Summary of project results by Mikko Peltoniemi
Finnish forests grow faster than ever before and they are regarded as sinks of carbon. Forests play also a role in international agreements as a mitigation tool for climate change. But how confident we are that forests are sinks of carbon at all?
United Nations' Framework Convention on Climate Change (UNFCCC) aims at reducing anthropogenic greenhouse gas emissions and in the Kyoto Protocol legally binding commitments were set to reduce emissions. The estimates of emissions and removals have to be appended with estimates of uncertainties.
The recent study shows that annual changes of carbon stocks (sinks or sources) in forest vegetation and soil are remarkable in comparison to GHG emissions in other sectors. Their role in the overall uncertainty of GHG inventory is even larger. To improve overall quality of the national GHG inventories it is necessary to reduce uncertainties of forestry sector sinks (Fig. 1).
Fig. 1 GHG emissions and removals in Finland in 2003 (error bar is 95% CI). From: Monni, Peltoniemi, Palosuo, Lehtonen, Savolainen, Mäkipää. 2006. Climatic Change in press.
The annual changes of carbon stocks in vegetation and mineral soils of Finland for 1989 - 2004 were based on a calculation combining forest inventory data, models of biomass, litter, and decomposition. Precision of the estimates was assessed with Monte Carlo analyses of uncertainty.
Factors that make forest sinks uncertain
The results showed differing factors controlling the uncertainty of vegetation and soil sink. This means that different actions are needed to improve the precision of these estimates.
Uncertainty of the estimates of growth variation and drain (harvests) dominated the uncertainty of vegetation carbon stock change. The consequent effects of growth variation on biomass allocation and forest litter production are completely unknown, and they were not part of the assessment.
Major source of uncertainty for the changes of soil carbon was the current estimate of carbon residing in soil. This uncertainty decreases with time elapsed since model initialization, and can be thus controlled to some extent. Thereafter, most important factors causing uncertainty were related to decomposition temperature, harvest residues, net forest area change, and soil model parameterization.
Large stocks, small changes
There are two profound challenges in forest carbon accounting. First, it is difficult to measure annual changes of carbon that are small in comparison to existing stocks of carbon. Second, there exists strong inter-annual variability within forest ecosystems, of which the data are limited.
Forest ecosystems have been studied in detail for decades and the understanding of processes has increased dramatically. However, intensive ecosystem studies do not provide regional up-scaling needed in carbon accounting. Forest inventories, on the other hand, concentrate on monitoring growing stock and growth of forests. They do not provide data, such as data on biomass allocation, litter or soil carbon, which could be used in forest carbon accounting
Greenhouse gas reporting
The uncertainties of the annual forest carbon stock changes are large in comparison to estimates themselves, and year-to-year variation of estimates is high. This means that high quality forest inventory data are needed to show that forests are not sources of carbon during the years of commitment period, as required by the Kyoto Protocol.
If forest carbon sinks are selected to compensate emissions, only a limited part of the sinks ('cap') is allowed to be credited. According to our results, in 1989-2004, the average carbon sink of Finnish forests on mineral soils (4.5 TgC / year) was safely over the cap (0.16 Tg C / year), even with 95% confidence interval. However, the low limits of annual estimates did not exceed the cap every year.
Annual reporting may be difficult to do in most countries but long term or average estimates calculated with the inventory-based method may still be available with reasonable precision. Due to the annual variability inherent in forest ecosystems, it would seem reasonable to report long term averages instead of annual estimates.