During silver electrodeposition on Au nanoparticle (NP)-covered highly oriented pyrolitic graphite, a transition from an initial growth of microsized particles to the growth of dendrites with pine tree shape (nanotrees) is observed, which is an advancement for material growth with hierarchical surface roughness. Using kinetic Monte Carlo simulations of an electrodeposition model, those results are explained by the interplay of diffusive cation flux in the electrolyte and relaxation of adsorbed atoms by diffusion on quenched crystal surfaces. First, simulations on NP-patterned substrates show the initial growth of faceted silver particles, followed by the growth of nanotrees with shapes similar to the experiments. Next, simulations on electrodes with large prebuilt particles explain the preferential nanotree growth at corners and edges as a tip effect. Simulations on wide flat electrodes relate the nanotree width with two model parameters describing surface diffusion of silver atoms: maximal number of random hops (G) and probability of hop per neighbor (P). Finally, simulations with small electrode seeds confirm the transition from initially compact particles to the nucleation of nanotrees and provide estimates of the transition sizes as a function of those parameters. The simulated compact and dendritic deposits show dominant (111) surface orientation, as observed in experiments. Extrapolations of simulation results to match microparticle and nanotree sizes lead to G = 4 × 1011 and P = 0.03, suggesting to interpret those sizes as diffusion lengths on the growing surfaces and giving diffusion coefficients 2 to 3 × 10-13 m2/s for deposited silver atoms. These results may motivate studies to relate diffusion coefficients with atomic-scale interactions.