Same Temperature, Different Microbes: 13-Year Baltic Sea Study Reveals Hidden Seasonal Hysteresis

For thirteen years, researchers at the Linnaeus Microbial Observatory in the Baltic Proper pulled seawater samples from the same spot, essentially every two weeks, through freezing winters, cyanobacterial summers, and everything in between. What they found challenges one of the most basic assumptions in ocean ecology: that if you know the temperature, you know roughly which microbes should be there. As it turns out, the Baltic Sea’s bacterial communities at 12 degrees Celsius in spring look strikingly different from those at 12 degrees in autumn, even though the thermometer reads the same. The way from winter to summer, it seems, is not the same as the way from summer to winter.

The study, published in the journal Ocean Microbiology, analyzed 726 samples of 16S rRNA gene amplicons collected between 2011 and 2023, detecting more than 67,000 distinct prokaryotic amplicon sequence variants, or ASVs. The team, led by Carina Bunse of Linnaeus University and the University of Gothenburg together with Jarone Pinhassi and colleagues, split each sample into two size fractions: particle-associated prokaryotes larger than 3 micrometers, and free-living cells in the 0.2 to 3 micrometer range. This dual-fraction approach, sustained over more than a decade, is rare in oceanography and allowed the researchers to ask questions that shorter studies simply cannot answer.

The sampling site itself experiences dramatic seasonal forcing. Winter water temperatures drop below 5 degrees Celsius while inorganic nutrients accumulate to concentrations of 2 to 3 micromolar nitrate plus nitrite and up to 1 micromolar phosphate. By summer, temperatures climb to 18 degrees or higher, and nutrients are stripped to near detection limits. Two major phytoplankton blooms punctuate the year, one typically in April dominated by dinoflagellates and diatoms, and a cyanobacterial bloom in July and August. Prokaryotic cell counts swing roughly six-fold, from about half a million cells per milliliter in winter to more than three million in summer.

One of the study’s clearest findings concerns biodiversity. Particle-associated communities consistently harbored higher diversity than free-living ones, with estimated and observed richness significantly elevated in the particle fraction across all seasons, peaking in winter. This makes ecological sense: marine particles, whether aggregates of living and dead plankton, fecal pellets, or condensed organic material, offer a patchwork of microscopic habitats, each with its own chemistry and resources. Free-living seawater, by contrast, selects for streamlined cells adapted to a more homogeneous environment. Yet the evenness of the communities, measured by Shannon and inverse Simpson indices, differed between fractions only in winter, suggesting that while particles host more species, the balance of abundance remains broadly similar.

The researchers went further, using a permutation-based particle-association niche index to determine which taxa genuinely prefer particles and which prefer open water. The results revealed remarkable fine-scale specialization. Among the Bacteroidia and Gammaproteobacteria, most families showed strong particle association, including Alteromonadaceae, Pseudomonadaceae, and Shewanellaceae. Among the Alphaproteobacteria, families split roughly evenly, with Rhizobiaceae and Caulobacteraceae favoring particles while Rhodospirillaceae and the famously abundant Pelagibacteraceae favored the free-living fraction. Most strikingly, some families such as Flavobacteriaceae, Burkholderiaceae, and Rhodobacteraceae contained species spread across the entire spectrum, meaning that even within a single family, closely related organisms have carved out opposite lifestyles.

The filamentous cyanobacteria provided a natural test case. Genera like Dolichospermum, Nodularia, and Aphanizomenon form the Baltic Sea’s notorious summer blooms, thriving on nitrogen limitation and residual phosphorus. In the dataset, Cyanobacteriia dominated the particle fraction in summer, sometimes exceeding 90 percent of relative abundance, while contributing at most around 30 percent to the free-living fraction, largely because their filaments are physically excluded from the smaller filter. Non-filamentous cyanobacterial families, by contrast, showed low particle-association indices, confirming that the fractionation captures real biology rather than an artifact of filtration.

Perhaps the most unexpected finding involves the rare biosphere. Most ASVs in the dataset were consistently rare, with mean relative abundances below 0.7 percent. But the abundant taxa, the so-called core species that appear in high numbers every year, turned out to be surprisingly ephemeral. A majority of the 50 most abundant ASVs in each fraction collapsed into the rare biosphere for extended stretches of the year, sometimes falling below detection entirely, before rebounding in their characteristic season. Recruitment from the rare biosphere, previously viewed as an occasional event following disturbance, appears here to be an annual routine for most dominant taxa in this strongly seasonal environment.

The dataset also captured rare microbes responding to dramatic events. Campylobacteria, formerly known as Epsilonproteobacteria, normally inhabit the oxic-anoxic interface in deeper Baltic waters. Yet in the winters of 2011/12 and 2014/15, coinciding with major North Sea inflow events that pushed saline, oxygenated water into the Baltic’s deep basins, a Sulfurimonas ASV surged to relative abundances exceeding 80 percent in surface waters in December 2011. These episodic blooms of typically deep-water specialists offer a microbial fingerprint of large-scale oceanographic disturbances, lingering for months before the taxa retreated back into rarity.

The centerpiece of the study, however, is the demonstration of direction-dependent dynamics, a phenomenon ecologists call hysteresis. Using generalized additive models that compared a single temperature-abundance relationship against models allowing separate smooths for warming and cooling phases, the researchers found that most co-occurring taxon clusters behaved differently depending on the direction of environmental change. One free-living cluster dominated by Bacteroidia reached roughly 50 percent relative abundance at 7 degrees Celsius in spring but only about 15 percent at the same temperature in autumn. Another cluster, rich in Acidimicrobiia and Actinomycetes, accounted for more than 40 percent of the community at 10 degrees in autumn but a mere 5 percent at 10 degrees in spring. The same held for day length: one cluster exceeded 60 percent abundance at 15 hours of daylight in spring but contributed less than 10 percent at 15 hours in fall.

What does this mean for how we understand the ocean’s microbial engines? Temperature undeniably shapes microbial metabolism, and day length defines the seasonal calendar at any latitude. But the study’s authors argue that neither variable directly determines community composition. Instead, the identity of the microbes at any moment reflects the community’s history, the direction of environmental change, the quantity and quality of organic and inorganic nutrients, and a web of biotic interactions with phytoplankton, grazers, and viruses. Spring waters carry fresh labile organic matter from the phytoplankton bloom; autumn waters carry a different cocktail of substrates after a summer of production and degradation. The microbes respond to this context, not just to the thermometer. Given that seasonal shifts in dominant microorganisms ripple through marine food webs and biogeochemical cycles, the finding suggests that climate-driven warming may alter ecosystems in ways that simple temperature correlations cannot predict. For a brackish inland sea already stressed by eutrophication and hypoxia, that is a sobering message, and one that only long-term, high-frequency, size-fractionated time series like this one could have delivered.

Subject of Research: Seasonal succession and direction-dependent dynamics of particle-associated and free-living prokaryotic communities in the Baltic Sea

Article Title: Comparable temperatures but different microbiomes: long-term time series exposes seasonally divergent Baltic Sea prokaryotic communities

Article References: Bunse, C., Farnelid, H., Lindehoff, E., Martínez-García, S., Fridolfsson, E., Di Leo, D., Martínez, C. P., Pontiller, B., Lindh, M. V., Sjöstedt, J., Lundin, D., Legrand, C., & Pinhassi, J. (2026). Comparable temperatures but different microbiomes: long-term time series exposes seasonally divergent Baltic Sea prokaryotic communities. Ocean Microbiology, 2(1), Article 5. https://doi.org/10.1186/s44375-026-00011-7

Image Credits: AI Generated

DOI: 10.1186/s44375-026-00011-7

Keywords: Baltic Sea, prokaryoplankton, microbiome, seasonal succession, time series, rare biosphere, hysteresis, particle-associated bacteria, free-living bacteria, cyanobacteria, 16S rRNA, marine microbial ecology