<jats:title>Abstract</jats:title><jats:sec><jats:label/><jats:p>The Arctic Ocean is characterized by substantial seasonal and inter‐annual variability, of which the sources and impacts are not yet fully understood. Here, we analyze how much of the variability found in in situ observations of biogeochemical and ecological variables collected at the Long‐Term Ecological Research Observatory HAUSGARTEN can be explained by differences in the physical conditions in the water masses passing through the Fram Strait (FS). Employing a size‐based plankton ecosystem model with nine distinct size classes of protist phyto‐ and zooplankton, we simulate standing stocks and fluxes within the nutrient, phytoplankton, zooplankton, and detritus pools in water parcels that follow trajectories tracing the opposing East‐Greenland and West‐Spitsbergen currents through the FS. Our model results agree with in situ observations of biogeochemical tracers, plankton size measurements, climatological data, and remote sensing observations. They show distinct temporal developments in plankton size composition, growth, and export in trajectory ensembles, highlighting how variable physical conditions affect the communities' specific growth histories. Our study indicates that 10%–72% of the variability in upper water column tracer concentrations observed in the FS can be attributed to differences in water parcel trajectories. The maxima of net primary production and vertical export along the trajectories occurred in some (spatial and temporal) distance upstream of the sites of in situ sampling. This study shows that Lagrangian modeling helps clarify complex biogeochemical‐ecological relationships in highly dynamic systems such as the FS, which is urgently needed to understand the role of climate change in the Arctic carbon cycle.</jats:p></jats:sec>