Comparison of uncertainties in carbon sequestration estimates for a tropical and a temperate forest

We compare uncertainty through sensitivity and uncertainty analyses of the modelling framework CO2FIX V.2. We apply the analyses to a Central European managed Norway spruce stand and a secondary tropical forest in Central America. Based on literature and experience we use three standard groups. to express uncertainty in the input parameters: 5%, 10% and 20%. Sensitivity analyses show that parameters exhibiting highest influence on carbon sequestration are carbon content, wood density and current annual increment of stems. Three main conclusions arise from this investigation: (1) parameters that largely determine model output are stem parameters, (2) depending on initial state of the model, perturbation can lead to multiple equilibrium, and (3) the standard deviation of total carbon stock is double in the tropical secondary forest for the wood density, and current annual increment. The standard deviation caused by uncertainty in mortality rate is more than 10-fold in the tropical forest case than in the temperate managed forest. Even in a case with good access to data, the uncertainty remains very high, much higher than what can reasonably be achieved in carbon sequestration through changes in forest management. (C) 2008 Elsevier B.V. All rights reserved.