Timber supply has traditionally been modelled using aggregate data. In this paper, we build aggregate supply models for four roundwood products for the US state of North Carolina from a stand-level harvest choice model applied to detailed forest inventory. The simulated elasticities of pulpwood supply are much lower than reported by previous studies. Cross price elasticities indicate a dominant influence of sawtimber markets on pulpwood supply. This approach allows predicting the supply consequences of exogenous factors and supports regular updating of supply models.