An objective reconstruction of the Mediterranean sea carbonate system

Abstract

An objective estimation of the current distribution of carbonate system variables for the Mediterranean Sea is proposed using empirical relationships derived from ship-based observations and combined with monthly climatological fields of hydrographic parameters. The high quality data of METEOR84/3 cruise were used to fit multiple linear regression models of Dissolved Inorganic Carbon (DIC) and Total Alkalinity (TA) from other hydrochemical parameters. These algorithms provided a robust estimation of DIC and TA, with corresponding Root Mean Squared Errors of 7.66 and 5.09µmol/kg, by accounting only for potential temperature, salinity, pressure, and nitrate concentration. After the application of the identified regression models to a set of publicly available climatological fields, an objective assessment of the reconstructed carbonate system monthly distributions was derived and compared against different ship-based surveys. Results showed that the Mediterranean Sea interior was well reproduced with errors <14µmol/kg, whereas the near surface layers still exhibited large uncertainties. The lower degree of confidence of this approach at the surface does not allow the direct application for studying anthropogenic CO2 trends, but some qualitative considerations were drawn from the comparison between the estimated inorganic carbon system and the available observational datasets. Most importantly, the present work showed that the estimated inventories are able to capture the linkages with the physical oceanic features of the system and we propose this method as an inexpensive solution to support the design of monitoring activities in the Mediterranean Sea, which is still poorly constrained by direct observations.