Probabilistic connectivity assessment of road networks exposed to spatially correlated rainfall-triggered landslides
Zhengying He, Mitsuyoshi Akiyama, Putri S. Firdaus, Yu Huang, Dan M. Frangopol, Koki Aoki
Abstract
• Framework assesses road network connectivity under rainfall-induced landslides. • Correlated rainfall intensities estimated via Kriging and distance-based model. • Proposed method optimizes prioritization strategies for slope reinforcement. • Slopes reinforced with piles enhances probabilistic connectivity of road networks. Rainfall-triggered landslides frequently lead to significant traffic disruptions and compromised road network connectivity. Accurate connectivity assessment of a wide-area road network with multiple vulnerable slopes must consider the probability of simultaneous landslide occurrence due to rainfall. Quantifying road network connectivity allows the development of effective strategies for prioritizing slope reinforcement. However, the current research on road network connectivity assessment lacks consideration of spatially correlated landslide occurrence probability. This study presents a novel methodology for assessing the probabilistic connectivity of road networks exposed to landslides, taking into account the spatial correlation associated with rainfall hazard intensities assessed using the ordinary Kriging method and distance-based intensity relationship. Vulnerable slopes affecting road accessibility are identified based on landslide susceptibility assessment and spatially correlated states of these slopes are estimated by fragility assessment. Probabilistic connectivity of road networks is estimated by leveraging geographical information system and graph theory. By using the proposed method, improved road network performance is achieved through an optimal strategy in slope reinforcement prioritization. As an illustrative example, the proposed framework is applied to a hypothetical road network in Hiroshima Prefecture, Japan, which is susceptible to rainfall-triggered landslides. The results underscore the significant impact of spatial correlation associated with rainfall hazard intensities on the connectivity probability of road networks exposed to landslides, highlighting the importance of incorporating such correlations in probabilistic road network connectivity assessment.