An Adaptive Tale: Optimizing Resilient Development

The Oxford-Cambridge Arc (OxCam), in the south of England (see Figure 1), is home to 3.7 million people and supports 2 million jobs. It has been identified as an important area of economic growth for the UK, with up to one million new homes expected over the next 30 years, alongside major infrastructure investment. However, there is great uncertainty in the effect of climate change and the shape and scale of development, creating a challenge in understanding how best to investment in flood resilience. The OxCam Flood Risk Investment Study sought to understand the 'optimum' level and timing of investment in the flood protection elements of flood resilience across three major catchments, but crucially doing so in a way that is mindful of future uncertainty and adaptive to different future scenarios. The OxCam Flood Risk Investment Study project used Flood Platform, Jacobs' data storage, management, simulation, optimization, and analysis platform, to automate the process of building 700 flood models, running 45,000 simulations, and completing a comprehensive impact analysis (Figure 2). Using an innovative 'simulation library' of outputs and novel code to interrogate it (Figure 3), 27 possible future scenarios were represented (Figure 4), spanning a range of possible climate change and development scenarios (Figure 5), along with billions of combinations of flood risk interventions (Figure 6). The project included developing a first-of-its-kind 'adaptive optimization', which explored a highly complex decision tree of billions of possible investment paths to identify investment decisions that are robust across futures and identify where and when to invest. The project served an important dual purpose, firstly providing evidence to support the Environment Agency's strategic engagement with stakeholders, and secondly as a pilot study to learn more about bringing the concepts of adaptive planning and economic optimization together at a regional scale. The project also produced a deep well of outputs and addressed the challenge of distilling them into meaningful key messages that provide evidence to support a strategic approach to flood resilience investment in the OxCam. The outputs are supporting the Environment Agency's engagement with a range of stakeholders — from national and local government to housing developers. For example, the project team found that without investment, the number of properties at risk could double between now and 2050 without new investment in flood resilience (Figure 7). This projected outcome is mainly due to flood zones expanding and encompassing existing housing. An average of between £140 and £206 million annual investment (with a benefit-cost ratio of around 5:1) may be needed to manage flood risk to existing properties and future development in the OxCam. The broad range shows how dependent investment is on future changes in risk, particularly how sensitive the area is to climate change. As part of delivering the work, significant advances have been made in the technology platform (Flood Platform), including automated model development, simulation parallelization, post-processing routines, and optimization integration which is already being used and further enhanced for the benefit of other projects. We have learned a great deal about the complexity of representing broad portfolios of interventions (including carbon) and a wide range of impacts in a large-scale analysis (Figure 8), in particular the challenge of bringing together optimization and adaptation by balancing investment needs across multiple futures. This learning is currently being applied by Jacobs to a project being delivered to the UK government (completion in 2025) which is developing long-term investment scenarios, at a national-scale, to provide the evidence base the Government is seeking to set investment levels (budgets) for flood and coastal erosion risk management. This presentation will outline the project background, technical approach and findings, and will explore the transferrable lessons learned about how adaptive planning and technology and data-driven optimization can be brought together to understand future risk and investment.