Disaster planning based on historical events is like driving forward while only looking in the rear-view mirror. To expand our field of view, we use a large ensemble of weather simulations to characterise the current risk of extreme weather events in case study locations in the Southeastern United States. We find that extreme temperature events have become more frequent between 1981 and 2021, and heavy precipitation events are also more frequent in the wettest months. Combining a historical analysis of people’s recent experience with the rate of change of extreme events, we define four quadrants that apply to groups of case studies: Sitting Ducks”, “Recent Rarity”, “Living Memory”, and “Fading Memory”. A critical storyline is that of the “Sitting ducks”: locations where we find a high rate of increase in extreme events and where the most extreme event in recent memory (1981-2021) has a low return period in today’s climate. We find that these locations have a high potential for surprise. For example, in Montgomery County, Alabama, the most extreme temperature event since 1981 has a return period of 13 years in the climate of 2021. In these places, we offer unprecedented synthetic events from the large ensemble for use in disaster preparedness simulations to help people imagine the unprecedented. Our results not only document substantial changes in the risk of extremes in the Southeastern United States but also propose a generalizable framework for using large ensembles in disaster preparedness simulations in a changing climate.