Abstract: Identifying and protecting key ecological elements to construct a regional ecological security pattern can provide stable spatial structure for ecosystems, with consequences on regional ecosystem stability. Taking Gansu-Qinghai section of the upper Yellow River as a case, we integrated ecological sensitivity analysis with multi-source data, including elevation, slope, land use type, and NDVI by using the minimum cumulative resistance model and gravity model to delineate the ecological security pattern and propose optimization strategies. Eighteen ecological sources were identified in the Gansu-Qinghai section of the upper Yellow River, primarily comprising grasslands, water body, and forests, accounting for 10.6% of the total area and forming the core space of the regional ecological barrier. We constructed a composite ecological network, consisting of 23 important corridors, 130 general corridors, 82 important nodes, and 378 general nodes, with a structurally complete yet optimizable configuration. The Gansu-Qinghai section of the upper Yellow River exhibited a typical plateau resistance surface, requiring focused monitoring grassland degradation zone in the western region, the interlaced zone of water conservation and urban development in the central region, and the significant area of urban expansion in the eastern region. The functional units of Gansu-Qinghai section of the upper Yellow River could be classified into ecological conservation zones, core restoration zones, ecological management zones, and ecological preservation zones. We should clarify spatial governance principles and implement differentiated ecological protection strategies to achieve regional ecological restoration and spatial optimization.

Key words: Gansu-Qinghai section of the upper Yellow River, ecological security pattern, gravity model, minimum cumulative resistance model, sensitivity analysis