Oceanic flows have been traditionally decomposed into two main components, "large-scale" and "mesoscale/small-scale" ("eddy"), which is in part motivated by the fact that most numerical ocean models under-resolve the eddy component. The eddy fluxes then have to be parameterized, and the most common approach is to use the flux-gradient relation with the eddy diffusivity (tensor) coefficient. The simplicity of this relation is appealing, but the main challenge lies in finding the appropriate diffusivity tensor, which varies with geographical location and depth (i.e., inhomogeneous) and time, and is direction-dependent (i.e., anisotropic), as evidenced by observation- and GCM-based estimates. The overarching goal of the proposed study is to explore properties of the inhomogeneous and anisotropic eddy-induced transport at mid-latitudes, and to examine their importance for tracer distribution. The main hypotheses are that the eddy-induced transport can be succinctly quantified by a spatio-temporal map of the eddy diffusivity tensor, and that this complexity can be systematically reduced to its most essential properties. Specifically, we will objectively calculate and analyze the corresponding eddy-diffusivity tensor maps and explore the dependence in the results on the spatial and temporal scales. This diffusivity map will then be used to produce tracer distributions in flows that do not fully resolve the eddy-driven tracer advection, and the resulting skill will be quantified using relevant metrics. We will employ a hierarchy of eddy-resolving numerical simulations, real-ocean drifter trajectories and a wide range of scale-aware flow decompositions, as well as several novel methods for estimating and interpreting the eddy diffusivity tensors for both fundamental understanding and practical purposes. The study will capitalize on the synergy of existing intensive and efficient collaboration between the US and UK members of the research group.