem multifunctionality, as a whole or separately, because they can be easily measured. The study carried out by Yan et al. (this issue) provides a seminal contribution to biodiversity and ecosystem multifunctionality research by systematically investigating the roles of species richness, phylogenetic and functional diversity, CWM and ecosystem traits, climate, and soil nutrients for ecosystem multifunctionality (Figure 1). The data used in this study were reliable, and required impressive teamwork in the study lasting for eight consecutive years (2013–2020). Information on the biodiversity (species richness, phylogenetic and functional diversity), CWM trait (the first principal component [PC] of three leaf economic traits: specific leaf area [cm2/g], leaf N concentration [mg/g], and leaf P concentration [mg/g], as well as two leaf size traits: leaf area [cm2] and leaf dry mass [g]), as well as ecosystem trait (the first PC axis of leaf N concentration [g/m2], leaf P concentration [g/m2], leaf area [m2/m2], and leaf dry mass [g/m2]) and soil nutrients (the first PC axis of total C and N and C/N in the top 10 cm soils) was based on inventory data obtained from 840 plant community plots and 2500 plant species with standard sampling protocols during their peak periods of plant growth in 72 typical natural ecosystems across all typical terrestrial biomes in China. The data on climate and ecosystem multifunctionality (per unit land area), represented by GPPmax, WUE, CUE, and LUE, along with climatic factors, were obtained from verified online resources and were further calculated with robustness examination. Moreover, because of the great variations in biotic and abiotic factors, Yan and colleagues transformed every ecological variable (mean = 0, standard deviation = 1) on a comparable scale, thus enhancing the robustness of their findings. First, they found that the three main facets of biodiversity were significantly associated with GPPmax, LUE, and WUE, but not with CUE, when both CWM and ecosystem trait were significantly related to all the four ecosystem functions. Second, species richness alone could only explain 5% of the variation in ecosystem multifunctionality (p = .04), whereas phylogenetic and functional diversity explained an additional 7% (p = .014) and 9% (p = .007) of the variation, respectively. The CWM and ecosystem trait were better predictors than any diversity facets, but the two trait indices were highly collinear. The goodness of fit was better in structural equation modeling (SEM) with ecosystem trait than when using the CWM trait because of the low correlation between the CWM and ecosystem trait for leaf nutrient (i.e., N and P) concentrations. This suggested that the role of ecosystem trait, which is standardized per unit land area and contributes to ecosystem multifunctionality, is more robust than that of the CWM trait, which might constitute various units (e.g., mg/g for leaf nutrient concentration, g for leaf dry mass, and cm2 for leaf area). Third, the SEMs revealed that despite the opposite influences of the three main facets of biodiversity on both ecosystem trait and multifunctionality, biodiversity had a significantly positive effect on ecosystem multifunctionality (total standardized effect: 0.08) only through the ecosystem trait (total standardized effect: 0.19). To our knowledge, Yan et al. (2023) is the first to reconcile the biodiversity paradoxes in ecosystem multifunctionality by creatively introducing the ecosystem trait. They seemingly built a convenient bridge between biodiversity and ecosystem multifunctionality (Figure 1), even with great variability in environmental factors, although this remains to be verified. Moreover, Yan et al. (2023) pointed out that the advantage of using ecosystem trait is that the relationship to ecosystem functioning quantified per unit land area is independent of the area as long as the surveyed quadrats are adequately representative of local biodiversity. This brings new inspiration to scaling theory, which is widely used by researchers in the field of biodiversity and ecosystem functioning (Barry et al., 2021). It is well known that most ecosystem-level functions, such as ecosystem productivity and resource use efficiency, are currently measured on a per unit land area basis, whereas on a per unit land area basis, functions should be constant with area. However, traditional measures of diversity (e.g., species richness) increase with increasing sampling area (i.e., quadrat size). Thus, increasing the sampling extent may decrease the strength (i.e., slope) of the diversity-ecosystem function relationship. Because the ecosystem trait, like ecosystem functions, is quantified on a unit land area basis, using it rather than traditional measures of diversity can avoid this problem. Some basic ecosystem functions (e.g., information transfer) have rarely been included in empirical research. This indicates the need for the integration of various important ecosystem functions, at least these basic functions, in quantifying ecosystem multifunctionality. Hopefully, the study carried out by Yan et al. (2023) will inspire further explorations of biodiversity and ecosystem multifunctionality relationships, as little is known about how extreme temperatures, soil water conditions, and other environmental factors regulate species survival, adaptations, and ecosystem multifunctionality in response to human-induced global changes. The author is thankful to the National Natural Science Foundation of China (32122055 and 32071603) and the National Key R&D Program of China (2022YFF1302801) for providing support and to Professor Xingguo Han for provided helpful comments. The authors declare no conflict of interest. Data used for this commentary can be found in Yan et al. (2023).