The response of plants to climate change is among the main questions in ecology and evolution. Faced with changing conditions, populations may respond by adapting, going extinct or migrating. Fine-scale environmental variation offers a unique mosaic to explore these alternatives. In this thesis, I used ecological surveys, field experiments and molecular methods to study the range of possible responses at a very local scale in the alpine dwarf willow Salix herbacea L. Since gene flow may impact the potential for adaptation and migration, I first explored whether phenological divergence driven by snowmelt patterns impacts gene flow. I found that sites with late snowmelt work as sinks of the genetic diversity, as compared to sites with early snowmelt. I also used a combined approach that looked at the selection, heritability and genomic architecture of ecologically-relevant traits, as well as genomic divergence across the snowmelt mosaic. In this way, I was able to understand which genomic regions may relate to phenological, growth and fitness traits, and which regions in the genome harbor genetic variation associated with late- and early- snowmelt sites. I found that most of the genomic divergence driven by snowmelt is novel and is localized in few regions. Also, Salix herbacea has a strong female bias. Sex bias may matter for adaptation to climate change because different sexes of many dioecious species differ in several functions that may fluctuate with changing conditions. I found that the bias is uniform across environments and is already present at seeds and seedlings. A polygenic sex determination system together with transmission distortion may be maintaining the bias. Overall, fast-evolving microhabitat-driven genomic divergence and, at the same time, genetically-based trait variation at a larger scale may play a role for the ability of S. herbacea to persist in diverse and variable conditions. / SNSF Sinergia Salix