This work is devoted to study numerically the self-diffusion of spherocylindrical particles flowing down an inclined plane, using the discrete element method (DEM). This system is challenging due to particles being non-spherical and because they are subjected to a non-uniform shear rate. We performed simulations for several aspect ratios and inclination angles, tracking individual particle trajectories. Using the simulation data, we computed the diffusion coefficients D, and a coarse-graining methodology allowed accessing the shear rate spatial profiles (y)over dot (z). This data enabled us to identify the spatial regions where the diffusivity strongly correlates with the local shear rate. Introducing an effective particle size d(perpendicular to), we proposed a well-rationalized scaling law between D and (y)over dot. Our findings also identified specific locations where the diffusivity does not correlate with the shear rate. This observation corresponds to zones where has non-linear spatial variation, and the velocity probability density distributions exhibit asymmetric shapes.