Analogues-based hydrodynamic modelling for local scales: a case study in Santander bay, Spain

Abstract

The analogues method, rooted in k-Nearest Neighbours (kNN) regression, has been successfully applied in meteorology and extended to oceanography for variables like waves and sea lev- el. This study applies it to modelling currents at a local scale in Santander Bay (Spain). By using clustering to select representative analogues, the method provides accurate predictions with a lower computational cost than traditional numerical modelling. Results indicate that the analogues method offers a robust, interpretable, and resource-efficient alternative to numerical modelling, with potential for coastal management applications.