Citizen science and AI reveal latitudinal flowering trends and adaptation mechanisms in perennial grasses

This article presents research that combines citizen science observations with computer vision AI to study phenological adaptation in warm-season perennial grasses across North America. The researchers developed AI to process citizen science data and discovered a consistent latitudinal trend of earlier flowering at higher latitudes. They then conducted common garden experiments with switchgrass, which revealed the opposite latitudinal flowering-time trend, allowing them to juxtapose in situ and ex situ observations to identify major genetic and environmental determinants of adaptation. The work uncovers mechanisms of adaptation that shape haplotype distribution across native habitats.