Abstract:

[Objective] This study aimed to reveal the spatiotemporal evolution characteristics and driving mechanisms of cropland net primary productivity (NPP) in the Huangshui River Basin under rapid urbanization, with particular attention to the ecological vulnerability of high-altitude valley agricultural areas. By identifying the long-term changes, spatial heterogeneity, persistence, and the major natural and anthropogenic controls of cropland NPP, the study sought to provide a scientific basis for cultivated land protection, optimization of land-resource allocation, ecological conservation, and coordinated regional agricultural development. [Methods] The Huangshui River Basin was selected as the study area, and multisource remote-sensing and spatial datasets from 2001 to 2024 were integrated to construct a long-term cropland NPP analysis framework. The datasets included MODIS-derived NPP, land-use/land-cover data, a digital elevation model, population and GDP data, and annual mean temperature, precipitation, and evaporation data. To ensure spatial comparability among datasets with different native resolutions, all variables were projected, clipped, and resampled to a common 500 m grid referenced to the NPP dataset. Continuous variables were resampled using bilinear interpolation, whereas categorical land-use data were processed using a nearest-neighbor method. Cropland was extracted from the land-use dataset and used to constrain the subsequent NPP analyses to cultivated areas. Based on the annual cropland NPP series, the coefficient of variation was used to characterize temporal stability and interannual variability; the Mann-Kendall trend and mutation tests were applied to identify the direction, significance, and possible turning points of long-term NPP changes; the Hurst index was used to evaluate the persistence of historical trends and infer the potential continuation or reversal of cropland NPP changes; and the geographical detector was used to quantify the explanatory power of environmental and socioeconomic factors and examine the enhancement effects resulting from interactions among multiple drivers. The analysis focused on four aspects: the temporal evolution of basin-wide cropland NPP, the spatial pattern of NPP and its variability, the spatial distribution and persistence of increasing or decreasing trends, and the relative contributions and interactions of natural and human factors. [Results] From 2001 to 2024, the annual mean cropland NPP in the Huangshui River Basin showed an overall fluctuating upward trend, increasing from 81.0 g·m⁻²·a⁻¹ to 238 g·m⁻²·a⁻¹, indicating a substantial improvement in cropland productivity over the study period. Spatially, cropland NPP exhibited a stable pattern characterized by higher values in the southeast and lower values in the northwest. High-value areas were mainly located in the middle and lower reaches of the basin, where hydrothermal conditions are relatively favorable for agricultural production, whereas relatively low values were concentrated in the colder and more environmentally constrained northwestern part of the basin. Variability analysis showed that 67.7% of the cropland area maintained relatively stable NPP dynamics, whereas areas with greater fluctuations were mainly concentrated in urban-rural transition zones around Xining, where topographic complexity and intensive human disturbance coexist. Trend analysis further demonstrated that cropland NPP was dominated by increasing trends, with 72.4% of the cropland area showing increases. These areas were highly clustered in valley agricultural zones in the middle and lower reaches, especially in Ledu District and Minhe County, where comparatively favorable water and heat conditions support sustained crop growth. In contrast, areas with decreasing trends accounted for a much smaller proportion and were scattered mainly along the main river corridor and at the margins of urban expansion, suggesting that locally intensified land-use change and urban encroachment may weaken cropland productivity. The mean Hurst index of cropland NPP was 0.650, and 81.7% of the cropland area showed positive persistence, indicating that the historical improvement trend is likely to continue across a large part of the basin. Nevertheless, some cropland near towns and along river corridors showed the potential for trend reversal or degradation, revealing localized ecological risks that require continued attention. Driver detection showed that mean annual evaporation, elevation, and mean annual temperature were the dominant natural factors shaping the macroscale spatial differentiation of cropland NPP. These results indicate that hydrothermal conditions and terrain constraints remain the fundamental environmental controls on agricultural productivity in the high-altitude basin. At the same time, gross domestic product (GDP) and population size, representing the intensity of human activities, exhibited stronger explanatory effects when interacting with topographic factors than when acting alone. These interaction effects were characterized by nonlinear enhancement, demonstrating that the spatial response of cropland NPP cannot be explained by a single environmental or socioeconomic factor. Instead, the combined influence of terrain, climate, urbanization, and population pressure amplifies local spatial heterogeneity, particularly in areas where high-quality cropland overlaps with zones of rapid socioeconomic development. [Conclusion] The spatiotemporal evolution of cropland NPP in the Huangshui River Basin reflects a coupled response to regional climate warming and wetting and increasing human activities. The overall rise in NPP and the large proportion of cropland showing positive persistence indicate that the basin has generally maintained or improved agricultural productivity over the past two decades, but this positive regional tendency coexists with localized risks associated with urban expansion, river-corridor development, and concentrated human disturbance. Therefore, cropland management should shift from uniform protection toward differentiated spatial regulation. In the southeastern valley areas and the middle and lower reaches, where high-NPP cropland is concentrated, priority should be given to strengthening the protection of high-quality cultivated land, controlling excessive conversion to construction land, and maintaining the stability of productive agricultural ecosystems. In the northwestern alpine areas, where environmental constraints are stronger and productivity is relatively low, management should be adapted to local ecological conditions and emphasize ecological restoration, rational land use, and the prevention of further degradation. For urban-rural transition zones and riverside areas showing unstable or potentially reversing NPP trends, closer monitoring and more precise land-use control are necessary to reduce cumulative human disturbance. These findings demonstrate that an integrated assessment combining long-term remote sensing, trend analysis, persistence evaluation, and geographic detection can effectively identify both broad spatial patterns and local risk areas of cropland productivity change, thereby providing a practical scientific reference for sustainable cropland utilization, ecological security, and high-quality agricultural development in high-altitude, ecologically fragile river basins.

Key words: cropland, spatio-temporal evolution, net primary productivity (NPP), driving forces, Huangshui River Basin

摘要：

【目的】揭示城市化背景下湟水河流域耕地净初级生产力（NPP）的时空演变及驱动机制，为高寒河谷区耕地保护提供依据。【方法】基于2001－2024年多源遥感与空间数据，采用变异系数、Mann-Kendall趋势与突变检验、Hurst指数和地理探测器，分析耕地NPP的时空分异、变化趋势、持续性及驱动因素。【结果】2001－2024年耕地NPP年均值由81.0 g·m⁻²·a⁻¹增至238 g·m⁻²·a⁻¹，空间上呈“东南高、西北低”的格局，全域67.7%的耕地NPP相对稳定，较高波动区主要位于西宁市周边城乡交错带；NPP呈增加趋势的面积占72.4%，高度集聚于流域中下游水热条件优越河谷地带的乐都区、民和县，呈减少趋势的区域较少，零散分布于干流沿线的城市扩张边缘；Hurst指数均值为0.650，正向持续面积占81.7%，局部城镇及河流沿岸存在潜在退化风险。年均蒸发量、高程和年均气温是主要自然驱动因子，GDP和人口规模与地形因子交互后呈非线性增强。【结论】耕地NPP演变是气候“暖湿化”与人类活动耦合作用的综合响应，应实施差异化分区管控，东南部河谷区强化优质耕地保护，西北部高寒区推进生态修复。

关键词: 耕地, 时空演变, 净初级生产力（NPP）, 驱动力, 湟水河流域