Mélanie Juza has joined SOCIB's Modelling Facility to work on the assessment of numerical simulations

Mélanie Juza has joined the SOCIB Modelling Facility, since December 2011, to work on the assessment of numerical simulations performed at SOCIB using available observations, such as satellite products (Sea Level Anomaly, Sea Surface Temperature) and in-situ data (temperature and salinity from Argo floats, CTD, XBT, buoys, drifters and gliders). More precisely, this work aims to evaluate the capacity of SOCIB's Western Mediterranean Operational model (WMOP) to improve the current MFS and Mercator forecast simulations at submesoscale.

Mélanie Juza obtained a PhD from the University of Grenoble, France. This work was done at LEGI (Laboratoire des Ecoulements Géophysiques et Industriels) in the MEOM team (MultiscalE Ocean Modelling). This study aimed to develope statistical approaches to systematically and quantitatively assess 50-year global ocean simulations against altimetric and hydrographic observations. Melanie and her collaborators quantified how the resolution of their ocean models (2º, 1º, 1/2º and 1/4º) affects the realism of their solutions with respect to both observational datasets. They quantified how broadening the range of resolved space scales significantly improve the representation of the mean circulation, the thermohaline structure, the seasonal cycle of mixed layers, as well as the sea-surface height variability at most space and time scales, especially at the interannual time scale. The simulations were also used to evaluate the accuracy of observational systems (like Argo): they show that the spatio-temporal dispersion of the Argo floats induces overestimations of the mixed layer depths and heat contents, and the geographical restrictions of the actual Argo array induces biases in amplitude on the seasonal and interannual variabilities of the global ocean heat content.