Forestry has made an important contribution to the Swedish economy for many years, which is why the Swedish NFI was started in 1923. The importance of forestry

differs among countries and if it is not a key-category, the IPCC (2003) accepts a higher uncertainty (Tier 1) for reported carbon stock changes. Our evaluation of the consequences of using BEFs relies on the assumption that biomass functions result in good (close to unbiased) results. This assumption rests on the ability of biomass functions to adapt to different conditions (through the measured independent variables) in a manner that BEFs cannot do. Although BEFs are assumed to be constants our results show that they vary substantially over time, and we think that this is an important message to selleck chemicals people and countries involved in greenhouse gas reporting based on NFI-type data. Although not studied here, the default method might be an alternative approach to the stock change method (IPCC, 2003). When using the default method, changes in biomass for the living biomass pool may be estimated by applying BEFs to growth and drain. We argue that the risk of bias is probably higher when using the default method and will now try to discuss why: The Swedish NFI provides

estimates of stem volume and growth based on bore IDH tumor cores extracted from sample trees (on temporary sample plots). To obtain acceptable accuracy, the estimated growth is based on the last five fully developed annual year rings combined

with average data for 5 years. This means that the growth for recent years has to be extrapolated. The drain is probably underestimated as it is difficult in the field to judge whether the harvest occurred within the last year, and a proportion of stumps are usually unidentified; however, we have tried to eliminate Cyclin-dependent kinase 3 this underestimation by calibration from stock changes on permanent plots. Alternatively, harvests may be estimated indirectly from consumption or production statistics of harvested wood products. For both growth and drain we expect a large potential bias when converting volume to biomass. This bias may be reduced if separate BEFs are derived for growth and harvest. One advantage of using harvest statistics is the data is reasonably up to date but disadvantages include (i) both legal and illegal export/import need to be considered, (ii) the proportion of pulp that is biomass has to be known, (iii) the data does not account for natural mortality and (iv) harvest cannot be correlated with land use (harvest should be reported and recorded for several KP-activities). Thus, it is likely that the risk of systematic errors is higher using the default rather than the stock change method.