In March 2001, FAO's FRA program presented a proposal to start a new approach to support National Forest Assessments through in-country capacity building activities. The long term objectives of the new approach was set to "contribute to the sustainable management of forests and trees outside forests (TOF) by providing decision makers and stakeholders with the best possible, most relevant and cost effective information for their purposes at local, national, and international levels" (Saket et al, 2002). The Committee on Forestry, FAO's governing body in the forestry sector, supported the idea and recommended in its final statement that "FAO continue its efforts to carry out broad assessments that included various aspects of forest resources, such as biological diversity, forest health, and resource use" (FAO, 2001).
According to the guidelines for the FAO approach to National Forest Assessments (hereafter FAO-NFA) the inventory component is based on a low-intensity, systematic sampling method, and includes the establishment of permanent field plots and the collection of data on the biophysical status as well as the management and uses of forest resources. Depending on the country's objectives with the assessment, the variability of both biophysical and demographic characteristics as well as available financial resources, a sampling design with the appropriate sampling intensity is developed by creating a nation-wide sampling grid. Each intersection in the sampling grid constitutes a site or "tract", where the field measurements are carried out. Each tract has a total area of 1.0 km2 and consists of a cluster of four sample plots. In the four sample plots, field personnel contracted by the host government measure several biophysical characteristics according to a FAO-NFA field manual, developed jointly between FAO and the corresponding host government.
Unlike many traditional forest inventories the FAO-NFA does not limit itself to measurements of tree and forest resources in forests, but includes measurements of such resources outside of forests. In fact, it is the explicit effort not to distinguish between forest and non-forest areas, that allows the FAO-NFA to get more accurate estimates of land use changes over time, whether these involve forests or not.2 The main bio-physical variables measured include: ecological zone, primary land use classes, forest types, soil, topography, as well as the location, structure, condition, dbh, and height for all trees on the plot.3
The measurement of the variables associated with the human use of tree and forest products in the site follows a different procedure. According to the FAO-NFA guidelines, field personnel should document what goods and services are derived from the site; what the relative importance of each of those are for different local people; who has the right to harvest what products - when and how; what their end use and purpose are; and whether the demand for and supply of these products are stable, increasing or decreasing.
To capture this information, the field personnel conduct interviews with local forest users who either extract resources from the site measured, or who have information about the products extracted. After interviewing a number of users for each site, the field team interprets the information obtained and enters the information into a form. As we shall see later in this report, there are several methodological challenges in ensuring that these interview variables are measured in a consistent and accurate way across the entire country. These critical issues will be discussed in some depth in sections 4 and 5 of this report.
Once fieldwork is completed, the gathered information is compiled and analyzed at the national level. The measurements of each variable form the basis for statistical estimates of a series of national level parameters, which are then reported to the government and to several international processes, including the global Forest Resource Assessment coordinated by the FAO.
The NFA initiative has the potential to contribute to better forestry policies both at the national and international levels. Because of its low costs and accompanying technical advice, NFAs have now become possible for many countries that have never carried out forest assessments based on field measurements. Before effective policy responses to forestry-related problems can be designed, policy makers need information about the problems' relevance, prevalence, scope and particular characteristics. NFAs help meeting the existing information needs at the national and international levels
As the publication of past FAO-FRA reports reveal, analysis of national and global forest change can be a polemic issue, and the credibility of the new FAO-FRA approach is likely to be critically assessed by the different FRA information-users. The reliability of the NFA methods and accuracy of its results are likely to be questioned by those actors who feel threatened by the possible implications of the NFA report. To ensure that these users perceive the new approach as an improvement regarding the quality of national and global level information, it is important that the new reports explicitly account for how the NFA methods contribute to more reliable and accurate data. This report argues that the credibility of the FAO-NFA approach would be strengthened if FAO incorporated a series of quality tests into the new approach, particularly as it relates to the interview component of the NFA protocol. In this report, I show how such a testing procedure could be developed and carried out.
The next section provides a more in-depth discussion of what indicators that the global FRA process hopes to produce in the future and how this list of variables corresponds to the data actually collected. I will use the case of Guatemala for this analytical exercise.
2 The non-distinction between forest and non-forest lands is particularly useful for estimates of national forest area, which may be the most frequently used indicator of forest condition. A recent pilot study in Costa Rica produced standard error of estimates for the national forest area of about 10% (Klein and Ramirez, 2002).
3 Each individual TOF with dbh >10 cm is measured, and for trees inside the forest, each tree with dbh > 20 cm. In addition, sub-plots are established in forest areas where trees with dbh between 10-20 cm are measured in smaller, circular and rectangular sub-plot seedlings with dbh < 10cm and height > 1.3 m are registered.
