Spatio-temporal dynamics of biogeochemical processes and air-sea CO2 fluxes in the Western English Channel based on two years of FerryBox deployment

From January 2011 to January 2013, a FerryBox system was installed on a Voluntary Observing Ship (VOS), which crossed the Western English Channel (WEC) between Roscoff (France) and Plymouth (UK) up to 3 times a day. The FerryBox continuously measured sea surface temperature (SST), sea surface salinity (SSS), dissolved oxygen (DO), fluorescence and partial pressure of CO2 (from April 2012) along the ferry track. Sensors were calibrated based on 714 bimonthly surface samplings with precisions of 0.016 for SSS, 3.3 ?M for DO, 0.40 ?g L- 1 for Chlorophyll-a (Chl-a) (based on fluorescence measurements) and 5.2 ?atm for pCO2. Over the 2 years of deployment (900 crossings), we reported 9% of data lost due to technical issues and quality checked data was obtained to allow investigation of the dynamics of biogeochemical processes related to air-sea CO2 fluxes in the WEC. Based on this unprecedented high-frequency dataset, the physical structure of the WEC was assessed using SST anomalies and the presence of a thermal front was observed around the latitude 49.5°N, which divided the WEC in two main provinces: the seasonally stratified northern WEC (nWEC) and the all-year well-mixed southern WEC (sWEC). These hydrographical properties strongly influenced the spatial and inter-annual distributions of phytoplankton blooms, which were mainly limited by nutrients and light availability in the nWEC and the sWEC, respectively. Air-sea CO2 fluxes were also highly related to hydrographical properties of the WEC between late April and early September 2012, with the sWEC a weak source of CO2 to the atmosphere of 0.9 mmol m- 2 d- 1, whereas the nWEC acted as a sink for atmospheric CO2 of 6.9 mmol m- 2 d- 1. The study of short time-scale dynamics of air-sea CO2 fluxes revealed that an intense and short (less than 10 days) summer bloom in the nWEC contributed to 29% of the CO2 sink during the productive period, highlighting the necessity for high frequency observations in coastal ecosystems. During the same period in the sWEC, the tidal cycle was the main driver of air-sea CO2 fluxes with a mean difference in pCO2 values between spring and neap tides of + 50 ?atm. An extraction of day/night data at 49.90°N showed that the mean day-night differences accounted for 16% of the mean CO2 sink during the 5 months of the study period implying that the diel biological cycle was also significant for air-sea CO2 flux computations. The 2 years of deployment of our FerryBox allowed an excellent survey of the variability of biogeochemical parameters from inter-annual to diurnal time scales and provided new insights into the dynamics of air-sea CO2 fluxes in the contrasted ecosystems of the WEC.