A New Global Ocean Climatology.

Global Ocean climatologies are fundamental for our understanding of climate variability and trends, essential for the initialization and validation of numerical models. The thesis aims to compute a new global ocean monthly climatology of basic physical climate state variables such as temperature, salinity, density and dissolved oxygen from in-situ based historical datasets collected in the World Ocean Database 2018. The novelty of these climatologies stems from the implementation of a new quality control procedure, called "Nonlinear Quality Control" (NQC) thereafter. NQC is applied to the database that is used to compute the climatology and the improvements in the analysis discussed. The climatologies presented in the thesis are processed by a statistical interpolation tool, the Data Interpolating Variational Analysis (DIVA) that is applied to the global domain for the first time. Two different versions of temperature and salinity climatologies are estimated based on the different temporal coverage of the data: a long-term average (1900 to 2017) using multiple platforms, and a shorter time estimate (2003 to 2017) using data from ocean drifting platforms such as profiling floats. Sensitivity experiments are carried out to choose the key parameters of DIVA. The computed climatologies show consistency with well-known reference climatologies such as World Ocean Atlas 2018 and World Argo Global Hydrographic Climatology.