Assessment of soil organic carbon (SOC) dynamics is often inadequately represented in empirical measurements because of the significant heterogeneity in soil structure and physico-chemical properties. Partitioning soil carbon (C) emissions into autotrophic and heterotrophic respiration is essential for understanding CO2 flux sources, but inconsistencies in their magnitude and responses reveal a knowledge gap in partitioning methodologies and their impact on respiration estimates. Utilizing data from an eight-yr field mesocosm study in a temperate oak forest, we computed C emissions from multiple components based on the metabolic theory. Our theoretical calculations of soil C emissions from various treatments were validated against periodic field measurements of soil respiration over an eight-year period. The optimized computations, which included annual precipitation data and accounted for biomass C from litter, roots, and microbes, closely aligned with field measurements of soil respiration across varying treatments. These results showed that fine root and microbial biomass jointly drove temporal variations in soil C emissions, while interannual precipitation variability plays a secondary role. This study confirms the feasibility of using metabolic theory to quantify soil C emissions and highlights the critical role of fine roots and soil microbial biomass, emphasizing the need for a deeper understanding of these factors in SOC budget assessments.
The present study aims to determine the potential impact of recent past, present-day and future climate conditions—along with silvicultural interventions—on the “intrinsic Water Use Efficiency” (iWUE). iWUE, defined as the amount of carbon assimilated per unit of water lost through stomata, is a valuable metric that reflects the combined effects of climate change and forest management on carbon and water balance in forest ecosystems. We studied these effects on a European beech (Fagus sylvatica L.) forest, one of the most common tree species in Europe, in a unique pre-Alpine site in Italy subjected to different silvicultural treatments in the past. Therefore, we analyzed iWUE derived from the δ13C measured isotope for the period 2013–2019 under three different silvicultural schemes observed at the study site. Opposite to what was expected, no statistically significant differences were found on iWUE between the treatments (ANOVA: p-value = 0.21) with a mean value for all treatments ranging from 94 μmol mol–1 and 98 μmol mol–1. To explore future dynamics, we used a validated process-based biogeochemical model to simulate iWUE under two climate scenarios (RCP4.5 and RCP8.5) and the same three silvicultural treatments. Again, silvicultural practices showed little effect on iWUE, while differences were evident between climate scenarios and time periods. iWUE increased between the first (2019–2029) and last (2040–2050) decades of simulation by 20.9%, 20.5% and 19.5% for the “Control”, “Traditional” and “Innovative” treatments, respectively. In conclusion, in the past and for the next half-decade, silvicultural treatments, at least at the study site, may not influence much the iWUE of beech forests even if it will increase remarkably under climate change.
Research on the sensitivity of carbon isotope composition (δ13C) dynamics to temperature (Q10) and its influencing factors in the process of ecosystem respiration (Re) can accurately predict the trend for ecosystem carbon release with global warming for assessing ecosystem carbon sequestration capacity. We used stable isotope techniques to monitor canopy CO2 concentration and δ13C in a cold-temperate Larix gmelinii forest in Northeast China. δ13C values were also analyzed in plant and soil samples across five stand types. The sensitivity of δ13C dynamics to temperature during Re and the main factors affecting the variation in Q10 values were determined. Carbon isotope composition of ecosystem respiration (δ13CRe), autotrophic respiration (δ13CRa), and heterotrophic respiration (δ13CRh) decreased with increase in temperature, and δ13CRa < δ13CRe < δ13CRh. The contribution of Ra and Rh to Re were 51.57 and 48.42%, respectively. Temperature and precipitation had inhibitory effects on Q10, whereas soil organic carbon and total nitrogen had stimulatory effects. Autotrophic respiration is the dominant pathway for carbon release in this ecosystem. Heterotrophic respiration, and particularly maintenance respiration, are more temperature-sensitive. Rising temperatures and precipitation reduce the δ13C sensitivity to temperature.
Bark beetles are one of the greatest threats to coniferous forests in Europe. Pheromone traps are currently the most effective method of controlling the mass infestations that some of them are known to cause. However, the efficiency of pheromone traps has not yet been sufficiently researched, especially in relation to other important variables that are influenced by the current global changes. Ips typographus L. and Pityogenes chalcographus L. are two economically important bark beetle species that cause major damage to conifers in Serbia. In the present study, we evaluate the efficiency of two commonly used pheromone traps set during a three-year experiment in the Tara National Park in Serbia. During this period, 672,934 ind. of I. typographus and 2,597,578 ind. of P. chalcographus were caught. Our results show that wet traps were about 1.8 times more efficient than dry traps for both species studied. Furthermore, our results indicate that the optimal temperatures for bark beetle flight are between 22 °C and 26 °C, with substantial swarming behaviour occurring at 16.5 °C. At the same time, the data also show a negative correlation between the number of individuals caught and temperatures above 16 °C, suggesting that temperature is probably not the only key factor influencing bark beetle activity.
The Hengduan Mountains, a biodiversity hotspot on the Tibetan Plateau, are undergoing rapid elevation-dependent warming, profoundly altering alpine forest dynamics. Larix species in this region dominate high-altitude treelines and represent the world’s lowest-latitude natural populations of larch forests. However, climate-growth relationships of Larix species at the whole regional scales have received little attention. To address this gap, we investigated the spatiotemporal variability of radial growth in Larix species and their climatic sensitivity across this region using a network of 26 tree-ring chronologies spanning 1960–2022. Hierarchical clustering identified three distinct geographical clusters (southwestern, central, and northeastern), revealing divergent growth trajectories and climate responses. Results demonstrated that growing-season temperature was the primary climatic driver of Larix radial growth, but its influence varied spatially: southwestern populations correlate strongly with May–August mean temperatures, central populations with May–August mean temperatures, and northeastern populations with April–June maximum temperatures. Each cluster exhibited unique growth trends and thermal sensitivities before and after the rapid warming since 1990. Meanwhile, an inter-individual response divergence became apparent under warming, reflecting intensified competition and microhabitat-scale stress, which highlights the limitations of traditional population-level climate response models that assume a uniform response. Spatial heterogeneity in climate-growth relationships reflected synergistic thermal-hydrological effects and species-specific adaptations, with warming enhancing carbon sequestration in moisture-sufficient areas but threatening high-elevation ecosystems through growth suppression and treelines instability. These findings underscore the need for regionally tailored conservation strategies to address climate-driven ecological imbalances in alpine forests.
Temperate secondary forests play an important role in climate change mitigation and the global carbon cycle, but their variations and drivers of ecosystem carbon storage (ECS) during succession remain unclear. In this study, ECS (trees, shrubs, herbs, litter, coarse woody debris—dead or fallen trees and soil) and environmental factors (temperature, humidity and soil nutrients) were measured in four forest types: Abies nephrolepis and Pinus koraiensis, Fraxinus mandshurica and P. koraiensis, Tilia amurensis and P. koraiensis, and Quercus mongolica and P. koraiensis, in three successional stages (early shrub-grass lands, middle secondary forests and old-growth forests) in temperate Changbai Mountains, to reveal the dynamics of ECS and its allocation patterns during succession, and its formation mechanisms. The results show that: (1) ECS ranged from 49.0–66.1 to 153.8–197.0 and 308.2–446.4 Mg ha−1 in early, middle and late successional stages, respectively; (2) ECS of the four secondary forests recovered to 48.9% of old-growth forest levels after 40 years of succession; their ecosystem carbon sequestration potential ranged from 154.4 to 249.3 Mg ha−1, mainly contributed by vegetation (89.7–94.0%), whereas, soil contribution was smaller (6.0–10.3%).These secondary forests may take at least 100 year to recover to the level of old-growth forest ECS at the current recovery rate; (3) The proportion of vegetation increased with succession in ECS from 3.3–4.6% at the early succession to 74.2–82.8% at the late succession. Moreover, vegetation carbon storage mainly depended on a few pioneer tree species (49.1–66.4%) (middle succession stage) and the climax tree species P.koraiensis and 1–2 associated species (87.5–89.7%) (late succession stage). The contribution of dominant tree species to vegetation carbon storage was significantly greater than that of the tree species diversity; (4) The ECS and vegetation carbon storage were promoted by stand conditions (average DBH and stand density), while soil carbon storage was jointly driven by soil organic carbon and ammonium nitrogen and stand conditions. Our research indicates that temperate secondary forests have considerable carbon sequestration potential (mainly dependent on vegetation) during succession and strengthening the cultivation of the climax species P.koraiensis and associated tree species will help to realize this carbon sequestration potential and better cope with climate change.
Understanding the mechanisms of species diversity maintenance is crucial for appreciating community assembly and predicting responses to global climate change. Niche differentiation is one of the most important mechanisms underlying biodiversity maintenance across ecosystems. However, direct evidence for niche differentiation remains scarce in subtropical speciose forests. In this study, a 25-ha (500 m × 500 m) subtropical montane deciduous broadleaved forest dynamics plot in Shennongjia national park was developed to assess species-habitat associations across life history stages. Five habitat types were identified using multivariate regression trees and mapped to 625 20 m × 20 m quadrats. Torus-translation randomization tests identified species-habitat associations across life forms and life stages. Out of 105 species, 81 were significantly associated with at least one habitat type and 65 associated with elevation or convexity (49 species with elevation and 36 with convexity). Across all life stages, saplings were most strongly related to low elevation habitats, while juveniles and mature trees most often correlated with the “low convex slope” habitat type. Canopy and shrub species were positively correlated with the “high convex slope” and “low convex slope” habitat types, respectively. In conclusion, niche differentiation during regeneration (based on topographic heterogeneity) is essential for stable multi-species coexistence and the maintenance of biodiversity in subtropical speciose forests. Future studies are needed to examine how demographic rates shift along environmental gradients of convexity and elevation, providing a more in-depth understanding of niche differentiation in forest ecosystems.
Accurate individual tree species classification is essential for forest inventory, management, and conservation. However, existing methods relying primarily on single-source remote sensing data (e.g., spectral, LiDAR, or RGB) often suffer from insufficient feature representation and noise interference, particularly in subtropical forests with high species diversity, leading to increased classification errors. To address these challenges, we proposed the Multi-source Tree Species Classification Fusion Network (MTSCFNet), a novel deep learning framework that integrates RGB imagery, LiDAR-derived feature maps, and GF-2 satellite data through a modified UNet backbone, which incorporates a three-branch encoder and a Triple Branch Feature Fusion (TBFF) module within a middle fusion strategy. We evaluated the MTSCFNet in Chinese-fir mixed forests located in the Shanxia Forest Farm, Jiangxi Province, China. The results showed that: (1) MTSCFNet outperformed four baseline models, achieving Macro F1 (0.78 ± 0.01), Micro F1 (0.93 ± 0.01), Weighted F1 (0.93 ± 0.01), a Matthews correlation coefficient (MCC) (0.89 ± 0.01), Cohen’s ĸ (0.89 ± 0.01), and mIoU (0.69 ± 0.01), with respective improvements of 4.05% in Macro F1, 1.89% in Micro F1, 0.09% in Weighted F1, 1.67% in MCC, 1.64% in Cohen’s ĸ, and 5.92% in Mean IoU over the second best model, SwinUNet; (2) Compared to the best two-source combinations (R + S, R + L), MTSCFNet achieved up to 1.50%, 3.28%, 3.42%, 6.72%, 6.76%, and 3.51% higher Macro F1, Micro F1, Weighted F1, MCC, Cohen’s ĸ, and mIoU, and up to 8.11%, 2.63%, 2.88%, 5.01%, 4.99%, and 11.48% improvements over single-source inputs, while also exhibiting the lowest variability, indicating strong robustness; (3) Under different fusion strategies, MTSCFNet with middle fusion surpassed early and late fusion by up to 15.31%, 3.74%, 3.99%, 7.66%, 7.76%, 22.33% and 24.13%, 5.76%, 6.20%, 11.48%, 11.57%, 32.96% in Macro F1, Micro F1, Weighted F1, MCC, Cohen’s ĸ, and mIoU, respectively, validating the effectiveness of feature-level multi-modal integration; (4) In cross-region transfer experiments, MTSCFNet demonstrated strong spatial generalizability, achieving average scores of 0.78 (Macro F1),0.87 (Micro F1), 0.86 (Weighted F1), 0.59 (MCC), 0.59 (Cohen’s ĸ), and 0.68 (mIoU), and outperformed SwinUNet by up to 38.80%, 9.40%, 18.58%, 22.48%, 26.17%, and 33.00% in Macro F1, Micro F1, Weighted F1, MCC, Cohen’s ĸ, and mIoU across varying forest densities. Overall, MTSCFNet offers a robust, accurate, and transferable solution for tree species classification in complex subtropical forest environments.
The decomposition of litter by microbial communities is essential for ecosystem functioning. High nitrogen deposition, interacting with the flexible nutrient features of litter, can disrupt microbial succession. However, little is known about the specific links between microbial assembly and nitrogen addition, particularly in the in-situ litter layer. In a Korean pine (Pinus koraiensis) plantation, we investigated how eight years of nitrogen addition (0, 20, 40 and 80 kg ha−1 a−1) affect litter layer microbial community, assessing changes in abiotic properties and microbial community succession. The findings revealed complex influences of nitrogen addition on litter abiotic properties throughout the decomposition stage, such as contrasting influences on NH4+ and NO3−, where NH4+ was elevated but NO3− was decreased. The effect on microbial community structure and assembly was highly stage-dependent. In the early stage, bacterial assembly was driven by stochastic processes (dispersal limitation). During the middle and late stages, high nitrogen addition shifted bacterial assembly from predominantly deterministic processes (heterogeneous selection) to stochastic processes (drift). However, it did not affect the predominance of stochastic processes during fungal assembly (dispersal limitation and drift). Thus, the influences of nitrogen addition on bacterial and fungal networks were inconsistent, with the stage-specific sensitivity differences between bacteria and fungi. Specifically, high nitrogen addition decreased bacterial stability and complexity, but promoted fungal stability over time. However, it did suppress fungal niche differentiation in the late stage. These results demonstrate that high nitrogen conditions influence litter abiotic properties and the associated microbial traits, such as community assembly, particularly during the late decomposition stage.
Forest litterfall is a key contributor to soil carbon accumulation. However, existing studies have primarily foused on site-level observations or annual-scale assessments, while the intra-annual dynamics and spatial distribution of forest litterfall at the national scale remain poorly understood. In turn, this limitied comprehensive spatiotemporal assessments of forest carbon sequestration capacity. In this study, we compiled 4,223 monthly litterfall observations from 88 forest sites across China and integrated multi-source environmental variables to develop a Transformer-CatBoost hybrid prediction model for estimating the spatiotemporal patterns of forest litterfall across three representatibe years (2002, 2009 and 2018), corresponding to major stages of ecological restoration efforts in China. Model evaluation demonstrated strong predictive performance (R2 = 0.74), effectively capturing the nonlinear relationships driving litterfall dynamics. By incorporating national forest area changes in 2002, 2009, and 2018, the study further revealed the spatiotemporal evolution of forest structure under large-scale ecological restoration programs. Based on nationwide monthly-scale modeling results, we systematically characterized the spatial distribution and seasonal variation of litterfall production across China’s forests, with an anuual average of 547.04 ± 0.23 g m⁻2 (or 479.13 ± 0.20 g m⁻2 excluding January and December). Furthermore, using a fixed carbon conversion rate, we estimated national carbon content of forest litterfall at 290.4 Tg in 2002, 311.9 Tg in 2009, and 354.1 Tg in 2018, indicating a clear increasing trend. This study represents the nationwide, monthly-scale modeling and prediction of forest litterfall in China.
Precise phenological regulation is critical for temperate trees to survive the winter. However, the underlying mechanism is still unclear. Here, we found that Pag4CL3 coordinately modulates lignin biosynthesis and melatonin accumulation in 84 K poplar (Populus alba × P. glandulosa). Overexpression of Pag4CL3 or Pag4CL5 increased the lignin content in stem but reduced plant growth. In contrast, knockout of either gene reduced stem lignin monomers, promoted growth, and improved cold tolerance, with Pag4CL3 mutants (4cl3) exhibiting more pronounced resistance. PagSNAT2, which encodes a key enzyme in melatonin (MT) biosynthesis, is markedly upregulated in the 4cl3 mutant. Consistent with this, overexpression of PagSNAT2 promoted MT accumulation in 84 K poplar, and the 4cl3 mutant exhibited significantly higher MT levels in both autumn dormant and spring sprouting buds compared to the wild-type. Yeast two-hybrid (Y2H) and luciferase complementation assays further confirmed that Pag4CL3 directly interacts with PagSNAT2. Additionally, low temperature inhibited the binding of transcription factor PagDRS1 to the Pag4CL3 promoter and attenuating its suppression of melatonin synthesis. This study thus unveils a cold-responsive PagDRS1–Pag4CL3–PagSNAT2 regulatory module that balances structural formation and stress adaptation in trees, providing a theoretical basis and breeding strategy for developing poplar varieties with enhanced biomass and winter resilience.
Despite the critical role of plant resource allocation and soil characteristics in plant survival across different altitudes in subtropical China, the detailed dynamics of these interactions had not been previously well-documented. This investigation endeavored to examine the linkage between the characteristics of plant community functions and the physical and chemical attributes of soil at differing elevations within the mountainous forests of the Jinhua Mountain area. The results indicated that as altitude increased, most soil nutrient and moisture traits showed a declining trend, with Bulk Density (SBD) and Total Phosphorus (TP) initially increasing, then stabilizing or slightly rebounding. Specific Leaf Area (SLA) increased from 101.5 to 153.8 cm2 g−1 with altitude increase, but Leaf Dry Matter Content (LDMC) and Potential Maximum Height (Hmax) decreased (from 22.3 to 6.1 m). High-altitude shrub communities preferred environments with high SBD (6.8 g cm3) but limited moisture and nutrients, exhibiting traits of rapid nutrient uptake and photosynthesis, indicative of a fast-growing ecological strategy. In contrast, low-altitude tree communities displayed more conservative strategy traits. Redundancy Analysis (RDA) revealed that climate variables accounted for 53.09% of the variance in RDA1, highlighting the significant impact of mean annual temperature and precipitation on plant community traits. Soil variables, in contrast, explained 47.99% of the variance in RDA2. The Structural Equation Model (SEM) confirmed that the raised altitude enhanced plant nutrient acquisition capabilities while suppressing the plant's ability to retain soil nutrients, significantly reducing soil nutrient content. Furthermore, the decline in soil moisture retention capacity further promoted plant acquisition strategies, exacerbating soil nutrient scarcity in high-altitude regions. The findings of this study contributed to a more nuanced comprehension of the multifaceted interactions between plant communities and soil in subtropical ecosystems, providing a robust foundation for ecological management and strategies for preserving ecosystems.
Elevated ozone (O3) levels pose a threat to tree physiology, with responses varying among genotypes. This study investigated O3 tolerance in two poplar clones (107 and 546) exposed to elevated O₃ (E-O3), focusing on photosynthetic performance and stomatal dynamics. Clone 107 demonstrated remarkable resilience, with only minor late-season reductions in light-saturated photosynthesis (Asat), maximum photosynthetic rate (Amax), and chlorophyll content under E-O3. However, clone 546 exhibited severe impairment, including dysfunctional photosynthetic light-response curves with a reduction in stomatal limitation and increase in biochemical limitation, indicating shifted constraint dominance. Moreover, there was stomatal paralysis characterized by sluggish kinetics (prolonged t90gopening, reduced SLmax) and loss of bell-shape-like gs-PPFD response. Results also revealed collapsed instantaneous water-use efficiency (iWUE) due to decoupled stomatal conductance (gs) and CO2 assimilation. The gs response remained static under elevated O3 without corresponding photosynthetic benefit, leading to simultaneous carbon starvation and water loss. E-O3 also induced chlorophyll degradation and premature senescence in clone 546, suggesting chronic oxidative damage. These findings demonstrate that the sensitivity of clone 546 stems from systemic failure in stomatal regulation and biochemical compensation, while the tolerance of clone 107 reflects maintained photosynthetic stability. The results indicate that SLmax and iWUE dynamics can be utilized as key diagnostic indicators of O3 stress in trees and highlight the importance of selecting resistant genotypes such as clone 107 for O3-polluted regions.
The improper handling of outliers in the analysis of variance (ANOVA) presents a persistent challenge in forestry research, which may lead to biased results, inflated Type I error rates, and obscured scientific signals. The current practice is often an ad hoc method, potentially driven by a need to achieve statistical significance rather than principled scientific reasoning. This Editorial paper addresses this systemic issue by proposing a structured, step-by-step framework for the diagnosis and management of outliers. The framework guides researchers to first investigate the cause of an outlier (data error, measurement error, or genuine extreme value), then statistically assess its impact on ANOVA results and assumptions, and finally, make a transparent decision on its treatment. We strongly advise against the statistically problematic practice of replacing outliers with the mean of other replicates, as it violates data integrity and obscures true variability. Instead, we recommend robust alternatives, including data transformation, non-parametric tests, or the use of trimmed means. This approach aims to uphold statistical robustness and scientific integrity, thereby improving the rigor of forestry research and its publications.
The therapeutic effects of forests have become a highlighted research focus in global health studies in recent years. The distinctive microclimate conditions and landscape aesthetics of forests have already been demonstrated to be beneficial for the rehabilitation of people’s physical and mental health. To advance the development of local forest therapy programs in the southern Taihang Mountain Range of North China, we investigated the therapeutic potential of four typical local forests: species-rich Pinus tabulaeformis forest (PTF) and Pinus bungeana forest (PBF), and species-poor Platycladus orientalis forest (POF) and Robinia pseudoacacia forest (RPF). This was done by assessing microclimate conditions and landscape aesthetics related to human health in these forests, while key stand characteristics determining landscape aesthetics were also explored. Microclimate conditions were monitored from June to October 2023, with monthly data collected from the 1st to the 4th of each month between 09:00 and 11:00 AM and diurnal variations recorded every two hours from 08:00 to 20:00 during September 1–4, while temperature, humidity and wind speed variables were integrated into a comprehensive climate comfort index (CCCI) as a surrogate for human comfort. Landscape aesthetics were evaluated with both eye-tracking technology and Scenic Beauty Estimation (SBE) method, which served as proxies for landscape visual perception. A structural equation model (SEM) was also developed to identify key forest characteristics determining landscape visual perception. The results showed that PM2.5 and bacterial concentrations in PTF and PBF were relatively lower than those in POF and RPF during summer and autumn. The CCCI values of PTF and PBF were 3.87 and 4.51, respectively, indicating superior air quality. Eye-tracking analysis revealed that PBF had the longest total gaze duration and highest number of gazes, possibly due to its trails, which provide high accessibility. SEM revealed that green coverage, tree height, shrub density, and herbaceous diversity jointly determined landscape visual quality (path coefficients: 0.98; 0.61; 0.71; –0.91), suggesting that integrated stand structure optimization is key to optimizing and enhancing these forests for their therapeutic values. These findings not only characterized the microclimate and aesthetic features of these forests but also identified key stand characteristics that affect their landscape aesthetics, providing a scientific basis for optimizing forest management works for forest therapy planning in the Taihang Mountains.
Understanding how the radial growth of major conifer species responds to climate change on the Tibetan Plateau is increasingly important. However, relevant studies in the middle of the Hengduan Mountains, eastern Tibetan Plateau have not been carried out. In this study, increment cores of Abies georgei, Picea likiangensis and Larix potaninii at the treeline (4260 m), and those of Pinus densata (3730 m), A. georgei (3560 m, 3330 m) and P. likiangensis (3560 m) from subalpine stands, were sampled. Radial growth responses of these conifers to climate change were analysed via dendroecological methods, and the impacts of climatic factors were investigated via partial least squares path modelling (PLS-PM) during the period 1952‒2023. Our results reveal that a rapid warming occurred after 1988, dividing the study period into two intervals: 1952‒1987 and 1988‒2023. A. georgei at the treeline showed improved growth whereas P. densata at the 3730 m site showed reduced growth during 1988‒2023. At the treeline, radial growth of the three conifers were significantly positively correlated with precipitation and the standardized precipitation evapotranspiration index (SPEI) in May/June, and A. georgei and P. likiangensis was negatively correlated with Tmax and Ped in May during 1988‒2023. Radial growth of P. likiangensis was negatively correlated with Tmax in the previous December during the same period. In the subalpine stands, tree ring chronologies of A. georgei at the 3560 and 3330 m sites, and P. densata at the 3730 m site were significantly positively correlated with May precipitation, P. likiangensis at the 3560 m site had significantly negative correlations with Tmax, Tmean, and Ped of the previous October over the period 1988‒2023. According to PLS-PM results, the warming-induced promotion of radial growth for A. georgei at the treeline was greater than the suppression by warming-induced drought. In contrast, warming had no significant effect on the radial growth of L. potaninii at the treeline and P. likiangensis at the 3560 m site. Early growing season drought caused by climate warming is the primary factor limiting the radial growth of subalpine conifers, and continued warming is expected to impact carbon sequestration and community composition in the Hengduan Mountains.