Grasslands, which occupy 25% of the world’s total land area and 66.7% of agricultural land (FAO, 2023), play a crucial role in global food supply (Boval and Dixon, 2012). Additionally, grasslands fulfill important ecological functions, including biodiversity maintenance, water supply, soil conservation, and climate regulation (Bengtsson et al., 2019). However, nearly 50% of global grasslands have experienced some degree of degradation due to global changes (Bardgett et al., 2021). Recent studies have increasingly focused on grassland degradation in alpine regions, particularly in the Third Pole, i.e., the Qinghai-Tibetan Plateau. Evidence suggests that climate change and inappropriate anthropogenic activities have led to extensive (exceeding 50%) grassland degradation in this alpine region (Harris, 2010; Cao et al., 2019). Such degradation has severely compromised the ecosystem serves provided by grasslands (Xu et al., 2021; Wang et al., 2022). Consequently, various efforts to reverse this adverse trend have been undertaken (Fu et al., 2023).

Restoration measures for degraded grasslands include grazing exclusion using fences, fertilization, irrigation, and revegetation (Li et al., 2023; Lyons et al., 2023; Ji et al., 2024). However, the restoration of degraded grassland ecosystems may yield high costs and low benefits if restoration measures do not align with the specific degradation conditions (Dong et al., 2023). This misalignment appears to be a pervasive issue. On one hand, current methods for diagnosing grassland degradation are often empirical, relying on qualitative assessments or assumed degradation thresholds (Yao et al., 2016; Zhu et al., 2023), or mechanistic approaches that classify grassland ecosystems into degradation types based on equidistant divisions of plant indicators (Wen et al., 2010; Zhang et al., 2025). Such cursory methods can lead to inaccuracies in classifying degradation levels and may fail to identify restoration factors pertinent to specific grassland ecosystems. On the other hand, the prevailing definition of grassland degradation tends to agricultural productivity-oriented, utilizing net primary productivity-related indices (e.g., the normalized difference vegetation index and plant community cover) to evaluate degradation conditions (Hilker et al., 2014; Wang et al., 2019, 2022). However, recent restoration practices have increasingly emphasized the multiple ecological functions of grassland ecosystems, such as biodiversity maintenance and carbon storage (Hu, 2023). This shift further exacerbates the mismatch between restoration measures and the degradation conditions, potentially hindering effective restoration efforts. Therefore, there is a pressing need to explore accurate, field data-based approaches that consider multiple ecological functions for diagnosing degradation in grassland ecosystems.

One of the most evident performances of grassland ecosystem degradation is the deterioration of the living status of vegetation (LSV). Specifically, the cover, height, species richness, and biomass of plant communities, which strongly contribute to the LSV, diminish as grassland ecosystems degrade (Wang et al., 2014; Tang et al., 2015; Zhang et al., 2018; Dong et al., 2023). These indices are extensively investigated in field work, as they not only reflect the multifunctional nature of grassland ecosystems but also provide cornerstones for a comprehensive assessment of the LSV (Zhang et al., 2015). Identifying the thresholds of LSV deterioration using reliable mathematical methods may offer a novel approach for accurately diagnosing the conditions of a grassland ecosystem.

The Mann-Kendall abrupt analysis is commonly employed to detect change points in meteorological factors (e.g., temperature and precipitation) (Liu et al., 2008; Zhao et al., 2008; Xia et al., 2023), runoff (Zhang et al., 2014), and epidemic transmission (Chen et al., 2022). For data series to be suitable for Mann-Kendall abrupt analysis, they must be independent (i.e., each data point should be random) and continuous (i.e., the data series should follow a specific order, such as a chronological order) (Mann, 1945; Kendall, 1949; Liu et al., 2008). In field studies, data from plant communities (i.e., cover, height, species richness, and biomass) are treated as independent within a random sampling framework (Schweiger et al., 2015; Henrys et al., 2024). However, the continuity (or order) of these data is often undefined. Fortunately, a geo-coding technique can establish continuity based on the coordinates of the sampling plots.

This study, using subalpine meadows in the Dashanbao National Nature Reserve in China as a case study, aims to align field-collected data on plant communities with the requirements for Mann-Kendall abrupt analysis through the application of a geo-coding method. Subsequently, a rule system is proposed to identify abrupt changes in LSV, facilitating the assessment of degradation conditions in grassland ecosystems.

Grassland degradation is a nonlinear process that typically occurs progressively rather than abruptly, often under stable disturbances such as livestock grazing (Tiscornia et al., 2019; Zhang et al., 2020). An effective diagnostic method for assessing grassland ecosystem degradation should reflect this characteristic. Many studies that classify grassland degradation levels based on empirical or mechanistic approaches emphasize a linear decline in cover, height, and biomass of plant communities (Li et al., 2023), creating a simplistic perception of degradation for stakeholders involved in grassland management. In this study, the subalpine meadows of the Dashanbao National Nature Reserve were classified into four relative degradation levels, i.e., healthy, light degradation, moderate degradation, and severe degradation, using the GAAB method. We observed that changes in the structure of the plant community (including α diversity, composition, and dominant species) were nonlinear and progressive, while functional indices (cover, height, and aboveground biomass) exhibited a linear reduction with increasing degradation levels. This asynchronous degradation process between the structural and functional aspects of plant communities underscores the nonlinear nature of grassland degradation. Furthermore, the non-equidistant degradation thresholds of LSV reflect a nonlinear degradation process as well. We found that the thresholds changed from 1.27 to 0.96 (Δ LSV = 0.31) from healthy to light degradation, with a Δ LSV of 0.17 from light to moderate degradation, and 0.28 from moderate to severe degradation. This indicated that LSV deterioration was rapid in the initial and final stages of degradation, while transitioning more smoothly during the intermediate stage. The response of plant communities to environmental disturbances may contribute to this phenomenon. For example, the height and cover of grassland plants declined sharply at first, followed by a gradual reduction during trampling experiments (Cole, 1995; Hill and Pickering, 2009).

The criterion for defining the “healthy status” of a grassland ecosystem is a critical consideration when assessing degradation levels. This “healthy status” should be contextualized by geographical location and types of disturbances, as there is no universal criterion applicable to all grassland ecosystems. Two strategies can be employed to define the “healthy status”. Firstly, it can be established based on historical records or inferred from model simulations, especially in theoretical discussions at larger temporal and spatial scales. For instance, ecologists might predict potential productivity and plant species distribution under varying climatic conditions (Gao et al., 2016; Han et al., 2019), with optimal model outcomes serving as a benchmark for “healthy status”. However, a more pragmatic approach is needed for effective grassland restoration practices. According to the law of distance attenuation, the impact of environmental disturbances on grassland ecosystem diminishes with distance from the disturbance source (Huang et al., 2021). Therefore, minimally disturbed grassland sections can serve as factual benchmarks for defining “healthy status”. By conducting systematic field surveys across the study area, ecologists may delineate these reference conditions that can guide the classification of degradation levels. This “healthy status” can also guide stakeholders in setting restoration targets. In the present study, the “healthy status” of subalpine meadows in the Dashanbao National Nature Reserve was determined by identifying plant communities with the highest LSV. We found that metrics such as cover, height, aboveground biomass, species α diversity, and dominant species under this “healthy status” consistently outperformed those in other degradation levels. Thus, restoration targets for grasslands in the Dashanbao National Nature Reserve could be established based on the plant indices characteristic of this “healthy status”.

Locking in priority restoration factors can significantly enhance the efficiency of grassland restoration efforts. Within the GAAB method, we can identify key restoration factors based on the contributions of various plant indices to the LSV. In this study, we observed a decrease in the contributions of cover, height, and aboveground biomass of plant communities to the LSV, while the contributions of α diversity indices increased with the degradation levels. In line with the diversity-stability relationship in grassland ecosystems (Tilman et al., 2006), it is crucial to prioritize the enhancement of ecosystem stability by increasing plant α diversity, particularly in LD and SD areas within the Dashanbao National Nature Reserve. Furthermore, selecting native plant species, such as Elymus sinicus and Eragrostis nigra, is essential to mitigate issues related to species invasion during restoration practices. Conversely, in regions experiencing high levels of degradation, the focus should shift toward improving cover and biomass of the plant community. Soil nutrient regulation strategies, such as fertilization, can effectively promote these parameters (Holatko et al., 2023; He et al., 2024; Pittarello et al., 2024). It should be noted that it is imperative to develop a specific restoration plan that outlines the number and abundance of plant species, as well as the type and quantity of fertilizers to be applied in these restoration efforts.

Every ecological method must be applied under specific conditions. Due to the heterogeneous distribution of ecological elements (Scheiner and Willig, 2007; Huang et al., 2023), the GAAB method should be applicable only to the monitoring of a particular grassland ecosystem or to historical assessments of the same grassland ecosystem. It is not suitable for evaluating degradation in grassland ecosystems characterized by high heterogeneity or across diverse types. Additionally, increasing the sample size or the number of observations will undoubtedly help improve the accuracy of the GAAB method. Furthermore, the LSV, which integrated plant community attributes such as cover, height, aboveground biomass, species richness, and the Pielou index to detect degradation thresholds in this study, can be flexibly tailored to diverse research contexts (Ganjurjav et al., 2022). In this study, all plant indices were assigned equal weights during the construction of the LSV, as determining the relative importance of individual indices for characterizing plant community status proved challenging. However, future research could implement differential weighting of standardized indices in LSV construction, provided that reliable methods for quantifying their relative ecological significance become available.

The GAAB method, which utilized the LSV as a monitoring index, effectively captured the nonlinear and progressive degradation processes, identifies the “healthy status”, and prioritized restoration factors for the subalpine meadows in the Dashanbao National Nature Reserve. This approach offers a non-empirical framework for diagnosing degradation conditions. The implementation of this novel method could enhance the understanding of the complexities associated with grassland degradation and contribute to the development of practical knowledge for the restoration of degraded grassland ecosystems.