The Qinling Mountains serve as a vital ecological barrier for China, needing regular monitoring for sustainable progress. The present study analyzed the Qinling Mountains (sections in Shaanxi province) via the Google Earth Engine (GEE) and Landsat images, constructing a Remote Sensing Ecological Index (RSEI) for quantitatively analyzing the spatiotemporal evolution of the ecological environment quality from 1990 to 2020. Additionally, this study integrated geodetector and multiscale geographically weighted regression (MGWR) to discern the driving factors influencing ecological quality. The results indicated the following: (1) over the 30 years, the RSEI in the study area initially declined (1990–2005) and then began to improve (2005–2020), demonstrating an overall upward trend; (2) RSEI varied with topography: lower values corresponded to lower elevations (<1200 m) and gentle slopes (1600 m) and steeper slopes (>20°); and (3) NDVI, DEM, land use, and the presence of protected areas emerged as the main driving forces of ecological quality. The interaction between ecological conservation policies and socioeconomic factors enhanced the explanatory strength for the differentiation in ecological environment quality. In conclusion, factors like increased vegetation cover and the establishment of protected areas have significantly promoted the improvement of ecological environment quality in the Qinling region, making this study a scientific reference for providing ecological environment protection and promoting high-quality development.

Soil erosion is a global environmental problem. The rapid monitoring of the coverage changes in and spatial patterns of photosynthetic vegetation (PV) and non-photosynthetic vegetation (NPV) at regional scales can help improve the accuracy of soil erosion evaluations. Three deep learning semantic segmentation models, DeepLabV3+, PSPNet, and U-Net, are often used to extract features from unmanned aerial vehicle (UAV) images; however, their extraction processes are highly dependent on the assignment of massive data labels, which greatly limits their applicability. At the same time, numerous shadows are present in UAV images. It is not clear whether the shaded features can be further classified, nor how much accuracy can be achieved. This study took the Mu Us Desert in northern China as an example with which to explore the feasibility and efficiency of shadow-sensitive PV/NPV classification using the three models. Using the object-oriented classification technique alongside manual correction, 728 labels were produced for deep learning PV/NVP semantic segmentation. ResNet 50 was selected as the backbone network with which to train the sample data. Three models were used in the study; the overall accuracy (OA), the kappa coefficient, and the orthogonal statistic were applied to evaluate their accuracy and efficiency. The results showed that, for six characteristics, the three models achieved OAs of 88.3–91.9% and kappa coefficients of 0.81–0.87. The DeepLabV3+ model was superior, and its accuracy for PV and bare soil (BS) under light conditions exceeded 95%; for the three categories of PV/NPV/BS, it achieved an OA of 94.3% and a kappa coefficient of 0.90, performing slightly better (by ~2.6% (OA) and ~0.05 (kappa coefficient)) than the other two models. The DeepLabV3+ model and corresponding labels were tested in other sites for the same types of features: it achieved OAs of 93.9–95.9% and kappa coefficients of 0.88–0.92. Compared with traditional machine learning methods, such as random forest, the proposed method not only offers a marked improvement in classification accuracy but also realizes the semiautomatic extraction of PV/NPV areas. The results will be useful for land-use planning and land resource management in the areas.

Understanding the integrated effects of vegetation types on the physicochemical properties and quality of sandy soils is crucial for guiding vegetation reconstruction and ecological restoration in desertified areas. This study selected three vegetation types at the southern edge of the Mu Us sandy land, including fenced Leymus secalinus Tzvel. grassland (LS), natural Hedysarum mongolicum Turcz (HM) forest land, and Salix cheilophila Schneid. (SC) forest land, as well as sandy land as a control (SD). The differences in the soil physicochemical properties were investigated by collecting soils from three layers within 0–60 cm. The soil quality index (SQI) was calculated using principal component analysis to comprehensively evaluate the soil quality. This study found that the soil physicochemical properties differed significantly among the plots and layers, and the soil properties exhibited a vertical distribution, with chemical indicators concentrated in the surface layer. As depth increased, differences in soil properties between the vegetation and control plots diminished, with vegetation influence mainly in the 0–20 cm layer. Among all the sample sites, the 0–20 cm layer of LS had the highest organic matter content (5.98 g/kg), which was 2.25, 2.28, and 4.71 times that of HM, SC, and SD, respectively. Moreover, LS had the lowest bulk density (1.35 g/cm³), which was 0.89, 0.91, and 0.86 times lower than that of HM, SC, and SD, respectively. The effects of different vegetation restoration types on the comprehensive quality of soil were different, as shown in LS (0.15) > HM (0.11) > SC (0.10) > SD (0.08). In conclusion, all three vegetation restoration types significantly affected the soil physicochemical properties and led to different degrees of variability of soil indexes in the vertical profile, and the fenced grassland restoration type may be preferable for ecological restoration and reconstruction in this region.

Remote sensing technology has been widely used to estimate fractional vegetation cover (FVC) at global and regional scales. Accurate and consistent field spectral measurements are required to develop and validate spectral indices for FVC estimation. However, there are rarely any experimental studies to determine the appropriate times for field spectral measurements, and the existing guidelines or references are rather general or inconsistent, it is still not agreed upon and detailed experiments are missing for a local research. In this experiment, five groundcover objects were measured continuously from 07:30 a.m. to 17:30 p.m. local time in three consecutive sunny days using a portable spectrometer. The coefficients of variation (CV) were applied to investigate the reflectance variation at wavelengths corresponding to MODIS satellite channels and the derived spectral indices used to estimate FVC, including photosynthetic vegetation (PV) and non-photosynthetic vegetation (NPV). The results reveal little variation in the reflectance measured between 10:00 a.m. and 16:00 p.m., with CV values generally less than 10%. The CV values of FVC spectral indices for estimating PV, NPV and bare soil (BS) are generally less than 3%. While more experiments are yet to be carried out at different locations and in different seasons, the findings so far imply that the in situ spectrum measured between 9:00 a.m. and 17:00 p.m. local time would be useful to discriminate FVC objects and validate satellite estimates-based indices using visible, near-infrared and shortwave infrared channels.

Accurate monitoring of the coverage and distribution of photosynthetic (PV) and non-photosynthetic vegetation (NPV) in the grasslands of semi-arid regions is crucial for understanding the environment and addressing climate change. However, the extraction of PV and NPV information from Unmanned Aerial Vehicle (UAV) remote sensing imagery is often hindered by challenges such as low extraction accuracy and blurred boundaries. To overcome these limitations, this study proposed an improved semantic segmentation model, designated SegFormer-CPED. The model was developed based on the SegFormer architecture, incorporating several synergistic optimizations. Specifically, a Convolutional Block Attention Module (CBAM) was integrated into the encoder to enhance early-stage feature perception, while a Polarized Self-Attention (PSA) module was embedded to strengthen contextual understanding and mitigate semantic loss. An Edge Contour Extraction Module (ECEM) was introduced to refine boundary details. Concurrently, the Dice Loss function was employed to replace the Cross-Entropy Loss, thereby more effectively addressing the class imbalance issue and significantly improving both the segmentation accuracy and boundary clarity of PV and NPV. To support model development, a high-quality PV and NPV segmentation dataset for Hengshan grassland was also constructed. Comprehensive experimental results demonstrated that the proposed SegFormer-CPED model achieved state-of-the-art performance, with a mIoU of 93.26% and an F1-score of 96.44%. It significantly outperformed classic architectures and surpassed all leading frameworks benchmarked here. Its high-fidelity maps can bridge field surveys and satellite remote sensing. Ablation studies verified the effectiveness of each improved module and its synergistic interplay. Moreover, this study successfully utilized SegFormer-CPED to perform fine-grained monitoring of the spatiotemporal dynamics of PV and NPV in the Hengshan grassland, confirming that the model-estimated f_PV and f_NPV were highly correlated with ground survey data. The proposed SegFormer-CPED model provides a robust and effective solution for the precise, semi-automated extraction of PV and NPV from high-resolution UAV imagery.

Net primary productivity (NPP) is a key metric for assessing ecosystem functionality and sustainability. This study utilized MOD17A3 NPP data in conjunction with trend analysis, a gravity center model, and the Geodetector method to examine the spatiotemporal evolution and driving mechanisms of NPP across the Yellow River Basin (Shaanxi section) from 2000 to 2022. Results revealed: (1) The average NPP over the study period was 353.01g C m⁻² with an upward trend of 9.7 g C m⁻²yr⁻¹; spatially, NPP increased from north to south, with significant variability in northern Shaanxi and a 17.89 km northeastward shift in NPP’s gravity center. (2) Areas exhibiting significant NPP increases (slope > 0, p < 0.01) comprised 97.83% of the region, while declines were mainly observed in Guanzhong. (3) Normalized Difference Vegetation Index (NDVI) was the dominant factor, with the strongest synergistic, nonlinear interaction with land use type reflecting human activities (q = 0.831), indicating that the combined influence of climate factors, land surface factors, and human activities amplifies the explanatory effect on NPP variability. The study demonstrates an overall improvement in NPP, although local declines occurred, and its spatial distribution was influenced by a combination of natural and human factors. These findings will provide data support for the high-quality development of the Yellow River Basin.