Rapid evolution predicts demographic recovery after extreme drought

Populations that are declining as a result of climate change may need to evolve to persist. Although evolutionary rescue has been demonstrated in theory and in the laboratory, its relevance to natural populations facing climate change remains unknown. Here we link rapid evolution and population dynamics in scarlet monkeyflower, <i>Mimulus cardinalis</i>, during exceptional drought. We leverage whole-genome sequencing across 55 populations to identify climate-associated loci. Simultaneously we track demography and allele frequency change throughout the drought. We establish range-wide population decline during the drought, geographically variable rapid evolution, and variable population recovery that is predictable by standing genetic variation in, and rapid evolution at, climate-associated loci. These findings demonstrate the possibility of evolutionary rescue in the wild, showing that genetic variation at adaptive, but not neutral, loci predicts population recovery.