Abstract. Oceanic particulate organic carbon (POC) is a small but dynamic component of the global carbon cycle. Biogeochemical models historically focused on reproducing the sinking flux of POC driven by large fast-sinking particles (LPOC). However, suspended and slow-sinking particles (SPOC, here < 100 µm) dominate the total POC (TPOC) stock, support a large fraction of microbial respiration, and can make sizable contributions to vertical fluxes. Recent developments in the parameterization of POC reactivity in PISCES (Pelagic Interactions Scheme for Carbon and Ecosystem Studies model; PISCESv2_RC) have improved its ability to capture POC dynamics. Here we evaluated this model by matching a global 3D simulation and 1D simulations at 50 different locations with observations made from biogeochemical (BGC-) Argo floats and satellites. Our evaluation covers globally representative biomes between 0 and 1000 m depth and relies on (1) a refined scheme for converting particulate backscattering at 700 nm (bbp700) to POC, based on biome-dependent POC / bbp700 ratios in the surface layer that decrease to an asymptotic value at depth; (2) a novel approach for matching annual time series of BGC-Argo vertical profiles to PISCES 1D simulations forced by pre-computed vertical mixing fields; and (3) a critical evaluation of the correspondence between in situ measurements of POC fractions, PISCES model tracers, and SPOC and LPOC estimated from high vertical resolution bbp700 profiles through a separation of the baseline and spike signals. We show that PISCES captures the major features of SPOC and LPOC across a range of spatiotemporal scales, from highly resolved profile time series to biome-aggregated climatological profiles. Model–observation agreement is usually better in the epipelagic (0–200 m) than in the mesopelagic (200–1000 m), with SPOC showing overall higher spatiotemporal correlation and smaller deviation (typically within a factor of 1.5). Still, annual mean LPOC stocks estimated from PISCES and BGC-Argo are highly correlated across biomes, especially in the epipelagic (r=0.78; n=50). Estimates of the SPOC / TPOC fraction converge around a median of 85 % (range 66 %–92 %) globally. Distinct patterns of model–observations misfits are found in subpolar and subtropical gyres, pointing to the need to better resolve the interplay between sinking, remineralization, and SPOC–LPOC interconversion in PISCES. Our analysis also indicates that a widely used satellite algorithm overestimates POC severalfold at high latitudes during the winter. The approaches proposed here can help constrain the stocks, and ultimately budgets, of oceanic POC.

No abstract

Abstract. When working with Earth system models, a considerable challenge that arises is the need to establish the set of parameter values that ensure the optimal model performance in terms of how they reflect real-world observed data. Given that each additional parameter under investigation increases the dimensional space of the problem by one, simple brute-force sensitivity tests can quickly become too computationally strenuous. In addition, the complexity of the model and interactions between parameters mean that testing them on an individual basis has the potential to miss key information. In this work, we address these challenges by developing a biased random key genetic algorithm (BRKGA) able to estimate model parameters. This method is tested using the one-dimensional configuration of PISCES-v2_RC, the biogeochemical component of NEMO4 v4.0.1 (Nucleus for European Modelling of the Ocean version 4), a global ocean model. A test case of particulate organic carbon (POC) in the North Atlantic down to 1000 m depth is examined, using observed data obtained from autonomous biogeochemical Argo floats. In this case, two sets of tests are run, namely one where each of the model outputs are compared to the model outputs with default settings and another where they are compared with three sets of observed data from their respective regions, which is followed by a cross-reference of the results. The results of these analyses provide evidence that this approach is robust and consistent and also that it provides an indication of the sensitivity of parameters on variables of interest. Given the deviation in the optimal set of parameters from the default, further analyses using observed data in other locations are recommended to establish the validity of the results obtained.

Abstract. Oceanic particulate organic carbon (POC) is a relatively small (~4 Pg C) but dynamic component of the global carbon cycle with fast mean turnover rates compared to other oceanic, continental and atmospheric carbon stocks. Biogeochemical models historically focused on reproducing the sinking flux of POC driven by large fast-sinking particles (bPOC). However, suspended and slow-sinking particles (sPOC) typically represent 80–90 % of the POC stock, and can make important seasonal contributions to vertical fluxes through the mesopelagic layer (200–1000 m). Recent developments in the parameterization of POC reactivity in the PISCES model (PISCESv2_RC) have greatly improved its ability to capture sPOC dynamics. Here we evaluated this model by matching 3D and 1D simulations with BGC-Argo and satellite observations in globally representative ocean biomes, building on a refined scheme for converting particulate backscattering profiles measured by BGC-Argo floats to POC. We show that PISCES captures the major features of sPOC and bPOC as seen by BGC-Argo floats across a range of spatiotemporal scales, from highly resolved profile time series to biome-aggregated climatological profiles. Our results also illustrate how the comparison between the model and observations is hampered by (1) the uncertainties in empirical POC estimation, (2) the imperfect correspondence between modelled and observed variables, and (3) the bias between modelled and observed physics. Despite these limitations, we identified characteristic patterns of model-observations misfits in the mesopelagic layer of subpolar and subtropical gyres. These misfits likely result from both suboptimal model parameters and model equations themselves, pointing to the need to improve the model representation of processes with a critical influence on POC dynamics, such as sinking, remineralization, (dis)aggregation and zooplankton activity. Beyond model evaluation results, our analysis identified inconsistencies between current estimates of POC from satellite and BGC-Argo data, as well as POC partitioning into phytoplankton, heterotrophs and detritus deduced from in situ bio-optical data. Our approach can help constrain POC stocks, and ultimately budgets, in the epipelagic and mesopelagic ocean.