The study was conducted at the Changling Grassland Research Station of the Chinese Academy of Sciences, located in Jilin Province, on the southwestern part of the Songnen Plain. Geographically, the site is positioned at 121°30′-123°44′-E, 44°44′-48°40′N, situated in the eastern part of the Inner Mongolian Plateau. The altitude of the area ranges from 138 to 145 meters above sea level. Since 2001, the site has been fenced to prevent grazing, thereby minimizing its impact on the vegetation (Huang et al., 2016). The Songnen grassland, situated near the Mongolia Plateau, experiences a typical continental semi-arid monsoon climate. The study area receives an average annual rainfall of approximately 470 mm, which varies significantly from year to year. In contrast, the annual evaporation rate exceeds 1,500 mm, more than three times the average annual precipitation. The mean annual temperature hovers around 4.9 °C, and the frost-free period lasts between 120 and 150 days. Characterized as a typical meadow grassland, the plant community is predominantly composed of Stipa baicalensis in areas with lower soil salinization. However, the vegetation is complex and diverse due to the influence of soil salinity. In more saline areas, the dominant species are mainly salt-tolerant graminoids such as Leymus chinensis and Puccinellia tenuiflora, and Suaeda glauca, which appear in the area of heavier salinity (Huang et al., 2019).

We selected 7 community types along the soil gradient ( Table 1 ), and ten 1 m×1 m plots were selected for each community. Within each plot, 1–2 A. ramosum plants were randomly selected. Thus,10–20 individuals were taken from each community. A total of 122 individuals of Allium ramosum from 70 plots were sampled.

For each Allium ramosum individuals from different communities, we measured the height and separated the organs into roots, stems, leaves, flowers and bulbs. They were weighed after drying at 65 °C for 48 hours. At the same time, we took three soil samples at 0–15 cm depth in each plot and pooled them as a composite sample to measure soil nutrients. Some of the sieved samples were used to quantify available nutrients. Soil pH and electrical conductivity (EC) were measured using pH meter and conductivity meter in a soil:water suspension (1:5) after 5 min of shaking, respectively. We used potassium dichromate oxidation external heating method to determined soil organic carbon (SOC) content. The total nitrogen content (TN) and total phosphorus (TP) was measured by spectrophotometry. Soil available P was extracted by NaHCO₃ and determined by Molybdate colorimetric method (Murphy and Riley, 1962). Available N was determined by the alkali diffusion method.

The assessment of soil fertility based on individual soil nutrients may not be comprehensive, as the co-limitation of multiple nutrients is a common occurrence. Consequently, we employed principal component analysis (PCA) to condense the complexity of soil nutrients (soil available nitrogen, phosphorus, organic carbon, total nitrogen, total phosphorus, and the N/P ratio) into fewer dimensions using the ‘psych’ package. The use of multiple unrelated nutrients in PCA may reduce the explanatory power of the first axis, but the principal components along the axes better capture nutrient variation along a soil gradient. To select an appropriate axis for the soil nutrient gradient, we also analyzed the correlations between the first four principal axis and each soil indicator.

Community-weighted means were used to describe the height of the plant community.

For each species in the community, we randomly selected five plants to measure height, and subsequently calculated plant community-weighted heights based on the ratio of species biomass in the community.

Coefficients of variation were used to describe the stability of plant communities and species as soil salinity increased. Standard principal axis regression analyses were used to obtain coefficients of biomass partitioning among Allium ramosum organs by ‘smatr’ package, and linear regression models were used to analyze the relationships between soil principal components and community and Allium ramosum traits. Finally structural equation modelling was used for analyzing the direct and indirect effects of soil salinity on plant biomass allocation by ‘lavaan’ package. All analyses were carried out in R 4.2.0.

Soil salinity, as indicated by pH and electrical conductivity (EC), demonstrated an inverse relationship with nutrient content. Specifically, an increase in soil salinity corresponded to a decrease in soil organic carbon (SOC) content and overall nutrient availability ( Figure 1 ). The first principal axis explained 77% of the variation in soil salinity and nutrient indicators ( Figure 1 ). The pH and EC exhibited a negative correlation with the first principal axis, while SOC, total nitrogen (TN) and other nutrient indicators exhibited a positive correlation with the first principal axis ( Supplementary Table S1 ). Consequently, the first principal axis can be interpreted as a representation of the soil salinity gradient, ranging from high salinity and low nutrients to low salinity and high nutrients. A significant and positive correlation was identified between plant community above-ground biomass and soilPC1, suggesting that as soil salinity decreased and nutrients increased, plant community biomass and stability increased ( Figures 2A, B ). The trend of species richness and community above-ground biomass exhibited consistency ( Figures 2C, D ). Soil salinity did not alter the relationship between plant community biomass and diversity; however, species evenness demonstrated a significant decrease as plant community aboveground biomass increased ( Supplementary Figure S1 ). The Allium ramosum population biomass and stability remained relatively unresponsive to variations in soil gradient, although the size of individual Allium ramosum plants exhibited an increasing trend with increasing soil gradient ( Figure 2E, F ). However, the variability in individual size did not demonstrate a significant linear relationship with soil gradient ( Figure 2H ). Concurrently, the height of individual Allium ramosum plants and the biomass of each organ increased with soil gradient, particularly the aboveground biomass. Notably, the flower biomass exhibited no significant change with soil gradient ( Supplementary Figure S2 ).

With soil gradient, the allometric exponent between leaf and flower biomass and between bulb and flower biomass decreased (i.e. plants tended to allocate more biomass to flowers relative to plant growth and storage organs (leaves, bulbs)) ( Figures 3C, G ). Similarly, the allometric exponent between leaves and stems decreased with soilPC1 (more biomass was allocated to stems) ( Figure 3A ). Significant allometric partitioning among other organs of Allium ramosum existed, but did not vary significantly with soil gradients or plant community aboveground biomass ( Supplementary Table S3 , Supplementary Figure S3 ).

The results of the structural equation model showed that the direct effect of soilPC1 on the leaf-flower scaling exponents was not significant, but there was a significant positive effect on the plant community-weighted height, i.e. the plant community-weighted height increased as soil salinity decreased, but the community-weighted height had a significant negative effect on the leaf-flower scaling exponents, and thus there was an indirect negative effect of soilPC1 on the leaf-flower scaling exponents. SoilPC1 and plant community-weighted height together explained 81% of the variation in leaf-flower scaling exponents ( Figure 4A ). Similarly, soilPC1 indirectly influenced the scaling exponents between bulbs and flowers by influencing the aboveground biomass of the community, explaining 94% of the variance ( Figure 4B ).

Soil salinization is a synergistic process of salt accumulation and nutrient loss. Studies have shown that salinity indicators (e.g. pH) are significantly negatively correlated with soil nutrients (organic carbon, nitrogen and phosphorus) in several ecosystems (Corwin, 2021; Garcia-Franco et al., 2021; Huang et al., 2024). Principal component analysis (PCA) further indicated that pH and EC could be the core indicators of soil degradation on the salinization gradient, while organic carbon and nitrogen and phosphorus became sensitive parameters for nutrient loss. The PCA showed that the salinity indicators (pH and EC) showed significant negative correlation with the nutrient indicators (organic carbon, nitrogen and phosphorus) ( Figure 1 ), which revealed the close coupling between soil degradation and nutrient loss in the salinization process. This result is consistent with the general pattern of saline grassland degradation in the global scale (Pan et al., 2013; Heng et al., 2022; Li et al., 2022). For example, in saline soils in western Jilin Province, organic matter and quick potassium content decreased significantly with increasing sodium ion (Na⁺) content and EC, while total phosphorus content remained at a low level for a long time (Cao et al., 2021). This negative correlation can be explained by the following mechanism: ion competition and nutrient fixation: under high salinity conditions, Na⁺ displaces nutrients such as calcium (Ca²⁺) and magnesium (Mg²⁺) in the soil colloid by cation exchange, leading to their leaching or fixation (Zhou et al., 2023). Salt reduces organic matter mineralization and nutrient release by increasing soil osmotic pressure and toxicity and inhibiting microbial decomposition functions (Zhang et al., 2015; Chen et al., 2024). Experiments on the improvement of saline soils showed that the organic matter content of unimproved soils was only 50% of that of improved soils, verifying the negative effect of salts on soil carbon and nitrogen cycling (Xiao et al., 2025). Salinization leads to soil sloughing and reduced porosity, further hindering nutrient uptake by the root system (Garcia-Franco et al., 2021). The process forms a positive feedback of ‘salt accumulation-nutrient loss-soil degradation’, which ultimately leads to a systematic decline in soil fertility.

Natural environmental gradients shape patterns of plant diversity and are important factors influencing plant community dynamics. For example, along large-scale gradients (latitude, elevation), biodiversity tends to increase with energy, temperature and environmental stability, while within regional scales, such as degradation gradients and soil nutrient gradients, also significantly influence plant community structure and function (Zhou et al., 2021; Heng et al., 2022; Wu et al., 2024; Yang et al., 2025). We showed that plant community biomass and diversity decreased significantly with increasing salinization, while community variability increased ( Figure 2 ). Salinization not only directly stresses plant physiology, but also affects community biomass and diversity by changing community structure and resource competition (Zhang et al., 2015). This result is consistent with studies in saline grassland, where salinization led to a large reduction in microbial community diversity (Heng et al., 2022). This change may be related to the fact that salinity inhibits plant photosynthesis and water uptake, leading to a decrease in overall community productivity (Corwin, 2021). At the same time, nutrient deprivation increases competition between species, and only a few salt-tolerant species can maintain population stability through phenotypic plasticity (e.g., reduced individual size, increased reproductive investment) (Tang et al., 2022). On the salinity gradient, local differences in soil physicochemical properties (e.g. pH and salinity patchy distribution) led to increased community variability (Zhang et al., 2015).

There was no significant trend observed in the biomass or stability of Allium ramosum populations along the soil gradient ( Figure 2e ). This phenomenon may be attributed to the physiological plasticity exhibited by Allium ramosum. Individuals within the species demonstrate a capacity to adapt to stressful environments through a reduction in body size, which leads to a decrease in resource requirements, and an increase in phenotypic variation, which can be considered a non-linear response (Huang et al., 2019; Ren et al., 2020; Matesanz et al., 2021). The impact of soil features on Allium ramosum populations may also be influenced by species interactions within the community. Our results demonstrated that individual size, plant height and the size of organs of Allium ramosum increased significantly with soil gradient ( Figure 2 , Supplementary Figure S2 ). These correlations indicate that natural selection may have prevented independent evolution of traits, suggesting a need for coordination among traits contributing to the same function (He et al., 2020; Zhou et al., 2022; Bin et al., 2024). The observed correlations among these traits can be regarded as a facet of the broader coordination between traits and organs in plants (Matesanz et al., 2021; Sanchez-Bermejo et al., 2023). Furthermore, in accordance with other studies, environmental stress did not modify the allometric distribution of biomass among organs, yet the allometric indices and constants exhibited a significant trend with soil salinity (Eziz et al., 2017; Peng et al., 2022). Growth and reproduction represent two of the most fundamental processes in plants, with the biomass of leaves, stems and roots determining the ability to capture light and access soil resources to provide photosynthetic products and nutrients for reproduction (Poorter et al., 2012; Yan et al., 2016; White et al., 2022). Allium ramosum exhibit an adaptive strategy for biomass allocation priority across salinity gradients. As soil salinity increase, the allocation of biomass to sexual reproductive organs (e.g. flowers) exceeds that allocated to nutritive organs (leaves, bulbs) ( Figure 3 ).

Based on a field study conducted in the Sugan Lake wetland on the Qinghai-Tibet Plateau. Soil salinity conditions across three distinct habitats: inland salt marsh, oasis wetland, and seasonal river wetland significantly influenced Saussurea salsa biomass allocation and morphology. Under high salinity, plants developed large, thick leaves with low specific leaf area (SLA) and formed roots with moderate diameter and length, allocating minimal biomass to roots to mitigate water stress and ion toxicity by enhancing water storage in leaves and reducing root exposure (Li et al., 2021). This strategic allocation of biomass ensures the survival of offspring by increasing reproductive output under stressful conditions (Zhang et al., 2021a; Van de Walle et al., 2023; Stott et al., 2024). This variation in the strategy for biomass allocation is also subject to change in response to climate change and across large climatic and soil gradients, where increased resource stress (e.g. drought, low soil nutrient content) is experienced. Artemisia spp. exhibit increased biomass for reproduction allocation and decreased biomass allocation to leaves (Tsogtsaikhan et al., 2025). The decrease in leaves may be due to individual size limitation and reduced water consumption from transpiration, since more saline soils have poorer water-holding capacity (Zhou et al., 2023). However, it is important to note that the results of structural equation modelling suggest that this allocation pattern does not constitute a direct response to soil features, but is mediated through changes in community above-ground biomass and structure ( Figure 4 ). The effect of soil salinization on the pattern of biological allocation between leaves and flowers of Allium ramosum was found to be indirectly influenced by affecting the community-weighted height of the plant. The decrease in community weighted height was attributed to salinization, resulting in the replacement of dominant species with low salt-tolerant species (Janousek and Folger, 2014; Token et al., 2022). At higher community heights, light competition is increased and plants allocate more biomass to growth, favoring plant extension for more light (Xiao et al., 2021). One reason why plants allocate fewer resources to reproduction may be mechanical limitations of mechanical support when plants are taller (West et al., 1999). According to Corner’s rules, larger inflorescences require thicker stems to support them (Fajardo et al., 2020). Hence, for Allium ramosum, taller stems is required so that the flowers can reach a sufficient height to favor pollination and fruit set. If plants devote more biomass to producing taller and thicker stems, this can lead to a less competitive plant, with stems being considered luxury organs (Xiao et al., 2021), especially in Allium ramosum plants, which only serve to support the inflorescence. Thus, as the community weighted height of the plant increases, the plant tends to allocate more biomass to leaves rather than flowers ( Figure 4a ).

It is also noteworthy that Allium ramosum is a perennial plant and that there will be a trade-off between storage and reproduction (Zhang et al., 2022). Our results also indicated that as soil salinity increased, plants similarly allocated more biomass to reproduction than to storage ( Figure 4b ). This strategy may reduce competition for light resources due to declining biomass in the plant community, prompting Allium ramosum to divert resources to reproduction in order to extend their dispersal advantage (Bonser, 2013). Conversely, in communities with less saline soils, increased aboveground biomass caused plants to allocate resources inside the storage organ - bulbs, to allow plants to have larger plants in the next growing season, as bulb size was significantly correlated with individual plant size (Howard and Cellinese, 2020; Ashagrie et al., 2021).

Notably, phosphorus has been found to be an essential element for plants, influencing the entire reproductive process, from bud differentiation to seed maturity (Fortier and Wright, 2021; Velez-Mora et al., 2024). At the intraspecific level, P-limited conditions resulted in later flowering onset, shorter individual flowering duration, and reduced flower/inflorescence production per plant. Interspecifically, species adapted to P-limited environments exhibited earlier flowering onset, longer seed stalks and panicles, but also shorter flowering periods and fewer flowers per plant (Wang et al., 2022). Critically, P limitation consistently constrained investment in sexual reproduction (e.g., reduced flower production, shorter flowering periods), potentially impairing dispersal capacity. Significant confounding effects of soil pH and moisture were also revealed which covaried with nutrient regimes—on reproductive traits, complicating the interpretation of N:P effects (Wang et al., 2022).

Based on the analysis of 599 Eurasian herbaceous sites, phosphorus (P) limitation (indicated by high N:P ratios in plant biomass) strongly influences plant reproductive strategies, with significant implications for endangered species. Plants in P-limited communities exhibit markedly reduced investment in sexual reproduction compared to nitrogen (N)-limited communities, manifested through lower seed production, diminished seed mass, shorter flowering periods, delayed flowering onset, and greater reliance on vegetative propagation and perennial lifespans (Fujita et al., 2014).

The experimental study investigated how absolute and relative nitrogen (N) and phosphorus (P) supply affect sexual reproduction traits in a common grass (Holcus lanatus) and an endangered forb (Parnassia palustris) demonstrated that the effect of N:P supply ratio on sexual reproduction investment is critically dependent on the absolute nutrient supply level: at low absolute nutrient supply, N:P ratio had minimal impact on reproduction traits, whereas at high absolute nutrient supply, a high N:P ratio (indicating low relative P availability) significantly reduced investment in sexual reproduction for H. lanatus (Wang et al., 2019). Low relative P-supply (high N:P ratio) restricted the positive response of sexual reproduction traits to increased absolute nutrient supply, essentially limiting the potential benefits of higher nutrient availability. While data for the endangered P. palustris were limited due to high mortality, its survival patterns mirrored the reproduction response of H. lanatus, suggesting similar constraints under low relative P-supply at high nutrient levels (Wang et al., 2019). This is consistent with our results that, as a perennial nondominant species, reproduction of Allium ramosum may be more strongly limited by phosphorus along the soil gradient ( Figure 1 , Supplementary Table S1 ).

The strategy for biomass allocation at the plant community level also produces adaptive variation in response to environmental change (Fant and Ghedini, 2024; Wu et al., 2024). A strategy for biomass allocation prioritizing survival is favored in arid environments, where more biomass is allocated to the root system for water, whereas investment in the above-ground fraction is prioritized in wetter areas with better soil nutrient (Wu et al., 2024). Global-scale studies indicate that the species composition and diversity of plant communities can buffer the effects of environmental gradients on biomass allocation (Skarpaas et al., 2016; Dolezal et al., 2021). For example, species turnover reduces the degree of variability in the overall allocation strategy of the community through functional trait complementarities (Zhou et al., 2021; Fang et al., 2023; Krak et al., 2025).

Recent studies of biomass allocation patterns have demonstrated considerable variation in this ratio across diverse environmental contexts (Chen et al., 2021; Beccari and Carmona, 2024; Dolezal et al., 2024). However, the majority of these studies have been conducted on individual plants, focusing exclusively on the abiotic environment and species traits, without considering the interactive dynamics among species within communities (Zhang et al., 2021a; Zhou et al., 2022; Krak et al., 2025; Tsogtsaikhan et al., 2025). Consequently, our understanding of how plants respond to the combined effects of soil salinity stress and species competition remains limited. The findings of this study contradict traditional, simplified models of direct soil-plant response and underscore the critical role of community-scale processes in individual adaptation strategies. At the community level, interspecific roles and environmental stresses may interact, complicating the individual to community level scaling transitions (Zhou et al., 2021; Fant and Ghedini, 2024). Consequently, we would like to propose that future studies include more characteristics of plant communities and populations when analyzing strategies for biomass allocation in response to environmental gradients or environmental changes.