Plant diversity is essential for regulating ecosystem functions, yet its global-scale relationship with soil respiration, a critical component of the carbon cycle, remains unexplored. While plant productivity modulates ecosystem processes, understanding how plant species richness regulates soil respiration across net primary production gradients is key to predicting carbon-climate feedbacks. Here, we integrate two global plant species richness datasets (tree species richness and vascular plant species richness, including woody and non-woody plants) with a global soil respiration dataset estimated using a deep learning model trained on 6355 field observations. We demonstrate that plant diversity enhances soil respiration in low- to mid-productivity forests (<1300 g C m-2 yr-1), but this effect diminishes in high-productivity forests. By stratifying forests along a net primary production gradient and controlling for climatic, soil, and vegetation covariates via structural equation modeling, we reveal a context-dependent role of biodiversity. Greater plant species richness amplifies soil respiration in resource-limited systems but contributes minimally in high-productivity forests, where abiotic factors exert stronger influence. Our findings provide insights into how biodiversity influences soil carbon fluxes, revealing its dynamic role in shaping ecosystem carbon dynamics across productivity gradients.

Net primary productivity is a critical component of terrestrial carbon cycling and an essential indicator of ecosystem carbon sequestration capacity. However, separating aboveground and belowground net primary productivity (ANPP and BNPP) and understanding the driving mechanisms of their spatial patterns remain challenging across global terrestrial ecosystems. Here, we used a modified multilayer perceptron network (MLP) built upon an updated database containing 5184 field observations to predict the spatiotemporal patterns of ANPP and BNPP and identify their driving mechanisms across global terrestrial ecosystems at 0.05 degrees resolution. Results indicated that the MLP model satisfactorily predicted ANPP (R-2 = 0.74) and BNPP (R-2 = 0.73). Spatially, both ANPP and BNPP exhibited strong spatial heterogeneity, with a decreasing trend from the tropics toward the poles. Temporally, ANPP showed an increasing trend of 0.02 Pg C yr(-2), with a global mean of 33.4 +/- 0.5 (mean +/- standard error) Pg C yr(-1) from 1981 to 2018. Similarly, the mean total BNPP was 19.2 +/- 0.73 Pg C yr(-1), with an increasing trend of 0.05 Pg C yr(-2). Quantitatively, significant trends were observed, with 65.4% and 60.8% of land areas showing increasing trend of ANPP and BNPP (P < 0.01), respectively. The spatial patterns of ANPP were mainly influenced by temperature and precipitation, while BNPP was controlled by soil properties. These findings highlight the importance of distinguishing ANPP and BNPP to better understand the driving mechanisms of carbon allocation strategies. These findings are crucial for advancing the understanding of vegetation dynamics in response to global climate change and improving terrestrial ecosystem carbon modeling.

The Qinghai-Tibet Plateau (QTP), known as the "Third Pole", plays an important role in regional and global climate regulation, but it is highly sensitive to human disturbances. Despite increasing tunnel constructions in this region, its effects on soil organic carbon (SOC) composition and stability in high-altitude ecosystems remain unclear. In this study, soil samples were collected from tunnel-affected (TA) and control (CK) areas across three elevations (3240, 3420, and 3600 m) and two soil depths (0-10, 10-30 cm) on the eastern QTP. A three-way ANOVA was used to evaluate the effects of tunnel construction, elevation, soil depth and their interactions on SOC and its active and recalcitrant fractions. Elevation and soil depth significantly affected SOC fractions, whereas tunnel construction did not significantly affect SOC (p = 0.281), easily oxidizable organic carbon (p = 0.507), dissolved organic carbon (p = 0.841), recalcitrant organic carbon (p = 0.423), or SOC functional indices, including the stability index (p = 0.078), carbon quality index (p = 0.304), and carbon pool management index (CPMI; p = 0.680). In contrast, the C/N ratio was significantly affected by tunnel construction (p = 0.002). Overall, elevation and soil depth exerted stronger controls on soil carbon fractions than tunnel construction, and short-term tunnel disturbance did not produce detectable changes in shallow-soil SOC stability or quality, although soil stoichiometric balance responded an early stoichiometric response.

Net primary production (NPP), comprising both aboveground (ANPP) and belowground (BNPP) components, serves as a critical indicator for evaluating terrestrial ecosystem responses to climate change. However, poorly characterized BNPP spatiotemporal dynamics limit carbon allocation understanding on the Tibetan Plateau (TP), and the impact of modeling uncertainties (on different models and their parameter settings) is not well constrained. Thus, we developed 97 data-driven models (including linear, machine learning, and deep learning), integrating field observations, and remote sensing datasets to produce 1-km-resolution ANPP and BNPP across the TP (1981–2018). Our results demonstrated strong model performance for both ANPP (R2 = 0.80 to 0.88) and BNPP (R2 = 0.89 to 0.95), with the Rborist (ANPP) and xgbLinear (BNPP) models showing optimal performance. Spatially, linear and machine learning models exhibited strong consistency in ANPP and BNPP estimates, while deep learning outputs showed marked divergence, particularly in eastern TP. Temporally, we observed a significant decline in total ANPP (–0.003 Pg C·a–1) but an increase in BNPP (0.001 to 0.003 Pg C·a–1), with intermodel variability in annual totals ranging from 0.13 to 0.32 Pg C·a–1. The discrepancies stemmed from varying factor influences across models, with evapotranspiration, Bio3 (isothermality), and leaf area index primarily driving ANPP, while BNPP was mainly governed by Bio3, Bio14 (driest-month precipitation), and leaf area index. Notably, the increasing contribution of BNPP to carbon sequestration on the TP highlights the necessity of its independent quantification in global carbon models. This study also reveals uncertainties in vegetation carbon allocation and provides key benchmarks for developing biogeochemical models in climate-sensitive regions.

Understanding the dynamics of soil respiration (Rs) and its environmental drivers is crucial for accurately modeling terrestrial carbon fluxes. However, current methodologies often lead to divergent estimates and rely on annual predictions that may overlook critical interactions occurring at seasonal scales. A critical knowledge gap lies in understanding how temporal resolution affects both Rs predictions and their environmental drivers. Here, we employ deep learning models to predict global Rs at monthly (MRM) and annual (ARM) scales from 1982 to 2018. We then consider three main drivers potentially affecting Rs, including temperature, precipitation, and a vegetation proxy (leaf area index; LAI). Our models demonstrate strong predictive capabilities with global Rs estimation of 79.4 ± 5.7 Pg C year-1 for the MRM and 78.3 ± 7.5 Pg C year-1 for ARM (mean ± SD). While the difference in global estimations between both models is small, there are notable disparities in the spatial contribution of dominant drivers. The MRM highlights an influence of both temperature and LAI, while the ARM emphasizes a dominant role of precipitation. These findings underscore the critical role of temporal resolution in capturing seasonal variations and identifying key Rs-environment relationships that annual models may obscure. High temporal resolution Rs predictions, such as those provided by the MRM, are essential for capturing nuanced seasonal interactions between Rs and its drivers, refining carbon flux models, detecting critical seasonal thresholds, and enhancing the reliability of future Earth system predictions. This work highlights the need for further research into monthly and seasonal Rs variations, as well as higher timescale resolutions, to advance our understanding of ecosystem carbon dynamics in a rapidly changing climate.

Grassland biomass is a key indicator for assessing the health and productivity of grassland ecosystems. Over the past two decades, China's grasslands have undergone substantial changes. However, accurately estimating grassland biomass and its response to driving factors remains challenging. In this study, we predicted the spatiotemporal patterns of aboveground and belowground biomass (AGB and BGB) and examined their driving mechanisms at a 500 m resolution from 2000 to 2022, using a modified multi-layer perceptron (MLP) neural network combined with 2233 AGB and 1093 BGB observations, along with environmental variables across China. The MLP model demonstrated strong predictive performance for AGB (R-2 = 0.79, RMSE = 20.02 g C m(-2) yr(-1)) and BGB (R2 = 0.87, RMSE = 223.65 g C m(-2) yr(-1)). Spatially, average AGB decreased from southeastern to northwest China, while BGB was highest along the eastern edge of the Qinghai-Tibet Plateau. Temporally, AGB increased by 1.27 Tg C yr(-1) over the past 23 years, covering 68.19 % of grassland areas, whereas BGB showed no significant trend. Projections indicate that AGB will increase in 48.6 % and BGB in 34.81 % of grassland areas in the future. Total AGB and BGB were estimated at 163.5 and 1044.5 Tg C, respectively. Temperature and precipitation were the primary drivers of both AGB (23.47 % and 20.21 % of grassland areas) and BGB (25.24 % and 21.05 %). For AGB, the remaining drivers, in descending order of influence, were vegetation, human activity intensity, soil physical properties, soil chemical properties, and terrain. For BGB, they were human activity intensity, followed by vegetation, soil physical properties, soil chemical properties, and terrain. This study offers critical insights for grassland biomass dynamics and their driving mechanisms, providing a scientific foundation for policy-making and adaptive management to ensure the long-term resilience of grassland ecosystems.

Abstract Background Soil organic carbon (SOC) is a critical component of the global carbon cycle, and an accurate estimate of regional SOC stock (SOCS) would significantly improve our understanding of SOC sequestration and cycles. Zoige Plateau, locating in the northeastern Qinghai-Tibet Plateau, has the largest alpine marsh wetland worldwide and exhibits a high sensitivity to climate fluctuations. Despite an increasing use of optical remote sensing in predicting regional SOCS, optical remote sensing has obvious limitations in the Zoige Plateau due to highly cloudy weather, and knowledge of on the spatial patterns of SOCS is limited. Therefore, in the current study, the spatial distributions of SOCS within 100 cm were predicted using an XGBoost model—a machine learning approach, by integrating Sentinel-1, Sentinel-2 and field observations in the Zoige Plateau. Results The results showed that SOC content exhibited vertical distribution patterns within 100 cm, with the highest SOC content in topsoil. The tenfold cross-validation approach showed that XGBoost model satisfactorily predicted the spatial patterns of SOCS with a model efficiency of 0.59 and a root mean standard error of 95.2 Mg ha−1. Predicted SOCS showed a distinct spatial heterogeneity in the Zoige Plateau, with an average of 355.7 ± 123.1 Mg ha−1 within 100 cm and totaled 0.27 × 109 Mg carbon. Conclusions High SOC content in topsoil highlights the high risks of significant carbon loss from topsoil due to human activities in the Zoige Plateau. Combining Sentinel-1 and Sentinel-2 satisfactorily predicted SOCS using the XGBoost model, which demonstrates the importance of selecting modeling approaches and satellite images to improve efficiency in predicting SOCS distribution at a fine spatial resolution of 10 m. Furthermore, the study emphasizes the potential of radar (Sentinel-1) in developing SOCS mapping, with the newly developed fine-resolution mapping having important applications in land management, ecological restoration, and protection efforts in the Zoige Plateau.

Microbial necromass carbon (MNC) is one of the primary sources of stable soil organic carbon (SOC), playing a crucial role in global carbon cycling. Despite its significance, the impact of tunnel construction, a commonly adopted human activity in railway development, on MNC and its underlying mechanisms remains underexplored. To address this gap, a field survey was conducted in shrubland ecosystems on the eastern edge of the Qinghai-Tibet Plateau. Our study aimed to investigate the influence of tunnel construction on MNC and its contribution to SOC along an altitudinal gradient (3240 m, 3420 m, and 3600 m) and at two soil depths (0-10 cm and 10-30 cm). Results indicated that tunnel construction did not exert a significant impact on SOC and MNC. However, it led to a substantial decrease in the contribution of MNC to SOC (15.6%-19.9% in tunnel-affected areas) compared to the control areas (19.04%-30.97%). This outcome was likely attributed to the decreased bacterial necromass carbon (BNC) and the increased input of underground plant-derived carbon. MNC at 10-30 cm was significantly lower than that at 0-10 cm, and it exhibited a notable decrease at 3420 m compared to both 3240 m and 3600 m. These observations were closely related to reduced carbon input and microbial necromass stability. Principal component structural equation modeling revealed that soil microbial activities were the most important for MNC and SOC accumulation. Consequently, tunnel construction influenced SOC formation by altering the relative contributions from plant and microbial sources. This study provides valuable evidence for assessing the impacts of engineering construction on SOC and its sources in the context of climate change. Moreover, it offers scientific support for ecosystem conservation and sustainable development in the QinghaiTibet Plateau region during engineering construction.

Due to the importance of trees in urban ecosystems, it is crucial to understand the distribution, prevalence, and intensity of parasitic mistletoe infection across urban landscapes. In this study, we aimed to understand how various tree-related factors (e.g., tree taxonomy, size, pruning history, nativity, foliage, density, and diversity) might influence mistletoe prevalence and infection intensity across a typical subtropical urban greenspace. We surveyed 6,012 trees representing 96 species, 76 genera, and 44 families across a large, urban university campus in a major Southwest Chinese city. In total, we found 353 mistletoe host trees (prevalence: 5.87%), largely concentrated on the eastern part of campus, partially due to the proximity of waterbodies and a large adjacent greenspace, as well as higher tree density. To understand how predictors affected the prevalence and intensity of mistletoe infection, we used a hurdle model fitting our zero-inflated count data. Four independent variables significantly affected the prevalence of mistletoe infestation, with three demonstrating positive correlations (crown width, tree density, and tree diversity). One variable (evergreen foliage) negatively correlated with mistletoe prevalence. However, only one factor was found to significantly affect mistletoe infection intensity (crown width, positively). Due to the variety of ecosystem services mistletoes provide, more studies are necessary to assess the impact of various mistletoe management strategies, especially on urban biodiversity.

Key message Evidence for the CFE was mixed. New dendrochronological experimental designs are needed, especially in tropical and boreal biomes. A weakness of current methodologies could impede the ability to detect a CFE. Abstract The effects of elevated atmospheric carbon dioxide on plant growth and the carbon cycle have stimulated extensive research. Over the last three decades, CO 2 enrichment experiments have tested the CO 2 fertilization effect (CFE) hypothesis. Yet, CO 2 enrichment studies have been criticized for having unrealistic designs. Multiple studies have sought to use tree-ring analyses to verify the CFE under more natural conditions. To assess the breadth of scholarship about CO 2 fertilization on tree radial growth, this study conducted a tri-lingual (English, French, and Chinese) literature review. Finding no French, just one Chinese, and 73 English articles directly related to this topic, all 74 relevant tree-ring studies (1984–2020) were reviewed in-depth. We found an underrepresentation of studies from key regions important to the global carbon cycle, particularly tropical and boreal biomes. Evidence for the CFE was mixed: about 32.4% of dendrochronological studies found a CFE, another 23% found radial growth trends explained by both CO 2 fertilization and other climatic variables, 5.4% attributed tree growth only to climate change, and 39.2% (primarily in tropical areas) did not find any evidence of a CFE. Synthesized results of global scholarship highlight the need for new dendrochronological experimental designs to exclude climatic and environmental variables and test whether CO 2 fertilization occurs in tropical regions. Furthermore, weakness of current methodologies could lessen the ability to detect a CFE. More focused research is necessary from under-represented regions, and to test the effects of increasing anthropogenic activities (such as localized elevated CO 2 emissions) on tree growth in nature settings.

Heavy metal(loid)s are widely regarded as important environmental pollutants and have attracted extensive attention. In addition to such areas with frequent human activities as cities and industrial areas, the heavy metal(loid) pollution in remote areas lacking long-term monitoring data also deserves attention. In this study, Chinese pine (Pinus tabuliformis) in Jiuzhaigou World Natural Heritage was sampled to analyze concentration of Pb, Zn, Cu, As, Cd, Co, Cr, and Ni and to reconstruct pollution history. Source analysis and evaluation of the potential ecological risk of heavy metal(loid)s were also performed. Results showed that Jiuzhaigou has been polluted by heavy metal(loid)s at a relatively high level and concentrations were the highest at the location with intensive human activities (Nuorilang). Tree growth was inhibited by increased heavy metal(loid) concentration and this situation was more pronounced at high concentrations. The increased heavy metal(loid) concentrations were attributed to human activities such as forest logging, infrastructure construction, and tourism development. Heavy metal(loid) pollution reached a very high level of ecological harm at Nuorilang and a high level of ecological harm in Shuzheng and Heye villages. Overall, even at low concentrations, we demonstrated that Chinese pine can be used not only as a bioindicator for heavy metal(loid) pollution but also to infer pollution history over a relative long-term period and to enhance our understanding of the biogeochemical cycle of heavy metal(loid)s in forest ecosystems.

China has experienced history's largest rural-to-urban migration. The social, economic, and environmental challenges brought about by urbanization are diverse and complex. Given China's national goal to achieve carbon neutrality by 2060 and commitment to urban sustainability, large cities have focused on urban greening initiatives. Yet, studies seeking to quantify ecosystem services and disservices only assess healthy, mature trees, rather than those with severe damage, declining health, or lack of vitality due to poor management. In this short communication, we conducted a case-study in one of China's major nursery stock-producing cities, Chengdu, on a common street tree, Ginkgo biloba, to assess the long-term impact of one of the most common yet extreme nursery transplant practices on tree growth (traumatic root-cutting of 'super-large' nursery stock). We used tree-ring data collected in a typical urban greenspace from 23 Ginkgo trees, including 18 trees transplanted as 'super-large' nursery stock and a control group (5 trees) transplanted as small-caliper trees. We found the trees transplanted as 'super-large' nursery stock experienced declining tree growth with decades of lost landscape potential likely due to traumatic root-cutting at the time of transplant from nursery to landscape. The control group allowed contrast between the growth patterns of 'super-large' transplanted trees with those that remained healthy, being transplanted as smaller-caliper trees. For the 'super-large' trees, we found a decrease in carbon sequestration from 7.6 kg C yr-1 on average per tree in 2001 to about 1.5 kg C yr-1 on average per tree in 2021, while no decreasing trends were observed among the control trees. This implies a negative impact on multiple expected ecosystem services including carbon sequestration, shade, canopy coverage, and pollutant mitigation. These results highlight the unrecognized costs of common Chinese nursery and transplant techniques on urban landscape trees, necessitating more research, science-based policies, and better management techniques.

Net primary production (NPP) is a pivotal component of the terrestrial carbon dynamic, as it directly contributes to the sequestration of atmospheric carbon by vegetation. However, significant variations and uncertainties persist in both the total amount and spatiotemporal patterns of terrestrial NPP, primarily stemming from discrepancies among datasets, modeling approaches, and spatial resolutions. In order to assess the influence of different spatial resolutions on global NPP, we employed a random forest (RF) model using a global observational dataset to predict NPP at 0.05 & DEG;, 0.25 & DEG;, and 0.5 & DEG; resolutions. Our results showed that (1) the RF model performed satisfactorily with modeling efficien-cies of 0.53-0.55 for the three respective resolutions; (2) NPP exhibited similar spatial patterns and interannual vari-ation trends at different resolutions; (3) intriguingly, total global NPP varied greatly across different spatial resolutions, amounting 57.3 & PLUSMN; 3.07 for 0.05 & DEG;, 61.46 & PLUSMN; 3.27 for 0.25 & DEG;, and 66.5 & PLUSMN; 3.42 Pg C yr-1 for 0.5 & DEG;. Such differences maybe associated with the resolution transformation of the input variables when resampling from finer to coarser resolution, which significantly increased the spatial and temporal variation characteristics, particularly in regions within the southern hemisphere such as Africa, South America, and Australia. Therefore, our study introduces a new concept em-phasizing the importance of selecting an appropriate spatial resolution when modeling carbon fluxes, with potential applications in establishing benchmarks for global biogeochemical models.

Environmental education (EE) is a concern in modern society. Plant blindness and nature-deficit disorder highlight how important nature contact and plant consciousness is for urban populations. Considering rapid urbanisation, these phenomena are of increasing concern in China. Integrating EE curricula in Chinese cities is necessary to help younger generations appreciate the environment and understand how to address environmental challenges. Based on our EE teaching experiences in Southwest China's largest urban conglomeration, Chengdu, we discuss anecdotes and propose strategies to improve Southwest China's EE, with guidance for future research. Although children are more drawn to animals, EE education should focus on ecosystem-wide approaches rather than entirely zoocentric or anthropocentric perspectives, since these could reinforce the problematic view that plants are merely 'backdrops'. Urban greenspaces should be utilised in active education schemes, with teachers trained via professional EE modules rooted in local environmental concerns and complementing national/regional priorities. EE programmes should actively engage parents, encouraging their interaction with schools during EE activity preparation and venue selection, as their participation can help overcome obsessive safety and liability concerns. Older generations should be encouraged to transmit their own environmental and ecological knowledge, with implications for both cultural resilience and biocultural diversity conservation efforts.

Though rising atmospheric CO2 concentrations (Ca) harm the environment and society, they may also raise photosynthetic rates and enhance intrinsic water-use efficiency (iWUE). Numerous short-term studies have investigated tree growth under elevated CO2 (eCO2) conditions, but no long-duration study has investigated eCO2 impacts on tree growth and iWUE under natural conditions. Utilizing a new dendrochronological experimental design in a heavily-touristed nature preserve in Southwest China (Jiuzhaigou National Nature Reserve), we compared tree growth (e.g., basal area increment) and iWUE in two biophysically and environmentally similar valleys with contrasting anthropogenic activities. Trees in the control valley with ambient CO2 benefited from increasing Ca, possibly due to the CO2 fertilization effect and optimal environmental conditions. However, trees in the treatment valley with intensive tourism experienced comparatively higher localized eCO2 and growth rate declines. While iWUE increased (1959–2017) in the control (25.3%) and treatment sites (47.8%), declining tree growth rates in the treatment site was likely because comparatively extreme CO2 exposure levels encouraged stomatal closures. As the first long-term study investigating eCO2 impacts on tree growth and iWUE under natural conditions, we demonstrate that increased forest iWUE is unlikely to overcome negative drought stress and rising temperature impacts. Thus, forest potential for mitigating eCO2 and global climate change is likely overestimated, particularly under dry temperate conditions.