Estimating the uncertainty of modeled carbon sequestration: The GreenCert™ system

The GreenCert (TM) system was developed to help farm and ranch owners to quantify, standardize, pool and market CO2 emissions offset (sequestration) credits derived from improved rangeland or cropland management. It combines a user-friendly interface with the CENTURY biogeochemical model, a GIS database of soil and climate parameters, and a Monte Carlo-based uncertainty estimation methodology. This paper focuses on uncertainty treatment, discussing sources of error, parameter distributions, and the Monte Carlo randomization approach, culminating in a sensitivity analysis of model parameters. Idealized crop and grazing scenarios were used to evaluate the uncertainty of modeled soil organic carbon stocks and stock changes stemming from variability in site and management parameters. Normalized sensitivity coefficients and an integrated index for relative sensitivity of the model to the ensemble of the tested variables indicate that environmental factors are the most important in determining the actual size of the soil carbon stock, but that management is a much more important determinant of short- to medium-term carbon fluxes. GreenCert (TM), uses the patented C-LOCK (R) approach to efficiently limit uncertainty in the most critical phase of the modelling process by maximizing the use of available management information, and quantifies the remaining uncertainty in an unbiased fashion using Monte Carlo parameter randomization. (C) 2010 Elsevier Ltd. All rights reserved.