Abstract
In this study, we use a quantitative genetics model of structured populations to investigate the evolution of senescence in avariable environment. Adaptation to local environments depends on phenotypic traits whose optimal values vary with age andwith environmental conditions. We study different scenarios of environmental heterogeneity, where the environment changesabruptly, gradually, or cyclically with time and where the environment is heterogeneous in space with different populationsconnected by migration. The strength of selection decreases with age, which predicts slower adaptation of traits expressed late inthe life cycle, potentially generating stronger senescence in habitats where selection changes in space or in time. This predictionis however complicated by the fact that the genetic variance also increases with age. Using numerical calculations, we found thatthe rate of senescence is generally increased when the environment varies. In particular, migration between different habitats isa source of senescence in heterogeneous landscapes. We also show that the rate of senescence can vary transiently when thepopulation is not at equilibrium, with possible implications for experimental evolution and the study of invasive species. Ourresults highlight the need to study age-specific adaptation, as a changing environment can have a different impact on differentage classes.