An Integrated Approach for Assessing Vulnerability and Potential Adaptation Options for a Coastal Water Supply and Demand System Subject to Climatic and Non-Climatic Changes

Abstract

Management of freshwater supply and demand systems in coastal areas faces many challenges given the high levels of uncertainty and complexity which follow from dynamic interactions and feedbacks amongst multiple climatic and non-climatic drivers such as sea level rise, changes in precipitation and river flows, and socio-economic development. Temporal and spatial variation among these driving factors further contributes to the highly complex management challenge. These issues are prevalent in coastal areas of developing countries which typically experience high rates of population growth and urbanization. To help inform management of coastal freshwater systems under conditions of high uncertainty and complexity, this thesis developed a coupled top-down and bottom-up modelling framework for a case study setting in close consultation with local stakeholders. A system dynamics (SD) model was applied as a top-down approach to assess the vulnerability of the system under climatic and non-climatic changes, and a Bayesian decision network (BDN) model was employed as a bottom-up approach to identify cost-effective adaptation options in the face of the same climatic and non-climatic changes. This decision-making framework was developed with an understanding of the strengths and weaknesses of top-down and bottom-up approaches, and SD and BDN models as well as in the light of the dynamics and uncertainties inherent in coastal freshwater supply and demand systems. A global systematic quantitative literature review found that Bayesian networks (BNs) have rarely been coupled with SD models in water resource management, and also that BNs have rarely been applied to prioritize cost-effective adaptation measures for managing water supply and water demand under climate change in developing countries and tropical regions. Equally importantly, the literature view found that only in very few instances has the performance of BN models been tested against other modelling approaches for cross-examining model types and outputs. The freshwater supply and demand system in the Da Do Basin in Hai Phong City, Vietnam was used as a case study in this thesis to develop the coupled top-down and bottom-up modelling framework. In addition to historical data collection, causal loop diagrams (CLDs) for the system were constructed during workshops with local stakeholders to better understand how interactions among climatic and non-climatic drivers affect system operation. Stakeholder consultations at these workshops were also used to identify key climatic and non-climatic drivers for inclusion in SD and BDN models of the system, and to select a short list of potential adaptation options to counteract adverse changes in these key drivers. The SD model was developed, calibrated and tested using historical data and stakeholder knowledge. SD simulations indicated that freshwater availability is sufficient to meet existing domestic, industrial and agricultural demands during the six-month dry season under representative current conditions, but that freshwater availability could collapse under some plausible future scenarios. Upstream flow decline was identified as the strongest threat to the system‘s vulnerability, with the consequent reduction in river water level and increase in salinity level severely restricting opening hours for the sluice gates which supply freshwater to the system. The BDN model was developed in close consultation with stakeholders to identify cost-effective adaptation options to counteract climatic and non-climatic changes in key drivers. The BDN model indicated that the cost-effectiveness of adaptation options differed depending on which future scenarios were considered. Building pumping stations individually, or in conjunction with increasing water prices, were identified as the most cost-effective adaptation options to counteract climatic and non-climatic changes in combination. Subsequent simulation of these options in the SD model showed that they should be effective and robust in increasing water availability and recovering system collapse during the dry season. The ultimate objective of this coupled top-down SD and bottom-up BDN modelling approach was to provide a learning tool for stakeholders to assess system vulnerability and identify appropriate adaptation options for this complex coastal freshwater supply and demand system subject to multiple threats. Subsequent applications of this approach are likely to be highly relevant for water resource management in other basins in Hai Phong City, as well as in urban estuarine settings elsewhere in the developing and developed world.