Precise estimation of shrub above-ground biomass (AGB) in arid regions is crucial for carbon cycle research and ecosystem assessment. Unmanned aerial vehicle (UAV) -borne light detection and ranging (LiDAR) has become a key tool for quantifying three-dimensional vegetation structure and estimating AGB. However, the short stature of arid zone vegetation, combined with sparse and low-quality point clouds acquired by UAV, limits high-accuracy shrub AGB estimation. To address this issue, this study selected Caragana korshinskii, a typical psammophytic shrub in Ordos City, as the research object. By integrating UAV-based multispectral and LiDAR data, a biomass estimation method based on a novel Shrub Structure Index (SSI) was proposed. The SSI workflow reconstructs the three-dimensional shrub structure under sparse point cloud conditions and improves AGB estimation accuracy. This workflow comprises Object-based image analysis (OBIA) classification for individual shrub extraction, Delaunay linear up-sampling, voxel-based partitioning, and dynamic stratification by height percentiles. Experimental results demonstrate that: (1) The individual shrub extraction method utilizing the large-scale mean shift (LSMS) segmentation algorithm and support vector machine (SVM) classification achieved a total quadrat segmentation accuracy of over 90.61 %, an overall classification accuracy of 91.51 % (Kappa = 0.86). (2) In SSI construction, the height-percentile stratification thickness, point-cloud sampling, and voxel edge length together set Caragana korshinskii stratification accuracy and density scale; the 5 % height percentile interval, a sampling size of 100 points, and 0.04 m voxel edge length proved optimal. (3) Comparative experiments showed that the three-dimensional feature integrated SSI significantly outperformed single-feature, two-feature, traditional allometric equation, and random forest (RF) models, with the SSI-based model achieving R-2, RMSE, MAE, and rRMSE of 0.90, 529.01 g, 432.58 g, and 26.54 %, respectively. These results indicate that SSI more effectively captures shrub spatial structure and improves AGB prediction under sparse UAV-LiDAR conditions.

Abstract Land degradation (LD) poses a major challenge to global sustainable development, with the attainment of land degradation neutrality recognised as a key indicator of Sustainable Development Goal 15.3 (SDG 15.3). This goal focuses on combating desertification, restoring degraded land and soil, and including land affected by desertification, drought, and floods, with the aim of creating a land degradation-neutral world by 2030. Eastern Inner Mongolia (EIM), a typical agro-pastoral transitional zone, has experienced increasingly severe LD in recent decades. Identifying its dominant drivers is essential for improving ecological governance and land use management capacity. This study employed the Global 30 m Land-Cover Dynamics Monitoring Product (GLC_FCS30D) to characterise the spatio-temporal patterns of LD across three periods: 1990–2000, 2000–2010, and 2010–2020. Twelve drivers were identified and categorised into four groups: natural, human activities, economic, and urbanisation. Using Partial Order Theory and the Hasse Diagram Technique, the influence intensity of each driver group was systematically evaluated and ranked for each period. The combination of POT and the HDT enables transparent, interpretable, and reproducible ranking of multiple drivers while accounting for nonlinear, hierarchical, and interdependent relationships, offering advantages over conventional statistical or machine learning methods. The results showed that: (1) from 1990 to 2020, approximately 23% of EIM experienced land cover change, with degradation rates of 0.78%, 0.36%, and 0.41% for the three periods, respectively, and restoration rates of 0.71%, 0.36%, and 0.29%, respectively, the net land degradation and restoration areas were 5.83 × 104 km2 and 4.62 × 104 km2, respectively; (2) the dominant drivers of LD were ranked and spatially distributed differently across each league and city; (3) In the first two periods, the order of influence was: urbanisation > natural > human activities > economic. In the third period, the order changed to: urbanisation > natural > economic > human activities. This study reveals the spatiotemporal dynamics and dominant drivers of LD in EIM, providing a scientific basis for formulating regional land management and ecological restoration policies. It also offers valuable references for optimizing land use strategies, curbing LD, and achieving the sustainable development goals.

Fractional Vegetation Cover (FVC) is a key biophysical parameter for characterizing vegetation dynamics and ecosystem functioning, and is essential for ecological monitoring in arid and semi-arid regions. However, sparse vegetation and strong soil background effects in these regions lead to substantial variability in the applicability and stability of different inversion models, thereby limiting further improvements in FVC estimation accuracy. To address these challenges, this study applied three FVC inversion models—Random Forest (RF), Fully Constrained Least Squares (FCLS), and the Dimidiate Pixel Model (DPM)—to Sentinel-2 data, and further proposed a consistency-constrained ensemble approach. The results show that the three models exhibit pronounced discrepancies in both spatial patterns and value ranges. Specifically, RF performs better in high-coverage areas, whereas FCLS is more sensitive to soil background effects under low-coverage conditions. The proposed ensemble approach effectively integrates the strengths of individual models, improving overall accuracy by approximately 2.1% and reducing estimation uncertainty relative to individual models. These results demonstrate that the proposed approach improves FVC estimation in arid and semi-arid regions and provides a robust framework for large-scale quantitative vegetation monitoring using remote sensing.

Understanding the spatiotemporal dynamics of ecosystem services (ESs) and their responses to social-ecological drivers is critical for restoration governance in ecologically vulnerable regions. However, how these drivers shape ES dynamics through nonlinear threshold effects across multiple temporal scales remains insufficiently understood. Here, we integrate multiyear trends with interannual variability to examine ES dynamics and their drivers in the Beijing–Tianjin Sandstorm Source Control Project region from 2001 to 2022, using an interannual fluctuation index and an extreme gradient boosting–Shapley additive explanations approach. Results showed that carbon sequestration, sand fixation, soil conservation, and water yield exhibited overall increasing trends, while their spatial hotspots displayed distinct interannual variability, with fluctuation indices of 13.31%, 17.31%, 18.48%, and 13.54%, respectively. ESs that were primarily driven by climatic factors (sand fixation, soil conservation, and water yield) showed greater interannual fluctuations than those dominated by vegetation (carbon sequestration). Although the top-ranked driver for each ES remained consistent across temporal scales, its relative importance and anthropogenic influence increased in short-term dynamics. Threshold analysis further identified normalized difference vegetation index of approximately 0.6 and annual precipitation around 400 mm as key transition points associated with ES hotspot formation and persistence. Areas exhibiting both high ES provision and high hotspot persistence represent priority targets for conservation and restoration, given their greater capacity to sustain stable service provision under climatic variability and human disturbance. These findings highlight the importance of taking short-term ES dynamics and threshold effects into account in restoration management to sustain ecosystem functioning and service provision under ongoing environmental change.

Accurately extracting the spatial distribution of shrubs is an important basis for scientific diagnosis, rational prevention, and control of shrub-encroached grasslands (SGs). In SGs, the landscape exhibits a high degree of spatial intermingling among shrubs, herbaceous vegetation, and bare soil, resulting in a serious issue of mixed pixels. The combination of UAV hyperspectral imaging and deep learning provides the most promising technical approach for shrub–grass separation at present. However, issues such as data redundancy and algorithmic adaptability urgently need to be addressed. In this study, Caragana microphylla Lam, a typical shrub species found in the SGs of Inner Mongolia, is the subject for extraction. This work aims to extract sensitive spectral bands from UAV hyperspectral data, construct a deep learning-based framework for shrub identification, and map the distribution of Caragana microphylla Lam in the study area. First, a dataset of hyperspectral images in Xilinhot City, Inner Mongolia, China, was constructed by manual visual interpretation for shrub identification. Second, a novel unsupervised band selection algorithm, US-BS-Net, was proposed to optimize the proxy task of BS-Conv-Net by introducing contrastive learning. It can be used to obtain sensitive spectral bands with discriminative feature representations. Finally, the US-BS-P-Net framework was proposed by combining the US-BS-Net with the prototypical network to construct high-precision shrub identification models in small-sample scenarios. Taking 80 bands as input, the highest overall accuracy reaches 93.15%, which is better than the full spectrum and other deep learning models. The deep synergy between hyperspectral technology and deep learning effectively overcomes traditional challenges such as spectral similarity between shrubs and grasses and their mixed spatial distribution. In particular, the proposed US-BS-P-Net, which combines the advantages of US-BS-Net in sensitive band extraction and the prototypical network in few-shot model construction, provides excellent fundamental data for precise monitoring of SGs.

Shrub encroachment significantly affects carbon cycling in grassland ecosystems. In recent years, the expansion of Caragana jubata in subalpine meadows has become a pressing ecological issue. Rapid and accurate estimation of its aboveground biomass (AGB) is crucial for understanding invasion mechanisms and guiding ecological management. However, scale effects and model complexity still constrain precise, rapid estimation of shrub AGB. To address this challenge, the study focused on subalpine meadows and proposed a UAV-satellite integrated cross-scale AGB estimation framework driven by the optimal single indicator. The results showed that: 1) At the UAV scale, two variants of the vegetation index-based weighted canopy volume model (CVMVI)-RGB_CVMVI and MS_CVMVI-outperformed individual spectral or structural metrics, with MS_CVMSAVI performing the best and achieving an accuracy of R-2 = 0.878 and RMSE = 355.47 g; 2) From the UAV scale to satellite imagery at different spatial resolutions, as the image resolution decreases, the kernel density curves of AGB distribution gradually became smoother and the peak progressively shifted leftward, indicating clear scale effects; 3) At the satellite scale, the unified evaluation based on generalized additive models showed that vegetation indices integrating red-edge, near-infrared, and red bands consistently outperformed SAR-derived indices, with the optimal index varying according to the spatial resolution of the imagery. The proposed cross-scale AGB estimation framework enables rapid and accurate estimation from UAV to satellite scales in subalpine meadows and provides a scalable approach for large-scale shrub encroachment monitoring.

Vegetation change serves as a comprehensive indicator for monitoring regional and global environmental conditions, exhibiting increased complexity under accelerated urbanization. As a typical example of rapid urbanization, China has witnessed the emergence of several large urban agglomerations, making vegetation change studies particularly critical. However, inadequate consideration of spatiotemporal heterogeneity and complex interactions among multiple factors poses challenges in analyzing vegetation change in China’s urban agglomerations. Additionally, few studies simultaneously identify factor interactions and further quantitatively assess the underlying processes, such as potential time lags. Focusing on the Beijing-Tianjin-Hebei (BTH) region from 2000 to 2020, this study characterizes long-term vegetation dynamics using fractional vegetation cover (FVC) derived from Landsat imagery and validated with manual samples. We introduced an analytical framework that combines geostatistical tools to address spatiotemporal heterogeneity with the optimal parameter-based geographical detector (OPGD) model to disentangle complex interactions among multiple factors. Our findings revealed distinct north–south differentiation in FVC across the BTH region, delineated by the Taihang-Yanshan Mountain Range as the primary biogeographic boundary. Vegetation cover exhibited a fluctuating yet increasing trend with 65.87% of the area showing improvement, accompanied by intensified spatial heterogeneity. Temporally, this improvement appears to be strongly associated with ecological policies. These spatial heterogeneous patterns underscore the region’s sensitivity to altitude-dependent climatic gradients and uneven policy implementation. Notably, we also identified stage‑dependent vegetation trends within urban agglomerations, indicating that later-stage urbanization does not inevitably lead to a net vegetation decline. The OPGD model identified land use type, which directly reflects the impact of human activities and policy, as the dominant overall factor. Other detection analysis demonstrated synergistic amplification between paired factors, with optimal vegetation growth thresholds identified. Long-term analysis specifically revealed the evolving influencing factors (e.g., the contribution of urbanization rate factors more than tripled between 2015 and 2020). Considering that both spatiotemporal analysis and driving model identified policy as a key role, we further quantify the time-lag effects of policy by employing the autoregressive distributed lag (ARDL) model. This analysis indicated that ecological policy in the BTH region significantly enhanced FVC, exhibiting a time lag of approximately three years. By qualitatively identifying and quantitatively analyzing vegetation, we offer empirical evidence and a methodological framework for balancing ecological preservation with sustainable development in urban agglomerations worldwide.

Studying the temporal and spatial variation characteristics and driving factors of carbon reserves in the middle reaches of the Yellow River is crucial for achieving sustainable development and regional ecological conservation against the backdrop of the "double carbon" plan. Based on the five-year interval, the land use data of the middle reaches of the Yellow River from 2000 to 2020 were selected, and the spatio-temporal evolution characteristics of carbon reserves were estimated and analyzed by coupling with the PLUS-InVEST-GeoDetector model, and the driving factors affecting the spatio-temporal differentiation of carbon reserves were discussed. Finally, the carbon reserves of the middle reaches of the Yellow River in 2030 were predicted under four developmental scenarios： natural development, ecological protection, economic development, and cultivated land protection. The findings indicate that： ① The middle reaches of the Yellow River's carbon storage showed a consistent growth trend between 2000 and 2020, exhibiting an increase by 5.75×107 t. The evolution of the spatial distribution was reasonably stable, exhibiting the characteristics of "southeast is higher than northwest." ② The middle reaches of the Yellow River's carbon storage differentiated both spatially and temporally between 2000 and 2020, with two-factor enhancement and nonlinear enhancement observed in the interaction detection of each driving element. The main driving force was the NDVI. ③ From 2020 to 2030, the carbon storage of the four scenarios in the Yellow River's middle reaches showed an increasing trend in comparison to that in 2020. Of them, the carbon storage of the ecological preservation scenario rose the highest at 3.93×107 t, while the carbon storage of the economic growth scenario increased the least at 4.8×106 t. The findings of the study will offer some evidence in favor of the middle reaches of the Yellow River's long-term development and ecological environment management.

The past five years are the five years when big model and general model of Artificial Intelligence(AI)are gradually integrated into people's daily work and life,and the five years when remote sensing+AI technology develops rapidly in the fields of land cover type identification,change detection,etc.It is also the first five-year for the implementation of the national strategy of"ecological civilization"and"beautiful China".Summarizing the progress made in the research,development and application of forestry and grassland remote sensing technology in these five years is of great significance for the country to formulate the development plan of forestry and grassland remote sensing in the future. The paper summarizes the main progress of the forestry and grassland remote sensing research and development in China in the past five years into four research directions,namely,change detection and classification of forest and grassland cover types,quantitative inversion/estimation of forest parameters by remote sensing,and that of grassland vegetation and early warning and monitoring of forest and grassland disasters.From a general point of view,the research on forestry and grassland remote sensing technology shows a rapid development trend from traditional shallow machine learning to deep learning,and from"data"-driven to"data+mechanism"-double-driven direction,and the deep learning method develops quickly and deeply in change detection and classification,but not in quantitative parameter inversion/estimation.The production technology of large-scale forest and grassland thematic products,such as global and national products,has also been developed rapidly. An analysis of the integration of remote sensing technology into existing technical standards and technical programs for forestry and grassland resources and ecological monitoring,disaster early warning monitoring and monitoring of nature reserves shows that forest and grassland cover type change detection/monitoring and classification technologies have been widely and deeply applied to various resource supervision and disaster early warning and monitoring operations in the forestry and grassland industry,but the degree of operational application of quantitative inversion/estimation technologies of forest and grassland quality parameter is still very low. In view of the challenges in promoting the comprehensive and in-depth application of forestry and grassland remote sensing technology,it is suggested that the forestry and grassland industry should vigorously integrate the"space-air-ground"multi-source earth observation resources,comprehensively apply remote sensing,artificial intelligence(AI),statistical inference and other cutting-edge technologies to build a"space-air-ground"integrated monitoring technology system,and greatly strengthen the investment in scientific research,technology exchange and talent exchange and cultivation.

The Beijing-Tianjin sandstorm source region (BTSSR) is an important ecological barrier in North China, which can prevent land desertification from spreading. However, the challenge of separating the impacts of climate change from those of human activities on ecological restoration remains a critical concern. This study addresses this issue by employing residual trend analysis to investigate long-term trends in net ecosystem productivity (NEP), water conservation (WC), soil erosion (SE), and habitat quality (HQ) in the BTSSR. The results showed that the ecological recovery of the study area was significantly improved. Notably, there is a significant increase in NEP, with affected areas constituting 68.88% of the total, alongside a 56.11 % expansion in WC. Conversely, HQ showed a modest increase in 25.59%, while SE experienced a significant decline in half of the study area. The strong positive correlation noted between NEP and summer precipitation, leading to the selection of NEP as a key index for assessing ecological restoration drivers. The research identifies engineering measures as the primary force propelling restoration efforts, contributing 38.18 %, followed by precipitation at 26.80 %. Predominantly, these impactful areas are situated in the southern region, where beneficial water and thermal conditions foster higher vegetation coverage. The methodology employed here enhances the precision of evaluating ecological engineering across various regions and climatic contexts, offering vital insights for national ecological governance and restoration initiatives.

Recently, grassland ecosystems’ productivity and ecological service capacity have declined due to human activities and natural changes, and ecological and environmental problems such as grassland degradation and land sanding have become hot issues of global concern. Therefore, timely and accurate information on grassland use based on remote sensing monitoring is of irreplaceable importance in coordinating the relationship between development needs and natural carrying capacity and protecting the ecological environment. Based on a single remote sensing data source, there are limitations in obtaining time-continuous data, which affects the accurate monitoring of grassland distribution, use type, and use intensity. The study proposed a method to integrate a daily land surface reflectance dataset based on Harmonized Landsat-Sentinel (HLS) data and GF-6 WFV data and constructed a grassland use intensity index based on time series data to estimate the use intensity. The results show that the daily land reflectance dataset integrated from HLS and GF-6 WFV data passed the consistency test, with a Correlation Coefficient (R) in all bands greater than 0.85 and a Root Mean Square Error (RMSE) in all bands less than 0.09, verifying the reliability of the fused data. This dataset can be used to solve quality problems, such as missing data, and to improve the observation frequency of the time series data. The identification of grassland use type based on the daily scale Normalized Difference Vegetation Index (NDVI) dataset achieved an overall accuracy with a Kappa coefficient of 0.80. Meanwhile, the grassland use intensity index constructed at the same time can better reflect the differences between different grazed intensities and provide a better estimation of the use intensity of grasslands.

The rapid urbanization and industrialization of the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) pose a severe challenge for rational land use. This study presents a multi-factor land-use suitability assessment system with economic, social, and environmental dimensions. System reliability and stability are confirmed by a Cronbach’s α coefficient (>0.7). We innovatively integrate the PS-DR-DP model with the Monte Carlo and Markov models. The Markov model analyzes transition probabilities between different land capacity states. The Monte Carlo method quantifies key parameter uncertainties through extensive random sampling, while the Markov chain-Monte Carlo approach dynamically evaluates and predicts land capacity. From 2002 to 2022, overall GBA land-population carrying capacity is stable above 0.6 and keeps rising, reflecting improved regional land capacity and successful coordinated development. However, the forecast results indicate that land capacity will first increase and then decrease between 2023 and 2042, with most cities reaching a peak carrying capacity (S-value approaching or exceeding 2) in 2027. This peak is followed by a projected decline, and by 2042, the overall land capacity may drop to around 0.5, signaling a significant long-term risk of overload. If current development trends continue, the region faces significant long-term risks of declining carrying capacity, particularly if the transition to a sustainable, innovation-driven economy is not managed effectively. This highlights the profound challenge of balancing economic growth, urbanization, and ecological protection. These recommendations offer scientific evidence and decision-making support for sustainable GBA development.

Drylands, as highly vulnerable ecosystems, support environmental functions and human well-being. Nevertheless, widespread land degradation and desertification present significant global and regional environmental challenges, with limited consensus on their area and degree. This study used time-series vegetation productivity and meteorological data from 2000 to 2020 to quantify global land degradation trends and driving factors in drylands. The results show a notable restoration of land degradation in drylands worldwide, with the area of improved land exceeding the degraded area by 1.4 times, although the threat of degradation persists. India and China emerge as pioneers in effective land improvement strategies, offering valuable experiences for other regions. Combined effects, as quantitatively distinguished by our established model, dominate the degradation and improvement processes. Notably, human activities play a decisive role in influencing land degradation trends, with the potential for either exacerbation or reversal. This study provides new perspectives on environmental health and human activities from global and regional observations. Finally, our research provides scientific support for desertification control and contributes to the overall advancement of the SDGs globally.

Evaluating forest ecosystem services (FES) is crucial for comprehensively recognizing forest value and for formulating targeted forest management plans. However, hurdles persist in traditional FES evaluations that are based on conventional data (e.g., statistical yearbooks and survey data), such as a coarse evaluation scale and difficulty in formulating refined and spatially continuous evaluation results. Forest canopy cover, canopy height, and forest aboveground biomass (AGB) are the core fundamental inputs of a robust FES evaluation. Their accuracy and degree of refinement will influence the final evaluation results obtained. To overcome the above issues, this study first explored accurate estimation methods for all 3 parameters above and then evaluated FES multidimensionally, by using these results combined with other remote sensing products and applying various principles and algorithms. Our results show that a high estimation accuracy (>80%) of the 3 key parameters is achievable for coniferous to broad-leaved forest stands and that FES evaluation results are obtainable with a high resolution and spatial continuity. The service functions, such as nutrient retention, carbon sequestration and oxygen release, and product supply are stronger while others relatively are weaker. It is worth noting that carbon storage by the AGB carbon pool surpasses that of other carbon pools. Finally, the potential of FES varies according to forest type. Compared with broad-leaved forest, coniferous forest has a greater capacity for product supply, windbreak, and sand fixation services. This study offers a methodological reference for the formulation of policies related to the paid use of FES.

In the context of ongoing climate change, relationships between tree growth and climate present uncertainties, which limits the predictions of future forest dynamics. Northwest China is a region undergoing notable warming and increased precipitation; how forests in this region will respond to climate change has not been fully understood. We used dendrochronological methods to examine the relationship between climate and the radial growth of four tree species in a riparian forest habitat in Altai region: European aspen (Populus tremula), bitter poplar (Populus laurifolia), Swedish birch (Betula pendula), and Siberian spruce (Picea obovata). The results reveal that European aspen was insensitive to climate changes. In contrast, bitter poplar showed a positive response to elevated temperatures and negative to increased moisture during the growing season. Swedish birch and Siberian spruce were adversely affected by higher temperatures but benefited from increased precipitation. A moving correlation analysis suggested that, against a backdrop of continuous warming, growth patterns of these species will diverge: European aspen will require close monitoring, bitter poplar may likely to show accelerated growth, and the growth of Swedish birch and Siberian spruce may be inhibited, leading to a decline. These findings offer insight into the future dynamics of riparian forests under changing climate.

Grass yield (GY) is a critical component of the comprehensive analysis of the grass - livestock balance in grassland. Net primary productivity (NPP) conversion methods, such as the Carnegie - Ames - Stanford approach (CASA) model, are an important tool for remote -sensing -based estimations of GY. However, the application of such approaches is limited by the simplification of key vegetation growth processes. In this study, we integrated high spatial and temporal resolution normalized difference vegetation index (NDVI) data collected from Gaofen6 (GF-6) and the Moderate Resolution Imaging Spectroradiometer (MODIS), respectively, in 2020 with the climatic characteristics of grassland vegetation to derive a reasonable expression of the optimum temperature. We then improved the CASA model for the accurate estimation of GY for six different grassland types in Zhenglan Banner (sandy sparse forest grassland, sandy shrub grassland, sandy meadow, low hill steppe, gently sloping steppe, and lowland meadow) at high spatial and temporal resolution. The model estimations were evaluated using field data. The results reveal that adopting the optimum temperature to incorporate vegetation growth characteristics achieves a better theoretical basis and minimizes the influence of anomalous NDVI maxima compared with the original CASA model. This largely avoids the influence of the lagged response of grassland vegetation growth to temperature. The developed GY model has strong applicability, and the correlation between the measured and estimated GY before and after optimization reached 0.75. Moreover, the overall estimation accuracy was improved by nearly 15%. The spatial distribution of GY in Zhenglan Banner was found to be similar to the spatial distribution of grassland types with obvious seasonal differences, and summer was the critical period for GY, accounting for more than 80% of growth. The proposed model aims to provide scientific and technical guidance for the regulation of grassland resources and reasonable grazing utilization in Northern China.

Shrub encroachment in grassland has become an ecological issue of mounting concern. Accordingly, an accurate estimation of aboveground biomass (AGB) of shrub vegetation is the basis for a sound assessment and in-depth understanding of carbon cycling in shrub-encroached grassland ecosystems. Yet the relatively low stature of plants in the shrub community, coupled with the high spatial heterogeneity of their distribution, contributes substantially to greater uncertainty in remote sensing estimation of shrub vegetation’s AGB. This study proposes a space–air-ground integrated approach to accurately estimate the AGB of shrub vegetation in shrub-encroached grassland ecosystems. The results showed that, at the UAV scale, the estimation of AGB for a monoculture shrub was highly dependent on planar geometric features. Based on the orthorectified images obtained from unmanned aerial vehicles (UAVs), four planar geometric features of shrub plants, namely crown area (S), crown perimeter (C), long-to-short crown dimension ratio (A1, A2), were retained as the most crucial predictors for AGB estimation. Among the 102 features related to vertical structure extracted via Light Detection and Ranging (LiDAR), only the crown height variation and the first layer’s density variable were retained. Utilizing the mentioned features and a random forest regression, the AGB prediction model for the shrub Caragana microphylla performed remarkably well, in having an R2 value of 0.84 and an RMSE of 310.14 g/plant. At the satellite scale, there was significant nonlinear relationship between the AGB of the shrubs and the band, texture, and index features extracted from GF-6 imagery. The derived AGB estimation model based on the Random Forest method demonstrates higher accuracy (R2 = 0.81, RMSE = 14.61 g/m2, MAE = 11.26 g/m2) than the linear stepwise regression (SR) and partial least squares regression (PLSR) models. Notably, the green band reflectance was retained in all three modeling approaches despite pronounced differences in their selected features uses. Yet both NDVIre1 and NDREI indices with red-edge bands were more important, suggesting the red-edge bands of GF-6 can serve as an ideal tool for remote sensing investigations of the AGB of shrub in shrub-encroached grasslands. This study provides technical and scientific support for quantitative assessments of shrub AGB in arid and semi-arid grassland regions.

In the context of global change, the carbon budget in arid and semi-arid regions have changed significantly. Understanding these dynamic features and their response to climate change is essential for regional carbon cycle assessments. The Beijing-Tianjin Sand Source Region (BTSSR), a significant ecological project in China, is central to studies on carbon budget dynamics. While global change intensifies, understanding its carbon budget response to drought is crucial. By integrating data from the Net Ecosystem Exchange (NEE) and the Standardized Precipitation Evapotranspiration Index (SPEI), we analyzed the spatiotemporal attributes of NEE in this region and its response to drought. The BTSSR is currently in a carbon sink state, with the mean NEE being -45.13 gCm(-2) and a rate of change at -0.005 gCm(-2)a(-1). The maximum carbon sink occurs during the summer, whereas the winter is characterized by a net release of carbon to the atmosphere. In terms of spatial variation, the NEE shifts from positive to negative from southeast to northwest across the region. The impact of drought on summer NEE is notably significant, with the correlation and sensitivity being -0.717 (p < 0.05) and -3.911, respectively. As drought intensity increases, NEE changes from negative to positive, transforming BTSSR into a carbon source. NEE responded most significantly to strong drought and was less affected by light drought. The NEE of BTSSR showed a significant short-term response to drought, with a lag of 1-4 months. Considering different vegetation types, the temperate desert's NEE is most significantly affected by drought and is the most sensitive, with correlation and sensitivity values of -0.841 (p < 0.05) and -12.570, respectively. In contrast, the warm temperate deciduous broadleaf forest shows stronger resistance to drought. This study elucidates how the carbon budget of different vegetation types in a typical ecological engineering area changes under the background of climate change and human activities, and contributes to the understanding of the impacts of drought events on the carbon cycle of different vegetation types, as well as the response of the latter to the former, especially the lagged response. It provides different perspectives for the subsequent ecological engineering construction.

Under the context of global change, the carbon budget in arid and semi-arid regions undergoes significant alterations. Understanding these dynamic features and their response to climate change is essential for regional carbon cycle assessments. The Beijing-Tianjin Sand Source Region(BTSSR), a significant ecological project in China, is central to studies on carbon balance dynamics. While global change intensifies, understanding its carbon balance response to drought is crucial. By integrating data from the Net Ecosystem Exchange (NEE) and the Standardized Precipitation Evapotranspiration Index (SPEI), we analyzed the spatiotemporal attributes of NEE in this region and its aridity response. The BTSSR is currently in a carbon sink state, with the mean NEE being -45.13 gCm−2 and a rate of change at -0.005 gCm−2a-1. The maximum carbon sequestration occurs during the summer, whereas the winter season is characterized by a net release of carbon to the atmosphere. In terms of spatial variation, the NEE shifts from positive to negative from southeast to northwest across the region. The impact of drought on summer NEE is notably significant, with the correlation and sensitivity being -0.717 (p<0.05) and -3.911, respectively. As drought intensity increases, NEE changes from negative to positive, transforming BTSSR into a carbon source. The NEE of BTSSR showed a significant short-term response to drought, with a lag of 1-4 month. Considering different vegetation types, the temperate desert's NEE is most significantly affected by drought and is the most sensitive, with correlation and sensitivity values of -0.841 (p<0.05) and -12.570, respectively. In contrast, the warm temperate deciduous broadleaf forest shows stronger resistance to drought. This study elucidates the spatiotemporal dynamics and drought response of NEE in the BTSSR, providing a theoretical basis for regional carbon cycle management and ecological engineering construction.