The present study defines new interpolation functions for hydrological data. These functions are applied to generate climatological maps of temperature-salinity distribution with a 25 m depth interval and a 30 km space interval (MEDTRANS data set). The data undergo rigorous initial data quality control, having passed several filtering procedures. The gridding is done on neutral density surfaces, which allows better representation of thermohaline fronts for the same gridding radius. The multi-pass Barnes optimum interpolation procedure with spatially variable size of the gridding window is used. The shape of the window accounts for the dominant along-isobath direction of water mass transport over steeply sloping topography. A local ratio of ...