Land-use and forestry projects generally are easier to quantify and monitor than national inventories because of the clearly defined boundaries for project activities, the relative ease of stratification of the project area, and the choice of carbon pools to measure (Section 220.127.116.11). Techniques and methods for sampling design and for accurately and precisely measuring individual carbon pools in LULUCF projects are based on commonly accepted principles of forest inventory, soil sampling, and ecological surveys (Pinard and Putz, 1996, 1997; MacDicken, 1997a,b; Post et al., 1999; Winrock International, 1999; Hamburg, 2000). For example, there is a wealth of experience with well-developed and accepted methods to inventory forests for merchantable volume and growth; these methods can be and are being readily adopted to inventory forest biomass carbon. Likewise with regard to measuring soil carbon, standardized techniques are well established. Further descriptions of methods for estimating the carbon pool in live tree biomass, understory and herbaceous plants, roots, fine and coarse litter, and soil are described in Chapter 2. Standard methods have not been universally applied to all projects, however, and methods of accounting for carbon benefits have not been standardized-resulting in some difficulties in comparing results across different LULUCF projects.
For most LULUCF projects, it would be necessary to measure non-project reference or control sites as well. These sites must be sufficiently similar to the project area to serve as valid proxies under the assumption that the project was not implemented (Vine et al., 1999). To help overcome the difficulty of establishing proxy areas, non-project reference sites could be identified during the project design phase. The location of proxy sites as close as possible to the project would be the most desirable situation. For example, in projects in which many small landowners convert to no-till agriculture, proxy areas would be farmers in the area who do not practice no-till agriculture. Box 5-3 illustrates the types of measurements being taken to estimate with- and without-project cases and the resulting carbon benefits.
The total carbon stock has been measured to <10 percent of the mean with 95-percent confidence in several pilot LULUCF projects (e.g., Programme for Belize, 1997a; Hamburg, 2000; NKCAP-see Box 5-3). Although techniques and tools exist to measure carbon stocks in project areas to a high degree of precision, the same level of precision for carbon benefits may not be achieved. The carbon benefit per unit area of land is the difference between the carbon stocks in the with-project case-which is high if, for example, the project is conserving carbon in existing forests through an avoided deforestation project-and the carbon pools in the without-project case, which is low if the baseline is agricultural or degraded lands. In this case, the estimated carbon benefit is likely to be high (a small carbon stock subtracted from a large carbon stock), and the error estimate, expressed as a percentage of the mean difference, likely to be small and similar to that obtained for carbon stocks in the forests. As the difference between the with- and without-project cases decreases as, for example, in reduced impact logging projects, however, the percentage error of the carbon benefit increases. To reduce this error, monitoring can be designed to measure the change in carbon stocks directly, as in the NKCAP (see Box 5-3). Difficulties in establishing baselines and leakage effects also affect the precision of the carbon benefits.
Other reports in this collection