Framework for Analysis of Error
Propagation in National Carbon Budgets
C. J. Cieszewski, D. P. Turner, and D. L.
Phillips
Abstract:
Existing national-level carbon budget studies
typically do not conduct formal uncertainty analyses. In cases where attempts
at addressing carbon budget uncertainties have been made, they are generally
limited to informal variations of sensitivity studies. The reasons for limited
effort towards uncertainty analysis in carbon budget studies include: 1) unknown
or poor statistical properties of the input variables; 2) use of arbitrary
relationships and analytically intractable discrete relationships in carbon
budget models; 3) a high level of aggregation and complexity in calculations;
and at times, 4) the interactive nature of the developed carbon budget models
which does not lend itself to the application of a Monte Carlo approach.
Development of a generic uncertainty analysis framework for carbon budget
studies is proposed here. Two initial steps in this framework are illustrated.
First, equations are developed for marginal uncertainties of carbon budget
terms. This allows a sensitivity analysis to determine the most important
sources of error to include in Monte Carlo simulations. Second, equations are
developed to summarize changes in five forest carbon pool components with stand
age, for a large number of stand types and management practices in U.S.
forests. This is a necessary first step for application of the error
propagation analysis framework to the U.S. forest carbon budget.
Author Keywords:
KeyWords Plus:
Addresses:
Cieszewski CJ, Univ Georgia, Sch Forest Resources, Athens, GA 30602 USA
Univ Georgia, Sch Forest Resources, Athens, GA 30602 USA
D. P. Turner,Forest Ecologist, ManTech
Environmental Research Services Corp., Corvallis, OR.
D. L.
Phillips, Research Biologist, U.S. EPA
National Health & Environmental Effects Laboratory,
Corvallis, OR.
Publisher:
EPA
IDS Number:
ISSN: