Canadian Forest Service Publications

Uncertainty of inventory-based estimates of the carbon dynamics of Canada's managed forest (1990-2014). 2017. Metsaranta, J.M.; Shaw, C.H.; Kurz, W.A.; Boisvenue, C.; Morken, S. Canadian Journal of Forest Research 47(8):1082-1094.

Year: 2017

Available from: Northern Forestry Centre

Catalog ID: 38890

Language: English

CFS Availability: PDF (request by e-mail)

Available from the Journal's Web site.
DOI: 10.1139/cjfr-2017-0088

† This site may require a fee.


Canada’s National Forest Carbon Monitoring Accounting and Reporting System (NFCMARS) quantifies the carbon (C) dynamics and greenhouse gas (GHG) emissions and removals of Canada’s managed forest to fulfill reporting obligations under international climate conventions. Countries are also requested to assess the uncertainty associated with these estimates, which we report here. We used Monte Carlo simulation to quantify uncertainty of carbon stock and flux estimates from the Carbon Budget Model of the Canadian Forest Sector (CBM-CFS3), the core ecosystem model of the NFCMARS. We evaluated the impacts of model algorithms, parameters, and the input data used to describe forest characteristics and disturbance rates. Under our assumptions, 95% confidence interval widths averaged 16.2 Pg C (+8.3 and –7.9 Pg C, or ±15%) for total ecosystem C stock and 32.2 Tg C·year−1 (+16.6 and –15.6 Tg C·year−1) for net biome production relative to an overall simulation median of –0.8 Tg C·year−1 from 1990 to 2014. The largest sources of uncertainty were related to factors determining biomass increment and the parameters used to model soil and dead organic matter C dynamics. Opportunities to reduce uncertainty and associated research challenges were identified.

Plain Language Summary

This paper calculates the statistical uncertainty in estimates of the amount of carbon in Canada’s managed forest. It also shows how much this uncertainty changed every year between 1990 and 2014, as calculated by Canada’s National Forest Carbon Monitoring Accounting and Reporting System (NFCMARS). All statistical estimates have some uncertainty, but this must be explained carefully because people have a tendency to misinterpret uncertainty. This report describes the different factors causing uncertainty, calculates the total result of all of these factors put together, and describes why it is difficult to come up with the “confidence interval” or range for the estimate. The paper is important because calculating a confidence interval for the projections from forest models is rarely done, and this is one of the first times that anybody has tried to do this. Models are often used to decide which course of action to take, and we could make the wrong choices if we do not know what the uncertainties are. International climate agreements have strongly suggested that models that are used to calculate how much carbon is in forests become more open about how they make their calculations and what the uncertainty is. This paper shows that Canada is serious about following through on this suggestion.

Date modified: