Developing an Agent-Based Simulation Model to Forecast Flood-Induced Evacuation and Internally Displaced Persons

Abstract

Each year, natural disasters force millions of people to evacuate their homes and become internally displaced. Mass evacuations following a disaster can make it difficult for humanitarian organizations to respond properly and provide aid. To help predict the number of people who will require shelter, this study uses agent-based modelling to simulate flood-induced evacuations. We modified the Flee modelling toolkit, which was originally developed to simulate conflict-based displacement, to be used for flood-induced displacement. We adjusted the simulation parameters, updated the rule set, and changed the development approach to address the specific requirements of flood-induced displacement. We developed a test model, called DFlee, which includes new features, such as the simulation of internally displaced persons and returnees. We tested the model on a case study of a 2022 flood in Bauchi state, Nigeria, and validated the results against data from the International Organization for Migration’s Displacement Tracking Matrix. The model’s goal is to help humanitarian organizations prepare and respond more effectively to future flood-induced evacuations.