Measurements of NEP are demanding of resources, therefore still are not numerous. They also are not always representative for the whole growing season, and they are never representative for the whole life cycle of longer lived plants. Few observational series extend over more than 5 years. Within decades, a forest site that was chosen for measuring NEP will likely enter a different phase or be subjected to natural disturbances such as fire, pests, and wind-throw or management operations such as thinning and felling. Although measurements made from satellites may suggest that NPP has been increasing over the past decades in some regions and measurements of NEP may seem to indicate that many forests are significant carbon sinks, when the effects of disturbances are included the sequestration of carbon into these forests may be significantly less or even, in some areas, close to zero. Importantly, the disturbances themselves may be influenced by climate, and a change in the disturbance regime may turn a positive NBP into a negative one-or vice versa. For example, measurements of NEP in Canadian boreal forests suggest that many of these forests are carbon sinks (see Section 220.127.116.11), but estimates of NBP from measurement of changes in disturbance regimes and consideration of forest stand dynamics indicate that NBP may have declined significantly over the past 3 decades and that these forests, over large areas, are close to being carbon neutral (Walker et al., 1999). With the limited empirical data available, however, it is difficult to derive accurate local estimates of NBP for regions, biomes, countries, or continents from spatial and temporal integration of the constituent processes and disturbances (Schulze et al., 1999; Houghton et al., 2000). Our restricted ability to build estimates of NBP from its components at the present time defines a gap in our knowledge and the need to use other methodologies.
If the values in the preceding sections are considered representative for the major forest biomes-tropical, temperate, and boreal forests, respectively-and for the total area that they cover (approximately 40 x 106 km2), total NEP for these systems would be about 10 Gt C yr-1. However, global NBP, derived as the difference between the output resulting from fossil fuel burning, on the one hand, and the increase of atmospheric concentrations and net ocean uptake, on the other, is currently a little less than about 1 Gt C yr-1 (see Table 1-2). If the effects of land-use change are excluded from this estimate of NBP, the estimate of global NBP increases to about 2-3 Gt C yr-1-about five times less than the total NEP (see Section 18.104.22.168). Thus, NEP values as reported so far are clearly not representative of the large-scale, long-term storage of carbon. This fact emphasizes the importance of viewing the activities defined in Articles 3.3 and 3.4 of the Kyoto Protocol in a large-scale and long-term perspective.
Recent attempts to determine the large-scale distribution of terrestrial sources and sinks (i.e., NBP) indirectly, on the basis of the observed spatial variation of atmospheric carbon dioxide concentrations, have interpreted low values as indications of net uptake and high values as the presence of net sources (Fan et al., 1998; Bousquet et al., 1999; Rayner et al., 2000). In light of the relatively small magnitude of the regional NBP, this determination requires accurate knowledge about both the spatial distribution of carbon dioxide concentrations and its horizontal and vertical transport as a result of air motions. The inverse modeling analyses of Rayner et al. (2000) and Bousquet et al. (1999) agree reasonably closely in indicating a net terrestrial carbon sink in Siberia and comparable, smaller net sinks in North America and Europe, with small net sources in South America and Africa.
Other reports in this collection