This document describes the implementation of a Gaussian Markov random field sampler for forward uncertainty quantification in the Ice-sheet and Sea-level System Model (ISSM). The sampler generates realizations of Gaussian random fields with Matérn covariance to characterize spatially varying uncertain input parameters in ice sheet models as random fields. It is based on representing such random fields as solutions to a stochastic partial differential equation, which is then discretized using finite elements. This provides a computationally efficient way to generate random field samples on complex ice sheet model meshes. The implementation is tested on synthetic problems and applied to assess uncertainties in projections of Pine Island Glacier retreat.