Marine phytoplankton functional types exhibit diverse responses to thermal change

Phytoplankton are the dominant contributors to marine primary productivity (~45 Gt C fixed annually) and are sensitive to climate-driven warming. Despite their phylogenetic and ecological diversity, many studies assume universal thermal growth sensitivities across taxa, potentially misrepresenting community structure and productivity under present and future climates. Species-specific thermal traits vary, implying that warming could drive differential changes in growth rates, shifts in distributions, and future communities with no modern analog. This study asks how major phytoplankton functional types (PFTs)—diatoms, dinoflagellates, coccolithophores, and cyanobacteria—differ in thermal response, and how temperature change alone could reshape their growth and geographic ranges. The work characterizes PFT-specific temperature-growth relationships and projects impacts using historical and end-of-century sea surface temperature scenarios.

Prior work established that phytoplankton thermal traits vary among species and across environments, but models and remote-sensing applications often adopt a single exponential temperature dependence (e.g., Eppley curve, Q10≈1.88) for all phytoplankton. More recent analyses suggest Eppley’s sensitivity may overestimate average thermal responses and that trait differences can influence diversity, productivity, and biogeochemical outcomes. Studies also highlight potential poleward shifts, changes in dominance under warming, and the role of multiple stressors (nutrients, light, acidification). However, explicit taxonomic resolution of temperature coefficients and growth maxima dependencies across major marine PFTs has been limited.

Data: Compiled thermal growth rates for 243 marine phytoplankton strains (3246 measurements) across four PFTs: coccolithophores (n=30, N=202), cyanobacteria (n=32, N=502; diazotrophs excluded), diatoms (n=135, N=1794), and dinoflagellates (n=46, N=748). Data were aggregated from prior compilations and post-2012 literature, digitizing when necessary. Selection criteria ensured comparable culture conditions; fluctuating nutrient studies were excluded; cyanobacterial diazotrophs were analyzed separately but excluded due to insufficient data. Thermal reaction norms: For each strain, fitted adapted Norberg-type curves to growth vs. temperature (maximum likelihood) to estimate thermal traits including optimal temperature (Topt), maximum temperature (Tmax), and maximum growth (µmax). Defined a thermally viable range using a 20% performance breadth (temperatures where growth ≥20% of strain-specific µmax). Computed ascending/descending slopes between µ20%max and µmax. Exponential temperature dependence: For each PFT separately, fitted the maximum growth rate as an exponential function of temperature using 99th quantile regression on log-transformed growth rates: µmax(T)=a·e^{bT}. Confidence intervals were estimated via Markov chain marginal bootstrapping (10,000 iterations). Calculated PFT-specific Q10 and activation energy from fitted parameters. Model comparisons (AICc) supported a temperature×PFT interaction over pooled fits. Thermal capacity metrics: For strains of known origin with quality-controlled Tmax, computed thermal safety margin (TSM=Topt−Thab), distance to growth equivalence (DGE=Tuequiv−Thab, where growth equals that at Thab on the warm side of the curve), and warming tolerance (WT=Tmax−Thab). Assessed latitudinal patterns and PFT differences (Kruskal–Wallis; Dunn’s tests). Climate projections: Used CMIP5 ensemble mean SST (1.25°) under RCP8.5 to define baseline (1950–1970) and future (2080–2100) conditions. For each strain, estimated growth at all grid cells within its 20% performance breadth, assuming no dispersal limits and temperature-only viability. Computed proportional growth change where viable in both periods: (µfuture−µpast)/µpast. Aggregated median changes per PFT for maps and zonal means. For cyanobacteria, also quantified potential thermally viable range expansion.

• PFTs exhibit distinct thermal niches and curve shapes. Cyanobacteria could not survive below ~9.5 °C; only 17% of coccolithophore strains were viable above 30 °C, versus >60% for other groups. All PFTs showed negatively skewed reaction norms; dinoflagellates had flatter curves (shallower slopes) and lower, less variable µmax.
• Exponential temperature dependence differed strongly by PFT. Compared to the Eppley Q10=1.88, the collective Q10 across the four PFTs was 1.46. PFT-specific Q10 values: cyanobacteria 2.13 (highest thermal sensitivity), dinoflagellates 1.67, diatoms 1.55, coccolithophores 1.42. Cyanobacteria and dinoflagellates had lower intercepts than diatoms and coccolithophores.
• Absolute growth potential: Despite lower Q10, diatoms and coccolithophores had the highest modeled µmax at 20 °C (≈1.91 d−1 and 1.50 d−1, respectively), with diatoms exhibiting the greatest modeled µmax across temperatures, consistent with r-strategist behavior.
• Thermal capacity across latitude: Many mid- and low-latitude strains inhabit temperatures at or above Topt (negative TSM), yet most have substantial warming tolerance (WT) before lethality. The new DGE metric indicates additional warming can occur before growth drops below baseline in many cases. Cyanobacteria often occur further below their Topt (larger TSM) than diatoms and dinoflagellates.
• Projected changes under RCP8.5 (2080–2100 vs 1950–1970): Low latitudes show proportional growth decreases for most strains, while mid- and high latitudes show increases. Coccolithophores are most susceptible at low latitudes, with up to ~61% average decrease and 100% of strains negatively impacted within 10° of the equator (83% within 20°). Cyanobacteria are projected to gain substantially at mid-latitudes and expand thermally viable range by ~6.5% (~18.8 million km²), becoming newly viable in regions such as the Norwegian Sea and Gulf of Alaska.
• Regional projections: Southern Ocean shows ~7.2% average proportional growth increase (diatoms, dinoflagellates, coccolithophores) under temperature alone; North Atlantic shows ~21% increase. These changes imply shifts in competitive balance, community composition, and potential impacts on carbon cycling (e.g., reduced low-latitude coccolithophore growth could affect alkalinity).

The study demonstrates that applying a single thermal sensitivity across phytoplankton masks substantial taxonomic differences that determine growth responses to warming. By resolving PFT-specific Q10, intercepts, and reaction norm shapes, the work explains how temperature alone can differentially alter growth rates and distributions: declines at low latitudes, gains at mid–high latitudes, and significant cyanobacterial range expansion. These mechanistic differences help interpret and refine projections of community structure and biogeochemical function, including potential shifts from larger to smaller cell-size taxa under warming and nutrient changes. While many strains currently experience temperatures near or above their optima, buffering (WT, DGE) suggests some tolerance before lethal thresholds or declines below baseline growth, though equatorial regions are particularly constrained. Incorporating PFT-resolved thermal dependencies into models, satellite primary production algorithms, and growth standardizations can improve predictions of future productivity and diversity patterns. Potential evolutionary adaptation could modulate outcomes, but rates and trade-offs remain uncertain.

This study provides a taxonomically resolved characterization of thermal growth dependencies for four major marine phytoplankton functional types. It establishes distinct PFT-specific Q10 values, growth maxima, viable thermal ranges, and reaction norm shapes, and shows that temperature alone can drive divergent changes in growth and geographic range under end-of-century warming. Key implications include substantial low-latitude declines for coccolithophores, mid-latitude gains and range expansion for cyanobacteria, and overall increases in mid–high latitude growth potential. The results offer a baseline for integrating temperature-driven responses into Earth system models and observational algorithms. Future work should incorporate additional stressors (nutrients, light, acidification), competitive and dispersal constraints, explicit treatment of diazotrophs, and evolutionary adaptation to refine projections of community and biogeochemical change.

• Temperature-only focus: Projections exclude other key drivers (nutrients, light, acidification, grazing, mixing/stratification), which can amplify or counteract thermal effects.
• Viability criterion: Using a 20% µmax threshold may overestimate viable ranges (some strains at 20% may not persist), making change estimates conservative in some contexts.
• Data gaps and biases: Limited representation of low-latitude strains could affect equatorial projections; cyanobacterial diazotrophs were excluded due to insufficient data, limiting generality for that group.
• Methodological assumptions: Assumed no dispersal limits and no biotic interactions in mapping viable ranges; 99th quantile regression results can be sensitive to quantile choice and require large datasets; some dinoflagellate cultures could have had incidental mixotrophy.
• Statistical significance nuances: Some between-group comparisons (e.g., TSM differences) approach but may not meet conventional significance thresholds, warranting cautious interpretation.