Reliable and comprehensive flood forecasting is crucial to ensure resilient cities and sustainable socio-economic development in a future faced with an unprecedented increase in atmospheric temperature and intensified precipitation. Floodwaters from the areas surrounding a city can heavily affect flood cycle behaviour across urban areas, introducing uncertainties into the forecast that are often non-negligible. However, currently the extent to which we can predict flood hazards is limited, and existing methods cannot for example deal with inter-regional dependencies (e.g. as was seen when floods affected nine different countries across Central and Eastern Europe). Presently in the UK approx. 25% of yearly flood insurance claims are from areas outside the zones forecast to be at flood risk, and annual flood damage costs are already high (approx. £1.5 billion). Also more than 20,000 houses per year continue to be built on floodplains. The need to transform flood forecasting for a range of applications and scales has already been recognised by various parties. The UK Climate Change Risk Assessment 2017 Evidence Report prioritises flooding as the greatest direct climate change related threat for UK cities now and in the future, and urges urgent action to be taken, including the development of new solutions over the next 5 years. The hydraulic software industry and consultancy firms have expressed a desire for more reliable and sophisticated flood forecasting approaches, which can also reduce the manual labour required. In addition, mathematics and engineering research communities are still searching for forecasting models that are joined-up, reliable and efficient, as well as versatile and adaptable. To address this need, 'Multi-Wavelets' technology will be employed in this fellowship with a view to transforming flood forecasting routines from a disparate set of activities into a unified automatic framework. The applicant's vision is to exploit the innate capability of Multi-Wavelets technology to reformulate flood forecasting methods by providing a smart modelling foundation for the delivery of timely and accurate flood maps, alongside statistically quantified uncertainties. This research presents a unique opportunity for the applicant, UK academia and UK industry, to establish a world leading capability in a nascent field while addressing Living With Environmental Change (LWEC) priorities for improved forecasting of environmental change. The fellowship research will stimulate the creation of new software infrastructure capable of significantly improving our flood forecasting ability across length scales and under multiple uncertainties, helping us to better design infrastructure against flood risk and to plan for the consequences.