Microbial communities have a central role in many ecosystem functions such as biogeochemical cycling, filtering and climate regulation. However, while functional trait-based approaches are commonly used for at least two decades in microbiology, most microbial ecology studies are still carried out with the rRNA genes. This talk will highlight the advantages of trait-based approaches for studying microbial diversity and its relationship with ecosystem functions. For this purpose, denitrification, a microbial process involved in N-cycling, was selected as a model functional trait. Using both naturally assembled and artificially manipulated communities, we will show how the diversity of the denitrifier community relates to process rates under optimal conditions but also under broad environmental gradients. We will also show that the analysis of spatial patterns of traits can help bridging microbial community ecology and ecosystem process in terrestrial environments.