Drivers of the autumn phytoplankton development in the open Black Sea

The dynamics of the autumn development of phytoplankton in the Black Sea were investigated using satellite derived chlorophyll-a concentration (Chl), which was estimated for two regions in the deep sea over a 20-year period. We analysed 8-day composite Chl images along with changes in: (i) nutrient concentration obtained from in-situ measurements, (ii) sea surface temperature (SST), (iii) photosynthetically available radiation (PAR) obtained from satellite imagery, (iv) wind speed from the re-analysis of meteodata and (v) the depth of the upper mixed layer (UML) calculated from a 3D numerical model of the Black Sea. The peak in Chl was identified most frequently in the first half of November. A positive correlation between the duration of strong wind events and phytoplankton development was revealed, which was associated with the deepening of the UML, and replenishment of the photic zone with nutrients. The impact on phytoplankton was significant when the cumulative duration of strong wind (>8 m s(-1)) exceeded 60 h over the preceding 8 days. In such cases, the frequency of the Chl peaks increased up to 30-50% with an average of 20%. Strong wind was shown to determine the timing of the autumn bloom, but not its strength. From a positive relationship between the maximum Chl and nitrate concentration we found instead that the intensity of the autumn bloom was mainly defined by nitrate replenishment in the photic zone. On average, the timing of the seasonal maximum of Chl in the first half of November coincided with the deepening of the UML to the bottom of the seasonal thermocline (ca 25 m). Elution of nitrate from deeper layers, where its concentration is substantially higher, mitigated the nutrient limitation of phytoplankton growth. At the same time, a sharp decrease in PAR after mid-November resulted in the limitation of light for phytoplankton growth. Inter-annual variations of Chl in spring and autumn were shown not to be correlated. For example, the basin-wide autumn blooms were observed in some years when the spring blooms were absent. As the bloom cannot be based on regenerated nitrate, the amount of 'new' nitrate in the photic zone should have a positive trend in autumn. However, the sources and mechanisms of the basin-wide increase of nitrate concentration in the upper layer in autumn are not clear. (C) 2017 Elsevier B.V. All rights reserved.