Dryland ecosystem resilience is increasingly threatened by increasing aridity under climate change. While large-scale tree restoration offers a potential mitigation strategy, its effect integrated with climate change pressures on ecosystem resilience remains unquantified, hindering science-based restoration guidance in drylands. Our systematic review, analyzing 435 core publications, reveal three critical research gaps: (1) the lack of a unified definition of ecosystem resilience and an imbalance in its standardized assessment methods across its three core dimensions: resistance, recovery, and adaptability, a focus evidenced by publication ratios of over 7:1 and 12:1 for recovery and resistance versus adaptability, hindering a unified, three-dimensional quantification; (2) An pronounced imbalance in aridity thresholds research, with a predominant focus on vegetation attributes such productivity and cover (48%) contrasted by scant attention to holistic ecosystem properties like stability, recovery, resistance, or resilience (a mere 6%); and (3) the absence of integrated models capable of simulating the integrated impacts of climate change and tree restoration on ecosystem resilience dynamics. To address these challenges, this study proposes a comprehensive framework that synergizes theoretical mechanisms, technical tools, and ecological management. The core of this framework includes a three-dimensional assessment method for ecosystem resilience, and a coupled modelling approach that combines climate change (e.g., from CMIP6) with tree restoration modules (e.g., using Budyko models and atmospheric moisture tracking) to simulate future aridity. The application of this framework is designed to identify critical aridity thresholds and elucidate ecological regulation mechanisms, thereby providing a quantitative basis for optimizing tree restoration practices and enhancing dryland ecosystem resilience under climate change.
Open-pit mining and ecological restoration jointly reshape land cover, terrain, drainage structure, and local storage, making hydrological recovery difficult to quantify. This study developed a multi-scenario framework integrating remote sensing and the Soil and Water Assessment Tool (SWAT) for the Juhugeng mining area in the northeastern Qinghai–Tibet Plateau. Three land-surface configurations representing pre-mining (2006), post-mining (2014), and post-restoration (2024) conditions were simulated under a common stochastic climatic background. After a two-year warm-up, five complete simulation years were averaged. Cross-stage subbasin matching, a denominator-screened recovery index (RI), reference- and threshold-sensitivity tests, water-body response classification, and exploratory association and machine-learning analyses were used. Relative to the pre-mining configuration, the post-mining scenario reduced surface runoff and runoff volume by 5.85% and 1.19%, respectively, while lateral flow increased by 12.63%. Relative to the post-mining configuration, the post-restoration scenario further reduced surface runoff and runoff volume by 1.29% and 2.21%, while lateral flow increased by 7.46%. Recovery-index medians indicated weak surface-runoff recovery and near-zero or negative recovery for most other indicators. Mining produced more complex subbasin reorganization than restoration across all tested overlap thresholds, although the restored drainage system remained distinct from the pre-mining configuration. Water area expanded from 1.37 to 13.35 km2, but associated hydrological responses diverged among subbasins and were not explained by area alone. Most Random Forest and XGBoost models showed weak out-of-sample performance; therefore, their variable rankings were interpreted only as exploratory diagnostics. Overall, restoration produced a reorganized and spatially heterogeneous hydrological system rather than a uniform return to pre-mining conditions.
Vegetation is crucial for maintaining ecological balance and supporting sustainable development at global and regional scales. This study combined GIMMS and MODIS NDVI products to build a monthly NDVI dataset for the Mid-Temperate Zone of China (1982–2024), and systematically investigated spatiotemporal vegetation dynamics and their driving mechanisms under climate change and human activities. The results showed that during 1982–2024, growing-season average NDVI increased significantly at a rate of 0.92 × 10−3 yr−1 (p < 0.01), with a notable acceleration after 2011 (2.41 × 10−3 yr−1, p < 0.01), accompanied by a 893 km northeastward migration of the greening barycenter. Precipitation and shortwave radiation were the dominant climatic drivers of vegetation change in the MTZ, and after 2011, increased precipitation (4.45 mm) together with reduced shortwave radiation (5.08 W/m2) jointly promoted vegetation greening. Across the entire MTZ, 66.3% of the region exhibited an increasing NDVI trend, with human activities (56.0%) contributing substantially more than climate change (10.3%). Among human factors, GDP growth was the strongest positive contributor across the entire study area, whereas grazing intensity exhibited a significant negative effect in the eastern MTZ. Ecological restoration projects, such as the Three-North Shelterbelt Forest Program, improved vegetation cover while also contributing to GDP growth, creating a positive feedback between ecological enhancement and economic development. This study shows that human activities are the dominant driver of vegetation greening in mid-temperate zone, offering important scientific evidence for adaptive ecological restoration planning and policy-making under global environmental change.
Wetland ecological health is assessed using an indicator system that comprehensively considers aquatic environments, soil conditions, biological communities, and socioeconomic factors. In this study, the assessment unit basins and baseline wetland sections were established, and a Pressure-State-Response (PSR) framework comprising twelve indicators was applied to evaluate the ecological health of riverine wetlands. The assessment units were delineated as small watersheds at a spatial scale appropriate for riverine wetland management. In addition, alluvial zones were introduced as a geomorphic reference to delineate baseline wetland sections, defined as the potential wetland space against which human-induced degradation was assessed. The indicator grades may vary depending on the study area, and the state indicators were analyzed using three to five indicators per basin. The assessment results of 80 unit basins in the upper Geum River Basin, South Korea, revealed that one unit basin rated as Grade 1 (Very healthy), 29 unit basins as Grade 2 (Healthy), 47 unit basins as Grade 3 (Moderate), three as Grade 4 (Vulnerable), and none as Grade 5 (Highly vulnerable). Unit basins with higher population density, urbanization rate, and cultivated land ratio generally showed lower wetland ecological health grades. Unit basins adjacent to the mainstream tended to be relatively healthy whereas some tributary unit basins with high urbanization rates and pollution loads showed degraded ecological conditions. This study provides a scientific foundation for prioritizing restoration efforts and establishing sustainable wetland management and conservation policies.
Soil and water conservation measures mitigate erosion by modulating the rainfall-runoff-sediment cascade. Most traditional evaluations only measure the amount of sediment reduced. They rarely consider whether these practices fundamentally alter the statistical relationships within the erosion cascade, a phenomenon we term “statistical decoupling.” To address this gap, we propose a statistical decoupling framework as a novel dimension for assessing conservation efficacy in karst ecosystems. Using five years of natural rainfall data (158 erosive events) from five runoff plots in Southwest China (representing bare land, planted forest, shrubland, downslope tillage, and contour cultivation), we applied partial least squares structural equation modeling (PLS-SEM) to quantify treatment effects along the rainfall-soil moisture-runoff-sediment pathway. Results indicate that planted forests achieved the greatest sediment reduction (98.45% relative to bare land), followed by shrubland (86.06%) and contour cultivation (31.65%).Conversely, downslope tillage increased soil loss by 15.44%. Sediment generation was predominantly driven by low-frequency extreme rainstorms (Regime IV, representing 10.15% of events). Furthermore, PLS-SEM revealed stark variations in the explanatory power for sediment yield: bare land exhibited strong coupling (R2 = 90.6%), whereas planted forests demonstrated near-complete decoupling (R2 = 2.20%). Intermediate values were observed for shrubland (19.7%), contour cultivation (18.2%), and downslope tillage (35.70%). Path analysis suggested that vegetation-based measures appear to substantially attenuate the rainfall-sediment pathway, whereas downslope tillage maintained a weakened but intact linkage that exacerbated erosion. Additionally, soil moisture showed no significant mediating role in most plots, reflecting the shallow soils and rapid leakage characteristic of karst terrains. Ultimately, we categorize conservation effectiveness into magnitude and structural dimensions. We identify downslope tillage as maladaptive in both respects and propose statistical decoupling as a critical supplementary indicator for long-term stability. This framework provides a robust theoretical basis for optimizing ecological engineering in karst regions.
