The extensive artificially accelerated erosion of spoil heaps on newly engineered landforms is a key ecological management point requiring better understanding. Soil surface roughness is a crucial factor influencing erosion processes; however, study on spoil heap erosion with a view of surface roughness is lacking. This study investigated the erosion processes and the spatiotemporal variation of surface roughness on spoil heaps, and then, analyzed how the roughness affected the hydrological and sediment yield characteristics. Sequences of four artificial rainstorms with constant rainfall intensity (90 mm/h) were applied to cone-shaped spoil heaps (ground radius 3.5 m, height 2.3 m) of a loess soil containing 30 mass percent rock fragments. The surface elevation was sampled by a laser scanner. For the surface roughness indicators, the root mean square height (rmsh) and the correlation length (cl) increased sharply during the first rainfall event, and in the last three rainfall events, rmsh increased slightly and cl showed a relative decrease. The initial rmsh/cl of the whole slope surface ranged from 0.063 to 0.135, and increased with the rainfall sequence, thus, indicating that the spoil heap surface became rougher. Increasing soil roughness in the rainfall sequence delayed the initial runoff time and increased the runoff yield. The average runoff coefficient of the spoil heaps was 0.658. The average erosion rate of each rainfall event can be simulated by a regression equation of the corresponding average runoff rate and median cl (R-square of 0.816). Soil slumping with an average volume of 0.014 m³ occurred in the first two rainfall events, thus, significantly changing the roughness and peak instant erosion rate. Together, the results revealed the effects of surface roughness on the erosion of spoil heaps and would provide a useful reference for soil loss prediction and control.

Soil organic carbon (SOC) constitutes the largest terrestrial carbon pool and plays a crucial role in climate regulation, soil fertility, and ecosystem functioning. Understanding its spatiotemporal dynamics is particularly important in semi-arid regions, where fragile environments and extensive ecological restoration may alter carbon cycling. The Loess Plateau, the world’s largest loess accumulation area with a history of severe erosion and large-scale vegetation restoration, provides a natural laboratory for examining how environmental gradients influence SOC storage over time. This study used a random forest model with multi-source environmental data to quantify soil organic carbon density (SOCD) dynamics in the 0–100 cm soil layer of the Loess Plateau from 2005 to 2020. SOCD showed strong spatial heterogeneity, decreasing from the humid southeast to the arid northwest. Over the 15-year period, total SOC storage increased from 4.84 to 5.23 Pg C (a 7.9% rise), while the annual sequestration rate declined from 0.046 to 0.020 kg·m⁻²·yr⁻¹, indicating that the regional carbon sink may be approaching saturation after two decades of restoration. Among soil types, Cambisols were the largest carbon pool, accounting for over 44% of total SOC storage. Vegetation productivity emerged as the dominant driver of SOC variability, with clay content as a secondary factor. These results indicate that although ecological restoration has substantially enhanced SOC storage, its marginal benefits are diminishing. Understanding the spatial and temporal patterns of SOC and their environmental drivers provides essential insights for evaluating long-term carbon sequestration potential and informing future land management strategies. Broader generalization requires multi-regional comparisons, long-term monitoring, and deeper soil investigations to capture ecosystem-scale carbon dynamics fully.

Water stress represents one of the most critical limiting factors affecting plant distribution, growth rate, biomass accumulation, and crop yield across diverse growth stages. Variations in species’ drought tolerance fundamentally shape global biodiversity patterns by influencing survival rates, distribution ranges, and community composition under changing environmental conditions. This study investigated the physiological responses of six plant species (Haloxylon ammodendron (H.A.), Nitraria tangutorum Bobr. (N.T.B.), Sympegma regelii Bge. (S.R.B.), Tamarix chinensis (T.C.), Potentilla fruticosa (P.F.R.), and Sabina chinensis (Linn.) Ant. (S.C.A.)) to varying water stress levels through controlled water gradient experiments. Four treatment levels were established: W1 (full water supply, >70% field water holding capacity); W2 (mild stress, 50–55%); W3 (moderate stress, 35–40%); and W4 (severe stress, 20–25%). Height growth and leaf mass per area decreased significantly with increasing water stress across all species. S.C.A. consistently exhibited the highest leaf mass per area among the six species, while H.A. showed the lowest values across all treatments. Leaf water content declined progressively with intensifying water stress, with T.C. and P.F.R. showing the most pronounced reductions (T.C.: 16.53%, 18.07%, and 33.37% under W2, W3, and W4, respectively; P.F.R.: 19.45%, 28.52%, and 36.08%), whereas N.T.B. and H.A. demonstrated superior water retention capacity (N.T.B.: 2.44%, 6.64%, and 9.76%; H.A.: 1.44%, 4.39%, and 5.52%). Water saturation deficit increased correspondingly with declining soil moisture. Diurnal leaf water potential patterns exhibited a characteristic V-shaped curve under well-watered (W1) and mildly stressed (W2) conditions, transitioning to a double-valley pattern with unstable fluctuations under moderate (W3) and severe (W4) stress. Leaf water potential showed linear relationships with air temperature and relative humidity, and a quadratic relationship with atmospheric water potential. For all six species, the relationship between pre-dawn leaf water potential and soil water content followed the curve equation y = a + b/x. Under water-deficient conditions, S.C.A. exhibited the greatest water physiological changes, followed by P.F.R. Both logarithmic and power function relationships between leaf and soil water potentials were highly significant (all F > 55.275, all p < 0.01). T.C. leaf water potential was the most sensitive to soil water potential changes, followed by S.C.A., while H.A. demonstrated the least sensitivity. These findings provide essential theoretical foundations for selecting drought-resistant plant species in arid regions of the Qaidam Basin. This study elucidates the response mechanisms of six distinct drought-tolerant plant species under water stress. It provides critical theoretical support for selecting drought-tolerant species, designing community configurations, and implementing water management strategies in vegetation restoration projects within the arid Qaidam Basin. Furthermore, it contributes empirical data at the plant physiological level to understanding the mechanisms sustaining species diversity in arid ecosystems.

In order to evaluate ecological technology scientifically, we constructed a modular “three-stage evaluation method” based on qualitative evaluation, semiquantitative evaluation and quantitative evaluation, and established the theoretical models of the four kinds of ecotechnology, such as soil and water conservation technology, desertification governance technology, rocky desertification governance technology and ecological restoration technology. We gave the quantification criteria of the first-level and second-level index commonly shared by four kinds of ecotechnology and defined the quantification criteria of the third-level index of reflecting the heterogeneity of soil and water conservation technology. An ecotechnology evaluation model combining Analytic Hierarchy Process and Logistic regression was established based on soil and water conservation technology. The rationality of the evaluation method and model were verified by field investigation data of soil and water conservation technology in Gaoxigou. The evaluation method and model could provide scientific basis for the effective introduction and popularization of ecotechnology.