As the impacts of climate change become increasingly evident, the United Nations has launched the Early Warnings for All initiative, which aims to ensure universal protection from hazardous climatic and related environmental events through life-saving early warning systems, anticipatory action, and resilience-building efforts. In this context, wildfire simulation systems support the understanding and control of wildfire behaviour, enabling both prompt emergency response initiatives and the development of effective prevention strategies. This paper reports on a collaboration between academic researchers and experts from the CIMA Research Foundation, the primary provider of such systems for Italy’s National Civil Protection Department, focusing on the architectural evolution of Propagator, their wildfire simulation system. We discuss the system’s limitations and re-engineering needs, describe the elicited evolution requirements, and present the newly proposed system architecture. Additionally, we provide insights and lessons learned for researchers and practitioners working in the context of natural disaster monitoring and forecasting. The re-engineering process is currently underway, being implemented according to the proposed architecture to enhance system capabilities and performance.