Lead-lag statistical analysis of simultaneous high frequency oceanographic parameters in moored coastal systems

High-frequency sampling at fixed positions in oceanography are installed all over the World. These provide time series of different oceanographic parameters over large range of scales and can help obtain informations on the complex coupling existing between physical, biogeochemical and biological parameters.

Here we explore the lead-lag information existing between two quantities: this is done by extracting the dissymmetry in the cross-correlation, corresponding to the statistical lead or lag of one series with respect to the other one (it is not necessarily a causality information). This analysis is done for all available parameters, two by two, giving way to generate a network of lead-lag influences.

As example this new approach is applied to the MAREL buoy system installed in Boulogne-sur-mer (France) operated by Ifremer (https://www.seanoe.org/data/00286/39754/). It is a moored buoy equipped with physico-chemical and biological measuring devices working in continuous and autonomous conditions with measurement every 20 minutes. We consider here the measurements at high frequency of air temperature, sea temperature, salinity, dissolved oxygen, fluorescence and turbidity for all year from 2005 to 2013. The new method is applied to the whole data set and also to data every year, in order to see a time evolution of the lead-lag network of relations between all studies parameters.