AbstractCities are complex systems where social, ecological, and technological processes are deeply coupled. This coupling complicates urban planning and land use development, as changing one facet of the urban fabric will likely impact the others. As cities grapple with climate change, there is a growing need to envision urban futures that not only address more frequent and intense severe weather events but also improve day-to-day livability. Here we examine climate risks as functions of the local land use with numerical models. These models leverage a wide array of data sources, from satellite imagery to tax assessments and land cover. We then present a machine-learning cellular automata approach to combine historical land use change with local coproduced urban future scenarios. The cellular automata model uses historical and ancillary data like existing road systems and natural features to develop a set of probabilistic land use change rules, which are then modified according to stakeholder priorities. The resulting land use scenarios are evaluated against historical flood hazards, showcasing how they perform against stakeholder expectations. Our work shows that coproduced scenarios, when grounded with historical and emerging data, can provide paths that increase resilience to weather hazards as well as enhancing ecosystem services provided to citizens.