Quantifying uncertainty in studies of forests is important to establish the significance of findings, make predictions with known confidence, and guide investments in research and monitoring. This symposium will address sources of uncertainty in estimates of carbon and nutrients in forest soils, above- and belowground biomass, and ecosystem inputs and outputs.  Presentations will address sources of uncertainty in forest ecosystem studies, including natural spatial and temporal variation, measurement error, model uncertainty, and model selection error.  Examples include the importance of spatial variation in detecting change over time in soil stores and measurement error in forest inventory due to identifying or classifying trees, measuring them, and determining whether trees are live or dead and in or out of a plot.  Model uncertainty within and across models is important in biomass estimation and climate predictions. Presentations will also address how these uncertainties influence monitoring designs or affect management and policy decisions.