The timing of tree flushing follows strong phenotypic and genetic clines across environmental gradients. It may be seen as an adaptive response to abiotic (escape of spring frost and maximizing growing season length) and biotic (escape of pest and disease) hazards. However, few studies have investigated jointly both types of hazards. We assessed exposure to both abiotic (spring frost) and biotic (powdery mildew) hazards within and between sessile oak populations along elevation gradients, during the flushing period in several years. For each population and phenological phenotype (early- vs. late-flushing trees), we estimated safety margins, defined as the time period separating budburst from the hazard occurrence (spore emission, spring frost). We observed that powdery mildew phenology (initiation of spore release in spring) was less responsive to the elevation gradient than oak phenology (budburst) and that it was not correlated with tree phenology within populations. The spring frost and disease safety margins varied considerably between oak populations as a function of elevation and within populations in relation to tree phenological phenotype. For both hazards, safety margins decreased significantly with increasing elevation. The safety margin for spring frost was mostly positive (i.e. escape), whereas the safety margin for powdery mildew was mostly negative (i.e. exposure), leading to infection. The abiotic and biotic hazards interact in opposite directions with phenology, especially at low elevations (< 500m) where early-flushing enabled trees to escape disease while late-flushing provided a higher safety margin against late frost. Synthesis. The observed patterns suggest that oak populations are better adapted to escape spring frost than pathogen exposure all along the elevation gradient. The combination of the biotic and abiotic selective pressures may have contributed to the maintenance of phenological diversity within low elevation tree populations. As tree and pathogen respond differently to environmental cues, climate change is likely to affect the phenological (a)synchrony between host and parasite, both within and between populations.