Dynamic risk modelling of a coupled human-drought system under multi-drought conditions

Drought risk is modified through hazard, vulnerability and exposure, and exacerbated by management shortcomings. A quantitative understanding of the combined effects of these drivers is required to effectively lower risk. Yet, knowledge about the dynamics and effects of risk drivers over time and space and the human-natural feedbacks that steer them is largely lacking. In this study, we propose using GEB, the first large-scale coupled-agent based hydrological model that simulates all individual farmers at field scale, to systemically quantify the relations between dynamic hazard, vulnerability, exposure, management and drought risk over a multi-drought period in the Bhima basin, India. First, we parametrized the coupled hydrological model with meteorological and hydrological data to capture hydrological drought  conditions of different paired drought events. Next, we develop the agent based model part to simulate the drought management behavior of two million farmers and how they respond to drought events. To simulate this behavior, we applied the protection motivation theory, supplemented by theory of planned behavior, to describe farmer agent behavior. The parameters of these theories were parametrized with survey data of Indian farmers fitted to the statistical distribution of the two million Bhima basin farmers. To study the dynamic attribution of the three risk drivers, a Global Sensitivity Analysis of all factors was performed at consecutive time intervals, showing the interaction of drivers before and after each drought event, as well as between the two events. The results are expected to further the understanding of drought risk dynamics and what disaster risk reduction measures can optimally reduce impacts in the long term.