Statistical analysis of ocean currents in the eastern Mediterranean

Abstract

We examine the probability density function (pdf) of current speeds at the DeepLev station located in the eastern Mediterranean Sea near the central coast of Israel. The currents cover depths from the surface to over 1.3 km and span a period from November 2016 to March 2024. We estimate the parameters of three typical distributions that are usually used to model the pdfs of currents and wind speed: the Weibull, the general extreme value, and the generalized gamma. We find that the three-parameter generalized gamma distribution best describes the pdfs of the observed current-speed series. Still, the studied current-speed time series may not be long enough to assess the exact values of the underlying pdf as some years exhibit stronger currents that affect the distribution. We also study the time series of the difference between consecutive current speeds and find that the stretched exponential pdf describes better (than the normal distribution) their statistics. The comparison of the measured current-speed pdfs to current-speed pdfs of a high-resolution (1 km) regional circulation model (ROMS) and Copernicus Mediterranean reanalysis daily mean currents ($\sim$ 4.6 km resolution) indicates discrepancies from the data. Our results may help to improve statistical models for ocean currents and the estimation of extreme current events.