Spatio-temporal changes of thermohaline properties and nutrients in the Mediterranean Sea

Abstract

Aim: To analyze long-term mean monthly values for sea surface height (SSH) and nutrient concentrations (ammonia,
nitrate, phosphate, and chlorophyll) in a part of the Mediterranean Sea, with a focus on the influence of spatial-
temporal SSH patterns on nutrient concentrations.
Methods: We downloaded monthly mean reanalysis data covering the period from 1998–2020 from the European
Union Copernicus Marine Service. We applied a neural gas machine learning method to cluster SSH data and used
MATLAB for data processing, anomaly calculations, spatial mapping, and exploring relationships with biological parameters.
Results: The neural gas method identified distinct interannual signals in SSH anomaly data, linking them to changes
in surface and subsurface circulations in the Ionian Sea. Classification of SSH allowed for the association of SSH
anomalies with varying nutrient concentrations. The analysis highlighted the role of the Adriatic-Ionian Bimodal
Oscillating System in regulating SSH and nutrient patterns.
Conclusions: The neural gas method was established as a robust classification tool for detecting low-variability signals
in the spatio-temporal dynamics of SSH, enabling connections with nutrient concentrations. This underscores the
importance of employing nonlinear statistical methods to understand the relationship between biological and physical
parameters in the Mediterranean Sea.