“Were you here for Alicia?” This is practically a greeting for people who lived in Houston in the ’80s. Now those memories can help predict natural disasters, like hurricanes, made worse by climate change. Old photos of flooded streets and buildings fill in the datasets we rely on for forecasting, like from satellite imagery. By fusing these two kinds of information together, we learn more about past disasters to prepare for the future. To accelerate this data fusion, we are designing open-source tools that meld community stories with pictures and then link them back to models of disaster events to train artificial intelligence to improve predictions. Our approach combines human experience with computer precision to deliver a faster, more advanced natural disaster warning system.