The vertical flux of particulate organic carbon (POC) in the ocean is a critical process influencing global carbon cycling and climate. This flux is determined by POC production in the surface ocean, particle sinking speed, and POC consumption in the subsurface. Previous models have focused on physical and chemical processes like particle size, lability, and temperature, but have largely neglected the crucial role of microbial dynamics. This is a significant oversight, as microbial activity directly impacts many of these physical and chemical factors. For example, microbial degradation fundamentally influences particle lability and size, thereby affecting sinking speed. Heterotrophic microbes consume a substantial portion (70–92%) of POC, and grazers like zooplankton further influence particle aggregation, disaggregation, and consumption. Current biogeochemical models often use simplified, constant rates of POC consumption, which fail to capture the inherent dynamism of microbial communities. Particle-associated heterotrophic microbial communities exhibit rapid growth rates, even on recalcitrant organic compounds, capable of consuming particles within days. This contrasts with observations of persistent particles at depth, suggesting other factors are at play, such as temperature and pressure limitations on microbial activity. Microbial dynamics such as enzyme production, attachment, detachment, and mortality significantly influence POC degradation rates. Laboratory studies indicate that successful particle colonization requires microbes to surpass a critical population size to counteract diffusive losses of enzymes and degradation products. This suggests highly variable POC consumption rates, contrary to the assumptions of existing carbon cycle models. Therefore, to accurately predict the vertical carbon flux, a mechanistic understanding incorporating particle-associated microbial behavior is needed.

The literature extensively documents the importance of particulate organic carbon (POC) flux in ocean carbon cycling and climate regulation (Martin et al., 1987; Gloege et al., 2017; Guidi et al., 2015). Numerous studies have investigated the physical and chemical factors influencing POC flux, including particle size spectra (Omand et al., 2020; Kriest & Oschlies, 2008), lability (Aumont et al., 2017; DeVries & Weber, 2017), and temperature (DeVries et al., 2014; DeVries & Weber, 2017; Cram et al., 2018; Boyd et al., 2019; Cavan et al., 2019). However, the explicit inclusion of microbial dynamics in large-scale models remains limited. Existing studies highlight the role of microbes in POC remineralization (Cho & Azam, 1988; Giering et al., 2014; Church et al., 2021) and the impact of zooplankton grazing (Steinberg et al., 2008; Romero-Romero et al., 2020; Maas et al., 2020; Möller et al., 2012; Cavan et al., 2017). Laboratory research emphasizes the complex microbial dynamics on sinking particles, including colonization, growth, detachment, and mortality (Datta et al., 2016; Ebrahimi et al., 2019; Enke et al., 2018; Kiørboe et al., 2003; Grossart et al., 2003; Ploug & Grossart, 2000; Kirchman, 2016). The importance of reaching a critical population size for successful colonization has also been highlighted (Ebrahimi et al., 2019; Kaul et al., 2016). Despite this understanding at the microscale, its translation to large-scale oceanographic models remains a challenge. The impact of pressure and temperature on microbial activity has also been studied (Tamburini et al., 2009; Tamburini et al., 2002; Tamburini et al., 2013). The need for more mechanistic models incorporating these factors is evident. Studies have also investigated the effect of particle aggregation and disaggregation on carbon cycling (Briggs et al., 2020; Burd & Jackson, 2009; Karakaş et al., 2009) and the role of phytoplankton dynamics (Kostadinov et al., 2016; Jin et al., 2006; Mouw et al., 2016).

The authors developed a water-column model that explicitly incorporates micro-scale observations of particle-associated microbial dynamics. The model simulates the colonization of sinking particles by free-living microbes and the conversion of polymeric organic matter (POM) into low-molecular-weight organic matter (LMWOM). To account for the complexity of microbial communities and organic carbon diversity, the model employs 'lability' as an ecosystem property, defining the conversion rate of POM to LMWOM by specific microbial groups. A wide range of lability values, assumed to follow a log-normal distribution, are tested. The model simplifies the system by using a single microbial group per particle and a single lability value per particle. Sensitivity tests with more complex representations (multiple microbial groups per particle, broader ranges of microbial types) indicated these simplifications did not significantly impact the model's output. The model is initialized using a particle export depth (default 100m), with particle size distributions spanning observed ranges (power law with exponents -2, -3, or -4). Additional particle formation below the export depth is not considered for simplicity. Each particle is stochastically assigned an initial radius, lability, and microbial degrader type. Microbial groups are defined by enzyme kinetics, maximum growth rate, and abundance in the free-living pool, which decreases exponentially with depth. The model simulates continual colonization of particles as they sink, calculating sinking speed based on size and specific gravity. Microbial growth evolves prognostically as particles sink, with carbon consumption or loss determined by diffusion and advection. Particle radius shrinks due to microbial degradation, reducing sinking speed. The model accounts for temperature dependence of microbial growth rates using a typical water column temperature profile. A key feature is the model's capture of density-dependent growth, where the population density must surpass a critical threshold for successful colonization due to mortality, detachment, and diffusive/advective loss of LMWOM. The model uses a simplified mathematical model of POC degradation to demonstrate the emergence of population-dependent growth rate from basic microbial dynamics and particle properties. The model explores multiple biotic and abiotic factors influencing particle-associated microbial growth and POC consumption, demonstrating the non-predictability of consumption rates based on individual factors. The model assesses the impact of microbial dynamics on POC flux attenuation, calculating the degradation timescale for each particle. Model parameters such as maximum growth rate, particle lability, initial cell density, and the density of free-living microbes were varied stochastically. The POC flux at specific depths is calculated as the sum of individual particle fluxes. To compare model results to observations, the rate of POC flux attenuation is quantified using a power law function. Finally, to evaluate the model's performance, the authors compare the model's simulated POC flux distributions to empirical data from the global ocean. The model's predictions of microbial community composition changes with depth are compared to existing literature on particle-associated communities. The role of temperature in the model dynamics is further investigated using varied temperature profiles and temperature limitation functions, comparing with observed temperature-growth relationships for marine heterotrophs. The influence of particle formation depth is also studied.

The study's mechanistic model successfully captures the density-dependent growth of particle-associated microbial populations, demonstrating that populations must surpass a critical threshold to overcome losses due to mortality, detachment, and diffusive/advective loss of LMWOM (Fig. 2). The model shows that the timescale of particle degradation varies significantly, ranging from a few days to over 200 days, with most particles lasting around 30 days. The key determinant of degradation time is the lag phase of particle-associated microbial populations; once exponential growth is reached, the particle is rapidly consumed. Successful colonization and rapid consumption occur when population density, encounter rates, and lability are high (Fig. 3, E-G curves), whereas low values result in struggling communities and slow POC remineralization, leading to greater transfer of carbon to the deep ocean (Fig. 3, A-B curves). The model demonstrates how struggling communities can be rescued by recruitment from the free-living microbial pool as the particle sinks, if encounter rates are sufficient (Fig. 3, A vs. C). The model highlights that the depth of particle consumption depends on multiple interactive factors: lability, particle size, microbial biomass at formation, temperature, and exchange with the free-living community. This dichotomy between successful and struggling communities provides a mechanistic explanation for the commonly observed double exponential representation of POC flux. The model reveals that lability should be considered an ecosystem property, not solely a chemical one, defined by both organic carbon composition and the microbial community. Particle remineralization rates vary by orders of magnitude across particles and depths due to these dynamics. This dichotomy is robust even in a modified model where direct POC consumption occurs, without diffusive loss. The model predicts a shift in particle-associated microbial community composition with depth, with fast-growing microbes being more abundant in the upper water column and slower-growing ones at depth (Supplementary Fig. 17). This is consistent with documented changes in prokaryotic communities with depth. The model incorporates temperature dependence of microbial growth, but suggests that temperature is a secondary factor; its importance becomes more apparent for struggling communities. Variations in microbial dynamics alone are sufficient to generate the observed range and distribution of global POC flux profiles (Fig. 4). The model shows that changes in microbial dynamics, like growth rate or lability, can alter the POC flux as much as changes in particle size spectra (Fig. 4). The model provides a potential explanation for pulsed carbon export from oligotrophic regions, proposing that the microbial community's adaptation to particle consumption plays a crucial role. The study demonstrates the significance of microbial dynamics in setting POC fluxes and emphasizes the need for better measurements of microbial processes to improve global carbon cycle models.

The findings significantly advance our understanding of ocean carbon cycling by demonstrating the critical role of stochastic microbial dynamics in regulating POC flux. The model's capacity to reproduce observed POC flux patterns, including the long tail of persistent POC at depth, highlights the importance of accurately representing microbial growth dynamics, particularly the frequency of communities struggling to establish on particles. The model suggests that relatively small changes in microbial dynamics can trigger substantial shifts in POC flux, particularly for particles persisting at depth due to suboptimal microbial growth. The variability in POC consumption rates caused by these stochastic interactions between biological, chemical, and physical factors provides a mechanistic basis for the observed spatial and temporal variations in POC flux. The model offers a novel explanation for pulsed carbon export events in oligotrophic regions, attributing it to differences in microbial community adaptation to specific particle types generated by different phytoplankton assemblages. This work underscores the insufficiency of current, simplified POC degradation parameterizations in global models and emphasizes the necessity of integrating the highly dynamic nature of particle-associated microbial communities. Future research should focus on quantifying complex ecological interactions between microbial communities and zooplankton, as well as exploring the role of zooplankton bactivory and direct consumption of particles by zooplankton.

This study presents a novel mechanistic model that integrates micro-scale particle-associated microbial dynamics into a water-column model, offering valuable insights into POC flux attenuation. The model's key contribution lies in demonstrating the significant and often underappreciated role of stochastic microbial community assembly in shaping POC flux patterns, revealing the importance of considering lability as an ecosystem property. The model successfully reproduces the observed variability in POC flux, highlighting the limitations of existing simplified representations and emphasizing the need for improved parameterizations in global carbon cycle models. Future research directions include incorporating additional factors like zooplankton dynamics and obtaining better in situ measurements of key microbial processes to further refine model predictions and enhance our understanding of ocean carbon cycling.

While the model represents a significant advance in incorporating micro-scale microbial dynamics into large-scale carbon cycle modeling, some limitations exist. The model simplifies particle composition and microbial diversity, potentially overlooking interactions that could influence POC flux. The model does not explicitly include zooplankton grazing, particle aggregation/disaggregation, or in situ particle formation, factors that also affect POC flux. These omissions might influence the model's predictive accuracy in specific environments. Furthermore, the model relies on parameter values obtained from a combination of field measurements and laboratory experiments; the uncertainty associated with these parameters might affect the model's robustness.