The aim of this paper is to investigate the impact of the ground-material uncertainty on the properties of electromagnetic fields generated by lightning strikes. Although several analytical as well as numerical approaches have been proposed throughout previous years for the prediction of lightning-induced fields, these are mostly applied within a fully deterministic context, thus overlooking the degree of randomness that the corresponding problems may exhibit. Considering the rotational independence of setups with vertical lightning channels and flat terrains, we propose a stochastic finite-difference time-domain algorithm that takes advantage of the generalized polynomial-chaos theory. The latter enables the representation of random processes via expansions with respect to specific orthogonal polynomials and facilitates capturing the fundamental statistical properties of the produced fields. The resulting scheme preserves the structure of the original Yee algorithm and, although computationally heavier than the latter, is clearly more efficient than conducting brute-force Monte Carlo studies. The potential of the proposed stochastic method is assessed via various numerical results, calculation of statistical metrics, and comparisons with reference solutions. KEYWORDS finite-difference time-domain method, lightning, polynomial chaos, uncertainty Int J Numer Model. 2018;31:e2238.wileyonlinelibrary.com/journal/jnm