Environmental drivers, most notably temperature, affect the biology of phyllosphere microorganisms but also induce changes in their population dynamics, even in their evolutionary trajectories. The impact of climate on foliar plant disease epidemics is usually considered in forecasting models to inform management strategies. Such models focus on averages of environmental drivers but disregard both individual variation within populations and the scale and extent of biologically relevant environmental changes. These simplifications are glossing over substantial levels of individual variation that may have important consequences on the capacity of a population to adapt to environmental changes, and thus on the dynamics of epidemics in a fluctuating or changing climate. To examine the range of validity and consequences of these simplifying assumptions, I investigated how individual variation and environmental heterogeneity jointly affect fitness, phenotypic composition and resilience of populations of a foliar pathogen (Zymoseptoria tritici) inhabiting wheat canopies. Three complementary ways of exploration were adopted in this case study. First, an in vitro high-throughput phenotyping framework was developed, validated, and used to characterise the diversity in patterns of thermal responses existing across Z. tritici populations that were sampled over contrasted scales (spatial and seasonal variation of temperature). Second, the spatio-temporal thermal variations encountered in a wheat canopy, considered as a habitat exerting fluctuating selective pressures on these differential thermal sensitivities of individuals, were investigated in depth. Third, the way selection of “thermotypes” (functional groups of individuals displaying a similar thermal sensitivity) occurs and drives dynamics of Z. tritici populations was examined. To this end, both empirical (in vitro, in planta and in natura) and theoretical (in silico) competition experiments were conducted under increasingly complex selective environments. This research work demonstrates that glossing over the natural extent of individual phenotypic diversity in a phyllosphere microbial population and over the heterogeneity of selective pressures – from phyllo- to mesoclimate – leads to underestimate the resilience of this population, and thus its adaptive potential to environmental variations. In doing so, the results of this thesis, at the interface between epidemiology, micrometeorology, and ecology, improve our understanding of how important is individual variation to population dynamics and how environmental heterogeneity allows to maintain population diversity. Finally, this thesis provides insight into how large-scale patterns and local population processes are interlinked and display a “two-tier” adaptive dynamics.