2.1 Study Region and Field Sampling

This study region (27.4°–48.7° N, 76.2°–109.8° E) is located in western China, which covers an area of 4.9 × 10⁶ km² (Figure S1). Western China is considered an amplifier for climate change and an ecologically fragile area in response to water resource changes (Song et al. 2022). Topographically, this region spans a wide altitude (195–5127 m) and is surrounded by high mountains, resulting in an alpine and arid climate (Yao et al. 2022), with mean annual air temperature (MAT) and mean annual precipitation (MAP) ranging from −4.7°C to 12.5°C and from 17.2 to 685.4 mm, respectively. Due to the complex topographies and unique climates, this region is abundant in lakes, rivers, wetlands and numerous small water bodies (e.g., ponds, ditches), which provide an ideal platform for the formation, development and evolution of wetland plants (Li et al. 2020; Wang et al. 2015).

Field sampling was conducted for 84 plant species in 232 wetland sites across western China during the growing season of 2018–2019. The 84 plant species were divided into three life forms: emergent plants (grow in waterlogged soil with their roots submerge in water and leaves in air), floating-leaved plants (leaves float on the surface of water with their roots in sediments) and submerged plants (whole bodies are submerged in water with their roots in sediments). All species nomenclature in this study followed the Flora of China (https://www.iplant.cn/foc). For the emergent plant species of Hippuris vulgaris and Schoenoplectus triqueter, their aboveground parts are sometimes submerged in water. In this case, we added the suffix ‘_sub’ to those emergent species to indicate their submerged eco-type. For each species in each site, we sampled 30–50 intact leaves (broad-leaved species) or segments (20 cm from the tips of submerged plants with filiform leaves) and mixed them as the leaf sample for measuring bioelements in each site (Zuo et al. 2022). We then set 1 m × 1 m quadrats in 225 sites to investigate the community and collected the whole aboveground biomass for each species. Three samples from the upper 0–20 cm of sediments were collected randomly in each site. After removing roots, plant residues and rocks, these samples were used as sediment samples for the site (Zuo et al. 2024). Finally, we collected 1058 plant samples for the 84 species, 232 sediment samples in 232 sites and 683 biomass samples for 59 species in 225 sites.

2.2 Chemical Analysis and Data Collection

Plant leaf samples were oven-dried at 70°C for 72 h and then ground into fine powder passing through a 0.25 mm mesh sieve. Plant biomass samples were sorted by species and oven-dried at 70°C for 72 h as the dry weight of the species. Sediment samples were air-dried, then ground and sieved through a 0.15 mm mesh sieve.

Total nitrogen (N) and sulfur (S) concentrations of both plants and sediments were measured by an elemental analyser (UNICUBE, Elementar, Germany). The total concentrations of the other 14 bioelements (P, phosphorus; K, potassium; Ca, calcium; Mg, magnesium; Na, sodium; Al, aluminium; Fe, iron; Mg, manganese; Si, silicon; B, boron; Zn, zinc; Cu, copper; Ni, nickel; Co, cobalt) of plants and sediments were measured by inductively coupled plasma mass spectrometry (ICP-MS-2030, Shimadzu, Japan). We provided the statistical summary of sediment elemental composition of 232 wetland sites in this study (Table S1).

The climate variables of each site were extracted from the climatic raster database that was downloaded from the world climate website (http://www.worldclim.org), with a finest spatial resolution of 1 km (30 s). We introduced growing season mean temperature (GST) and growing season precipitation (GSP) to explain the effects of climate on plant growth and community biomass. Growing season in the study region was defined as the 5 months from May to September (Wang et al. 2015).

2.3 Data Analysis

3.1 Homeostatic Regulation Coefficients (H) of Bioelements at the Species and Community Levels

For all of the 16 bioelements, leaf elemental concentrations positively correlated to nutrient supply in sediments, that is, bioelements showed statistically significant H values as described by homeostasis models at both species and community levels (p < 0.01, Figure 1 and Table S4). At both levels, four of six macroelements (N, S, K, Ca but except P and Mg) exhibited ‘homeostatic’ and ‘weakly homeostatic’, while all 10 microelements (Na, Fe, Al, Mn, Si, B, Zn, Cu, Ni and Co) exhibited ‘weakly plastic’ and ‘plastic’ (Figure 1). The H values showed an increasing trend as the elemental concentrations increased at the species level (p = 0.10, Figure 2a) and were significantly negatively correlated with CV at both species and community levels (p < 0.01, Figure 2b), indicating that bioelements with higher concentrations showed lower variability and higher elemental homeostasis.

For the 12 common species, the relationships between leaf P, K, Ca and Na concentrations and their corresponding sediment nutrients could be significantly described by the homoeostasis model, with mean H values of P (H_P), K (H_K), Ca (H_Ca), and Na (H_Na) as 1.61 (0.79–2.52), 1.84 (0.99–2.70), 2.88 (1.57–4.41) and 1.59 (1.03–2.24), respectively (Table 1 and Figures S2–S5). H_P, H_K, H_Ca, and H_Na of 6 submerged plant species, excluding N. marina, were higher than those of the four emergent plant species (Table 1). Then, we defined the species with H_P, H_K, H_Ca and H_Na higher than the average of the 12 species as ‘high-H species’ (S. pectinata, M. spicatum, P. perfoliatus, R. bungei, P. natans and P. pusillus), while those that had lower H values (S. filiformis, H. vulgaris, N. marina, H. tricuspis, T. palustris and E. dulcis) were ‘low-H species’.

3.2 Relationships Between Elemental Homeostasis and Species Performance

At the species level, we identified significant positive relationships between species H_P and species performance, including species abundance, biomass, dominance and stability (Figure 3a,e,h,i). Similar positive relationships were observed for H_K, H_Ca and H_Na (Figure 3), suggesting that species with higher elemental homeostasis exhibited better species performance. Furthermore, the homeostasis coefficients of these bioelements were significantly positively correlated with each other (Figure S6). Additionally, species abundance, biomass, dominance and stability were positively correlated with each other (Figure S7).

3.3 Relationships Between Elemental Homeostasis and Community Structure and Biomass

At the community level, elemental homeostasis (H_P, H_K, H_Ca and H_Na) was positively correlated with community biomass of submerged plants, but not in other community types (Figure 4a–d). Additionally, total biomass in communities dominated by high-H species was higher than that in communities dominated by low-H species (366.6 g m⁻² vs. 243.6 g m⁻², p < 0.05). SEM explained 31% of the variation in community biomass (Figure 4e). Community H had the largest direct effect (path coefficient = 0.35), followed by community elemental concentration (0.22) and climate (0.19). Climate, sediment nutrients and community elemental concentration also regulated community biomass through indirect effects (Figure 4e).

4.1 Multi-Elemental Homeostasis of Wetland Plants: From Species to Community

Stoichiometric homeostasis (H) has been considered a reliable indicator that reflects the form and function of species, along with the ecological strategies they follow (Julian et al. 2020; Yu et al. 2015). However, only a few studies have documented the homeostatic level of bioelements beyond N and P, especially at the community level (Zhao et al. 2020). This study applied the homoeostasis model to 16 bioelements of 84 plant species from 232 wetland ecosystems in western China. Consistent with our hypothesis H1, the homeostatic level of bioelements in plants enhanced with increasing concentrations (Figure 2), confirming the similar pattern Karimi and Folt (2006) found in zooplankton.

Macroelements, which have higher requirements in plants, often serve as key limiting factors for ecosystem productivity (Elser et al. 2007; Han et al. 2011), and maintaining their concentrations is more advantageous for the stability of species and communities (Yu et al. 2010). Physiologically, as important components of organic compounds and macromolecules, the higher stability of macroelements also contributes to the stable supply of energy (Kumar et al. 2021). However, microelements show lower physiological requirements and looser functional coordination with other bioelements and always fluctuate with the environmental supply without significantly affecting the whole elementome in organisms (Han et al. 2011; Kumar et al. 2021; Zuo et al. 2024). Correspondingly, some microelements (e.g., Fe, Al, and B) showed H < 1 at both species and community levels (Figure 1), reflecting element-specific stoichiometric response patterns. These patterns are driven by the mechanisms that maintain the stability of different bioelements, involving processes such as acquisition, assimilation and recycling or turnover (El-Sabaawi et al. 2023; López-Sepulcre et al. 2024). H < 1 indicates that these microelements are highly plastic and vary more in lockstep with environmental nutrient. In this situation, organisms may rely more on external nutrient absorption rather than merely on internal regulation to maintain the homeostasis of microelements. In contrast, the high homeostasis and strong phylogenetic conservatism of macroelements are evolutionary consequences of maintaining specific biological structures and functions, which further supports the stability of limiting elements hypothesis (Han et al. 2011) from the perspective of stoichiometric homeostasis. Based on these findings, we proposed a conceptual framework for multi-elemental homeostasis of wetland plants and suggested that bioelements with higher concentration exhibit lower variability and higher homeostasis (Figure 5). Using this framework, we can explore the strong potential for homeostasis in some cases. One potential is to help determine the optimal elementome of species, and further test the effects of environmental changes on species biogeochemical niches (Peñuelas et al. 2019). The variations in those bioelements with less homeostasis may be the key in leading to the contraction or expansion of biogeochemical niches under environmental changes. Another potential lies in plants utilising the homeostatic differences among bioelements to regulate their nutrient uptake. For example, species with high-H_N may have an advantage in balancing N uptake, thus avoiding toxicity in N-rich environments. Notably, we used sediment total elemental concentration as environmental nutrient supply to calculate the elemental homeostasis coefficients, although total concentration cannot separate the structural and metabolic/available components. Nevertheless, given that total concentrations reflect the long-term nutrient conditions and supply capacity of soils, this method is reasonable and practical for large-scale field studies (Cheng et al. 2023; Ci et al. 2022; Su et al. 2019).

4.2 Linking Multi-Elemental Homeostasis With Species Nutrient Strategies

At species level, significant H_P, H_K, H_Ca and H_Na values were detected for all the 12 common species, indicating that these four bioelements showed significant stoichiometric homeostasis in these species. Within these species, all H_P, H_K, H_Ca and H_Na of submerged plants were basically higher than that of emergent plants (Table 1), reflecting the different nutrient utilisation strategies between the two plant life forms in their unique habitats. Submerged plants, whose whole plant is submerged under water, can absorb nutrients both from sediment via roots and from water via leaves (Denny 1972; Su et al. 2019), thus reducing their dependence on sediment nutrient availability and enhancing their potential to maintain a relatively stable elementome. Generally, natural freshwater ecosystems are considered to be P-limited (Elser et al. 2007). Long-term selective pressures have resulted in submerged plants possessing stable P acquisition and storage capacities, thus improving their adaptability to P-limitation habitats (Güsewell and Koerselman 2002; Li et al. 2018). Besides P, submerged plants also tend to evolve higher homeostasis of K and Na to maintain ion balance and osmotic regulation under water (Rubio et al. 2020). In addition, plant Ca is extensively bound as Ca-pectate (Pectate₂Ca₁₀) to participate in cell wall construction (Marschner 2012; Xing et al. 2021), and thus, the higher Ca homeostasis of submerged plants could improve cell wall load-bearing capacity and their adaptation to hydraulic forces.

Distinct from terrestrial ecosystems (Yu et al. 2010), H_P, rather than H_N, significantly correlated with species biomass, dominance and stability (Figure 3), indicating that P may control ecological processes more closely than N in freshwater ecosystems (Su et al. 2019). As a fundamental component of DNA, RNA and ATP, P directly regulates genetic information transfer, protein synthesis and energy metabolism, thereby influencing the growth rate of species (Isanta-Navarro et al. 2022; Sterner and Elser 2002). From the perspective of species nutrient economics (a continuum of plant strategies from nutrient conservative to nutrient acquisitive), plants with higher elemental homeostasis are conservative nutrient users and less dependent on resource availability (Yu et al. 2011), that is, high-H species could conservatively use resources to achieve stable growth in resource-limited environments by reducing leaf nutrients, increasing nutrient use efficiency and elementome plasticity (Ci et al. 2022; Fernández-Martínez 2022). In contrast, low-H species are luxurious nutrient users which tend to survive in resource-rich environments (more opportunistic). Additionally, we found that higher species H_K, H_Ca, and H_Na significantly enhanced species performance. Functionally, K serves as a co-factor for over 60 enzymes, including ATP synthase and Rubisco. Species with high H_K promotes the coordination between light reactions (ATP production) and dark reactions (CO₂ fixation), thereby enhancing plant productivity (Johnson et al. 2022). Ca plays critical roles in cell wall structure, membrane stability and signal transduction, with its homeostasis contributing to mechanical resilience and stress resistance of plants (Xing et al. 2021). Species with high H_Na values maintain species stability by effectively regulating osmotic pressure and avoiding ionic toxicity in saline environments (Kaspari 2020). Furthermore, the homeostasis coefficients of these bioelements were positively correlated with each other (Figure S6), suggesting their coordinated interactions may constitute a ‘multidimensional buffering mechanism’ (e.g., P-mediated K uptake via H⁺-ATPase, K-activated Rubisco relying on P substrates for C fixation and P-Ca signalling pathway interaction, Marschner 2012). These integrated mechanisms allow high-H species to alleviate metabolic constraints and optimise resource use efficiency in fluctuating or limited nutrient environments, ultimately dominating community structure and functioning.

We also found that the higher the species H_P, H_K, H_Ca and H_Na, the greater the regional species abundance (Figure 3a–d), which is consistent with the findings in woody plants (Ci et al. 2022). High-H species (S. pectinata, M. spicatum, P. perfoliatus) are widely distributed in western China, even in stressful habitats such as oligotrophic, cold, and saline water bodies (Wang et al. 2015; Zuo et al. 2022). Conservative nutrient utilisation strategies and high homeostasis should be the key drivers in forming the species distribution and productivity patterns. In agreement with our hypothesis H2, high elemental homeostasis increases the species dispersal potential by alleviating nutrient limitation and strengthens the species performance in the community by enhancing the stability and biomass of the species.

4.3 Mechanism and Framework Linking Elemental Homeostasis With Community Structure and Functioning

When scaling up from species to community level, community H of P, K, Ca and Na of submerged plants was positively correlated with community biomass as expected (Figure 4a–d), which proved that stoichiometric homeostasis is an important mechanism in maintaining plant community structure and functioning. Firstly, the strength of species homeostasis determines their nutrient economics, thus affecting interspecific competition and coexistence (Peñuelas et al. 2019; Reich 2014). Species with higher elemental homeostasis may better sustain competitiveness and dominance in environments where nutrient availability fluctuates greatly (Ci et al. 2022). Secondly, dominant species occupy most resources and niches within the community, leaving other species reliant on remaining resources for survival (Hooper et al. 2005; Yu et al. 2010), and thus control the community structure and biomass (Wang et al. 2013). Beyond community H_P (Su et al. 2019), for the first time, we identified the significant roles of community H_K, H_Ca and H_Na in indicating community productivity. The irreplaceable functions of K, Ca and Na in nutrient absorption, osmotic regulation and physiological metabolism are conducive to establishing the primary productivity and dominance of wetland plants (Kaspari 2020; Rubio et al. 2020). Especially for submerged plants, leaf hydraulic economy and osmotic homeostasis regulated by K⁺ and Na⁺ were strongly correlated with plant growth and biomass accumulation (Rubio et al. 2005; Sardans and Peñuelas 2015).

Previous studies have shown a decreasing trend in the strength of relationships between community productivity and community H from short-term to long-term experiments (Yu et al. 2010). This suggests that environmental factors, such as climate and nutrient, should be additionally considered when linking instantaneous community H with long-term community productivity (Bai et al. 2004), especially in highly heterogeneous habitats. Consistent with our hypothesis H3, SEM identified that climate, sediment nutrient, community elemental concentration and community H co-regulated community biomass of submerged plants (Figure 4e). These findings suggested that multi-elemental homeostasis generates positive feedback in driving community structure and functioning: high-H species exhibited survival advantages over other species within the community, accompanied by more conservative nutrient strategies and wider regional distribution, thus resulting in higher species stability and community productivity (Ci et al. 2022; Yu et al. 2015). This feedback expands the linkages between multi-element homeostasis and ecological performance (e.g., species performance and community functioning). Therefore, we incorporated Ca, K, and Na into the stoichiometric homeostasis-productivity framework on the basis of P (Figure 5), and strongly advocate for a further comprehensive assessment of the potential influence of H_Ca, H_K and H_Na on species and community ecological processes.