This paper describes an advanced bottom-up approach for modelling the energy-environment sector to study greenhouse gas abatement. Three new features are described which give significant new capabilities to this class of models. These are: endogenization of end-use demands which allows computation of partial equilibria in energy markets; modelling future uncertainties using multi-stage stochastic programming; and, combining several bottom-up models as a multi-region model to explore issues of cooperation and burden sharing. Each of these new features is illustrated by results taken from large scale Extended MARKAL models of Québec and Ontario. The focus of the paper is on the nature of issues which can be addressed by this methodology, rather than on specific conclusions drawn from the discussed examples. We beleive that a very promising avenue of research lies in exploring the role of multiple advanced botton-up models in the integrated assessment of climate change.