Models for the elastic material properties of wood aim to predict the mechanical response of wooden structures to external loading. Traditionally, the variability of these properties in trees is described by a taxonomy of growth defects that is typically based on visual indicators in the material. This includes curvature, knots, and spiral grain models. Existing meso- and macro-scale models fail to describe the uncertainty connected to the local heterogeneity of the material. In this paper, we propose a novel meso-scale model that describes the natural variability of Norway spruce morphology and material properties based on random field theory. Our approach removes the need for a taxonomy of growth defects and enables uncertainty quantification of the stiffness and density in a straightforward fashion using simulations. This may enhance confidence for stiffness-graded applications, where the dynamic resonant behavior of wood structures is relevant and growth anomalies are present. Further, our stochastic models can generate images that realistically mimic wood patterns, which is relevant for applications like synthetic wood panels and flooring.