Evaluation of ecological network resilience using OWA and attack scenario simulation in the Gansu section of the Yellow River Basin, NW China

Abstract Regional ecosystem quality has been rapidly declining as a result of rapid urbanization, which has also fragmented landscapes and reduced ecosystem connectedness. Strengthening ecological network resilience helps improve the ecological environment’s quality, protect biodiversity, and maximize ecological benefits. To quantitatively assess ecological network resilience based on 2020 data, we have selected the Gansu region of the Yellow River Basin as a case study in this study due to its significant ecological condition and sensitive vulnerability. Firstly, based on four ecosystem services (ESs)—water yield, carbon storage, soil conservation, and habitat quality—the ordered weighted averaging (OWA) method is introduced to determine ecological sources by calculating ordered weights under different risk coefficients to weigh multiple ESs. Two node attack simulations—random attack and deliberate attack—are used for quantitatively evaluating ecological network resilience, which can simulate the impact of external interference. The concept of a resilience threshold is introduced into the evaluation process of network resilience to improve the objectivity and accuracy of results. The findings indicate that (1) 156 ecological corridors spanning a total distance of 6,569.3 km and 73 ecological sources totalling 20,840 km 2 were found in the study area. These findings generally demonstrate a concentrated and contiguous configuration in the southwestern region and a broken and scattered configuration in other regions. (2) With a mean degree of 4.27, a mean path length of 4.08, and a clustering coefficient of 0.47, an undirected and unweighted complex network with improved connectivity and no discernible clustering characteristics was established. (3) The ecological network is more resilient to perturbations from natural disasters when compared to assault simulations in two scenarios; the resilience threshold is 0.34. Considering the final results of node global features and resilience curves, ecological protection suggestions are proposed.