The dynamics of physico-chemical processes in the bay of Brest are studied to better understand the impact of local events on the coastal ocean. We are analyzing long-term in situ data (since 2000), high and low frequencies, from the COAST-HF (buoy Marel Iroise) and SOMLIT sites, located at the entrance to the bay of Brest. This study is divided into two main stages: the detection and characterization of extreme events, followed by a numerical simulation of these events to understand the underlying oceanic processes. The focus is on desalination events during the winter months (December, January, February and March), considering the season with the greatest number of extreme events. We show the relationship between local extreme events and variability at larger scales, considering climate indices such as the North Atlantic Oscillation (NAO). A comparison between the low frequency data of the SOMLIT program and the high frequency data of the Marel buoy is carried out, highlighting the interest of high frequency measurements for the detection of extreme events. A comparison between in situ data and two numerical simulations of different resolutions is also performed over salinity time series. This study also show an interannual variability of extreme events interesting in a context of climate change. We make the link between these extreme desalination and the nitrate level in the bay of Brest. To finally study the relationship between nutrients and chlorophyll in order to look at the correlation that it can exists between extreme events and biology in the coastal environment.