An inverse method to derive surface fluxes from the closure of oceanic heat and water budgets: Application to the north-western Mediterranean Sea

The large amount of data collected during DeWEX, MOOSE, and HyMeX campaigns in the north-western Mediterranean in 2012-2013 allowed to implement an inverse method to solve the difficult problem of heat and water budget closure. The inverse method is based on the simulation of the observed heat and water budgets, strongly constrained by observations collected during the campaigns and on the deduction of adjusted surface fluxes. The inverse method uses a genetic algorithm that generates 50,000 simulations of a single-column model and optimizes some adjustable coefficients introduced in the surface fluxes. Finally, the single-column model forced by the adjusted fluxes during 1 year and over a test area of about 300 x 300 km(2) simulates the daily mean satellite bulk SST with an accuracy/uncertainty of 0.0110.072 degrees C, as well as daily mean SSS and residual buoyancy series deduced from wintertime analyses with an accuracy of 0.0110.008 and 0.030.012 m(2) s(-2), respectively. The adjusted fluxes close the annual heat and rescaled water budgets by less than 5 W m(-2). To our knowledge, this is the first time that such a flux data set is produced. It can thus be considered as a reference for the north-western Mediterranean and be used for estimating other flux data sets, for forcing regional models and for process studies. Compared with the adjusted fluxes, some operational numerical weather prediction models (ARPEGE, NCEP, ERA-INTERIM, ECMWF, and AROME), often used to force oceanic models, were evaluated: they are unable to retrieve the mean annual patterns and values.