*We introduce a class of noisy quantum cellular automata on a qubit lattice that includes all classical Markov chains, as well as maps where quantum coherence between sites is allowed to build up over time. We apply such a construction to the problem of excitation transfer through 1-d lattices, and compare the performance of classical and quantum dynamics with equal local transition probabilities. Our discrete approach has the merits of stripping down the complications of the open system dynamics, of clearly isolating coherent effects, of allowing for an exact treatment of conditional dynamics, all while capturing a rich variety of dynamical behaviours. PACS numbers: 03.65.Yz, 03.65.Aa, Quantifying the extent to which quantum coherence enhances the performance of antennae or communication systems is a timely [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19], yet often controversial, subject. Typically, a classical system is compared to the analogous, quantised model [12,13,[20][21][22], although such a correspondence is not straightforward when open, dissipative systems are considered, as it should be in most cases of interest.In this manuscript, we introduce a strictly local model of energy transfer via a noisy quantum cellular automaton construction [23,24] on a qubit lattice. Tuning one real parameter of such a model will allow us to range from a classical Markov chain, where quantum coherence is systematically suppressed at each time-step of the automaton, to dynamics where quantum coherence is allowed to build up over time, while keeping, by construction, the local transition probabilities constant. Thus, a "fair" comparison between classical and quantum energy transfer may be carried out, where the effect of quantum interference is singled out with no ambiguity. Our model, restricted to the first excitation subspace, can be studied exactly for very large systems, also including conditional dynamics due to measurements (which will model the absorption of the excitation at the end of the energy transfer process).The plan of the paper is as follows. We shall first consider the problem of constructing a class of one-qubit completely positive (CP) maps that, in a certain limit, reproduce all classical Markov transition matrices on dichotomic probability distributions. We will then apply our construction to the first excitation subspace of a partitioned quantum cellular automaton structure (where one-qubit maps will be applied to the two-dimensional space spanned by excitations at neighbouring sites), obtaining a global dynamics on a lattice which is capable of describing excitation transfer. We will then present a study of the performance of classical versus quantum maps, showing by how much and under what conditions does quantum coherence improve the probability of excitation transfer through the lattice. Finally, we shall draw some conclusions and discuss outlook of this work. Let us remind the reader that the matrix (and map) T p,q is referred to as doubly stochastic if p = q. Any classical...