The Ili River Basin is the area with the highest precipitation in the Tianshan Mountains and frequently experiences heavy precipitation. Studying the fine characteristics of the precipitation at different time scales can advance meteorological forecasting and flood control. Using the hourly precipitation data from nearly 300 regional automatic stations in the Ili River Basin during the warm season (May-August) from 2010 to 2021, taking the latest generation of the Global Precipitation Measurement Mission satellite precipitation product as the initial field, and taking the actual precipitation as the observation field, this study combines the bias-corrected IMERG precipitation estimates with the rain gauge observations through optimal interpolation (OI). The probability density function-OI two-step fusion correction method yielded high-resolution (0.1°×0.1°) hourly multi-source precipitation fusion data, from which the extreme precipitation characteristics in the Ili River Basin were analyzed. The results show that: (1) extreme hourly precipitation events (22.5-38.9 mm) occur in the northern and eastern high-altitude mountainous areas of the basin, whereas the extreme precipitation intensity is stronger in the mid-altitude plain areas at 2000 m. The spatial distribution of the extreme intensity is similar to that of the extreme precipitation thresholds but exhibits a different frequency distribution; specifically, a gradual increase of frequency with altitude. (2) The diurnal variation of hourly extreme precipitation is largest in the northern part of the basin, second-largest in the Wusun Mountain area, third-largest in the eastern section, and smallest in the southern part of the basin. In contrast, the diurnal variation of extreme precipitation frequency is largest in the southern part of the basin, followed by the Wusun Mountain area, the eastern section, and the northern part. (3) The hourly extreme precipitation (EP_{1 h}) threshold ranges from 1.2 to 8.5 mm. The three-hourly extreme precipitation (EP_{3 h}) threshold is distributed similarly to that of EP1h, showing a gradual decrease from the mid-altitude area to the high-altitude mountainous area of the basin. The thresholds of the EP_{6 h} and EP_{12 h} gradually decrease from 2000 m and 3000 m to the high-altitude areas.

Based on the daily precipitation data from 1972 to 2022 and the ERA5 high-resolution monthly reanalysis data, the empirical orthogonal function “decomposition” and synthetic analysis methods were applied to analyze the interannual and interdecadal variability characteristics of extreme precipitation in northern Shanxi Province. Among the 10 extreme precipitation indices in northern Shanxi Province, the spatial distribution characteristics were greater in the southwestern region than the northeastern region, while the duration of the wet period was greater in the central region than the eastern and western regions. After 2010, northern Shanxi Province experienced frequent extreme weather events with alternating droughts and floods. Extreme precipitation occurred during the summer over many years. Weather events in northern Shanxi Province are controlled by the bottom of an abnormal trough area in the Baikal Lake region, and water vapor is transported from the Western Pacific Ocean, the Sea of Japan, and the South China Sea. There is a divergence anomaly at 200 hPa in the upper troposphere and a significant upward movement anomaly in the vertical direction, resulting in an increased frequency and stronger intensity of extreme precipitation in summer. Extreme precipitation is expected to occur in northern Shanxi Province. Affected by the strong high-pressure ridge of Lake Baikal moving southeastward, dry cold air from the north moves southward, blocking the movement of water vapor. The dynamic conditions of the troposphere are not conducive to the generation of extreme precipitation. The decadal time scale revealed a spatial distribution characteristic of more extreme precipitation in the southwest region and less in the northeast region of northern Shanxi Province. During the period of extreme precipitation and heavy rainfall (2009-2016), the convergence of cold and warm air masses generated abnormal northerly airflows and local cyclone circulation in northern Shanxi Province. In the southwestern region, there were strong low-altitude convergence and significant upward movement anomalies, which were conducive to more frequent extreme precipitation. The characteristics of the period with less rainfall (1982-1990) were the opposite.

Utilizing short-duration heavy precipitation data of 109 stations in Shanxi Province (1981-2023), minute precipitation data of 980 regional stations in past 12 years, and socioeconomic data, this study analyzes the spatial distribution characteristics of probability density for short-duration heavy precipitation with different durations and intensity levels through probability density calculations. The fuzzy analytic hierarchy process (FAHP) and semi-trapezoidal distribution model were employed to conduct disaster risk assessments for two high probability density areas: Yuanqu County and Jincheng Urban District. The results show that: (1) The probability density of the short-duration heavy precipitation of 1 hour and 6 hours has the same distribution characteristics: It is higher in the south than in the north, higher in the eastern and western mountainous areas than in the central basin at the same latitude. (2) Yuanqu County is the region with the maximum probability density for 1-hour rainfall in the ranges of 40 mm≤R(1 h)<60 mm, 80 mm≤R(1 h)<93 mm, and for 6-hour rainfall in the ranges of 110 mm≤R(6 h)<130 mm, 130 mm≤R(6 h)<218 mm; Jincheng Urban Area is the region with the maximum probability density for 1-hour rainfall in the ranges of 20 mm≤R(1 h)<40 mm, 60 mm≤R(1 h)<80 mm, and for 6-hour rainfall in the range of 80 mm≤R(6 h)<110 mm. (3) For the evaluated area, disaster risks are classified into five levels: low risk, sub low risk, medium risk, sub high risk, and high risk. Based on the weight vector of the target layer and the fuzzy comprehensive evaluation results of the indicator layer for the evaluated area, the maximum membership degrees of Yuanqu County and Jincheng City are calculated to be 0.3195 and 0.7005, respectively, with fuzzy comprehensive evaluation results indicating secondary high risk and high risk. The evaluation results are consistent with the disaster investigation findings. (4) The provincial disaster risk zoning indicates that the high-risk areas for short-duration heavy precipitation disasters encompass most of Jincheng City, the central and eastern regions of Changzhi City, the north-central sections of Yuncheng City, southeastern counties and districts of Linfen, Liulin County in Luliang, Xiaodian District in Taiyuan, as well as specific counties and townships within Jinzhong.

Urban agglomeration on the northern slope of the Central Tianshan Mountains ranks among the most serious air-pollution problems in China. The impact of temperature inversion over local complex terrain on air pollution has been poorly understood for many years. Here, we derive the wintertime temperature profiles from data obtained by triplicate ground-based microwave radiometers (MWRs) deployed along an elevational transect (613-935 m above sea level) in the Urumqi Valley, Middle Tianshan Mountains, from November 2023 to March 2024. Based on concurrent radiosonde profiles and air-quality monitoring data, we verify the vertical credibility of MWR temperature retrievals, analyze the elevational dependence of the temperature inversion characteristics, and relate the temperature inversion to pollution regime. The MWR and radiosonde data are consistent below 2000 m (R>0.94) and the MWR retrievals can capture the vertical evolution of temperature stratification. In winter, the temperature difference between the top and bottom of the inversion layer and the thickness of the inversion layer in the urban area from south to north increase by 2.1 ℃ and 186 m, respectively, for each 100 m decrease in altitude. Diurnal evolution shows that nocturnal surface-based inversions dominate the valley with maximum inversion at 08:00 Beijing time (top-to-bottom temperature difference of ΔT=9.1℃; inversion layer thickness=1300 m) at low-elevation sites. Post-sunrise, southern Urumqi experiences inversion dissipation while the central and northern sections of the city experience elevated and surface-based inversions, respectively, indicating post-sunrise heterogeneity. As the air quality index (AQI) rises from grade I to grade IV, the thickness of the inversion layer, ΔT, and intensity of the temperature inversion synchronously increase at different altitudes. The ΔT increases by 2-6 ℃. At AQI of grade V or higher, the above inversion parameters slightly decrease from those at grade IV. This change indicates that temperature inversion on the northern slope of the Tianshan Mountains aggravates the pollution problem. The most serious stage of pollution may be contributed by additional factors such as regional transport and chemical transformation.

To investigate the mechanisms by which urban lakes in arid regions locally regulate climate at extremely high temperatures, this study simulated a high-temperature event in the Yuehai Lake area of Yinchuan during July 29-30, 2021. Simulations were performed in the Weather Research and Forecasting (WRF) Community Land Model coupled model configured with four nested grids. The WRF_CLM model accurately characterized the diurnal temperature variations at the Yinchuan station and effectively reproduced the spatial distributions of physical fields such as the near-surface temperature field and wind field, demonstrating its suitability for simulating lake microclimates in arid regions. Under high-temperature conditions, Yuehai Lake, situated in an arid region, can effectively mitigate urban heat-island intensity through synergistic thermal and dynamic effects. The low humidity and strong radiation in the arid region significantly enhance evaporative cooling during the daytime; consequently, the 2-meter air temperature is approximately 4 ℃ lower near the lake than in the surrounding urban areas. During the nighttime, this temperature difference narrows to 1 ℃ because the strong radiative cooling effect over land combines with the thermal inertia of the cooled water body. This “stronger during the day, weaker at night” pattern reduces the diurnal temperature range in the lake area. Simultaneously, radiative evaporation raises the relative humidity in the lake area by 4%-12% from that in urban areas, with the humidity peaking on the downwind lakeshore. Furthermore, the lake-land thermal contrast effectively drives the lake-land breeze circulation. During the day, a horizontal divergence center forms over the lake surface and extends its influence to approximately 0.05 toward the lakeshore; vertically, it induces a closed circulation cell centered at an approximate height of 2200 meters, creating a core zone of specific humidity. At night, the evaporation weakens and the urban background wind system dominates the moisture diffusion, diminishing the local humidification effect around the lake. This study quantifies the local climatic effects of a typical lake in an arid region, providing a scientific basis for ecological urban planning.

Intensifying climate change has led to a surge in extreme precipitation events, posing unprecedented challenges to basin-wide flood prevention and mitigation. The unique combination of meteorological, topographical, hydrological and vegetation factors in the Shule River Basin makes it highly prone to flooding, yet existing forecasting models in the region exhibit low predictive skill. To address the flood control needs of the Shule River Basin，a digital twin platform was developed, and its precision in prediction and simulation was enhanced，this study establishes a coupled flood-prediction framework that centers on an improved SCS rainfall-runoff model and a Snowmelt Runoff Model (SRM), integrated with the Muskingum channel-routing scheme. In data-scarce areas, the SCS-CN loss method is employed to compute net rainfall, and a triangular unit hydrograph is adopted for efficient flow concentration. For alpine snowmelt-driven floods, the SRM is introduced and refined from a daily to a sub-daily time step. The upper Shule River is discretized into 121 sub-basins, and a digital-twin architecture comprising a “data foundation-model platform-knowledge platform” is built, enabling model registration, parameter calibration, real-time forcing and rolling forecasts. Model parameters were calibrated against the July 11, 2023 flood event; the simulated peak discharge showed an error of 6.8% and a timing lag of 2 h relative to observations. During the operational forecast for the July 15, 2024 flood, the platform issued an alert 72 h in advance, predicting a peak discharge of 438 m³·s^-1 arriving at 12:00 on July 15, while the observed values were 491 m³·s^-1 and 14:00, respectively. All errors meet the tolerance criteria specified in the Chinese national standard SL250-2000 for hydrological forecasting. The system provided timely support for reservoir pre-release and downstream evacuation, effectively reducing flood losses. The study advances flood-forecast accuracy and offers an intelligent solution for smart water management and resilient basin development.

The Ji Zi Bend of the Yellow River—a critical sediment source area and water-wind erosion ecotone within the Yellow River Basin—exhibits highly complex water-sediment relationships driven by a combination of aeolian activity and extreme precipitation events. Based on hydro-meteorological data from the Maobulakongdui, Xiliugou, and Huangfuchuan watersheds within the bend for the years 1964-2020, scholars have used various methods(e.g., Mann-Kendall test, double mass curve analysis, and cumulative anomaly method) to elucidate patterns in the spatiotemporal variation of runoff and sediment transport and their influencing factors.The results demonstrate (1) significant downward trends in annual runoff and sediment load for the three watersheds. However, decline rates partially deviated from the west-to-east geographical gradient. Notably, the transitional Xiliugou watershed exhibited the lowest decline rates but the highest anthropogenic contributions, which implies that human interventions override natural geographical gradient. (2) Abrupt change years observed in runoff and sediment series closely aligned with the implementation timelines of major ecological projects. Specifically, changes in Maobulakongdui and Xiliugou corresponded with the implementation periods of the Grain for Green Program and Large-scale Check Dam Construction, while the change in Huangfuchuan was strongly associated with the enforcement of a Grazing Exclusion Policy. (3) Human activities predominantly drove the drastic reduction in water and sediment within these typical bend watersheds, which accounted for an average contribution rate of more than 90% to reductions in runoff and sediment load. Furthermore, this regulatory effect has significantly intensified with the ongoing advancement of ecological management projects. These results can provide a scientific basis for differentiated ecological governance strategies in the Yellow River Basin.

Through controlled field-scouring experiments, we examined the soil erosion dynamics and hydrodynamic mechanisms on slopes hosting different vegetation in the southern Ningxia mountainous area. Specifically, we quantify the synergistic effects of slope gradient (10°, 15°, 20°, 25°), vegetation cover type (barren land, grassland, shrubland), and scouring intensity (4, 8, 16 L·min^-1) on the sediment yield. They key findings are as follows: (1) on slopes with the same vegetation type, the effects of slope gradient and scouring intensity on sediment yield intensify with increasing slope steepness, with slope gradient exerting the greatest promotional effect on the erosion process. The sediment yield was maximized (15.64 kg) on barren land with a 25° slope gradient and a scouring intensity of 16 L·min^-1 and minimized (0.03 kg) on shrubland with a 10° slope angle and a scouring intensity of 4 L·min^-1. (2) The hydrodynamic parameters are significantly correlated with sand production (P<0.05), but the dominant factor depends on the environmental flushing conditions. The water-flow power per unit can contribute up to 44.19% of the sand production on grassland and scrubland at low flushing intensity (4 L·min^-1), whereas each energy parameter contributes nearly 20.00% at high flushing intensity (16 L·min^-1). The factor interactions evolve nonlinearly: the contributions from direct single-factor effects dominate at low scouring intensity (4 L·min^-1) but weaken with increasing scouring intensity (16 L·min^-1) as the multifactor interactions gain prominence. This study can theoretically guide the optimization of slope configurations for regional soil and water conservation purposes and practically guide the precise prevention and control of erosion in ecologically fragile areas.

The seasonal difference in satellite time phase directly affects the accuracy of studies of multiyear changes in glaciers. In this study, we used the daily MODIS images for the warm season (June-September) from 2000 to 2023 to establish a unified comparison standard for the daily snow-area sequence and the annual minimum snow area in glacier areas, evaluate the impact of time-phase deviation on determinations of the temporal variations in glaciers, and analyze the multi-year dynamic-variation characteristics of the snow area in the glacier areas. The results show that: (1) The maximum deviation of glacier-area estimations caused by the time-phase deviation of satellite images is -23.25%, and the deviation is still as much as 7.75% even when the period is in the range from June to September. (2) After eliminating the influence of the time-phase deviation, we found that the existing research results for the changes in the Geladandong and Animachen glacier regions have significantly underestimated the downward trend of glacier changes in the past 20 years. In addition, the rate of reduction in the area of the Animachen glacier is 17.03%, which is greater than the original finding. (3) Based on the time series for the minimum area of snow cover from 2000 to 2023, the annual rate of reduction of the Geladandong glacier over the past 24 years is 0.24%·a^-1 (a cumulative reduction of 5.42%), and that of the Animachen glacier is 0.43%·a^-1 (a cumulative reduction of 9.82%). The annual minimum snow cover area of the Geladandong glacier exhibits a significant downward trend, while that of the Animachen glacier is not obvious, which is different from the current understanding. (4) Taking these two glacier regions as typical examples, we find that the snow-cover area in the glacier regions in the hinterland of the Qinghai-Xizang Plateau usually reaches its minimum during the first ten days of August. The time-phase deviation is generally less than 0.5% when the satellite images in this period or near 20 days are selected. If instead satellite images in early June or late September are selected, the time-phase deviation may may account for more than 12% of the resulting total deviation, and the time-phase deviation of the images in June is generally less than that in September.

To investigate the dynamics of plant species diversity and soil nutrients in artificial Hippophae rhamnoides forests with different restoration years in the Shuanglonggou coal gangue area, this study examined plots restored for 3 a, 6 a, and 11 a, using unrestored gangue wasteland as control. Field vegetation surveys were conducted, soil nutrients were estimated, and plant-soil correlations were analyzed using the space-for-time substitution method. The results showed that: (1) During restoration, 65 plant species from 48 genera and 24 families were recorded, with Asteraceae being the most dominant family. (2) The number of plant individuals increased with restoration time, reaching a maximum of 1234 in the 11 a plot. Species richness, Simpson dominance index, and Margalef richness index increased, reaching the peak values at 34, 0.277, and 3.213, respectively, in the 11a plot, whereas the Pielou evenness index and Shannon diversity index showed overall declining trends. (3) Soil total nitrogen, total phosphorus, organic matter, ammonium nitrogen, and available phosphorus decreased significantly with soil depth, showing obvious surface aggregation, whereas total phosphorus, nitrate nitrogen, and ammonium nitrogen increased significantly with restoration time. (4) Plant species diversity was closely correlated with soil nutrients; plant individual count was significantly positively correlated with nitrate nitrogen and ammonium nitrogen and significantly negatively correlated with available phosphorus; and species richness was significantly positively correlated with soil pH, total phosphorus, nitrate nitrogen, and ammonium nitrogen. Simpson dominance and Margalef richness indices were significantly positively correlated with total phosphorus, nitrate nitrogen, and ammonium nitrogen. However, Pielou evenness index was significantly positively correlated with available phosphorus. This indicated that these soil nutrients were key factors influencing plant species diversity in this region. The study demonstrated that prolonged artificial restoration significantly enhanced plant species diversity and soil nutrient content in the Shuanglonggou gangue area, providing a scientific basis for ecosystem restoration in abandoned mining regions.

Applying nitrogen and phosphorus fertilizers effectively can significantly improve the growth of Pinus sylvestris var. mongolica in sandy regions. To investigate this, we conducted a study on P. sylvestris var. mongolica trees in the Mu Us Sandy Land. We included an unfertilized control (CK) and tested eight different fertilization treatments: N1 (46 g N per plant), N2 (92 g N per plant), P1 (18 g P per plant), P2 (36 g P per plant), P1N1 (18 g P and 46 g N per plant), P1N2 (18 g P and 92 g N per plant), P2N1 (36 g P and 46 g N per plant), and P2N2 (36 g P and 92 g N per plant). Specifically, the present study investigated the physiological and biochemical responses of P. sylvestris var. mongolica needles to varying N and P fertilization regimes. We quantified nutrient elements, chlorophyll, malondialdehyde, and superoxide anion concentrations. Correlations and comprehensive ranking analyses of these indicators were conducted to establish a theoretical framework for sustainable fertilization in sandy habitats. Our results show that both single and combined applications of nitrogen and phosphorus fertilizers significantly increased the concentrations of total nitrogen, phosphorus, potassium, calcium, and magnesium in the needles (P<0.05), confirming enhanced nutrient uptake. While higher nitrogen or phosphorus rates did not significantly increase chlorophyll content and sometimes led to a decline, certain treatments reduced oxidative stress markers. Specifically, malondialdehyde content was significantly lower in the P1, P1N1, P1N2, and P2N1 treatments relative to CK (P<0.05), indicating that phosphorus enhances stress tolerance. Similarly, P1N1, P1N2, and P2N1 treatments significantly reduced superoxide anion levels (P<0.05), demonstrating that combined nitrogen and phosphorus application confers greater resistance to oxidative damage. A comprehensive analysis, incorporating a membership function value and fertilization costs, identified P1N1 as the most effective fertilization strategy.

It is of great scientific significance and practical value to quantitatively evaluate the coupling coordination degree and mutual response relationship between new quality productivity and ecological resilience in the Yellow River Basin for promoting ecological protection in the basin. Based on the construction of the coupling coordination evaluation index system of new quality productivity and ecological resilience in the Yellow River Basin, this study comprehensively used Kernel density estimation, Dagum Gini coefficient decomposition method, coupling coordination model, and panel VAR model to explore the spatiotemporal distribution characteristics and interactive response relationship between the coupling coordination degree of new quality productivity and ecological resilience in the Yellow River Basin from 2013 to 2022. The results showed that: (1) During the study period, the coupling coordination level of new quality productivity and ecological resilience in the Yellow River Basin significantly improved, with still much room for improvement in terms of high quality coordination. (2) The contribution rate of inter-basin differences was 60%-70%, primarily causing regional imbalance, especially a “fault gap” in the upstream, middle, and lower reaches. (3) The new quality productivity and ecological resilience in the Yellow River Basin had a “self-reinforcing effect,” and there was a dynamic interaction response of “short-term inhibition and long-term promotion” between the two but a virtuous circle was not yet formed. This study provides a certain reference value and scientific basis for promoting the development of new quality productivity and ecological resilience in the Yellow River Basin based on local conditions.

The Shule River Basin belongs to a typical arid zone oasis-desert ecosystem, with a fragile ecological environment. Scientific regulation of land use is crucial for regional development and ecological protection. Based on land use data obtained from the interpretation of high-resolution remote sensing images in 2012, 2017, and 2022, combined with the Markov-PLUS model, three scenarios (natural development, cultivated land protection, and ecological protection) were set to predict land use in the ecologically fragile areas of the middle and lower reaches of the basin in 2035 and assess changes in the basin’s habitat quality. The results show that: (1) During the period 2012-2022, the land use in the river basin exhibited a distinct positive transition. Saline-alkali land and unused land displayed a notable declining trend, with their areas decreasing by 484.08 km² and 654.61 km² respectively. Over the same period, the areas of wetlands and shrub-covered land increased by 228.69 km² and 502.33 km² respectively. (2) The overall habitat quality of the basin was relatively low. From 2012 to 2022, the average habitat quality was 0.2799, with low habitat quality being dominant, accounting for 54.28% of the study area. However, it showed an overall improving trend. (3) Among the three scenarios in 2035, the area of shrubland and wetlands will increase, while the area of saline-alkali land and unused land will continue to decrease. Habitat quality will mainly show a transition from “low level to high level”. The habitat quality under the ecological protection scenario is the optimal, increasing by 7.25% compared with that in 2022. Under the continuous warming and wetting trend in Northwest China, efforts should be made to strengthen the optimal allocation and scientific management of land use.

To analyze the vertical variations of particle size distribution and atmospheric particulate concentrations in the middle and lower atmosphere of the Qinghai-Xizang Plateau and clarify the characteristics of the background aerosol spectra, this study utilized the aircraft-borne observational data from five flights in 2020 to analyze the size distribution and number concentration of atmospheric aerosols in the northeastern part of the Qinghai-Xizang Plateau. (1) The scattered cloud systems in the research area moved from west to east, with the cloud top height ranging from 8000 to 10000 m. The backward trajectories demonstrated that the altitude range of 6000-7000 m was primarily affected by the surrounding environmental transmission. The high-altitude aerosol particles were associated with long-distance atmospheric transport, and they originated from the edge of the Qinghai-Xizang Plateau. (2) Aerosols in the middle and lower atmospheric layers over the northeastern Qinghai-Xizang Plateau were predominantly small particles with an average effective diameter (r_e) of 0.68 μm and an average concentration (N_a) of 10²-10³ cm^-3. On average, N_a and r_e reached 507.16 cm^-3 and 0.40 μm, respectively, in the vertical profile, exhibiting a multi-peak structure. (3) The aerosols could be categorized into three modes, with Mode I having mean N_a of 10² cm^-3 and Mode II having mean N_a of 10¹ cm^-3. For Mode III, mean N_a ranged 10^-1-10⁰ cm^-3. N_a decreased with increasing altitude. N_a gradually decreased with decreasing temperature, indicating a significant positive correlation. As relative humidity increased, the aerosol spectral type remained stable, but the concentration gradually increased. (4) At vertical heights of 7000-8000 m, the relative dispersion of cloud droplet (ε) value was relatively high at approximately 0.4, suggesting that cloud droplet sizes vary more significantly at this altitude range. Linear fitting and power-law both indicated a negative correlation between ε and N_a.

Sustainable utilization of cultivated land in the Hehuang Valley, a core grain-production area in the northeastern margin of the Qinghai-Xizang Plateau, is crucial for guaranteeing regional food security and ecological security protection. Utilizing long-term high-resolution land use data from 1980 to 2020, we integrate a topographic gradient analysis, kernel density estimation, the PLUS model, and standard deviational ellipse theory to quantify the temporal dynamics, spatial agglomeration characteristics, topographic differentiation, and driving factors of cultivated land changes in the region. The results revealed (1) a net loss of 119.24 km² of cultivated land in the Hehuang Valley from 1980 to 2020. Most of the lost land was converted to grassland, water area, and built-up areas. (2) A distinct spatial pattern emerged, revealing a high cultivation density in the river valleys and a sparse distribution in mountainous zones. The spatial centroid shifted 2.277 km northwestward, accompanied by progressive spatial contraction. (3) The cultivated land was concentrated on the sunlit slopes of hills and plains below 2772 m altitude, where the slope gradient is 15° and human activities have significantly changed the land use. (4) The population density dominantly drives the changes in cultivated land area, with a contribution degree reaching 0.11, followed by distance to railway, annual average wind speed, altitude, and gross domestic product per unit area. The cultivated land has expanded along the Datong River and Yellow River corridors, contrasting with rapid contraction in the Huangshui River section of Xining. Topographic gradients significantly regulate the spatiotemporal patterns of cultivated land dynamics in the Hehuang Valley. Location-specific management strategies are imperative for optimizing the resource allocation and ensuring sustainable agricultural development under heterogeneous environmental constraints.

The northern slope economic belt of Tianshan Mountains faces multiple challenges in agricultural resource utilization, including water scarcity, ecological vulnerability, and an unbalanced production structure, all of which are closely linked to patterns of agricultural land use. To address these issues, a systematic analysis and assessment of land use patterns and their sustainability is urgently needed. This study examines the evolution and transformation of agricultural land use in the region from 1990 to 2020, identifying 36 distinct utilization patterns. An evaluation framework integrating economic, ecological, and management dimensions is developed to assess the sustainability of two representative models: efficient water-saving irrigation and agricultural mechanization. The results indicate that, driven by socioeconomic progress and technological advancements, the efficient water-saving irrigation model has undergone continuous optimization, achieving the highest sustainability rating (Level I) by 2020, with notable economic, social, and ecological benefits. Similarly, agricultural mechanization has experienced rapid growth since the early 21st century, also reaching Level I by 2020, with the enhancement of mechanization capacity as its primary driver. Looking ahead, agricultural land use in the Northern Slope Economic Belt should transition toward an intelligent, efficient, and ecologically sustainable high-quality development pathway.