The southern Xizang Plateau, China, lies within the transition zone between the Asian summer monsoon and mid-latitude westerlies, where complex climate interactions drove extensive paleoglaciation during the Last Glacial Maximum (LGM). However, quantitative paleoclimatic constraints remain limited in this area. To address this gap, we compared two valleys that were glaciated during the LGM, the Pengnongcuo Valley on Mount Jiare in the eastern Gangdisê Mountains and the Shiduchong Valley on Mount Pamana in the central Himalaya Mountains. Using a two-dimensional numerical model constrained by Google Earth imagery and digital elevation model (DEM) data, we reconstructed paleoglacier geometries in the two valleys to constrain equilibrium line altitude (ELA) changes and to infer paleoclimatic conditions during the LGM. Precipitation-temperature (PT) and lapse-rate (LR) models were used to reconstruct paleoclimatic conditions during the LGM. The Pengnongcuo Valley had a maximum LGM ice thickness of 194 m, covered an area of 12.2 km² with a volume of 0.94 km³, and had an ELA depression of 331 m; the Shiduchong Valley had a maximum ice thickness of 234 m with an area of 14.1 km², a volume of 1.34 km³, and an ELA depression of 559 m. Despite having nearly identical modern ELAs (5819 m and 5835 m a.s.l. for Pengnongcuo Valley and Shiduchong Valley, respectively), the two glaciers exhibited dramatically different responses during the LGM. By constraining paleo precipitation to 60.0%-80.0% and 63.0%-79.0% of modern values, respectively, based on regional proxies and distinct moisture sources, reconstructed summer temperature depressions were 2.0°C-3.3°C and 3.4°C-4.7°C respectively. Glacial expansion during the LGM in the monsoon-westerlies transition zone of the southern Xizang Plateau resulted from the combined effects of a regional temperature depression, enhanced fluctuations in moisture supply from the Asian summer monsoon, and intensified mid-latitude westerly circulation, which enhanced cold-season precipitation. This study provides quantitative constraints on the LGM glacier-climate interactions in the monsoon-westerlies transition zone and highlights the critical role of coupled temperature and moisture forcing in shaping paleoglacial evolution on the southern Xizang Plateau.

Aerosol optical depth (AOD) is an important indicator for evaluating atmospheric pollution. Xinjiang Uygur Autonomous Region, China and its surrounding areas including Kazakhstan, Uzbekistan, Turkmenistan, Kyrgyzstan, and Tajikistan are well known as the typical arid and semi-arid areas mainly affected by dust aerosols, but the researches of different aerosol types, particularly the anthropogenic emissions are not enough. Moreover, the applicability of Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2) reanalysis AOD data remains unclear and the long-term trends of AOD in the study area have not been thoroughly investigated. Therefore, the accuracy of MERRA-2 reanalysis AOD data was evaluated against Aerosol Robotic Network (AERONET) observations. Subsequently, the contributions of five aerosol types (dust, sulfate, organic carbon, black carbon, and sea salt) derived from MERRA-2 AOD data from 2012 to 2022 were qualitatively and quantitatively explored. Results showed that AOD based on MERRA-2 reanalysis exhibited reasonable agreement with 19 AERONET sites (R²=0.49; bias= -0.07), confirming its regional reliability. The overall trends of MERRA-2 AOD revealed that aerosol concentrations in the Taklimakan Desert showed a significant increase (Theil-Sen slope≥0.005; standardized test statistic (Z)≥1.96) and higher variability (standard deviation (SD)>0.05), mainly owing to strong influence of natural disturbances, such as dust transport events. Conversely, low AOD variability (SD<0.01) appeared in the high-altitude and sparsely populated areas, such as eastern Tajikistan. The contributions of five aerosol types showed that dust AOD was the highest (0.120), followed by sulfate (0.067) and organic carbon (0.022), whereas black carbon and sea salt contributed relatively little, with values of 0.006 and 0.003, respectively. Spatially, dust AOD exhibited significant variation and gradually decreased from south to north, with high values mainly concentrated in the Taklamakan and Karakum deserts, whereas northern Kazakhstan was weakly affected by dust aerosols (AOD<0.060), but showed relatively high sulfate aerosol loading (AOD>0.100), mainly associated with anthropogenic emissions. From the perspective of land use types, spring dust aerosols made the largest contribution under the influence of dust sources of arid areas, with the highest contributions observed in barren land (75.28%; AOD=0.255), followed by cropland (55.21%; AOD=0.128), grassland (53.45%; AOD=0.150), and forest land (34.17%; AOD=0.068). Although greatly affected by dust and organic carbon at four seasons, sulfate remained the dominant aerosol type in forest land, driven by long-distance transport of industrial emissions, with percentages exceeding 46.00% (AOD>0.080). These findings indicate that aerosol variability in Xinjiang Uygur Autonomous Region, China and its surrounding areas is jointly controlled by natural dust emissions, anthropogenic sulfate inputs, and land-use-dependent seasonal processes, highlighting the importance of distinguishing natural and anthropogenic aerosol contributions in regional air quality and climate assessments.

Shallow groundwater may influence surface soil salinization in arid irrigated plains, but the timing and spatial form of this association remain uncertain. Consequently, this study aimed to identify the monthly timing, seasonal asymmetry, and spatial differentiation of surface soil salinization responses associated with shallow groundwater dynamics. We combined monthly groundwater depth records from 116 monitoring wells (January 2021-August 2024) in the Yinchuan Plain, northwestern China, with multiple other datasets, including groundwater hydrochemical data, surface soil salinity samples, multi-source remote sensing images, model interpretability diagnostics, response-time sensitivity tests, and representative profile observations. A hierarchical classifier was trained using surface soil salinity samples collected in August 2022 and tested externally using independent samples collected in August 2025. The resulting monthly salinization class sequence was used as a model-generated regional indicator. Shallow, relatively saline groundwater was concentrated in the lacustrine-alluvial plain (shallow), whereas severe salinization was mainly found in the localized shallow groundwater areas. A regional class area analysis showed a decrease in the severe salinization area proportion accompanied by an increase in the moderate salinization area proportion within 1.00-3.00 months after the onset of summer irrigation-related groundwater shallowing. Approximately 4.00-5.00 months after the onset of winter irrigation-related groundwater shallowing, a delayed winter-spring resalinization signal was detected. Sensitivity tests showed that the separation between earlier summer responses and later winter responses persisted across the tested cycle subsets and operational settings. The water-salt coupling zone accounted for 56.97% of the classified area, and three response-based groundwater-control forms were identified: threshold-like depth, water-salt sensitive response zone, and groundwater-control instability form. These findings support salinity management that considers both the subregional setting and seasonal stage. The predicted 4.00- to 5.00-month delay in winter-spring resalinization should be understood as an integrated winter-spring resalinization indicator rather than a direct groundwater transit time. More broadly, this response-based paradigm extends regional research on groundwater-related surface soil salinization beyond static depth thresholds by explicitly accounting for temporal lags, seasonal stages, and hydrogeological heterogeneity.

The efficient use of limited water resources remains a critical challenge in arid and semi-arid areas, where land-use structure exerts strong control over eco-hydrological processes. The Jinghe River Basin, located in the central Loess Plateau of China, is a typical semi-arid watershed that has undergone pronounced land-use transitions under the Grain-for-Green Program. This study developed a sub-basin- scale analytical framework to optimize the land-use structure in the Jinghe River Basin to improve water-use efficiency. Multi-source remote sensing data (2002-2022) were used to classify the dominant land-use structures based on K-means clustering, and to quantify the key hydrological indicators, including the runoff coefficient, evapotranspiration ratio, and water yield. The eco-hydrological thresholds and trade-off-synergy relationships were identified through boundary-line and correlation analyses, and were then incorporated into the land-use optimization model that was solved using LINGO and spatially simulated in R v.4.2.2 software. The results showed that the basin can be classified into grassland-dominated, cropland- dominated, and mixed cropland-forestland-grassland sub-basins with distinct hydrological responses. Moreover, the optimization results based on these boundary-line thresholds indicated structure-specific optimal land-use proportions for water-use efficiency: the grassland proportion was approximately 0.85 in grassland-dominated sub-basins, and the cropland proportion was approximately 0.63 in cropland-dominated sub-basins. In mixed cropland-forestland-grassland sub-basins, the optimal cropland and grassland proportions were approximately 0.33 and 0.41, respectively, with a composite utilization index of 0.35. These thresholds represented the balance points among the runoff regulation, evapotranspiration control, and water-yielding capacity under limited water resources in the region. By explicitly linking the land-use configuration with the hydrological efficiency through threshold-based constraints, this study provides a data-efficient operational framework for coordinated land-water management in the Loess Plateau, as well as other semi-arid river basins.

Shoulder-lines are topographic features of loess landforms, and their dynamic changes visually reflect differences in gully erosion status and processes on the Chinese Loess Plateau. Therefore, accurate shoulder-line extraction is important for understanding loess landforms. However, existing extraction methods cannot sufficiently record loess shoulder-line details, reducing extraction accuracy. Based on digital elevation model (DEM) data acquired in 2013 and 2016, this study systematically analyzed the loess shoulder-line morphology in a small watershed on the Chinese Loess Plateau. We integrated theoretical approaches (e.g., signal processing) with bidimensional empirical mode decomposition (BEMD) combined with edge detection (Laplacian of Gaussian (LOG) operator) to automatically extract loess shoulder-lines, and assessed the extraction accuracy through a comparison with three methods (LOG, illumination model, and positive-negative terrain) from existing studies and manual extraction results. The proposed method transformed loess landform research from the spatial domain to the frequency domain. BEMD combined with edge detection effectively revealed shoulder-line features while maintaining the local completeness of shoulder-lines, specifically in areas with high terrain complexity or curvature. The extraction accuracy was 90.11% in the study area, with a mean absolute error (MAE) of 0.8040 m and a median absolute error (MedAE) of 1 m. When applied to different loess landform types, the method demonstrated broad universality and feasibility. The extraction accuracy exceeded 95.00% in the loess tableland watershed, and reached 91.90% in the loess hill watershed. Thus, our proposed method provides a novel perspective for shoulder-line extraction, which can effectively characterize changes in loess landforms at different scales and ensure the automation and high accuracy of shoulder-line extractions. Ultimately, this study is expected to advance gully erosion research on the Chinese Loess Plateau from a frequency-domain perspective.

Topography plays a crucial role in regulating vegetation distribution through its control over solar radiation, soil moisture, and erosion-deposition dynamics. However, fine-scale vegetation-terrain relationships remain insufficiently understood because previous studies have relied on coarse- or single-source data and lacked an integrated analytical framework. This study integrated unmanned aerial vehicle laser scanning (ULS) and structure-from-motion (SfM) photogrammetry by mapping SfM orthophoto texture onto the ULS point cloud, to generate a unified high-resolution vegetation-terrain dataset for the Qiaogou catchment in the hilly and gully Loess Plateau, China. Nine topographic factors were derived from ULS digital elevation models (DEMs), while texture and vegetation indices were extracted from SfM imagery. The vegetation-terrain relationship was quantified using a hierarchical framework that sequentially evaluated the linear effects using ordinary least squares (OLS), nonlinear effects using generalized additive models (GAM), random forest (RF), and extreme gradient boosting (XGB), spatial non-stationarity using geographically weighted regression (GWR), and interaction among explanatory factors using Geodetector. Results showed that vegetation cover exhibited pronounced spatial heterogeneity, with dense vegetation primarily located on moist valley floors and sparse cover on steep and rugged slopes. Elevation and slope imposed the primary topographic constraints on vegetation cover, while terrain roughness index (TRI), topographic position index (TPI), topographic wetness index (TWI), and slope length factor (LS) provided complementary descriptors of local terrain variability and moisture conditions. Nonlinear models found that slope (45°) and wetness thresholds determined ecological stability limits, while spatial models revealed a significant non-stationarity (Moran's I=0.352) in topographic influence. Geodetector interaction analysis indicated that terrain effects were expressed mainly though factor combinations, with several slope-related pairs providing additional explanatory power beyond individual factor, while the single-factor explanatory power for overall vegetation cover remained modest. By fusing ULS and SfM data and integrating multi-level analytical models, this study provided a process-based and scale-consistent understanding of how geomorphic structure and hydrological connectivity co-regulated vegetation heterogeneity. The proposed framework offered a robust reference for high-resolution vegetation monitoring, eco-geomorphic modeling, and restoration planning in complex loess landscapes.

Meeting the escalating global food demand while maintaining sustainable ecosystem development poses a pressing global challenge. Since 1990, China's newly reclaimed farmland has been predominantly distributed in arid northwestern regions, where farmland expansion boosts food production but intensifies water consumption and threatens oasis ecological stability. However, conventional continuous models are incapable of identifying nonlinear threshold effects in water-carbon trade-offs, which are essential for targeted ecosystem management. To fill this research gap, this study developed a water-carbon trade-off degree (WCTD) indicator, discretized the WCTD indicator into ordered categorical variables, and adopted the Ordered Logit Model to quantify the impacts of farmland expansion on water-carbon trade-offs of the Heihe River Basin, China. Results showed that in the middle reaches of the Heihe River Basin, farmland expanded by 307.11 km² from 2001 to 2020, mainly due to conversion from grassland and unused land. The regional annual mean net primary productivity (NPP) and evapotranspiration (ET) both showed significant increasing trends. However, the relatively faster growth rate of ET resulted in a pronounced decrease in water use efficiency (WUE), with a linear trend slope of -0.0072 g C/(mm•m²•a) (P<0.050). The Ordered Logit Model results suggested that increased well density and temperature were both significantly positively associated with the deterioration of the water-carbon trade-off, and that grassland-to-farmland conversion imposed a stronger negative impact than unused land-to-farmland conversion. This is mainly because unused land has extremely low baseline WUE, and its reclamation improves soil and hydrological conditions, thereby greatly enhancing WUE. In comparison, grassland reclamation disrupts ET balance; water consumption increases more rapidly than carbon sequestration, which restricts WUE improvement. This study provides a scientific reference for rational farmland planning in the Heihe River Basin and constructs a discretized modeling analytical framework that can be applied to other arid regions to explore the nonlinear interaction mechanisms of human-driven water-carbon coupled systems.

Farmland shelterbelts play a crucial role in protecting arable land in arid and semi-arid areas, but they may also compete for soil water with crops in agroforestry systems. Film-mulched drip irrigation in adjacent farmland substantially alters the water supply dynamics of shelterbelts, yet studies on the hydrological interactions between farmland shelterbelts and crops under this irrigation regime in the arid oasis areas along the desert margins of Xinjiang Uygur Autonomous Region, China remain limited. This study conducted an experiment in an oasis at the southern margin of the Gurbantunggut Desert in Xinjiang over the 2022-2024 growing seasons. We characterized the evapotranspiration (ET) patterns of the cotton field and adjacent Populus shelterbelt, quantitatively evaluated their interspecific water relationships, and examined fine root distribution to identify the zones and intensity of water competition. The results indicated that during the seedling and flowering-bolling stages, cotton field ET exceeded shelterbelt ET by 56.0%-66.1%; however, during the boll-opening stage, shelterbelt ET was 24.2% higher than cotton field ET. Within the 0.1H-0.5H into the cotton field from the shelterbelt-cotton field interface (where H is the mean tree height), Populus exhibited a competitive advantage in the 20-40 cm soil layer, with a competitiveness index of 0.083-0.146 (versus 0.039-0.051 for cotton), while competition was weakest in the 40-60 cm layer. At 0.1H, Populus root biomass density (RBD) and root length density (RLD) were 1.0-15.2 and 6.0-50.0 times those of cotton, respectively; at 1.0H, cotton RBD accounted for 63.0%-93.5% of total root biomass; and no Populus roots or evidence of water competition were detected at 2.0H and 3.0H. Supplemental irrigation should be applied to cotton fields adjacent to shelterbelts to offset water loss from interspecific competition, particularly during the boll-opening stage. The findings provide a practical basis for optimizing agricultural water management in oasis-desert agroforestry systems under film-mulched drip irrigation and offer regional empirical evidence for global research on water competition in arid-zone agroforestry systems.

Mining activities often lead to drastic alterations in surface structure, and the subsequent vegetation recovery process exhibits significant spatiotemporal heterogeneity. The Changping mining area in Shanxi Province, China was selected as the study region. Leaf area index (LAI) data derived from Sentinel-2A/B multispectral imagery were subjected to time series decomposition. Based on the decomposed series, we applied an improved critical slowing down (CSD) model to calculate two indicators—the first-order autoregressive (AR(1)) coefficient and the detrended fluctuation analysis (DFA) scaling exponent—to assess vegetation resilience. The spatiotemporal variation of vegetation resilience and its correlation with mining-induced disturbances were subsequently analyzed in the six panels in the Changping mining area from 2016 to 2024. The results indicated the following: (1) Overall, LAI in the Changping mining area exhibited a spatial pattern of higher values in the west and lower values in the east. Within naturally vegetated areas, the annual mean LAI increased from 0.891 to 1.172, showing a gradual upward trend over time; (2) Overall vegetation resilience in the Changping mining area exhibited a declining trend. Compared with 2016-2020, the proportion of areas with enhanced resilience increased by 33.97% during 2021-2024, with the most pronounced changes observed in the panels 4 and 5, indicating significant impacts of mining activities. At the working face scale, the negative effects of mining on vegetation resilience were particularly evident, as resilience during the mining period was substantially lower than that in the corresponding control areas; And (3) mining activities exerted significant disturbances on vegetation resilience in the Changping mining area, with both temporal and spatial gradients. Owing to differences in mining intensity and operational schedules among zones, the responses of vegetation resilience were spatially heterogeneous. Vegetation resilience tended to increase with distance from the actively mined areas, and this spatial difference diminished after mining ceased, indicating a delayed and partially reversible disturbance effect. These findings provide a theoretical basis and a decision-making reference for the protection and restoration of vegetation ecosystems in mining areas.