Long-term studies reveal major environmental factors driving zooplankton dynamics and periodicities in the Black Sea coastal zooplankton.

Background: The development and management of shelf-sea ecosystems require a holistic understanding of the factors that influence the zooplankton structure and ecosystem functions. The Black Sea is an example of such areas influenced by eutrophication, overfishing, climate variability, invasions of the ctenophores followed by . Thus, there is a set of principal factors which may influence and explain periodicities in the Black Sea ecosystem.

Methods: We analysed a total of 918 samples taken from 1991 to 2017 with intervals of 10 days. Taxa were identified to species, their abundance and biomass were calculated. We tested 12 environmental factors, which may explain zooplankton distribution: temperature, productivity-linked factors (surface chlorophyll as a proxi), wind, turbidity, lowest winter temperature, and concentration of the ctenophore . We used canonical correspondence analyses to find the dominant environmental factors and further regression analyses to retrieve dependences of plankton biomass on the major factors. Periodicities were assessed with the use of the Continuous wavelet transform and tested with use of One-way ANOSIM and PERMANOVA. The distances between ecosystem states in different years were assessed using non-metric multidimensional scaling.

Results: Currently, temperature and productivity are the major environmental factors driving zooplankton dynamics. Not long ago, before 1999, abundance of was one of the major factors explaining the zooplankton variance. Spectral analysis of species abundances revealed a 4-year transitional period in 1999-2002 (not reported before) when ecosystem adapted to a new invader . Statistically robust 2- and 3-year periodicities were retrieved for most plankton taxa and some benthic larvae. We found robust correlations between temperature and surface chlorophyll concentration on one side and plankton abundances and biomass on the other, and retrieved multivariate regressions, which may have a prognostic value.