Modeling flood resilience and food security pathways in river basins

Many river basins were developed in the past century with large multi-purpose reservoirs for maximizing agricultural production and managing floods. Increasing frequency of extreme precipitation and floods, however, have highlighted that existing food production and flood protection systems may not remain viable in a changing climate. Motivated by these challenges, this new study uses a network modeling approach to analyze the Indus River basin in Pakistan – a region with large-scale irrigation-based agricultural production and subject to intensifying floods. This work builds upon our earlier network analysis of food production in this region by conducting a new analysis of flood risk reduction. The food production system is modeled with five types of components – human, technical, environmental, institutional, and knowledge – deemed necessary for sustainability-relevant analysis. The human (H), technical (T), and environmental (E) components are modeled as nodes, and their interactions (mediated by institutions and knowledge) are modeled as links. A set of networks, each representing the system in discrete time periods demarcated by major institutional changes from 1947 to 2022, are used to trace how pathways relevant for flood resilience and food production evolved. An empirical analysis of flood damage impacts on food production and associated food and flood policies during the study period is conducted, with an emphasis on catastrophic flood events in 1976, 1988, 2010, and 2022. The flood event of 2022 was reported to have affected 33 million people and led to damages of over $40 billion (~ 11.5% of the national GDP). A set of future policies are simulated in this work to study potential changes in system (network) structure. These policies include construction of small, distributed reservoirs for retaining flood waters, establishing flood insurance programs for small farms, and inter-provincial food exchange. A set of metrics related to network structure, including node centrality and pathway lengths, are used to quantitatively characterize the potential impacts of modeled interventions. Additionally, nodes and links that may affect both flood resilience and food security pathways are evaluated as potential leverage points for simultaneously affecting flood and food related objectives.