Ecosystem models are useful tools for evaluating environmental controls on carbon and water cycles under past or future conditions. In this paper we compare annual carbon and water fluxes from nine boreal spruce forest ecosystem models in a series of sensitivity simulations. For each comparison, a single climate driver or forest site parameter was altered in a separate sensitivity run. Driver and parameter changes were prescribed principally to be large enough to identify and isolate any major differences in model responses, while also remaining within the range of variability that the boreal forest biome may be exposed to over a time period of several decades. The models simulated plant production, autotrophic and heterotrophic respiration, and evapotranspiration (ET) for a black spruce site in the boreal forest of central Canada (56 ¿N). Results revealed that there were common model responses in gross primary production, plant respiration, and ET fluxes to prescribed changes in air temperature or surface irradiance and to decreased precipitation amounts. The models were also similar in their responses to variations in canopy leaf area, leaf nitrogen content, and surface organic layer thickness. The models had different sensitivities to certain parameters, namely the net primary production response to increased CO2 levels, and the response of soil microbial respiration to precipitation inputs and soil wetness. These differences can be explained by the type (or absence) of photosynthesis-CO2 response curves in the models and by response algorithms of litter and humus decomposition to drying effects in organic soils of the boreal spruce ecosystem. Differences in the couplings of photosynthesis and soil respiration to nitrogen availability may also explain divergent model responses. Sensitivity comparisons imply that past conditions of the ecosystem represented in the models' initial standing wood and soil carbon pools, including historical climate patterns and the time since the last major disturbance, can be as important as potential climatic changes to prediction of the annual ecosystem carbon balance in this boreal spruce forest. ¿ 2001 American Geophysical Union