Future phytoplankton diversity in a changing climate

Marine ecosystems provide critical socio-economic services and regulate climate through CO2 absorption and sequestration. Maintaining biodiversity underpins resilience to climate change and extremes, yet biodiversity loss is driven by human activities and marine ecosystems may be experiencing rapid reorganization rather than simple local declines. Anthropogenic climate change—via warming, altered circulation/stratification, acidification, and deoxygenation—is expected to force phytoplankton community reorganization, affecting productivity, size structure, and biogeochemical functions. Predicting future changes is challenging due to sparse sampling, intrinsic variability, and uncertainties in combined stressor effects. With future oceans projected to be ~2–4 °C warmer, more acidic, and less oxygenated, species must adapt, migrate, or face extinction, with consequences for ecosystem functioning and services. As the base of marine food webs, phytoplankton functional groups (e.g., diazotrophs, diatoms) and size structure strongly influence trophic interactions and carbon export. This study asks how phytoplankton diversity, community composition, turnover, and size structure will change under a high-emissions climate scenario. Using a mechanistic marine ecosystem model with 35 phytoplankton types and 16 zooplankton size classes, the authors examine global and regional reorganizations of phytoplankton communities over the 21st century and assess implications for stability and resilience.

Correlative and niche models often assume stationarity of relationships between present-day environmental conditions and phytoplankton diversity, limiting their ability to project responses to novel conditions. Earth System Models typically represent only 2–3 phytoplankton types, emphasizing nutrient-driven responses that favor small species under increased stratification, but missing broader trait and competitive dynamics. Prior work shows that diversity is shaped by limiting nutrient supply, relative supply ratios (e.g., nitrate:iron; silica:nitrate), grazing pressure, and transport/mixing. Studies using correlative approaches have suggested decreasing diversity in the tropics and increases at high latitudes, including Southern Ocean ‘tropicalisation’ signals, but such methods cannot capture mechanistic niche loss or dynamic shifts in competitiveness. The present work leverages a complex mechanistic model to capture functional group and size-class diversity, enabling assessment of niche loss, changing coexistence, and turnover under climate forcing beyond correlative assumptions.

A global marine ecosystem model (GUD) configured per prior publications includes 35 phytoplankton types (covering 7 functional groups: 2 prokaryote, 2 picoeukaryote, 5 coccolithophore, 5 diazotroph, 11 diatom, 10 mixotrophic dinoflagellate) spanning 0.6–2425 µm equivalent spherical diameter, and 16 zooplankton size classes. Biogeochemical cycles of C, P, N, Si, Fe, and O2 are included. Phytoplankton are modeled with constant C:N:P:Fe stoichiometry using Monod kinetics; parameters for growth, grazing susceptibility, and sinking are size-related and vary by functional group. Maximum growth rates and grazing scale with cell size; smallest groups have lowest nutrient affinity while fastest growth occurs near 3 µm. Zooplankton graze following a Holling III functional response, preferring prey ~10× smaller (range 5–15× smaller), and differ by size only. Mixotrophy is represented only in dinoflagellates. The physical framework is the MIT Integrated Global System Model (IGSM) with ocean resolution 2°×2.5°, 22 vertical layers (10 m near surface to 500 m at depth). The simulation spans 1860–2100: 1860–1990 with observed greenhouse gas emissions; 1990–2100 under a high-emissions scenario similar to RCP8.5. Analyses focus on 2006–2100. Model validation in prior studies shows reasonable agreement with satellite and in situ observations of functional types and size classes. Analyses use depth-integrated biomass across the full ocean depth, and are implemented in Matlab 2019a. Species contributing <0.1% of total biomass at a location and time are excluded. Functional richness is the number of coexisting phytoplankton types (functional group × size class) at each location and time. The Shannon diversity index is defined as Shannon = −∑ p_i ln(p_i), where p_i is the biomass proportion of type i; evenness is Shannon/ln(richness). Community turnover between two periods is turnover = (N_g + N_l)/N_T, where N_g and N_l are numbers of types gained and lost, and N_T the total observed in both periods. Turnover is computed between means of 2005–2024 and 2081–2100; change in turnover rate is computed as turnover(2061–2080 to 2081–2100) minus turnover(2011–2030 to 2031–2040). Size spectrum slopes are computed by summing biomass within each of 16 phytoplankton size classes and regressing log(–biomass) versus log(equivalent spherical diameter) using the Theil–Sen estimator; the spectrum slope is the regression slope plus 3, assuming spherical phytoplankton.

• Biomass: Projected decreases over much of the tropical and subtropical ocean due to reduced nutrient supply; increases at high latitudes due to sea-ice retreat, longer growing seasons, and higher growth rates in warmer waters.
• Richness: Declines by 2100 across large parts of northern hemisphere subtropical and temperate regions (64% of area between 23–55° N declines). Increases in polar and some equatorial regions (69% of area poleward of 55° or within 23° of the equator increases). In some tropical regions, up to 30% of modeled phytoplankton types become locally extinct; in polar regions, colonization exceeds extinction with richness increases up to 30%.
• Functional groups: Reduced nitrate:iron supply favors diazotrophs, expanding poleward (notably in the northern hemisphere). Decreased silicic acid relative to nitrate reduces diatom richness, with up to 30% of diatom types going locally extinct in some regions. Some mixotrophic dinoflagellates disappear along subtropical gyre boundaries but colonize much of the Southern Ocean where intermediate nutrient supply and increased small-prey abundance improve competitiveness. Picoplankton distributions change little by 2100.
• Evenness and Shannon diversity: Evenness decreases almost universally (93% of ocean area), concentrating biomass in fewer types; Shannon diversity declines in 92% of the ocean.
• Turnover: By 2100, community turnover relative to 2005–2024 reaches ~20% in parts of the temperate northern hemisphere and the South Pacific subtropical gyre; elsewhere, <10%. The turnover rate increases over time across 63% of the ocean, indicating increasing variability (decreasing stability) in community composition.
• Size structure: The phytoplankton size spectrum slope decreases (shift toward smaller types) in most subtropical regions (69% of area) and in the Southern Ocean (90% of area). In subtropics, driven by loss of larger types; in the Southern Ocean, driven by greater increases in small types. Some regions (33% of ocean area) show increased overall size (e.g., North Atlantic influx of larger dinoflagellates and loss of diatoms).
• Mechanisms: Changes in limiting nutrient supply and ratios, grazing pressure (including loss of larger zooplankton), and increased stratification/mixing changes jointly determine patterns of richness, evenness, and functional group shifts.
• Implications: Trend toward smaller phytoplankton suggests less productive food webs and reduced carbon export. Increasing turnover implies reduced niche diversity, more ephemeral species, fewer persistent dominants, and loss of ecological resilience. Relocation of communities does not necessarily prevent extinction by 2100, particularly at low latitudes, and diatoms/larger phytoplankton decline globally.

The mechanistic ecosystem model reveals that climate-driven changes in nutrient supply rates and ratios, grazing pressure, and physical transport/mixing will reorganize phytoplankton communities this century. Declines in biomass and richness in many lower-latitude and temperate regions reflect reduced macronutrient supply and grazer-mediated reductions in coexisting size classes, while high-latitude biomass and richness increases result from longer growing seasons and warming-enhanced growth. Functional shifts, including poleward expansion of diazotrophs and declines in diatom richness due to silica limitation, demonstrate how resource ratio changes restructure coexistence. Almost ubiquitous declines in evenness lead to lower Shannon diversity, even where richness increases (e.g., Southern Ocean), indicating dominance by fewer types. Turnover intensifies and accelerates over time, signaling decreased community stability and resilience, with potential consequences for higher trophic levels that rely on consistent, specific phytoplankton prey. The projected shift to smaller phytoplankton implies reduced trophic transfer efficiency and carbon sequestration. Comparisons with correlative studies show broad agreement on tropical decreases and high-latitude increases in diversity, but the mechanistic approach captures dynamic niche loss and altered competitiveness under novel conditions. Regions at polar edges may act as refugia as niches expand without excluding extant types. Although species migration may track changing climates, analogous dietary environments may not align spatially, potentially disrupting food webs. The increasing variability and reduced evenness indicate a loss of ecological resilience and greater instability in marine ecosystems under continued high emissions.

Under a high-emissions scenario, mechanistic modeling projects widespread reorganization of marine phytoplankton communities by 2100: tropical to temperate richness declines, high-latitude richness increases, nearly global decreases in evenness and Shannon diversity, intensified turnover, and a shift toward smaller cell sizes. These changes imply reduced ecological resilience, altered food-web productivity, and diminished carbon export, with potential trophic amplification and socio-economic impacts, especially in low-latitude regions reliant on fisheries. The study advances understanding by linking nutrient supply changes, resource ratios, and grazing dynamics to functional group and size-class diversity shifts, moving beyond correlative projections. Future research should incorporate additional trait dimensions (e.g., thermal niches), adaptation and evolutionary responses, finer spatial resolution to include coastal and sea-ice communities, and broader anthropogenic stressors beyond climate, to refine projections and assess potential mitigation or resilience pathways.

• Trait representation is limited to functional group and size; thermal norms and other traits (morphology, colony formation) are not included, so biogeography shifts arise indirectly via nutrient and competitiveness changes rather than explicit thermal niche responses.
• Coarse spatial resolution (2°×2.5°) limits representation of coastal regions; sea-ice communities are not explicitly modeled.
• Only climate-driven forcings are considered; other anthropogenic impacts (e.g., runoff, pollution, habitat reduction) are excluded.
• No evolutionary adaptation or acclimation is represented; plankton do not adapt to changing conditions, potentially yielding a worst-case scenario.
• Zooplankton diversity is represented only by size (no functional differences), focusing analyses on phytoplankton diversity.
• Mixotrophy is only represented in dinoflagellates and is constrained by limited observational data.