The last two decades have witnessed unprecedented changes in beta diversity, the spatial variation in species composition, from local to global scales. However, analytical challenges have hampered empirical ecologists from quantifying the extinction and colonization processes behind these changing beta diversity patterns. In this presentation, we demonstrate a novel numerical method to additively partition the temporal changes in beta diversity into components that reflect local extinctions and colonizations. By applying this method to empirical datasets, we revealed spatiotemporal community dynamics that were otherwise undetectable. In mature forests, we found that local extinctions resulted in tree communities becoming more spatially heterogeneous, while colonizations simultaneously caused them to homogenize. Applications of the method to a Swedish fish community dataset revealed biotic homogenization. Species-level partitioning further showed that the homogenization was largely caused by the increased population sizes of brown trout (Salmo trutta) in sites where it was already present but in low abundances, resulting in a more spatially-uniform abundance distribution. Partitioning the dynamic processes that underlie beta diversity can provide more mechanistic insights into the spatiotemporal organization of biodiversity.