Understanding how changes in community composition and function are driven by environmental change remains a fundamental challenge—to which trait-based approaches offer a mechanistic perspective. To examine links between functional and compositional changes we introduce a metric—the “principal trait”—relating the mean traits of a community to principal components (PC) of community composition. We demonstrate the benefits of this approach by analyzing nearly five decades of phytoplankton monitoring data from Lake Constance. Based on principal traits, PC-scores, and PC-rotations we identify the transition of a predation susceptible winter-spring to a defended summer community and distinguish patterns of compositional and functional change during eutrophic and oligotrophic conditions. Remarkably, this approach uncovered the independent emergence of a strong tradeoff between the resource acquisition traits phosphate and light affinity at two temporal scales: once in the response of the predation susceptible winter-spring community to seasonal changes in light and phosphorus availability, and once in the response of the defended summer community to long-term changes in the lake’s nutrient status. The analysis identified community members and mechanisms involved in functional changes, demonstrating how the interplay between multiple traits determined the responses of phytoplankton community composition and function to environmental change.