From City to Region: Spatial Structure and Evolution of Resilience Networks in the Yangtze River Delta

Abstract

Urban agglomerations are increasingly exposed to shocks that extend across administrative boundaries through infrastructure systems, industrial chains, population mobility, and shared ecological spaces. Existing urban resilience research has predominantly assessed city-level capacity, while comparatively less attention has been paid to the relational structures associated with differences in resilience capacity and geographic proximity. This study examines the model-implied spatial correlation network of urban resilience across 41 cities in the Yangtze River Delta from 2013 to 2022. A multidimensional resilience index covering economic, infrastructure, social, and ecological dimensions was constructed using the entropy weight method and directed intercity interaction potentials were estimated using a modified gravity model and analyzed through social network analysis. The results show that overall resilience improved during the study period, although substantial cross-city disparities persisted. Under the baseline row-specific mean threshold, network density remained broadly stable, fluctuating between 0.2476 and 0.2506 and decreasing slightly from 0.2506 in 2013 to 0.2482 in 2022. The model-implied network remained fully reachable but relatively sparse, with persistent positional differentiation. Nanjing and Hangzhou consistently occupied leading total-degree positions, while Hefei, Nanjing, Hangzhou, Huangshan, and Chuzhou repeatedly exhibited relatively high betweenness. The block-model structure was also comparatively stable, with only three cities changing block membership between 2013 and 2022. Robustness analyses further showed that broad positional differentiation was more stable than exact density levels, outward-oriented rankings, and brokerage positions. Overall, improvements in city-level resilience did not translate into substantial densification or reorganization of the regional network, highlighting the need to distinguish internal resilience capacity from model-implied network position when designing differentiated regional coordination strategies.