Statistical inference methods and stable isotopes as a tool for predicting the quality of marine bathing waters

The movement of water as a scarce resource through the soil is a physical, causal-mechanical process characterized by the ability to continuously transfer a marker in space and time. A process is capable of transferring a marker if the marker, once introduced at a particular location, persists at other locations without further interaction. In this sense, stable isotopes are markers that are transferred from one location to another over time. The analysis of multiple indicators across space and time is known in statistics as longitudinal data analysis or panel data analysis. We show how some of these relatively new inductive statistical inference methods, in conjunction with known deductive nomological models, can be useful in building a predictive model for the quality of marine bathing waters in the Kvarner Bay (Adriatic Sea, Croatia).

This work was supported by the University of Rijeka under the project numbers: uniri-pr-prirod-19-24, UNIRI CLASS – A1-21-8 34.