The research was carried out at the Zoige Alpine Wetland Ecosystem Research Station (3440 m, 33°51′52″N, 102°08′46″E) in the eastern Tibetan Plateau. The area has a typical plateau continental semi-humid climate with no frost-free period throughout the year. The annual average air temperature is 1.1°C, and there is around 600 mm of precipitation per year, mostly during the growing season from June to September.

Based on the categorization by the US Department of Agriculture, the soil type in the research area was identified as silt clay loam, consisting of 31.2% sand, 56.0% silt, and 12.8% clay in the top 30 cm of soil. Elymus nutans and Kobresia setschwanensis are the dominant species in the study area, while other associated species include Potentilla anserina, Roegneria nutans, Poa pratensis, Plantago depressa, Leymus secalinus, and Ajania tenuifolia, etc. We choose E. nutans and P. anserina as the focal species for investigating the impacts of warming on alpine plants. The two species representing two major plant functional groups (grasses and forbs) and naturally co-exist within our study site. Elymus nutans is the dominant perennial grass species in the alpine meadow. Potentilla anserina is a common companion species and is widely distributed in alpine meadows. Due to its wide ecological amplitude and vegetative reproduction ability, P. anserina is considered a prime species for ecological restoration in alpine regions (Wu et al., 2022). Our field investigation at the study site found that E. nutans and P. anserina together account for nearly 60% of the total vegetation coverage. E. nutans and P. anserina display distinct stratified structures. E. nutans predominantly occupies the upper part of the community and grow in full-sun conditions, while the leaves of P. anserina were restricted to the lower parts of the community close to soil surface. Consequently, they display distinct resource utilization patterns, particularly in terms of light and soil nutrients (Zhou et al., 2021). The upper leaf layer within the canopy typically absorbs light beyond its saturation point and dissipates excess energy through heat dissipation mechanisms. In contrast, lower layer leaves often face limitations due to insufficient available light.

In April 2015, we launched our in-situ warming experiment using open-top chambers (OTCs) to assess the impacts of warming on the alpine meadow ecosystem. The OTC had a 6.4 m² surface area, a height of 2 m, a bottom side length of 1.15 m, and a regular octagon form with an outside diameter of 3 m. Open areas (OAs) were developed as control regions with characteristics comparable to those of the OTCs. The detailed layout of the experimental design was mentioned in our previous articles (Zhou et al., 2021).

The ambient temperature (T _a, °C) and relative humidity (RH, %) were measured every 30 minutes by HOBO (U23-002, Pocasset, MA, USA), which was placed in the middle of the OTCs and OAs at a height of 1.5 m above the soil surface. Using an ECH₂O-TE sensor and EM50 data collecting system (Decagon Devices, Inc., USA), we automatically monitored soil temperature and soil moisture (volumetric soil moisture, V/V%) at a depth of 5 cm with 30-minute intervals during the experiment. Following four years of warming, the air temperature consistently increased year over year. In 2015–2018, the daily mean air temperature in the OTCs was 0.65°C, 0.74°C, 0.75°C, and 0.75°C higher than that in the OAs respectively. The average soil temperatures over the four-year period were recorded as 14.0°C in the OTCs and 13.2°C in the OAs. Warming led to a decline in soil moisture, with the OTCs experiencing reductions of 12.5%, 13.4%, 16.7%, and 10.1% during the growing seasons of 2015-2018. The lowest amount of precipitation during the growing season was recorded in 2015 (435 mm), while the highest amount was recorded in 2018 (669 mm) (see details in Supplementary Figure S1 ).

The gas exchange properties were measured using a portable photosynthetic system (LI-6400, LI-COR, Lincoln, USA). The experiments were conducted during the vigorous growth period (mid-July and mid-August) in each of the years spanning from 2015 to 2018. For comparative analysis, we selected 2015 (representing short-term warming) and 2018 (representing medium-term warming). The measurements were typically taken three times per month on clear days. Fully expanded and exposed leaves were selected, and measurements were taken from 09:00 to 13:00 with local time (which is 72 minutes later than Beijing time). For each replication, three individuals in a similar healthy state were selected. The following parameters were measured: net photosynthetic rate (P _n), transpiration rate (T _r), stomatal conductance (g _s), stomatal limitation (L _s), and intercellular CO₂ concentration (C _i). After that, the leaves were collected and their areas were scanned and precisely calculated using Image J software (version 1.47v, USA).

Leaf light use efficiency (LUE, mmol·CO₂·mol^-1·photons) was calculated based on the Equation 1:

Where PAR_i represents the incident photosynthetic active radiation (μmol·m^–2·s^–1).

A portable modulated chlorophyll fluorometer was used to assess the chlorophyll fluorescence characteristics immediately following photosynthesis measurements in the same plant (PAM-2100, Walz, Germany). The leaves were pre-adapted in the dark for 30 minutes before being measured hourly from 09:00-13:00 h. The maximal photochemical efficiency of PS II (F _v/F _m), photochemical quenching (qP), non-photochemical quenching (qN), and effective photochemical efficiency (Yield) of the chlorophyll fluorescence parameters were measured.

Fresh leaf samples were collected from the sunny side of each species, immediately weighed (fresh weight, FW), and then submerged in distilled water in the dark until saturated. After determining the saturated fresh weight (SW), the leaves were oven-dried at 70°C for 48 hours to estimate the dry weight (DW). The Equation 2 was used to obtain the relative water content (RWC, %) of the leaf:

The leaf area of the fresh leaves was measured using Image J, and specific leaf area (SLA, cm²·g^-1) was calculated as the ratio of one side’s area of each leaf in each set to its dry mass (Perez-Harguindeguy et al., 2013). The drainage method was used to compute the specific leaf volume (SLV, cm³·g^-1), which is the ratio of leaf volume to dry mass. Leaf dry matter content (LDMC, mg·g^-1) was computed as the ratio of leaf dry weight to leaf saturated fresh weight. In mid-July and mid-August, three repetitions of measurements were conducted for each leaf trait index of the two species.

After sampling, the leaves were mixed to form composite samples from ten to fifteen different individuals. The materials were crushed, sieved through an 80-mesh screen, and then dried at 70°C for 24 hours before being packed in plastic bags for measurement. A Vario Macro Cube Elemental analyzer (Elementar, Hanau, Germany) was used to assess total contents of carbon and nitrogen. Total phosphorus contents were measured using a molybdenum antimony resistance colorimetric method. For each species, the measurements were repeated three times.

The LI-8100 automatic soil CO₂ flux system (LI-COR, Lincoln, USA) and its 20 cm survey chamber (8100-103) were used to assess soil respiration (R _s). The soil collars (8100-103) were positioned with their tops 2-3 cm above the soil surface and installed one day before the measurements. The litter in the collar was cleared before each measurement, and the above-ground of the plant was subtracted.

After 24 hours of equilibration, the soil respiration rate (R _s, μmol·m^-2·s^-1) returned to its initial level before collar insertion. Measurements were taken once per hour for two minutes between 9:00 to 13:00 with three repetitions. The R _s was computed based on the Equation 3:

where S _C is the area of soil that the survey chamber covers (0.03 m²). V _C (m³) is computed as the sum of the volume of the 20 cm survey chamber (4.82×10^-3 m³) and the product of the chamber offset (the distance from the settled collar to the ground inside it) and the soil area (S _C), which represents the total volume of soil respiration system. ∂ C s ∂ t is the rate of change in chamber CO₂ during soil respiration measurements (μmol CO₂ mol^-1 s^-1), where P is the atmospheric pressure (P _a), T is chamber air temperature (°C), and R is the gas constant (8.314 Pa m³ mol^-1 K^-1).

The potassium dichromate oxidation titration was used to measure the content of soil organic carbon (SOC). The semi-trace Kjeldahl technique was used to measure the soil total nitrogen (TN), and the vanadium molybdate blue colorimetric method was employed to determine the soil total phosphorus (TP) (Xu et al., 2022). Alkaline hydrolysis was used to estimate soil available nitrogen (AN), and the molybdenum blue technique was employed to evaluate soil available phosphorus (AP) after extracting soil samples with sodium bicarbonate. After extraction with ammonium acetate, soil available potassium (AK) was measured using a flame photometric method. Using an autoanalyzer (SmartChem140, AMS Alliance, Italy), the content of soil NH₄ ⁺-N was determined in extracts of 2 M KCl (1:4, soil: extractant).

The data was statistically analyzed using SPSS 20.0 (SPSS Inc., Chicago, USA), and the results were presented as means ± standard error (SE). Two-way ANOVA analysis was used to compare the photosynthetic physiological characteristics of two different species under varying warming treatments and years, followed by post hoc Duncan multiple comparison for further analysis. Principal component analysis (PCA) was employed to compare the variance in leaf traits among species and treatments over the course of four years. The plspm package in structural equation modeling (SEM) was utilized to investigate the influence of leaf functional traits and environmental factors on leaf photosynthetic capacity. The analyses were conducted using R program v3.4.4, and figures were generated with Origin Pro 2021 (OriginLab Corporation, United States).

The data representing the interaction pathways between soil nutrients, atmospheric CO₂, leaf traits, and leaf photosynthetic capability in response to warming was adequately fit by structural equation models (SEM). It can be seen that soil nutrients have significant effects on leaf traits both in OTCs and OAs. The SEM analysis revealed that warming impacts leaf photosynthetic capability indirectly through soil nutrients, atmospheric CO₂, and leaf traits in both OAs and OTCs. Firstly, the temperature had a significant effect on atmospheric CO₂ and leaf traits in OTCs but not in OAs. Secondly, the photosynthetic capability of OTCs was more affected by temperature variations than that of OAs. Thirdly, the effects of temperature variation on leaf traits were greater in OTCs than in OAs. Furthermore, the correlation between leaf traits and photosynthetic capacity increased in OTCs, while the correlation between soil nutrients and photosynthetic capacity decreased ( Figure 5 ). Finally, atmospheric CO₂ had a significant effect on leaf photosynthetic capacity in OAs but not in OTCs, indicating that the effect of atmospheric CO₂ on leaf photosynthetic capacity decreased with increasing temperature.

A PCA-Biplot was used to compare the variance in leaf traits among different species and treatments over a period of four years. The first and second PC axes explained 27.3% and 20.1%, respectively. For the two species, E. nutans showed a strong correlation with L _s, SLA, RWC, TN, TC, NP, qN, Yield, F _v/F _m and LDMC. P. anserina exhibited a strong correlation with P _n, G _s, T _r, LUE, SLV, TWC and TP (see the Supplementary Figure S2A ). The warming treatment (OTCs) had a significant impact on the photosynthetic capability (photosynthesis and fluorescence) of species (see the Supplementary Figure S2B ).

Alpine plants are largely restricted by low temperatures, and warming might directly reduce the impact of low temperatures on plant growth, even change the community structure and species composition of alpine meadows (Li et al., 2018; Zhou et al., 2021). Photosynthesis can reflect the physiological adaptability of plants under specific environmental conditions. Several studies suggest that alpine plants exhibit higher photosynthetic capacities and leaf nitrogen concentrations compared to the global average (Wright et al., 2004; He et al., 2006). The majority research has shown that warming has positive effects on plant photosynthesis. Fu et al. (2015) discovered that warming significantly increased in the P _n of the alpine plants on the Tibetan Plateau, which was related to the increase of stomatal conductance (g _s), chlorophyll (Chl) content, yield, and non-photochemical quenching of Chl fluorescence. Carroll et al. (2017) found that the photosynthesis of three dominant species (Pinus contorta, P. ponderosa, and Populus tremuloides) had different responses to warming. Climatic warming also affects leaf photosynthetic physiological parameters, such as Chl fluorescence, g _s, and intercellular CO₂ concentration (C _i), which are all temperature-dependent (Ruiz-Vera et al., 2013). In our research, higher g _s was linked to better photosynthetic carbon absorption capability. The g _s affects CO₂ diffusion from the atmosphere into the intercellular space of leaves, and high g _s promotes plant photosynthesis and C assimilation (Sugiura et al., 2020). Moreover, we observed a significant difference in the g _s of both species between the two durations of warming. Specifically, compared to the short-term warming, the two species exhibited a decrease in g _s during the medium-term warming, which can be attributed to a simultaneous decrease in P _n.

Photosynthesis provides a comprehensive depiction of a plant’s physiological condition, which can be used to quantify growth differences between plants and the degree of environmental impact (Lin et al., 2017). The response of photosynthesis may differ between short-term and medium-term warming. Under short-term warming, the P _n of E. nutans decreased while that of P. anserina increased. However, the P _n of the two species had an opposite trend during medium-term warming. The P _n of E. nutans increased, while that of P. anserina decreased. Phenotypic plasticity is widely recognized as a primary mechanism by which plants adapt to variations in environmental factors, serving as an observed adaptation to short-term fluctuations in the environment (Gratani, 2014). The contrasting effects of short-term warming on P _n and T _r for E. mutans versus P. anserina, possibly attributed to phenotypic plasticity rather than adaptation. In the alpine meadow, E. nutans was the dominant species, while P. anserina was a common subordinate species. The height of a plant plays a crucial role in determining its ability to compete for light. Both of these two species have a distinct layered structure, and their photosynthetic capacity and LUE are significantly different. The competitive coexistence of different functional groups in alpine meadows is primarily attributed to their disparities in canopy photosynthetically active radiation, soil nutrient acquisition (Kleyer et al., 2019; Rao et al., 2023). Climate warming has resulted in a decline in soil nutrient and moisture, intensifying the competition between E. nutans and P. anserina. Consequently, this competition has led to changes in their leaf traits, ultimately impacting their photosynthetic capacity. In addition, climate change can directly influence plant photosynthesis through alterations in temperature and precipitation patterns. Ma et al. (2017) suggested that the two dominant species, E. nutans and Stipa aliena, were relatively insensitive to environmental changes, probably because of their greater ability to acquire nutrients and light. Warming increased the photoinhibition of E. nutans but decreased the photoinhibition of P. anserina, which is consistent with previous researches (Shi et al., 2010; Zhou et al., 2021).

The F _v/F _m value is a sensitive indicator of photoinhibition. In our study, the F _v/F _m value of the two species in the OTCs was higher than that of the OAs, indicating that the alpine plants were restricted by low temperatures, and warming improved their ability to resist stress ( Figure 2 ). Under medium-term warming, the F _v/F _m value of P. anserina was higher than that of E. nutans, indicating more effective carboxylation and quicker light-harvesting by the PSII antenna complexes. Changes in F _v/F _m value reflect up-or down-regulation of PSII, which is associated with changes in qP or qN in PSII (Wu et al., 2018). qN is an indicator of a plant’s ability to reduce heat dissipation in its photosynthetic membranes, thereby minimizing chloroplast damage (Ware et al., 2015). E. nutans in the OTCs activated photoprotection and reduced photoinhibition mechanisms, resulting in a greater qN to dissipate excess heat and sustain C assimilation capacity.

Understanding the LUE of alpine plants is crucial for enhancing their productivity and mitigating the degradation of alpine meadows. Zhang et al. (2015) believed that the changes in plant height and coverage under warming directly affect their competition for light energy and LUE. Our study revealed that warming enhanced the LUE of both two species, with significant variations observed among different species and over different years of warming ( Figure 1 ). In contrast to our findings, Zhou et al. (2016) suggest that simulated warming may reduce the LUE of alpine meadows due to the negative impact of warming-induced dry micro-environment on LUE, which masks the favorable effect of temperature rises. Overall, our results imply that both of the two species were impacted by climate change and that the short and medium-term impacts on various species varied. Warming improved the LUE of the two species, which was beneficial to the growth and productivity of alpine plants.

Leaf traits are important indicators of plant adaptation to environmental change because they are linked to the efficiency of plant resource acquisition and use (Li et al., 2018; Xu et al., 2018; Li et al., 2021). In addition, leaf traits account for the majority of the variation in ecosystem productivity (Sigdel et al., 2023). Numerous studies have demonstrated that leaf traits, such as leaf lifespan, leaf area, and leaf nutrient content, are highly sensitive to climate warming (Myers-Smith et al., 2019; Bjorkman et al., 2020). Warming directly affects leaf traits and indirectly affects plant photosynthesis and LUE through these traits. Our research indicate that the impact of warming on leaf photosynthetic capacity varies among the two species, which is related to their functional traits and soil nutrient availability ( Figure 5 ). Under medium-term warming, both species’ RWC and SLA increased. Additionally, the TWC and SLV of E. nutans increased and the LDMC of P. anserina also increased. Our findings support the hypothesis that phenotypic plasticity in certain plant traits can serve as a predictor for community performance under climate change.

SLA and leaf nitrogen content play a crucial role in carbon fixation (Díaz et al., 2016). According to Kattge et al. (2020), SLA exhibits sensitive to climate change and is closely associated with species-specific resource utilization. It plays a crucial role in influencing photosynthesis, light interception, and plant growth, while also serving as a predictor of competitiveness and environmental tolerance (Worthy et al., 2020). Plants with a larger leaf area are more efficient in capturing light and carbon. Our research showed that the SLA of the dominant species E. nutans is enhanced by warming, which implies higher resource acquisition in warmer climates. This finding is consistent with many other studies (Guittar et al., 2016; Liu et al., 2018). SLV is an important leaf functional trait, is influenced by factors such as leaf thickness, overall dimension, and dry matter content. It serves as an indicator of a plant’s ability to adapt to extreme environments like cold and arid conditions (Su et al., 2018). Leaf volume is determined by the combination of photosynthetic area and thickness, representing the entirety of photosynthetic organs. This trait facilitates better comparison among different plant species. Su et al. (2018) proposed the concept of SLV and suggested that alpine plants with higher SLV would exhibit greater resistance to harsh environmental conditions. In our study, we observed a strong positive correlation between leaf SLV and leaf P _n during the medium-term warming ( Figure 5 ). The higher the leaf P _n of plants, the greater the SLV; however, short-term warming exhibited an opposite trend. These results indicate that during the process of long-term adaptation to the environment, the photosynthetic capacity of plants and leaf traits are mutually adapted and coordinated.

Leaf nitrogen content is strongly correlated with photosynthetic capacity, as nitrogen is essential for the synthesis of Rubisco, which is the key enzyme in photosynthesis (Reich et al., 1994). We observed a positive correlation between leaf C and N content and LUE. Peng et al. (2020) suggested that warming slightly increased the coverage of legumes and the C: N ratio of all plants in the alpine meadow. Our investigation revealed that the C: N and N: P ratios of the two species’ leaves increased under warming, which is consistent with Peng’s findings. The alpine meadow subjected to experimental warming displayed higher leaf C: N and N: P ratios, indicating that plants were more efficient in utilizing nitrogen for growth. Plant communities exhibit both positive and negative interactions between different plant species (Callaway and Walker, 1997). Cui et al. (2023) confirmed that grasses exhibited a higher competitive ability compared to other functional groups, primarily attributed to their increased investment in roots and enhanced capacity for resource uptake.

The productivity of grasslands is influenced by the interactions between soil and climatic conditions (Shen et al., 2022). Our structural equation modeling revealed that warming indirectly impacts leaf photosynthetic capacity through factors such as soil nutrients, atmospheric CO₂, and leaf traits ( Figure 5 ). Our previous studies have demonstrated that the optimal temperature range for alpine plants is between 20°C and 25°C (Zhou et al., 2017). During the vigorous growth period in 2015 and 2018, between 09:00 and 13:00 hours, we used an LI-6400 portable photosynthesis system to measure average air temperatures of approximately 28.5°C and 31.5°C. These temperatures exceeded the optimal range for alpine plants, resulting in a negative impact on leaf photosynthetic capacity.

Soil plays a crucial role in providing the majority of nutrients necessary for plant growth, and these nutrients are closely associated with how plant leaves utilize resources (Gao et al., 2019). The soil nutrients (N) have a significant impact on leaf N and plant photosynthesis (Yu et al., 2019), and a low soil phosphorus (P) may result in reduced leaf P content, thereby limiting overall plant function (Sun et al., 2019). Climate change modifies the physical and chemical properties of soil, thereby impacting the functioning of the alpine meadow ecosystem. Our findings indicate that short and medium-term experimental warming have similar impacts on soil nutrients. Warming reduced soil nutrient contents at depths of 0-30 cm ( Table 2 ), indicating that warming stimulated soil nutrient cycling and organic matter decomposition, resulting in a decrease in soil C and N contents (Xu et al., 2022). However, some studies have shown inconsistent results regarding the responses of soil nutrients to climate warming. While some suggest that warming has no impact on soil C and N contents (Ning et al., 2019; Yang et al., 2022), others indicate an increase (White-Monsant et al., 2017). These contradictory results may be attributed to variations in temperature and duration of warming, as well as the diversity of grassland ecosystems. In our study, soil nutrient content did not change significantly under warming. This is because, in the case of relatively short-term warming experiments (< 5 years), it was challenging to significantly alter the vast soil carbon pool due to the substantial spatial heterogeneity in soil organic carbon among plots and the limited number of repeated warming experiments.

Several studies indicate a positive correlation between warming-induced changes in plant total biomass, above-ground biomass, and below-ground biomass with soil nutrient content. This subsequently impacts the LUE of plants (Song et al., 2019; Zhou et al., 2022). Shen et al. (2022) found that the change in soil pH and nutrient imbalance caused by N and P enrichment were the main factors impacting the photosynthetic characteristics of plant in the alpine steppe. In our study, the effects of soil nutrients on LUE were primarily mediated through their impact on leaf traits. The content of soil nitrogen significantly influences leaf N content and plant photosynthesis, while inadequate levels of soil P can lead to reduced leaf P content and limit leaf function (Gong and Gao, 2019; Sun et al., 2019).

Soil respiration (R _s) is expected to have positive feedback on global warming (Wang et al., 2021). Some studies suggest that short-term warming promotes soil respiration, but there is no consistent pattern under long-term warming and variations exist among ecosystems (Chen et al., 2016; Ganjurjav et al., 2018). According to García-Palacios et al. (2021), R _s is highly sensitive to temperature, and warming will stimulate R _s activation and accelerate its rate. Yu et al. (2019) found no significant effects of experimental warming on R _s. Our findings were similar to prior research that alpine meadow R _s increased in response to both short- and medium-term warming.