Interpolating single-beam data for sea bottom GIS modelling

The “Istituto Idrografico della Marina Militare” (IIMM) secures the Italian hydrographic service through the execution of bathymetric surveys, the production of nautical charts and publications, and the dissemination of nautical information, aimed at the safety of navigation and of human life at sea. The datasets of depth points acquired with bathymetric surveys are useful to model sea-bottom starting from the interpolation methods available in Geographic Information System (GIS) software. In this paper, a single-beam dataset from IIMM is used with the aim to compare different interpolation methods for sea bottom GIS modelling. The study area is the sea close to the east coast of Isola del Giglio in Tuscan Archipelago (Italy). The following nine different interpolation methods are selected and applied using ArcGIS software, version 10.3.1: Inverse Distance Weighting (IDW), Local Polynomial Interpolations of different orders (from the first to the fifth order), Ordinary Kriging, with three different Variogram models (Gaussian, circular, exponential). The result accuracy is tested via cross-validation leave-one-out, so statistical values (minimum, maximum, mean and root mean square error) are calculated for each interpolation method, taking into account the residual given for each sampling point between measured and interpolated value. Finally, a 3D model is created from the best interpolating algorithm. The results remark the role of the cross validation as preliminary way to select the most preforming interpolation method that is difficult to identify in other way.