1. Introduction

Sustaining ecological stability is the key to optimizing ecosystem performance, and ultimately achieving the goal of providing sustainable and reliable ecosystem services (Hautier et al., 2015; Pennekamp et al., 2018). Ecological stability is significantly impacted by extreme climate events, such as heavy rainfall and severe droughts (IPCC, 2021), which alter community structure and functions (He et al., 2022). The ecological stability of grasslands, one of the largest terrestrial ecosystems on the earth, is especially susceptible to extreme weather events. Thus, understanding how grassland ecosystems maintain stability in response to environmental changes is necessary to develop adaptive strategies and effective management of grasslands (Chen et al., 2023; De Keersmaecker et al., 2016; Zhang and Wang, 2023).

Ecological stability refers to the ability of ecosystems to resist environmental fluctuations and changes (Donohue et al., 2013). Ecological stability is multidimensional, mainly including resistance (i.e., the ability of an ecosystem to withstand external disturbance), resilience (i.e., the rate at which a function returns to its pre-disturbance status) and recovery (i.e., the difference in a function between a pre-disturbance year and a post-disturbance year) (Donohue et al., 2016; Xu et al., 2022). Previous studies have investigated factors (e.g., interspecific relationships, diversity and functional traits) that may contribute to ecological stability following extreme environmental events (Isbell et al., 2015; Yu et al., 2021). Plant diversity has been found to increase plant community resistance (Isbell et al., 2015; Wilcox et al., 2020), resilience, and recovery to extreme climate scenarios (Kreyling et al., 2017). However, plant diversity has also been found to negatively affect resilience and recovery of plant communities (Hossain et al., 2022). The resistance of plant community to extreme climatic events has been shown to increase due to asynchronous dynamics among species, although these dynamics might weaken resilience and recovery (Xu et al., 2022). Studies have also shown that the ecological stability of dominant species and canopy cover led to substantial variations in resistance, resilience and recovery of plant communities under extreme climate scenarios (Xu et al., 2014).

Previous studies have shown that appropriate grassland management methods (e.g., nitrogen enrichment, warming and grazing) may contribute to ecological stability (Hallett et al., 2017; Yan et al., 2023; Zhang and Wang, 2023). An additional grasslands management approach includes the use of plant litter. Plant litter plays an important role in terrestrial ecosystems (Zhang et al., 2023), regulating key ecological processes through decomposition, such as balancing ecosystem biomass production (Shen et al., 2016) and optimizing the compositional dynamics of plant communities (Ma et al., 2022b). Plant litter has also been shown to alter plant community canopy cover, as well as increase the stability and asynchrony of plant functional groups (Ma et al., 2021, 2022a). These changes, in turn, may affect ecological stability. However, the role of plant litter dynamics in shaping the resistance, resilience, and recovery of grassland plant communities under extreme climatic events is largely unknown.

The alpine grasslands of the Qinghai-Tibet Plateau (QTP) represent a typical alpine ecosystem located in the area most sensitive to environmental change in the world (Chen et al., 2013). Extreme precipitation in QTP, which is expected to increase, accounts for 15%–24% of annual precipitation since the 1970s (Ge et al., 2017; You et al., 2021). These extreme rainfall events have been shown to directly affect ecological barriers and ecosystem services (Luo et al., 2022). In addition, the composition of predominant functional groups among vegetation communities in alpine grasslands in QTP is constantly changing because of the varied management strategies, which likely alters plant litter composition and biomass accumulation. For example, one effective method for repairing grassland ecology (e.g., grazing enclosure) has been shown to increase plant litter accumulation (Zou et al., 2016). Plant litter accumulation is also caused by the widespread distribution of inedible vegetation in degraded grasslands (Wang et al., 2022). Research has shown that heterogeneous plant litter (both in terms of mass and species) substantially alters the biodiversity and ecological functions of alpine grasslands (Wang et al., 2010). In a previous study of alpine grasslands of the QTP, we found that plant litter addition increases plant community temporal stability and heterogeneity and decreases grassland ecosystem multifunctionality (Ma et al., 2022a, 2022b). However, it remains unclear how plant litter affects grassland resistance, resilience and recovery in response to disturbances such as extreme rainfall events.

In this study, we examined whether plant litter addition increases ecological stability after an extreme rainfall event in alpine grasslands. To answer this question, we determined whether plant litter mass and species (Elymus nutans, Kobresia setchwanensis, and Ligularia virgaurea) affects resistance, resilience and recovery of alpine grasslands following extreme rainfall. We then asked which direct and indirect factors contribute to resistance, resilience, and recovery in these plant communities. Specifically, we determined whether ecological stability (i.e., resistance, resilience, and recovery) of alpine grasslands was correlated with plant diversity, canopy cover, functional group biomass variability, asynchrony, and/or dominant functional group stability. Our findings suggest that plant litter accumulation may promote ecological stability in response to ongoing and future extreme rainfall event in alpine grasslands and other similar ecosystems and to provide reliable ecosystem services.

2. Materials and methods 2.1. Study site and experimental design

This study was conducted at the Qinghai-Tibet Plateau Research Base of Southwest Minzu University, an alpine grassland site on the eastern Tibetan Plateau in Hongyuan County (31°47′ N, 102°33′ E; 3485 m a.s.l), Sichuan Province, China. The region is characterized as a continental cool-temperate monsoon climate. The mean annual precipitation and temperature from 1961 to 2019 are 747 mm and 1.5 ℃, respectively. The soil is classified as Mat Gry-gelic Cambisols according to the Food and Agricultural Organization. This site is dominated by Elymus nutans, Carex thibetica, Saussurea hieracioides, Anemone trullifolia, and Anemone rivularis. The normal range of precipitation is generally determined based on the average of at least 30 years of precipitation data (Vepraskas et al., 2019). Extreme rainfall events are those that occur with a probability of less than 5% in the study area. The average precipitation during the growing season in this study area is 584 mm (based on precipitation data of the growing season over the past 60 years, 1961–2019). During the 2018 growing season, the study area experienced a heavy rainfall event, with a total of 787 mm rainfall, 212 mm of which fell in July. Precipitation during the growing season exceeded the 95% range of the probability distribution curve, indicating the occurrence of extreme rainfall events in the study area. By comparison, precipitation during the 2017 growing season (577 mm) and 2019 growing season (683 mm) were within the normal range (He et al., 2022).

The plant litter addition experiment was conducted from 2017 to 2019 in an area with uniform vegetation. We used a randomized block design with five litter addition levels (0 (control), 100, 200, 400, and 600 g m⁻²) crossed with three species (Elymus nutans, Kobresia setchwanensis, and Ligularia virgaurea). There were four replicates for each treatment. A total of 52 plots were laid out in each of the four blocks and randomly assigned to all treatments. Each 2 m × 2 m plot was separated from adjacent plots by 1 m and blocks were at least 2 m apart. Beginning in 2017, plant litter was added annually to plots at the early stage of seedling regreening, in late-April. Plant litter samples were collected from the previous year, air-dried and cut into about 1-cm length. Clipped litter of desired amount and species was evenly placed onto the plots each year. Dead tissue was removed from all plots before adding test litter (Ma et al., 2022a).

2.2. Plant community monitoring and sampling

From 2017 to 2019, plant community monitoring and sampling were conducted each year in mid-June (the early growing season), July and August (the peak of plant growth during the growing seasons). In each plot, plant species richness and canopy cover were measured in a permanent quadrat (0.5 m × 0.5 m). Species richness was quantified as the number of species recorded within each quadrat, and canopy cover was quantified as the percentage of the surface area covered by all plant species. One quadrat (0.5 m × 0.5 m) was randomly established in each plot and all plant tissues were harvested at the soil surface to estimate aboveground biomass. Living plants were sorted into functional groups (forbs, grasses, legumes, and sedges) based on their functional forms, oven-dried at 65 ℃ for 48 h, and then weighed.

2.3. Plant community ecology stability metrics and characteristics

Resistance, resilience and recovery of plant community to extreme rainfall events (2017–2019) were calculated according to Xu et al. (2022): Resistance=ln(B2018/B2017), Resilience=ln(B2019/B2018) and Recovery=ln(B2019/B2017), where B is the aboveground biomass of a plot in a sampling period. Resistance, resilience, and recovery of four functional groups (forbs, grasses, legumes and sedges) were also calculated accordingly. Plant asynchrony and dominance for each plot were estimated using the community-wide asynchrony index (Loreau and de Mazancourt, 2013) and Simpson's dominance index (Smith and Wilson, 1996), respectively, based on aboveground biomass of each plant functional group.

2.4. Data analyses

A linear mixed-effects model was used to determine the effects of litter mass, litter type and their interaction on plant community stability dimensions (resistance, resilience, and recovery), aboveground biomass of plant functional groups and plant asynchrony. Litter type, litter mass and their interaction were used as fixed effects, and time and block as the random effects. All statistical analyses were performed using the ‘lme 4’ packages in R v.4.1.0 (R Core Team, 2021). Differences were considered to be statistically significant at P < 0.05. A simple linear regression was used to assess how species richness, canopy cover, plant asynchrony, plant dominance and plant aboveground biomass affect each ecological stability dimension. The relationships between the resistance, resilience, and recovery of plant functional groups and those of the overall community were also calculated. Random Forest model was conducted using the ‘rfPermute’ R package (Cutler et al., 2007) to identify the important predictive factors. Finally, we used piecewise structural equation modeling (SEM) to explore the pathways through which litter addition affect the metrics of plant community ecological stability in response to extreme rainfall. We simplified the aboveground biomass of plant functional groups using Principal Component Analysis (PCA), and constructed the structural equation model using the first principal component (PC1). All data used in the SEM were standardized by z-transformation and the plot was used as the random effect (only variables with a variance inflation factor of < 5 were included in the same model). We first developed a priori conceptual model containing all cascading pathways between ecological factors (Fig. S1). Model adequacy and fitness were evaluated using directed separation tests on Fisher's C statistic and AIC (Akaike information criteria) (Shipley, 2013). The SEM construction and analyses were performed using the ‘piecewiseSEM’ package (Lefcheck, 2016) in R v.4.1.0 (R Core Team, 2021).

3. Results 3.1. Effects of litter addition on resistance, resilience, and recovery

Litter addition of 400 and 600 g m⁻² significantly increased the resistance, resilience and recovery of grasses (P < 0.05) (Table S1 and Fig. 1). However, compared to the control, litter addition only increased community resilience at 400 g m⁻², and had no significant effect on community resistance and recovery. Community recovery decreased much more in response to 100 g m⁻² treatment compared than in response to 400 or 600 g m⁻² treatments. Furthermore, litter addition had no significant effect on resistance, resilience, or recovery of sedges, legumes, and forbs (P > 0.05), although legume resilience was higher at 600 g m⁻² than at 200 g m⁻².

3.2. Effects of litter addition on aboveground biomass of plant functional groups

The effects of litter addition on aboveground biomass of plant functional groups varied (Figs. 2 and S2; Table S1). Addition of Ligularia virgaurea litter decreased aboveground biomass of grasses and legumes more than did that of Elymus nutans (P < 0.05, Fig. 2a, c). Litter addition with L. virgaurea increased aboveground biomass of sedges (P < 0.001, Fig. 2b), while the aboveground biomass of forbs was decreased (P < 0.001, Fig. 2d). Compared with the control, litter addition of 400 g m⁻² increased aboveground biomass of legumes (P < 0.05, Fig. 2c). Litter addition of 400 and 600 g m⁻² decreased aboveground biomass of forbs (P < 0.001, Fig. 2d), but had no significant effect on aboveground biomass of grasses and sedges (Fig. 2a, b). In addition, the plant litter type that had the strongest effect on the aboveground biomass of sedges and forbs was L. virgaurea (P < 0.01; Table S1 and Fig. S2).

3.3. Relationships of ecological stability with plant diversity, canopy cover, asynchrony, and aboveground biomass

Plant community stability (resistance, resilience and recovery) was negatively correlated with several ecological factors (species richness, canopy cover, plant asynchrony and aboveground biomass of grasses, sedges, legumes and forbs) (Fig. 3). In addition, the stability of plant communities was positively correlated with resistance, resilience and recovery of their functional groups (except for legumes) (Fig. S3). Resistance, resilience and recovery of plant functional groups (grasses, sedges, legumes, and forbs) were negatively correlated with species richness, canopy cover, aboveground biomass, and plant asynchrony (Fig. 4).

3.4. Pathways through which litter addition altered resistance, resilience and recovery

The Random Forests model identified resistance, resilience and recovery of grasses, sedges and forbs, plant asynchrony and aboveground biomass of plant functional groups as significant predictors of the resistance, resilience and recovery of plant community (Fig. 5). SEM shows that the direct and indirect effects of litter addition explained 86%, 89% and 93% of the total variation of plant community resistance, resilience and recovery, respectively (Fig. 6). SEM results showed that the impact of litter addition on plant community resistance, resilience and recovery was partly mediated through the change in aboveground biomass of plant functional groups and plant asynchrony. The change in aboveground biomass of plant functional groups simultaneously increased plant asynchrony and decreased resilience and recovery of forbs. Resistance, resilience and recovery of forbs, followed by grasses and then sedges had the greater direct contribution to plant community stability dimensions, while plant asynchrony played a secondary role (Fig. 6). Moreover, litter addition had a direct influence on plant community resilience and recovery (Fig. 6b, c), but did not affect plant community resistance (Fig. 6a).

4. Discussion

The importance of environmental disturbances in shaping ecosystem functions has been well recognized (Polazzo and Rico, 2021; Ma et al., 2024). Our study explores how plant litter affects resistance, resilience, and recovery of alpine grassland productivity following extreme rainfalls in the Qinghai-Tibet Plateau. Our findings suggest that plant litter mass, rather than litter type, increase the ecological stability of alpine grasslands. In our experiments, we found that plant litter (400 g m⁻²) increased community-level resilience and higher masses of plant litter (600 g m⁻²) increased both resilience and recovery of grasses. In addition, our findings indicate that plant litter directly and indirectly increases alpine grasslands resistance, recovery, and resilience in response to extreme rainfalls.

Plant litter directly increased resilience and recovery of alpine grasslands communities in response to extreme rainfall by altering the ecological stability of plant functional groups. In our plant litter experiments, the functional group with the strongest direct effect on resistance, resilience, and recovery was forbs. Forbs are the dominant functional group of grasslands communities and include multiple life forms and ecological functions. In addition, the functional traits (niche breadth) of forbs have been shown to be sensitive to environmental interference (Ma et al., 2022a). Forbs may have contributed most to ecological stability because the variability of forbs was more sensitive to litter addition than that of grasses or sedges.

Plant litter addition was also found to have a direct effect on the resistance, resilience and recovery of grasses. These grasses, in turn, positively contributed to the ecological stability of alpine grasslands communities following extreme rainfall. One explanation for this finding is that most grass species tend to regain productivity in the year following extreme heavy rainfall. We note that after the extreme heavy rainfall, resistance, resilience and recovery of grasses increased with higher litter addition. Given their unique resource acquisition and reproductive strategies (da Silveira Pontes et al., 2015), grass species under plant litter treatments may take advantage of heavy rainfall-induced (accelerating litter decomposition and releasing more nutrients) resource subsidies to stabilize and retain biomass production after extreme rainfall.

Changes in plant community compositional dynamics mediated by plant litter addition indirectly affected ecological resistance, resilience, and recovery of alpine grasslands. Our finding that plant litter addition decreased the aboveground biomass of forbs, but that of grasses and sedges remained stable indicates there were change in the distribution of plant functional composition and plant community heterogeneity (Ma et al., 2022a). These changes in community dynamics, in turn, indirectly affected the resistance, resilience and recovery of plant community. Our findings are consistent with previous research that showed litter addition-induced changes in compositional dynamics indirectly increases resistance, resilience and recovery of plant community following extreme heavy rainfalls by improving clonal reproduction in grasses and sedges (Ma et al., 2022a). These indirect effects of litter addition are also in agreement studies that have examined the effects of nitrogen enrichment in degraded grassland (Zhang and Wang, 2023).

Plant litter may have also indirectly increased resistance, resilience and recovery of plant functional groups by increasing plant asynchrony. Previous studies have found that nitrogen enrichment in response to an extreme climate event induces a weak relationship between species asynchrony and the ecological stability of dominant species (Xu et al., 2022). High plant asynchrony implies strong species compensation dynamics within the community (Ma et al., 2017), which helps plant communities to resist extreme rainfall events and enhance community resilience and recovery. Moreover, a community with high plant asynchrony would promote the summed increment of plant community productivity due to the varied growth of functional groups in response to environmental changes. For example, the aboveground biomass of forbs in the lower layers of community was inhibited by the growth of competitive grasses after extreme heavy rainfall. Conversely, inhibition of aboveground biomass in the upper layers of the community will promote the growth of grasses and sedges.

Plant litter also indirectly affected ecosystem stability by altering canopy cover and species richness. Consistent with previous research (He et al., 2022), we found that resistance, resilience, and recovery in the alpine grasslands community were negatively correlated with species richness. This finding suggests that litter addition increases plant asynchrony and, due to the ecological stability of plant functional groups, overrides the effect of species richness on resistance, resilience and recovery of plant community following extreme rainfalls (Ma et al., 2023). Previous research has also reported that litter addition promotes vertical growth of plants, and the mechanical shielding of uncompounded litter reduces the canopy coverage of plant communities (Ma et al., 2021), further promoting the resilience and recovery of plant community.

Our findings showed that litter addition directly affected plant community compositional dynamics, indirectly affected plant asynchrony, and thus directly affected plant community resistance, resilience and recovery (Fig. 6). This may be attributed to two potential mechanisms. Firstly, different plant species have specific functional traits that lead to different responses to environmental changes, thereby altering plant community compositional dynamics (Loreau and Mazancourt, 2008) that mediate the response of plant asynchrony to litter addition. Secondly, litter addition changed the species richness, canopy cover and compositional dynamics of plant community, in turn affecting plant asynchrony. Our findings highlight the need to consider plant litter dynamics of grasslands in future studies to better explore appropriate grassland management strategies for maintaining ecological stability under extreme climate scenarios. A reasonable management strategy may regulate the accumulation of plant litter to an appropriate level, which will then enhance the ecological stability of plant communities to resist and adapt to extreme environmental fluctuations in grassland ecosystems.

5. Conclusion

The current study demonstrates that plant litter mass, rather than litter type, plays a central role in regulating the ecological stability of grassland ecosystems under extreme rainfall events. By experimentally manipulating plant litter mass, this study shows that the effects of litter addition on the resilience and recovery of plant community productivity are mainly driven by the ecological stability of functional groups. Our findings emphasize the underlying mechanisms of ecological stability change with plant litter dynamics, and provide empirical knowledge for developing optimal strategies-such as strategic retention or removal of litter for maintaining the sustainability of grassland ecosystems and their ecological service functions in the face of ongoing global environmental change.

Acknowledgments

This study was supported by the National Natural Science Foundation of China (32161143028, U21A20242), the National Key Research and Development Program of China (2021YFD1300504), the Program of National Science and Technology Assistance (KY202002011), the Program for Innovative Research Team of Ministry of Education (IRT17R50), the Postdoctoral Fellowship Program (Grade C) of China Postdoctoral Science Foundation (GZC20241180), and the Sichuan Science and Technology Program (2024NSFSC2075). We thank the staff of the Qinghai-Tibet Plateau Research Base of Southwest Minzu University.

CRediT authorship contribution statement

Zhouwen Ma: Writing – original draft, Investigation, Formal analysis, Data curation. Lan Li: Software, Formal analysis. Yingxin Wang: Investigation, Data curation. Qingping Zhou: Supervision, Resources. Xinquan Zhang: Writing – review & editing. Fujiang Hou: Writing – review & editing, Methodology, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2025.06.005.