Most aerobic methane-oxidizing bacteria (MOB) harbor nitrogen-fixing genes and exhibit nitrogen-fixing activity, however, the mechanisms by which oxygen supporting aerobic methane oxidation while inhibiting nitrogenase activity remain unclear. We investigated two paddy soils subjected to contrasting long-term (28 years) water management regimes: an intermittently flooded rice-rape rotation soil (RS) and permanently flooded paddy soil (FS). We hypothesized that long-term differences in water management alter soil oxygen exposure history, thereby shaping MOB communities and functions, and ultimately regulating methanotrophy-coupled nitrogen fixation. We found that oxygen concentrations in FS were significantly lower than those in RS. Microcosm incubation with 13CH4 and 15N2 showed that net 13C-SOC accumulation was 550 μg g⁻¹ d.w.s in RS, slightly lower than that of 569 μg g⁻¹ d.w.s in FS following 16 days of methane oxidation. In contrast, methane oxidation stimulated significantly greater 15N-TN accumulation of 20.5 μg g⁻¹ d.w.s in RS than that of 12.5 μg g⁻¹ d.w.s in FS. DNA-based stable isotope probing (DNA-SIP) showed that Methylocystis-like type II and Methylosarcina-like type I MOB were the predominant phylotypes in RS and FS, accounting for 84.86% and 76.32% of the active MOB community, respectively. Metagenome assembly showed that the Methylocystis genome encoded genes of nitrogen fixation (nifHDK), together with a suite of antioxidant enzymes and high-affinity terminal cytochrome oxidases, which may collectively contribute to sustaining nitrogen fixation under fluctuating oxygen conditions. In contrast, the Methylosarcina genome lacked nitrogen fixation genes and possessed fewer antioxidant-related genes. This study demonstrates that long-term distinct water management regimes fundamentally alter the process of methanotrophy-coupled biological nitrogen fixation. Intermittently flooded paddy soils exhibit a higher nitrogen fixation potential, mediated by metabolically flexible type II MOB with greater tolerance to oxygen stresses.

The alpine grassland vegetation on the Qinghai-Tibet Plateau is composed of plant patches in varied sizes. It remains uncertain whether vegetation recovery following grazing exclusion (GE) in degraded grasslands is driven by increases in patches number (NP), patch size (PS), or both. We based our predictions on two hypotheses: GE intensifies plant competition, and facilitation prevails near patches while competition prevails in interpatch spaces. We predicted that the NP would remain stable or decrease and PS would increase under GE treatment. To evaluate these predictions, we conducted a study in six lightly degraded alpine grasslands under free grazing (FG) conditions in Bangor County, Tibet Autonomous Region, China, with corresponding GE treatments using transects in 2017 and 2018. Results revealed that four sites in 2017 and five sites in 2018 had reduced NP and increased PS, with probabilities of 0.033 (2017) and 0.004 (2018), respectively, and a joint probability of 0.0001 under the null hypothesis that GE does not affect NP or PS. The NP reduction was solely due to the decrease in small patch sizes. An increase in PS was common across species, and a predominant tendency for NP reduction was observed among species across the sites. The overall changes in NP and PS were primarily driven by the three most abundant species (contributing more than 60% in both years), rather than by shifts in floristic composition. Our findings highlight that vegetation recovery in Bangor alpine steppes following GE relies solely on the expansion of existing patches rather than the recruitment of new ones in interpatch gaps. We recommend prioritizing growth-promoting measures, such as nutrient or water management, over seed addition when assisting with GE for restoring lightly degraded grasslands.

Plant community stability is often hindered by nitrogen and phosphorus deficiencies in high-altitude alpine steppes. While nutrient addition can mitigate these constraints, the individual and interactive effects of these elements to community stability remain unclear. Here we elucidated the underlying mechanisms by decomposing asynchrony into compensatory and statistical-averaging effects across undegraded and degraded grasslands on the Tibetan Plateau, under conditions with and without nitrogen and phosphorus additions. In undegraded alpine steppes, phosphorus limits plant coverage, while nitrogen joins as a limiting factor under degradation. Phosphorus addition threatens stability in pristine areas, whereas nitrogen enhances degraded steppes' biodiversity and statistical-averaging effects. Despite varying nutrient scenarios, the compensatory effects among dominant species persist, primarily regulating community stability, rather than population stability or statistical-averaging effects. These results underscore the pivotal role of dominant species in community stability and their diverse compensatory dynamics in undegraded and degraded steppes under nutrient addition conditions.

Accurate estimation of soil organic carbon (SOC) is essential for monitoring carbon sequestration and mitigating climate change across agricultural and natural ecosystems. Mid-infrared (MIR) diffusereflectance spectroscopy is a cost-effective, high-throughput alternative to wet chemistry and dry combustion methods, yet reported model performance varies widely across studies and regions. This study assesses the effectiveness of MIR spectroscopy for SOC estimation using a global meta-analysis of 289 studies. Meta-analytic results indicate that fine grinding (≤53 µm) is associated with higher median R² (≈0.77 vs 0.72) and lower RMSE than ≤ 2 mm sieving alone, and that air-dried samples tend to show lower RMSE than oven-dried samples. Broader spectral ranges and higher resolution are generally associated with lower RMSE in SOC estimation, although these patterns are not consistent across all instruments and soil types. Among preprocessing options, multiplicative scatter correction and Savitzky–Golay derivatives are frequently associated with improved model fit, but their benefits are dataset-dependent. Across chemometric methods, partial least squares regression (PLSR) remains widely used (n = 278), whereas Cubist shows high median R² in a very small number of studies (n = 6) without statistically robust improvement over PLSR in this synthesis. These findings offer methodological insights for improving SOC monitoring and supporting global carbon accounting initiatives. We further provide reporting considerations and point to a companion paper (Part II) that prospectively evaluates these insights using a controlled modelling pipeline with global and national spectral libraries.

Abstract Though livestock grazing is the most common grassland management, its effects on species interactions and spatial patterns of soil microbial communities are still under-investigated. By using the nested sampling strategy, we investigated soil prokaryotic and fungal communities in alpine steppes being either grazed or fenced on the Qinghai-Tibet Plateau (QTP), focusing on their species interactions indicated by the co-occurrence network and spatial patterns based on changes in network node numbers along the increasing sampling area. Our results showed that grazing consistently enhanced the network complexity of soil prokaryotic community in different sites, possibly because grazing enhanced environmental heterogeneity and resource pulses while restricted microbial dispersal through physical disturbances. In contrast, grazing inconsistently influenced the fungal network complexity, likely due to the deterministic dominantly assembly of fungal community and heterogeneous grazing effects on environmental properties in different sites. Different from the spatial pattern for the community richness, the richness for interacted or co-occurred species tended to increase along the sampling area before 64 m2 for fungi and 256 m2 for prokaryotes, respectively, but then decreased after these tipping points, no matter with grazing or fencing. We find that the grazing effects on microbial network complexity depend on community assembly mechanisms, and changes in the network node number along the sampling area presented a reverse U shape spatially. These findings provide novel clues for a comprehensive understanding of grazing consequences in alpine grasslands.

The spatial pattern and community assembly processes of soil microbial taxa are critical for understanding biodiversity formation and maintenance mechanisms. While rare fungal taxa likely exhibit distinct biogeographic patterns and assembly processes compared to abundant taxa, such differentiations remain poorly characterized, particularly at continental scales. Here, we investigated distance-decay patterns and underlying assembly mechanisms for abundant and rare fungal taxa in 129 soil samples collected across 4,000 km in Chinese northern grasslands, based on high-throughput sequencing data. A total of 208 abundant operational taxonomic units (OTUs, relative abundance > 0.1%) and 5,779 rare OTUs (relative abundance < 0.01%) were identified. Both abundant and rare fungal taxa showed significant distance-decay relationships (P < 0.001), but the turnover rate for rare taxa (0.0024 per 100 km) was nearly half that of abundant taxa (0.0054 per 100 km) based on the binary Bray-Curtis distance. The lower turnover of rare fungal taxa was likely due to their community assembly mechanism dominated by stochastic processes, which were less influenced by environmental gradients. In contrast, abundant taxa assembly was dominated by deterministic factors like soil variables and plant traits, which varied significantly along the geographic distance. Consistently, rare fungal taxa were also less sensitive to environmental changes, with a lower turnover rate by environmental distance (0.0027 vs. 0.0099) than abundant taxa. Our findings revealed that rare fungal taxa—shaped mainly by stochastic processes—had lower spatial turnover compared to abundant taxa, which are dominated by deterministic processes. This deepens our understanding of rare microbial biogeography.

Shallow groundwater dynamics on the Qinghai-Xizang Plateau (QXP)-the "Asian water tower" supplying freshwater to billions downstream-remain poorly understood despite their critical role in buffering climate impacts. Integrating more than 8000 in-situ groundwater records with multi-source remote sensing data, we present the first high-resolution assessment of shallow groundwater across the QXP's non-permafrost plains. From 2000 to 2020, groundwater depth has been decreasing at a rate of 0.02 m year- 1, adding approximately 31.44 Gt (Gigatons) of freshwater storage-directly countering the prevailing narrative of widespread water loss. By combining the maximum capillary rise height, this rise sustains ∼53 500 km2 of alpine ecosystems and is closely linked to increasing NDVI in emerging groundwater-dependent vegetation. These results reveal the possibility of ecosystem regime shift under shrinking groundwater depth, highlighting the importance of groundwater dynamics in understanding alpine ecosystem changes. With an estimated 426.6 Gt of remaining storage capacity in the vadose zone, we identify managed underground reservoirs as a promising but still prospective opportunity for climate adaptation. Our findings reveal that the QXP's shallow aquifers function as a dynamic, growing freshwater reservoir, challenging surface-water-centric views of water-tower vulnerability and suggesting that current studies may underestimate the resilience plateau's water system to ongoing global warming.

The Qinghai-Tibet Plateau spans a vast area with highly complex natural conditions, making ecosystem classification and the delineation of field boundaries challenging. Ongoing climate warming has intensified glacial melting and thawing of permafrost, further complicating surface hydrological processes. The frequent alternations of redox conditions driven by soil water saturation make it difficult to distinguish the boundaries between wetlands and grasslands. Consequently, traditional wetland definitions and classification frameworks are inadequate for depicting the spatiotemporal patterns of alpine wetlands on the plateau, constraining a deeper understanding of the ecosystems. In this study, a novel multilevel framework for wetland classification is proposed. By integrating long-time-series Landsat 5/7/8 images, multisource environmental data, and field investigations, a 30-m annual dataset of alpine wetlands distribution on the plateau for 2000–2020 was constructed on the basis of a cloud computing platform to analyze spatiotemporal patterns and their driving factors. The results indicate that the overall classification accuracy is 0.92, which is highly consistent with that of high-resolution images. Over the past two decades, the total area of alpine wetlands on the plateau ranged from approximately 9.0×104 to 1.2×105 km2, accounting for approximately 4

Global warming is shifting rainfall patterns towards fewer but more intense events, leading to longer intervals between rainfall events. However, the effects of these changes on hydrological processes and elemental transport in the highly vulnerable periglacial ecosystems remain largely unexplored. This study used a rainfall manipulation experiment to examine how altering rainfall intervals with fixed seasonal total precipitation affects hydrological processes and carbon, nitrogen, and phosphorus export in a periglacial grassland on the Tibetan Plateau. Four treatments were applied on a uniform slope (similar to 11 degrees): natural rainfall (CK), 3-day interval (P1), 7-day interval (P2), and 11-day interval (P3). Our results revealed that altered rainfall patterns reshape hillslope water balance. Specifically, P1 increased surface runoff (SR), soil water storage (SWS) and soil evaporation while reducing deep percolation (DP), because frequent small rainfall events enhanced soil water retention in the 0-30 cm layer while limiting infiltration into deeper layers. In contrast, P2 and P3 significantly increased SR while reducing SWS, resulting in a non-significant effect on DP. Moreover, distinct thresholds were identified for the generation of SR and DP at 1.8 mm and 1 mm rainfall, respectively. Additionally, altered rainfall patterns significantly increased the losses of dissolved carbon, total nitrogen and nitrate in SR but did not affect ammonium and phosphate. Notably, nitrate losses exhibited a nonlinear response to changing rainfall patterns, which peaked at a 7-day interval instead of 11-day interval with the highest single rainfall amount. This suggests that the extreme rainfall events may not affect the dissolved nitrogen exports in the periglacial grasslands as it requires not only a large single rainfall amount but also high rainfall frequency. Therefore, future studies should consider the interactions between rainfall frequency and single event size to disentangle the responses of hydrological and biogeochemical processes to climate extremes in these fragile periglacial ecosystems, thus supporting sustainable watershed management for the Asian Water Tower.

Ecosystem stability is a fundamental attribute that underpins the delivery of essential ecosystem services. However, most research has primarily focused on the temporal stability of biomass, overlooking the multidimensional nature of stability that cannot be captured by a single metric. In this study, we investigated the effects of nitrogen (N) and phosphorus (P) additions on the stability of plant and microbial communities in alpine meadows of the Qinghai-Tibet Plateau. Our results demonstrated that N and P additions significantly increased plant community biomass but reduced the diversity of plants, prokaryotes and fungi. Although N and P additions did not significantly reduce biomass stability, a decreasing trend was observed. More importantly, compositional stability was significantly reduced by nutrient additions, with differing underlying mechanisms. Nitrogen addition primarily reduced community compositional stability by decreasing prokaryotic diversity, affecting plant diversity and the stability of subdominant species. In contrast, P addition mainly reduced the compositional stability of dominant species, thereby decreasing overall community stability. Furthermore, N addition significantly decreased the network stability of both prokaryotic and fungal communities. Importantly, microbial diversity and network properties were significantly correlated with plant community stability, highlighting the interconnectedness of above- and belowground communities. Our findings emphasize the need for future research to adopt a multidimensional approach to ecosystem stability, integrating both compositional and functional aspects of plant communities, and incorporating microbial diversity and network characteristics. Through the nitrogen and phosphorus addition experiments conducted in alpine meadows of the Qinghai-Tibet Plateau, this study found that while nutrient addition increased plant biomass, it significantly reduced the stability of plant and microbial communities. Nitrogen addition primarily affected the stability of secondary dominant species composition by reducing prokaryotic microbial diversity, while phosphorus addition directly weakened the stability of dominant species composition. Microbial network stability was significantly correlated with plant diversity and biomass stability, revealing the intrinsic link between aboveground and belowground communities. This study emphasizes that future assessments of ecosystem stability need to integrate multidimensional indicators and focus on the interaction mechanisms between plants and microorganisms. (sic)(sic):(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic)(sic)(sic)(sic)(sic)(sic)(N)(sic)(sic)(P)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic)(sic)(sic):1)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic),(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).2)(sic)(sic)(sic),(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic):(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic),(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic);(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).3)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic), (sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic)(sic).

Abstract Understanding how plant functional groups respond to nutrient enrichment is essential for predicting alpine ecosystem dynamics under global change. Although biomass responses to nitrogen (N) and phosphorus (P) additions have been widely investigated, shifts in plant N uptake strategies among functional groups under nutrient enrichment remain poorly understood. We conducted a 7‐year field experiment in an alpine grassland on the Qinghai‐Tibetan Plateau using factorial N (0, 7.5, 15 g m −2 yr −1 ; N0, N1, N2) and P (0, 7.5, 15 g m −2 yr −1 ; P0, P1, P2) additions. To evaluate N‐form uptake, we performed an in situ 15 N labeling experiment using ammonium (NH 4 + ), nitrate (NO 3 − ) and glycine, and collected plant and soil samples 4 h after label application. NH 4 + was the dominant N source taken up by the plant community, with community‐level uptake enhanced under moderate N (N1) but showing no further increase under higher nutrient inputs, indicating a non‐linear threshold response. Functional groups exhibited contrasting strategies: sedges primarily preferred NH 4 + and responded most strongly to N1, grasses showed plastic uptake across treatments, and forbs increased NO 3 − and glycine‐derived N uptake under N2P2, suggesting niche differentiation. Plant N uptake was weakly related to soil physicochemical variables but strongly associated with plant N:P stoichiometry, underscoring the predominance of intrinsic plant traits over soil conditions in driving functional group N uptake. Our results suggest that long‐term nutrient enrichment alters N‐form uptake partitioning among functional groups, with potential consequences for niche complementarity and community composition in alpine grasslands.

Climate change is driving unprecedented shifts in global precipitation patterns, yet how these hydrological perturbations alter the relationship between plant diversity and microbial biodiversity, both of which play integral roles in ecosystem functioning, remains a key unresolved issue in biodiversity-ecosystem function research. Here we conducted a large-scale biodiversity experiment in a semi-arid grassland, manipulating both plant diversity (monocultures to 16-species mixtures) and precipitation patterns (50%, 100%, and 150% of ambient rainfall) to investigate how plant diversity modulates soil prokaryotic and fungal biodiversity under altered precipitation. We found that high plant diversity enhanced prokaryotic diversity, network complexity, and stability under the ambient precipitation, consistent with insurance hypothesis. However, these positive effects were environmentally contingent. Under increased precipitation, the positive effect of plant diversity on prokaryotic diversity was strengthened (+100%). Under decreased precipitation, plant diversity no longer correlated with prokaryotic diversity or network stability; instead, it increased prokaryotic network complexity, consistent with stress gradient hypothesis. For fungi, plant diversity showed no correlation with diversity, however, it induced a U-shaped relationship with network complexity under ambient precipitation. Under decreased precipitation, both fungal diversity and network complexity exhibited similar U-shaped patterns as plant diversity increased. Furthermore, we revealed that soil total organic carbon positively drove prokaryote, while nitrate nitrogen negatively regulated fungi. Crucially, precipitation changes reshaped the plant-microbe relationship by changing soil carbon-nitrogen dynamics. These findings advanced our understanding of climate-plant-microbial feedback mechanisms and provided critical insights for sustainable grassland management under global change.

Hydrological connectivity (HC) is crucial for sustaining ecosystem structure and function, especially under conditions of degradation. However, the mechanisms linking HC to degradation have not been fully explored. We conducted a comprehensive ecological assessment across 22 sampling sites (110 plots total) with varying degradation and HC levels in the Yellow River source region during the peak growing season (August 2021), using a multi-method analytical approach including two-way ANOVA, NMDS, and structural equation modeling (SEM). HC significantly mediates degradation effects on plant community structure, productivity, and soil nutrients in alpine grasslands. Although degradation reduced most soil nutrients (p < 0.001) except total potassium (TK), high HC offsets total nitrogen (TN) loss in degraded grasslands (p < 0.01); however, it enhanced TK and reduced total phosphorus (TP) in undegraded grasslands (p < 0.01). High HC increased and decreased aboveground biomass (AGB) in degraded and undegraded grasslands (p < 0.001), respectively. Moreover, regardless of degradation status, HC mediated soil properties’ effects on plant diversity through elevated water and nutrient exchanges. In degraded grasslands, high HC-induced hydrological disturbance led to a negative biodiversity–productivity relationship. In undegraded grasslands, enhanced HC brought disturbances and shifted the negative biodiversity–productivity relationship to a positive one. Our results highlight that HC affects the mechanisms maintaining plant diversity and productivity in degraded alpine grasslands. Therefore, ecosystem restoration strategies in mountainous regions should be tailored to local hydrological conditions.

1. Plant species coexistence in nutrient-limited ecosystems is often maintained by niche differentiation in nitrogen (N) uptake. However, how this partitioning responds to changes in nutrient stoichiometry is still poorly understood. 2. We conducted an in situ N-15 labelling experiment in an alpine grassland on the Qinghai-Tibetan Plateau, using NH4+, NO3-, and glycine tracers under a 6-year factorial N and phosphorus (P) additions (0, 7.5, 15 g N m(-2) year(-1); 0, 7.5, 15 g P2O5 m(-2) year(-1)), thereby generating a gradient of soil available N:P ratios ranging from 1.2 to 9.0. Six coexisting plant species were assessed, including dominants (Kobresia pygmaea, Kobresia humilis, Stipa aliena) and subdominants (Poa annua, Potentilla bifurca, Potentilla nivea). 3. All species preferred NH4+, but dominant species intensified inorganic N uptake, while subdominants increased glycine use under nutrient additions. Specifically, in dominant species, NO3- uptake increased to similar to 2.5-5.9 times that in CK, whereas glycine uptake decreased by 75%-78% under high N and P addition levels (N2P2). In contrast, subdominants increased glycine uptake to similar to 5.8-9.6 times that in CK. This divergence enhanced N-form niche differentiation, which was positively correlated with species evenness in the alpine grassland ecosystem. Niche differentiation was most pronounced when soil available N:P ratios were below similar to 1.6, corresponding to the half-saturation point of the N-form differentiation curve. 4. Our findings highlight that changes in soil N:P stoichiometry induced by N and P addition drive below-ground N partitioning and promote coexistence through resource complementarity. These results provide a mechanistic basis for biodiversity maintenance in alpine grasslands under nutrient limitation.Read the free for this article on the Journal blog.

Ecosystem hysteresis, the occurrence of catastrophic transitions due to external disturbances, is a prevalent phenomenon in dynamic ecosystems. Understanding hysteresis is essential for predicting ecosystem responses and developing effective restoration strategies. However, the intrinsic dynamics quantifying the positive-negative feedback in driving hysteresis and its intensity remain undisclosed. We introduce a quantitative framework to address hysteresis by assessing ecosystem states and feedback loops, which diverges from prior phenomenological theories of hysteresis. Employing this framework, a generalized mechanism model is proposed to estimate positive-negative feedback strengths and defines the irreversible potential of hysteresis to determine its intensity. We identify a dimensionless critical constant that indicates whether hysteresis occurs. The model effectively captures both forward and backward trajectories of hysteresis across various ecological scales. The direction of state transitions may be predicted using unidirectional data. Our findings offer a universal framework for predicting and mitigating catastrophic state shifts of ecosystems.

Artificial intelligence (AI) has rapidly advanced soil organic carbon (SOC) estimation by improving spatial resolution and predictive accuracy. However, whether these advances support reliable carbon stock accounting under complex environmental conditions remains insufficiently examined. This commentary argues that mountain systems provide a natural stress test for evaluating the robustness and applicability of AI-based SOC estimation. Characterized by steep climatic gradients, strong topographic control, variable soil thickness, geomorphic instability, and sampling bias, mountain environments amplify structural weaknesses that may remain hidden in more homogeneous landscapes. We identify three failure modes revealed under mountain conditions: regime-dependent shifts in predictor meaning that undermine transferability, scale mismatch between point prediction and stock aggregation, and weak process constraints that lead to physically implausible spatial patterns. Building on this diagnosis, we outline the core capabilities required for next-generation AI frameworks, including spatially structured representation learning, context-conditioned inference across environmental regimes, process-aware constraints, and explicit applicability-domain assessment. We further propose a mountain-tested benchmark that links these failure modes to cross-regime generalization, scale-and-depth coherence, process plausibility, uncertainty attribution, and applicability-domain mapping. By positioning mountains as diagnostic environments rather than merely challenging terrain, this perspective reframes AI-based SOC estimation from tool driven prediction toward defensible, process-aware carbon accounting.

Grasslands, as dominant terrestrial ecosystems, significantly influence soil microbial communities through alterations in soil properties. However, their effects on spatial patterns of soil microbial communities are still under-investigated. To address this, we quantified taxa-area (TAR) and node-area (NAR) relationships for prokaryotic and fungal communities across temperate steppe (TS), alpine steppe (AS), and alpine meadow (AM). Our findings indicated that the spatial turnover of both prokaryotic and fungal communities were higher in alpine steppe and alpine meadow than in temperate steppe, mirroring the gradient of soil environmental heterogeneity. Notably, overall species richness increased logarithmically with sampling area in all grasslands; in striking contrast, co-occurring richness exhibited an increasing and then decreasing trend in AS and AM, but declined monotonically in TS, indicating that microbial interaction networks collapse once a critical spatial threshold is exceeded regulated by ecosystem type and environmental heterogeneity. In growing season, the stochastic dominance in prokaryotic assembly (Normalized stochasticity ratio = 0.71-0.89) and deterministic dominance in fungal assembly (Normalized stochasticity ratio = 0.23-0.37) can be explained by their differences in niche breadth and migration rate. These scale-dependent biogeographic patterns demonstrate that grassland type impacts distinct interactions and spatial patterns of microbial communities. These findings provide novel insights into a comprehensive understanding of how grassland type mediates soil microbial community.

Accurately reconstructing Holocene vegetation dynamics in ecologically sensitive regions is critical for understanding past climate-vegetation interactions and predicting future ecosystem responses. The Yanshan region, a key climatic and ecological transition zone in northern China, has experienced significant vegetation changes during the Holocene, yet previous reconstructions have largely relied on qualitative or semi-quantitative approaches, with limited spatial coverage and temporal reconstruction. In this study, we applied biomization of 33 high-quality fossil pollen records to quantitatively reconstruct the spatiotemporal vegetation succession of the Yanshan region over the past 12,000 years. Our results showed a general trend of steppe-forest-steppe transitions, primarily driven by variations in East Asian summer monsoon (EASM) precipitation. Spatially, vegetation characteristics and succession differ significantly across topographic units. A significant increase in forest cover occurred in the plain after similar to 5 ka BP, which contrasted with the similar Holocene pattern in the plateau and mountain. This difference may be attributed to transgression of the Bohai Sea into the plain before similar to 6 ka BP. Notably, a reversal of forest decline occurred after 3 ka BP, particularly in the plain and mountain areas, which is attributed to intensified human activities, including economic tree cultivation and secondary forest expansion. These findings highlight the dominant role of precipitation in regulating regional vegetation, while underscoring the increasing influence of anthropogenic factors in the late Holocene landscape transformation. This study provides a robust quantitative framework for understanding long-term vegetation dynamics and offers critical insights for ecological restoration and climate adaptation strategies in northern China.

Climate change reshapes global terrestrial ecosystems via rising temperatures and shifting precipitation, jointly regulating soil organic carbon (SOC) dynamics. Multi-factorial experiments are vital for unraveling climatic interactions, yet most only apply one warming level, leaving ecosystem nonlinear responses poorly explored. Based on a five-year field experiment in an alpine meadow, we examined the interactive effects of multi-level warming (+0, +1, +2 and +4 °C) and increased precipitation (+50%) on SOC and its nonlinear responses to warming. We found that +4 °C warming combined with increased precipitation led to a 55% increase in SOC in the topsoil (0–10 cm), whereas warming alone had negligible effects. This SOC increase is primarily attributed to the accumulation of mineral-associated organic carbon linked to fungal residues. Lower warming levels (+1 °C and +2 °C) had little impact on mineral-associated organic carbon, fungal residues, and bacterial biomass, whereas these variables increased markedly at +4 °C only under increased precipitation. This consistent gradient trend revealed a precipitation-dependent nonlinear response of the three belowground variables to warming, with identified functional thresholds falling within the 2–4 °C range. These findings highlight the need to account for interactive climate drivers and potential nonlinear ecosystem responses to refine carbon–climate feedback projections. Combined +4 °C warming and increased precipitation led to a 55% increase in soil organic carbon in the topsoil (0–10 cm), whereas warming alone had negligible effects, based on a five-year experiment in an alpine meadow on the northern Qinghai–Tibet Plateau, China.