To address the challenges of low detection accuracy and weak generalization in identifying mining-induced ground fissures from UAV imagery, caused by their slender and discontinuous morphology, complex background clutter, and multi-scale surface features, this paper proposes an enhanced deep semantic segmentation model, SAG-DeepLabV3+ (with Spatial Vision Transformer, Attention mechanisms, and Adaptive Gated Fusion). Specifically, to enhance global context modeling and fine boundary delineation, we introduce a Spatial Vision Transformer (SVT) branch within the Atrous Spatial Pyramid Pooling (ASPP) module. We further employ a dual attention mechanism, sequentially combining Squeeze-and-Excitation (SE) and a Convolutional Block Attention Module (CBAM), for progressive channel and spatial feature refinement. Moreover, an Adaptive Gated Fusion (AGF) module is designed to dynamically optimize the fusion of multi-level decoder features. Experiments on a dedicated UAV-based mining fissure dataset comprising 1280 annotated images show that SAG-DeepLabV3+ achieves a state-of-the-art mean Intersection over Union (mIoU) of 79.52% (with Xception backbone) and 79.19% (with lightweight MobileNetV2 backbone), surpassing DeepLabV3+, U-Net, and PSPNet by a significant margin. Furthermore, by leveraging transfer learning (pre-training on the public CrackVision12K dataset and fine-tuning on our mining fissure dataset), the model’s mIoU is further elevated to 82.04%, demonstrating superior generalization capability. The proposed SAG-DeepLabV3+ effectively balances high accuracy with operational efficiency, fulfilling the potential demand for lightweight automated fissure monitoring under resource-limited field deployments, and lays a foundation for subsequent real-time on-site deployment verification.

Soil pH is a fundamental geochemical parameter with direct implications for environmental quality, but its spatial drivers in geologically complex mountain regions remain poorly understood. This study investigated surface soil pH across 452 sites in Pingquan City, a semi-arid, lithologically heterogeneous mountainous area of Hebei Province, China. The results show that the soil in Pingquan City is predominantly alkaline, with higher pH in southwestern and northeastern areas and lower pH in the northwest. Soil pH ranged from 4.62 to 9.98, with strong positive spatial autocorrelation. Comprehensive quality assessment indicated that the overall soil quality is moderately low. GeoDetector analysis identified average annual temperature, soil texture, elevation, and bedrock lithology as dominant structural drivers, with bi-factor enhancement interactions. GeoSHAP further uncovered two local effects: precipitation exerts a positive influence on pH in carbonate-rock-dominated areas, reversing the leaching–acidification pattern; and temperature functions as a proxy variable integrating co-varying topography, parent material, and texture rather than a direct thermal driver. The combined application of spatial autocorrelation, GeoDetector, and GeoSHAP provides an effective framework for identifying spatial phenomena, discriminating dominant drivers, and explaining local variations. These findings support regional soil quality assessment and land management, and provide a geochemical baseline for safeguarding groundwater resources in mountainous regions.

Rapid urbanization has significantly altered land use patterns in Deqing County, placing increasing pressure on its ecosystem. This study establishes a Driving Force–Pressure–State–Impact (DPSI) framework and selects ten indicators to evaluate ecological sensitivity for 2014, 2019, and 2024. It then couples this framework with the PLUS model to simulate land use and ecological sensitivity changes under the following three 2034 scenarios: natural development (ND), ecological protection (EP), and urban development (UD). Results show that ecological sensitivity exhibits a “high west, low east” spatial pattern and it fluctuated with an initial decline followed by a rise from 2014 to 2024, though the overall trend was a slow decrease. Land use type (weight 0.342), distance to rivers (weight 0.191), and distance to roads (weight 0.146) were the dominant influencing factors, together depicting Deqing’s ecological landscape of “western forests, eastern farmlands, and dense water networks”. The PLUS model demonstrated good validation accuracy (Kappa = 0.81), and the contributions of driving factors shifted over time. Multi-scenario simulation indicates that the ecological protection (EP) scenario is more suitable for sustainable urban development. We recommend implementing differentiated ecological control zones, integrating sensitivity evaluation into planning decisions, and improving the differentiated ecological compensation mechanism.

The hilly terrain at the western foot of the Wanda Mountains boasts intricate geological conditions, compounded by long-term mineral extraction and farming activities. The multi-sourced heavy metals stored within stream sediments create substantial obstacles to accurate source apportionment and targeted risk mitigation. In this work, stream sediment samples were collected across the study watershed to test concentrations of eight heavy metals: As, Cd, Cr, Ni, Cu, Pb, Hg and Zn. Pollution and ecological risk assessments were implemented via the pollution load index (PLI), geo-accumulation index (I_geo) and potential ecological risk index (RI). Kriging interpolation was utilized to visualize and interpret the spatial variation patterns of heavy metals. Combined correlation analysis, principal component analysis (PCA) and the Absolute Principal Component Scores–Multiple Linear Regression (APCS-MLR) receptor model were adopted to qualitatively and quantitatively distinguish pollution endmembers of heavy metals. The main results are as follows: Cr, Cd, Pb and As show significant enrichment, whereas Ni, Cu, Zn and Hg appear depleted; all measured heavy metals present strong spatial variability. Weathering of basic bedrocks governs the widespread high-value zones of Cr, Ni, Cu and Zn. Patchy enrichment of Pb, As and Cd is concentrated in river valley farmlands, with no continuous anomalous zones detected for Hg. The watershed as a whole suffers mild contamination and low potential ecological risk, with only sampling points surrounding mines and agricultural lands experiencing moderate-to-severe pollution and medium ecological hazards. Three source endmembers are identified in the region: a natural geogenic endmember (Cr, Ni, Co, etc.), a mixed mining-associated anthropogenic endmember (Zn, Hg, Cu, Cd), and an agricultural non-point endmember (Pb, As, Sb). Quantitative APCS-MLR calculations confirm that mining disturbances serve as the predominant anthropogenic input for Zn, Hg, Cu and Cd; agricultural practices dominate the accumulation of As and Pb; and geogenic contributions stay marginal for all elements. This study corroborates that regional lithology sets the natural geochemical baseline of heavy metals, while anthropogenic interferences function as the primary driver of localized pollution anomalies. The conclusions can deliver theoretical support for heavy metal pollution remediation in similar hilly watersheds throughout Northeast China.