Skip to main content
SLU publication database (SLUpub)

Doctoral thesis2012Open access

Sink or source? : uncertainties in large scale model predictions of forest soil organic carbon dynamics

Ortiz, Carina

Abstract

Soil organic carbon (SOC) is the largest terrestrial carbon pool and small changes in this pool may affect the global carbon balance, especially atmospheric concentration of CO2. Within the context of climate policy, the quantification of these changes is important, as pool changes may affect a country's national greenhouse gas budget. The aim of this thesis was to assess and analyze uncertainty related to the up-scaling of modelled SOC stocks and change estimates to regional or national scale. Two process-based models Q and Yasso07, were used to estimate SOC stocks and changes at different regional scales in Swedish coniferous forests. The parameter uncertainty of the Q model was assessed and established with the Generalized Likelihood Uncertainty Estimation (GLUE) method through the Swedish Forest Soil Inventory (SFSI) data at county scale. The calibration resulted in a set of parameters that were used for further modeling at regional scale. The Q and Yasso07 models were used to assess the impact of different uncertainties in the SOC stocks and changes. The most important uncertainty source in the model estimates was litter production. Increased harvest residue extraction was analyzed with the Q model to study the effects on SOC accumulation. SOC accumulation decreased with increased harvest residue extraction, although there was temporal and geographical variation. However, increased emissions from changes in the SOC pool resulted in a net decrease in CO2 emissions due to the substitution of coal combustion with biofuels. The coherence of scales between large-scale inventory data and process-based simulation models was explored. Inventory data became more uncertain when going from national to regional scale, due to the smaller sample, whereas, model estimates became more uncertain when applied to larger areas, due to increased uncertainty in parameter determination at larger scales resulting from varying conditions. The magnitudes of the uncertainties for model and inventory estimates of SOC were comparable, but the origins of uncertainties differed and could not be compared. Both models and inventories can be used to estimate the carbon sink of Swedish forest soils at national level, but if the changes are small, a few ‰ yr-1 in the SOC pool, the uncertainty may prevent a definite answer, if there is a change in the SOC pool.

Keywords

SOC; GLUE; Swedish Forest Soil Inventory; models; uncertainty; GHG reporting; sensitivity analysis; coniferous forests

Published in

Acta Universitatis Agriculturae Sueciae
2012, number: 2012:25
ISBN: 978-91-576-7661-0
Publisher: Dept. of Soil and Environment, Swedish University of Agricultural Sciences