Inadequacies in the representation of sub-seasonal phytoplankton dynamics in Earth system models

Abstract. Sub-seasonal phytoplankton dynamics on timescales between 8 days and 3 months significantly contribute to annual fluctuations, making it essential to accurately represent this variability in ocean models to avoid distorting long-term trends. This study assesses the capability of Earth System Models (ESMs) participating in the Coupled Model Intercomparison Project Phase 6 (CMIP6) to reproduce sub-seasonal surface ocean phytoplankton variations observed in ocean color satellite data. Our findings reveal that, unlike sea surface temperature, all models struggle to accurately reproduce the total surface ocean phytoplankton variance and its decomposition across sub-seasonal, seasonal, and multi-annual timescales. Over the historical period, some models strongly overestimate sub-seasonal variance and exaggerate its role in annual fluctuations, while others underestimate it. Our analysis suggest that underestimation of sub-seasonal variance is likely a consequence of the coarse horizontal resolution of CMIP6 models, which is insufficient to resolve mesoscale processes—a limitation potentially alleviated with higher-resolution models. Conversely, we suggest that the overestimation of sub-seasonal variance is potentially the consequence of intrinsic oscillations such as predator-prey oscillations in certain biogeochemical models. ESMs consistently show a reduction in variance at sub-seasonal and seasonal timescales during the 21^st century under high-emission scenarios. The poor capability of CMIP6 models at simulating sub-seasonal chlorophyll dynamics casts doubt on their projections at these temporal scales and multi-annual timescales. This study underscores the need to enhance spatial resolution and constrain intrinsic biogeochemical oscillations to improve projections of ocean phytoplankton dynamics.