Soil salinization constitutes a critical environmental bottleneck that severely restricts sustainable agricultural development and jeopardizes ecological security in arid oasis regions. However, the effective long-term monitoring and precise governance of this phenomenon in large-scale oases are often hindered by the trade-offs between spatial and temporal resolution in available remote sensing imagery, the high computational costs of processing large-scale datasets, and an insufficient understanding of the complex driving mechanisms behind salt accumulation. To address these challenges, this study establishes a comprehensive technical framework applied to the Yarkand River Basin Oasis in Xinjiang, China. The research innovatively incorporates the compute unified device architecture-enhanced spatial and temporal adaptive reflectance fusion model (CU-ESTARFM), a GPU-accelerated algorithm designed to fuse Landsat and MODIS imagery. This approach effectively addresses the spatiotemporal discontinuity caused by sensor revisit cycles and cloud contamination, generating a high-quality, continuous spatiotemporal dataset. Furthermore, to enhance the physical interpretability of the model, the XGBoost-SHAP (shapley additive explanations) framework was utilized to quantitatively screen and identify the most sensitive environmental covariates from a multidimensional pool of topographic, climatic, and land-use variables. Based on the fused imagery and optimized variables, five distinct machine learning models—random forest (RF), extreme gradient boosting (XGBoost), convolutional neural network (CNN), back propagation neural network (BP), and support vector machine (SVM)—were constructed and systematically compared to reconstruct the spatiotemporal evolution of soil salinity during the critical spring sowing period (April) from 2019 to 2024. The results demonstrate that the CU-ESTARFM algorithm significantly improved computational efficiency, successfully solving the problem of data scarcity for long-term, high-precision monitoring in large-scale areas. Among the evaluated models, the random forest (RF) model exhibited the highest performance, achieving a coefficient of determination (R²) of 0.65 and a root mean square error (RMSE) of 0.16%, confirming the superiority of integrating spatiotemporal fusion with multi-source environmental data. Spatiotemporally, the soil salinity content in the study area displayed a clear increasing trend from the upstream alluvial fans to the downstream plains, characterized by a distinct "low in the southwest and high in the northeast" distribution pattern. Land use analysis indicated that salinity in bare lands experienced a dramatic increase of 86% to 157%, whereas farmland salinity remained relatively stable due to consistent irrigation and management practices, despite an overall intensification of salinization in the downstream regions. The attribution analysis revealed that wind-assisted evaporation and vegetation degradation show a significant correlation with the process of surface salt accumulation, particularly in the dry spring season where high wind speeds and temperatures synergistically accelerate evaporation in the absence of effective leaching. Consequently, a "partitioned policy + edge blocking" management strategy is proposed. This includes implementing a dual control strategy for water rights and groundwater levels in upstream areas to cut off salt sources, and strengthening salt drainage and bio-chemical reclamation measures in downstream areas. Crucially, the study advocates for constructing an ecological resilience barrier within the "farmland-grassland-bare land" ecotone. By adopting intensive management for marginal farmlands and deploying three-dimensional sand control measures, the vicious cycle of "land degradation inducing salt surface accumulation" can be effectively intercepted. This research provides a replicable paradigm for dynamic salinity monitoring and offers a scientific basis for precision governance in complex arid ecosystems.

This study aims to investigate the dynamics and distribution of soil water, heat, and salt in cotton fields under mulched drip irrigation in southern Xinjiang. The field experiment was conducted under different treatments of water and salt irrigation at Xinjiang Alar Modern Agriculture Academician and Expert Workstation in 2023. Three irrigation levels (75%, 100%, and 125% of crop irrigation water demand for cotton) and three levels of irrigation salinity (1.5, 3.5, and 5.5 g/L). The soil moisture content, soil salt content, and temperature of 0-80 cm layer were measured in the whole period of cotton growth. Then HYDRUS-2D model was used to explore the effects of irrigation amounts and salinity on the two-dimensional (2-D) migration and distribution of soil water, heat, and salt in the cotton field. Scenario simulation was implemented to reveal the mechanism of soil salt leaching/accumulation under different irrigation schedules. The appropriate irrigation schedule was proposed under brackish water irrigation. The main results were as follows. The horizontal moisture of the soil profile was unevenly distributed in the early stage of growth, indicating two-dimensional (2-D) distribution. The moisture content of the soil decreased gradually from the middle of the mulched area to the bare soil between the mulch in the horizontal direction. While there was a more uniform lateral distribution of soil moisture in the later stage. The wet range of soil and irrigation uniformity increased with the increase of irrigation quota. The higher irrigation quota, 125% of the irrigation water demand failed to increase the soil water content in the root zone, compared with 100% of irrigation water demand. The efficiency of soil salt leaching increased with the increase of irrigation amount. Salt accumulation position was shifted to the lower soil layer (40-80 cm) and bare soil under low irrigation salinity. While the moderate and high irrigation salinity levels increased the soil salinity beneath the drip tape and along the centerline of the plastic film. Higher water volumes under low irrigation salinity raised the salt content in the soil of the root zone under the mulch. Soil salt accumulation was dominated to decrease the depth of salt accumulation with the increase of irrigation salinity. The outstanding salt accumulation was observed at 60-80 cm depth under the mineralization degree of 5.5 g/L. Furthermore, the soil temperature under mulch was higher than that in the un-mulched area in the whole growth period, where soil temperature decreased with the increase of soil depth. Mulching has significantly enhanced the soil warming at 0-40 cm depth in the early growth stages. HYDRUS-2D platform reliably simulated the transport of water, heat, and salt in cotton fields under mulched drip irrigation. Specifically, excellent consistency was found in the simulated and measured values of soil moisture content, salt content, and temperature within 0-60 cm, where all the R² values were higher than 0.56. Scenario simulation indicated that the highest salinity of irrigation water was 3.2 g/L to prevent salt accumulation in the soil beneath the film (0-40 cm) under full irrigation conditions. Once the irrigation amount was 90% of the irrigation water demand, the salinity of 3.5 g/L was the maximum threshold for the salt balance in the root zone soil beneath the film.

With the improvement of point source pollution control and treatment technology, non-point source pollution has become an important source of water pollution. As the second largest tributary flowing into Qiandao Lake, it is important to quantify the non-point source pollution load in the Wuqiang River Basin, analyze the spatial and temporal distribution characteristics of non-point source pollution, and propose the best management practices (BMPs) suitable for reducing pollutants in the Wuqiang River Basin. Based on the Soil and Water Assessment Tool (SWAT), this study analyzed the temporal and spatial distribution characteristics of runoff and total nitrogen output load in Wuqiang River Basin, explored the reduction effects of different management measures and combinations, and proposed targeted treatment measures for non-point source pollution in Wuqiang River basin. The results showed that: 1) The SWAT model had good applicability for the simulation of runoff and total nitrogen pollution load in Wuqiang River Basin, and the coefficient of determination (R²) of runoff calibration period and verification period were 0.86 and 0.97, respectively. The Nash-sutcliffe coefficient (NSE) were 0.83 and 0.96, and the percent bias (PBIAS) were 15.8% and -6.3%, respectively. The coefficient of determination of total nitrogen calibration period and verification period were 0.87 and 0.74, respectively. The Nash coefficient was 0.63 and 0.66, and the PBIAS were 31.6% and 21.2%, respectively. 2) The runoff and total nitrogen load were mainly concentrated from March to July, accounting for 71.67% and 75.76% of the annual load, respectively. Considering the source and loss path of nitrogen, subbasins with large proportion of cultivated land and forest land and steep slope were set as the key pollution source areas of total nitrogen, and measures to reduce non-point source pollution such as adjusting fertilizer application rate, changing farming methods and setting vegetation buffer belts were taken into account to simulate the reduction efficiency of total nitrogen pollution load. The results showed that the 10-meter vegetation buffer zone was the best single control strategy to reduce the total nitrogen pollution load, and the total nitrogen reduction rate could reach 69.90%. The effect of integrated management measures on the reduction of total nitrogen pollution was better, and the reduction rate of total nitrogen can reach 74.79% when the 10-meter vegetation buffer zone and fertilizer amount were reduced by 20%. The results of this study can provide a theoretical basis for water quality management and control of Qiandao Lake.