A potential interaction-based approach for appraising robustness and identifying critical links of regional road networks exposed to repeated flooding: case study of Dibrugarh district, Assam, India

Abstract

The article presents a methodology for appraising the robustness and identifying critical links of regional road networks exposed to recurring flooding. A set of indices termed the Network Robustness Index is introduced to appraise the robustness of a regional road network by comparing its performance between normal and disrupted conditions due to inundation. The performance indicator in the Network Robustness Index is the aggregate 'potential interaction' within a study region, estimated with the inputs of centrality, population size and spatial separation of constituent settlements, based on the theoretical framework of the Gravity Model. A diminution of 'potential interaction' in disruption conditions quantifies the network's robustness. K-means cluster analysis technique is applied to identify the 'very critical', 'critical' and 'less critical' flood zones based on the criteria of relative diminution of aggregate ‘potential interaction’ resulting from inundation-induced serviceability loss of road links. The criticality of a road link corresponds to the criticality of its associated flood zone. The GIS platform is utilised for data extraction, processing, mapping and other analyses. The suggested methodology is demonstrated in Dibrugarh, one of the worst flood-affected districts in Assam, India. The findings indicate that approximately 34.67% of the study area experiences regular inundation, and the regional road network may suffer an estimated 18.23% performance loss in the worst possible flood scenario. Flood zones are categorised, and critical road links are identified. The study provides essential insights for prioritising pre-disaster mitigation, post-disaster retrofitting and disaster management planning. It also highlights opportunities for further research.