The development of forests in Pukaskwa National Park, Ontario, Canada, was simulated over 150 years to investigate boreal carbon dynamics and to test the feasibility of simulating large tracts of heterogeneous boreal forest. Pukaskwa National Park, located on the north shore of Lake Superior, encompasses 1835 km^2 of the Superior Section of the boreal forest. A patch model was developed called BOPAS (BOreal PAtch Simulator), to simulate the development of carbon pools in a function of environmental parameters. Using GIS techniques, the park was divided into patches defined by a unique combination of forest type, age, climatic variables, soil type and topography, and then a forest gap model was used to develop biomass-over-time relationships for each patch type. BOPAS uses these relationships to simulate the development of carbon pools for trees, moss and litter/humus. Results are reported for constant climate, but BOPAS can be easily adapted to changing climatic scenarios. Good results were obtained for predictions of carbon storage in trees. The initial value was 3.61 kg C m^-2, which agrees closely with literature values. With no disturbance, tree carbon increased to a maximum of 3.97 kg C m^-2 at 30 years then slowly declined. Carbon storage was stabilized by introducing fire as a disturbance with a return interval of 100 years. Predicted forest floor carbon density, however, was much lower than expected, being less than half that of trees. It was anticipated to be substantially higher than tree carbon density based on a preliminary survey in the park and values reported in the literature. Published data, however, are very limited in coverage and give such a wide range of values that it was impossible to draw any firm conclusions about the validity of the model. BOPAS also showed that the forest floor carbon pool was relatively constant over the timescales of the simulation, but no published data were available to test this prediction. In summary, this work has demonstrated the feasibility of the BOPAS approach, but has highlighted the necessity for more extensive data on forest floor carbon storage and dynamics.