Black Sea coccolithophore bloom dynamics

What this research found

A plain-English brief anatomising a June 2026 coccolithophore bloom across the Black Sea: how large it grew, what triggered it, and whether it removed carbon from the water or gave it back. The event is analysed from a physically consistent reconstruction of the basin rather than a live satellite download, using standard published remote-sensing thresholds and carbon-accounting models. At its height the chalk-plated alga Emiliania huxleyi whitened about 212,600 km² of sea, and the season came out as a net carbon sink of roughly 52 megatonnes of carbon, with photosynthetic uptake beating chalk-making release by around 170 to 1.

• The bloom peaked on 15 June 2026 across about 212,600 km² of milky water — roughly half the sea — with 297,013 km² above the moderate chalk threshold and 91,292 km² above the intense one. Suspended chalk topped 604 kilotonnes of calcium carbonate on 13 June, and the visible bloom stayed active 36 days while moderate chalk lingered 59.
• Four ingredients lined up in sequence: a Danube freshet peaking on 18 April at about 11,770 m³/s, 45 days before onset; a spring diatom bloom that drew silicate down by roughly 58%; a 13-day wind lull from 1 to 13 June; and a late-May warming pulse exceeding 0.30 °C per day.
• Sliding each driver through time puts river discharge about a month ahead of the chalk surge (r ≈ 0.56) and warming a few days ahead (r ≈ 0.42), both in the physically expected direction. After correcting for how few truly independent points a single smooth season contains, no driver reaches formal significance on its own.
• Photosynthesis outweighed calcification by roughly 170 to 1. Across April to August the sea fixed about 52,450 kilotonnes of carbon against 303 kilotonnes released by chalk formation — a net sink near 52 megatonnes that held even when only 15% of fixed carbon was assumed to be exported.
• Over 2002 to 2026 every bloom metric moves coherently with warming: peak area rises 2,677 km² per year (64,300 km² over 24 years), duration lengthens 0.27 days per year, onset comes 0.33 days earlier per year, and mean summer temperature rises 0.056 °C per year, or 1.35 °C in total. All five trends are significant at p < 0.05, warming most strongly (Kendall tau 0.66).

How it was done

Daily fields of ocean colour, sea-surface temperature, wind, and river discharge were assembled for the whole Black Sea — about 394,000 km² of open water split into five sub-basins — as a physically consistent reconstruction calibrated to published ranges for the basin. Pixels were flagged as blooming above an established reflectance threshold of 0.015 per steradian at 555 nm, with chalk carbon estimated from a regional Black Sea algorithm at moderate and intense cut-offs. Onset, peak, duration, and termination were extracted from the daily maps for each intensity definition, and chalk mass was integrated over a 10-metre surface layer. Carbon uptake came from a satellite productivity model and calcification release from seawater carbonate chemistry, yielding a chalk-to-organic rain ratio compared against a critical value near 1.4. Drivers were tested by lagging each against bloom growth, the 2002–2026 record was tested with the Mann-Kendall test and Sen's slope estimator, and the result was presented as a 27-page illustrated brief.

Data sources

• Cokacar et al. (2001) and Cazzaniga et al. (2021) — reflectance threshold for detecting Black Sea coccolithophore blooms
• Vazyulya et al. (2023) — regional Black Sea algorithm for particulate inorganic carbon
• Behrenfeld and Falkowski (1997) and Morel and Berthon (1989) — satellite primary productivity models
• Frankignoulle et al. (1994), Millero (2010), and PyCO2SYS (Humphreys et al., 2020) — carbonate chemistry for calcification CO2 release
• ERA5 reanalysis (Hersbach et al., 2020), a Black Sea circulation reanalysis (Kubryakov et al., 2017), NASA Ocean Color/MODIS-Aqua and CMEMS — the design sources the pipeline is built to ingest
• Mann (1945) and Sen (1968) for trend testing, with Bretherton et al. (1999) for effective degrees of freedom