Calcite carbonate sinks low-density plastic debris in open oceans

Marine plastic pollution has been a global concern, but the fate of low-density plastics in deep seas remains uncertain. Low-density plastics such as PE and PP comprise over half of plastic waste and fragment into smaller pieces, yet field measurements show far fewer low-density microplastics (LDMPs) at the sea surface than expected, especially below 5 mm. Deep-water and sediment studies since 2013 have detected considerable LDMPs from the twilight zone to abyssal depths, and mass-balance assessments suggest a large share has settled on the seafloor. Prior modeling has mainly focused on biofouling-driven density changes. Kooi et al. developed a 1D, depth-dependent model that explained low surface concentrations and accumulation within ~200 m but produced oscillatory behavior near the surface and could not account for abundant LDMPs in sediments. Subsequent work incorporating ocean processes allowed some particles to penetrate deeper, but only a negligible fraction reached the abyssal seafloor. This study asks whether microbially induced calcium carbonate precipitation (MICP), commonly occurring in calcium-rich, high-pH environments and potentially accompanying biofouling, can provide a sufficient ballast to alter LDMP vertical motion and help explain observed subsurface and seafloor accumulations.

Field evidence shows LDMPs throughout the water column and in sediments across basins, including samples down to >10,000 m. Traditional explanations for LDMP removal from the surface include biofouling, ingestion (fecal pellets), aggregation with marine snow, and frustule attachment. However, ingestion likely sinks only ~0.13–0.19% of LDMPs and marine snow 0.06–8.8%, insufficient to explain sediment enrichment. Biofouling-based models produce near-surface oscillations and do not deliver many particles to the seafloor; even with advection, turbulence, grazing, and mixing, only ~15 of 10,000 LDMPs reached >5000 m in prior simulations. MICP is known under calcium-rich, high-pH conditions; photosynthesis and urea hydrolysis can raise pH and precipitate CaCO3 near biofilms, forming exoskeletons (primarily calcite in typical oceanic temperatures). Calcite (density ~2.63 g cm−3) could act as an effective ballast, potentially overcoming limitations of purely biofilm-driven density increases.

The authors developed a one-dimensional, depth-dependent hydrodynamic model to simulate LDMP vertical settling under quiescent ocean conditions while explicitly including MICP and updated motion equations for irregular particles. LDMPs considered: densities 0.85–1.00 g cm−3; sizes 1 µm to 5 mm (exponential size distribution tuned to observations); shapes evenly distributed among spheres, fibers (length 57–13,000 µm; or 100 µm–1.3 mm in Methods), and films (thickness 0.1–100 µm). Monte Carlo simulations generated ensembles of LDMPs (e.g., 500 particles) to compute trajectories over up to 200–300 days.

• Vertical settling model: Uses fluid dynamics valid over a wide Reynolds number range, adopting a one-equation drag model for irregular particles (Dioguardi & Dellino, 2018). The drag coefficient depends on Reynolds number and a shape factor (Ψ), and settling velocity is computed iteratively with convergence on Re. Particle instantaneous density ρpt(t) accounts for the masses/volumes of the plastic core, attached biofilm (algae), and calcite precipitates.
• Shape factor and geometry: Ψ = Φ/χ, where sphericity Φ relates the surface area of an equivalent sphere to the particle surface, and circularity χ uses projected perimeter metrics. Ψ = 1 for spheres. Equivalent spherical diameter (ESD) is used for encounter kernels and diffusivities when needed.
• MICP module: Calcite precipitation at the plastic surface is driven by algal photosynthesis, modeled via a calcium flux (Fick’s law) and a precipitation rate scaled linearly with local light intensity: Pca = (I/Io) Pca0. Oversaturation near biofilms promotes calcite formation; dissolution is introduced below the regional calcite saturation depth (CSD), accounting for depth/pressure effects on solubility and oxygen consumption in deep waters. Regional CSD variability (∼1000–4600 m) is considered, with limited dissolution expected for fast-sinking particles.
• Biofouling module: Algae-only biofilm growth is modeled to simplify community dynamics; processes include collision/attachment (ambient algal concentration inferred from chlorophyll-a via temperature- and light-dependent conversion), growth (temperature- and light-dependent kinetics), mortality, and respiration. Encounter kernels include Brownian motion, differential settling, and shear. Light follows Lambert-Beer attenuation with surface light as a sinusoid; extinction includes water and chlorophyll-a components. Geometry-specific expressions estimate biofilm thickness and surface area changes for films, fibers, and spheres.
• Ocean profiles: Temperature and salinity profiles (nowcasts) are obtained from NOAA’s RTOFS (HYCOM-based), with linear interpolation between 33 depth levels. Seawater density is computed from T and S using established polynomials; viscosity uses IAPWS 2008 correlations adjusted for salinity. Simulations used specific profiles (e.g., tropical Pacific) and seasonal/arctic scenarios to compare with field distributions.
• Simulation design: Ensembles of randomly generated LDMPs (shape ratio 1:1:1; size/density distributions as above) are integrated via a Markov Chain time-stepping to produce vertical trajectories and statistics (densities, drag coefficients, settling velocities, depths) over 20–300 days. Sensitivity analyses examine size, shape, and initial density impacts.

• MICP as a dominant ballast: Calcite (2.63 g cm−3) provides a stronger, more persistent density increase than biofilms (1.15–1.18 g cm−3), frustules (~1.80 g cm−3), marine snow (1.02–1.03 g cm−3), or fecal pellets (1.02–1.06 g cm−3). A 10 µm PE sphere (ρ=0.92 g cm−3) needs only ~0.35 µm of calcite coating to exceed 1.1 g cm−3, surpassing maximum seawater density (~1.09 g cm−3).
• Two settling patterns emerge with MICP: (1) Damped oscillation in the epipelagic zone for larger, more spherical LDMPs with low drag; small biofilm losses can reverse buoyancy and cause rapid ascents, but retained calcite prevents full re-emergence, producing damped cycles. (2) Direct sinking to sediments for smaller (<500 µm) and/or non-spherical LDMPs (films, fibers) with higher drag, which allows time for heavier fouling; densities can reach up to ~2.40 g cm−3 before leaving the epipelagic zone.
• Size-selective sinking and accumulation: Sphere-shaped LDMPs of 100–500 µm can independently settle to the seafloor without aggregation under tropical Pacific profiles, with settling velocities ~100–300 m day−1. For 10–100 µm spheres, velocities drop to ~10–20 m day−1; <10 µm particles fall <1 m day−1 and can be retained longer in the upper water column. LDMPs of 1–5 mm generally do not reach the seafloor without aggregation. These size dependencies match observed enrichment of 10–200 µm LDMPs in sediments.
• Shape effects via drag: Fibers and films have higher specific surface areas and vertical drag coefficients than spheres, leading to deeper but broader subsurface distributions (100–400 m) and longer times to reach steady states; spheres tend to concentrate at 100–200 m and are more likely to accumulate on the seafloor for sizes <~500 µm.
• Limited impact of calcite dissolution: Despite CSD-driven undersaturation at depth (regional 1000–4600 m), modeled particles typically experience only slight calcite loss and velocity reduction because (i) high-pH microenvironments inhibit respiration-driven CO2 near the surface biofilm, (ii) sinking is relatively fast (often reaching the seafloor within <1 year; minimal dissolution for canonical 100 m day−1 sinking), and (iii) many seafloors lie above regional CSD.
• Two vertical accumulation zones reproduced: A subsurface accumulation zone (∼60–400 m) forms in subtropical settings with stable light and chlorophyll, and a seafloor accumulation zone forms as an ultimate sink. Arctic distributions differ seasonally: high summer light promotes rapid density gain and sinking through a nearly uniform density profile; winter darkness reverses buoyancy growth, increasing surface abundances and weakening subsurface peaks.
• Aggregation as a supplemental pathway: While not required for 100–500 µm spheres, aggregation can assist very small (<10 µm) and large (1–5 mm) LDMPs to sink. Oscillating large LDMPs enhance collision rates, and small, dense-fouled particles serve as effective ballast in aggregates.
• Model-data consistency: Simulations reproduce observed low surface concentrations, subsurface peaks in many basins, and high sediment abundances, especially of 10–200 µm LDMPs. Prior models without MICP delivered only ~0.15% of particles to >5000 m (15 of 10,000), whereas MICP-enabled simulations commonly brought most LDMPs to the seafloor within tens to hundreds of days.

Including MICP fundamentally changes LDMP vertical dynamics by providing persistent, high-density ballast that remains after biofilm decay. This mechanism explains the paradox of missing/lost LDMPs at the surface and their widespread detection in sediments, particularly for 10–200 µm particles. The two-pattern behavior (damped oscillation vs direct sinking) arises from interactions among size, shape (drag), seawater density stratification, and fouling/dissolution kinetics. Subsurface accumulation forms under steady light/chlorophyll conditions as some sizes/shapes cannot sustain negative buoyancy against density gradients or are slowed by high drag. Seafloor accumulation emerges as the ultimate sink for sizes (notably 100–500 µm spheres and larger equivalent films/fibers) that can gain and retain sufficient calcite. Regional differences, such as the Arctic’s weak density stratification, seasonal light extremes, and downward currents, modulate these patterns and align with observed higher surface and seafloor concentrations and diminished subsurface maxima. The results suggest that previous underestimation of MICP’s role led to models that failed to deliver LDMPs to sediments, while MICP addresses this gap and provides a coherent framework for observed distributions.

The study introduces a MICP-enabled, shape-aware hydrodynamic model that reproduces observed LDMP vertical distributions, including subsurface and seafloor accumulation zones. Calcite precipitation accompanying biofouling imparts durable ballast, driving size- and shape-selective sinking: sphere-shaped LDMPs of 100–500 µm (and larger equivalent fibers/films) can sink independently at high velocities, while very small (<10 µm) and large (1–5 mm) LDMPs often require aggregation. MICP induces damped oscillations for larger, low-drag particles and direct sinking for smaller, high-drag particles, and calcite dissolution exerts limited influence within typical sinking timescales and regional CSDs. These insights resolve aspects of the missing plastic problem and imply significant implications for benthic ecosystems and the marine carbon/calcite cycles. Future work should (i) expand field measurements of deep-water and sediment LDMPs, prioritizing identification of calcite-covered particles; (ii) refine MICP parameterizations with in situ rates and microbial community data; (iii) integrate full oceanographic dynamics (advection, turbulence, seasonality) at basin scales; and (iv) quantify ecosystem and biogeochemical impacts of calcite-laden plastics on the seafloor.

Key limitations include: (1) Quiescent, 1D framework without explicit 3D advection/turbulence (though profiles and seasonality are represented), potentially underrepresenting transport heterogeneity; (2) Simplifications of biofouling (algae-focused, reduced parameter set) and MICP (linear scaling with light, autotrophic pathway only; bacteria and heterotrophic processes neglected); (3) Assumed shape ratio (fiber:sphere:film = 1:1:1) and exponential size distribution tuned to prior observations, which may differ regionally; (4) Use of RTOFS nowcasts and parameterizations for T–S–ρ–μ that may not capture all mesoscale/submesoscale variability; (5) Limited treatment of aggregation dynamics (considered conceptually, not fully coupled), ingestion/egestion pathways, and resuspension; (6) Calcite dissolution below CSD represented simply and not fully constrained by site-specific carbonate chemistry; (7) Model validation focuses on reproducing distribution patterns rather than absolute concentrations due to unknown inputs. These factors may affect quantitative predictions and generalizability across ocean regimes.