Spatial and time scales of human-water feedbacks facing drought risk in megacities: a socio-hydrological approach in São Paulo

• Ano de defesa: 2023
• Autor(a) principal: Souza, Felipe Augusto Arguello de
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: aprendizado de máquina; drought risk management; game theory; gestão de risco de secas; machine learning model; políticas consumo de água; rios transfronteiriços; sócio hidrologia; sociohydrology; teoria dos jogos; transboundary river; water-saving policy evaluation
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/18/18138/tde-20102023-095009/

The Sao Paulo Metropolitan Region hosts more than 20 million inhabitants, who are supplied by several water sources wherein the Cantareira System, one of the largest water supply reservoirs in the world, is capable to deliver more than 30 m3/s of water. The reduced rainfall observed in 2013 and 2014 triggered the most severe water crisis in its recent history and raised several questions about the hazard intensity, the decision makers\' ability to handle with such events and the community role in reducing the water consumption when required. This thesis aims at investigating the interactions over time between water availability and human action in Sao Paulo to assess whether the crisis resulted from this drought event could have been avoided. First, historic records and key aspects related to drought risk management, such as hazard intensity, preparedness, exposure, vulnerability, disaster response and mitigation alternatives, are used to compare the 2013-2015 water crisis to the 1985-1986 drought, observed long ago, and contrast the evolution of those aspects so far. Therefore, the evidence suggests the greater hazard intensity and people\'s exposure to drought, in combination to both late water-saving policies\' implementation and the dependency of several service areas on a single reservoir, culminated in the disaster experienced in 2013-2015. Second, a machine learning model is employed to address the community response to water saving policies and to outline a hypothetical storyline considering the early implementation of such policies. The model outputs suggest stronger significance on the contingency tariff rather than the bonus tariff. Therefore, the penalty tariff would be required two years in advance to promote water conservation of local users and prevent the Cantareira System from reaching the dead pool level. Third, the water allocation from the Paraiba do Sul River Basin is evaluated upon the scenarios of transboundary interactions between upstream – Sao Paulo State – and downstream – Rio de Janeiro State – within the context of the 2013-2015 water crisis. Those scenarios address the impacts of i) water allocations from the Paraiba do Sul River Basin to the Cantareira System through a tunnel concluded after the drought, and ii) updated operation rules of the Paraiba do Sul River Basins as a response to drought impacts within its basin observed in 2013-2015 as well. The three scenarios show that the impacts on water availability and hydropower production does not satisfy the two players at the same time and, therefore, put then in a game where hydroelectricity would be reduced for both states at any scenario, while the water transfers to Sao Paulo would be equivalent to the supply of 1 million people downstream in the three years. The three working fronts explore the two-way feedbacks between water availability and humans\' behavior to better understand the coevolution of this coupled system in Sao Paulo and outline hypothetical storylines to improve the responses of futures drought events.