From Causes to Consequences: A Novel Aspect for Evaluating Anthropogenic Drought and Water Resource Management Policies

To explore the interaction and causality between management decision-making and 
the evolution of anthropogenic drought, we proposed a comprehensive decision
evaluating framework and analytical method. This framework consists of several key 
components: current status description, actions from a virtual agent, the consequences 
of these actions, policy objective design, and the identification of an optimal datum 
policy. In the context of anthropogenic drought, the modified water accounting and 
vulnerability evaluation plus (modified WAVE+) is employed to simulate socio
hydrological interactions, providing a detailed description of the current status. The 
consequences of actions are determined using the Monte-Carlo method, serving as 
conditional probabilities for anthropogenic drought occurrence. The proposed optimal 
objectives, which focus on maximizing supply capacity and minimizing water 
shortages, are achieved using a Q-learning mixed strategy integrated with the modified 
WAVE+. To further analyze the dynamics of anthropogenic drought, we decomposed 
the sources of change in conditional probability into two key factors: anthropogenic 
pressure and vis major. This decoupling of socio-hydrological information allows for a 
more nuanced causality analysis. By comparing the optimal datum policy with the 
quantified evaluations of anthropogenic pressure and vis major, we introduced a 
concept to determine whether drought dynamics are resistible or irresistible and 
whether there is potential for improvement in decision-making. Applying this 
evaluation framework and analytical method to the Shihmen water supply system in 
Northern Taiwan, we not only demonstrated how anthropogenic drought co-evolves 
with water resource management policies but also conducted an irresistible and 
causality analysis of historical drought events.