The approach of FRA 2000 by FAO was the reliance on the participation of individual countries for both supply and analysis of information. It is hoped that this approach will lead for further capacity building in countries (FRA 2000 –main report). While countries firmly support this approach, it has sometimes been criticised on the basis that country information may be inaccurate, incomplete, out-of-date, or internationally inconsistent (Stokstad, 2001; Czaplewski, 2002). According to the FRA 2000 main report, many countries still lack reliable primary information at the national level. Some examples of country level changes in FRA 1990 and FRA 2000, as well as reliability assessment of TBFRA 2000, support this concern. One goal of the future assessments will be to further strengthen country capabilities and participation. In this way, FAO intends to improve the information quality as well as to assist developing countries in their inventories. FAO should also work towards reducing the interval between successive assessments, or towards the establishment of continuous regional assessments.
FAO conducted a remote sensing study of tropical forests in FRA 2000 for assessing the area changes between 1980-1990 and 1990-2000. Stratified sampling with a sampling ratio of 10 % was applied. The purpose was to independently evaluate the data quality of country information and to improve FAO’s understanding of land-cover change processes in the tropics, especially deforestation, degradation, fragmentation and shifting cultivation, among others.
FAO also plans to continue to use country information and independent remote sensing surveys in future assessments, but also to emphasise field observations as a means of gathering broad and representative information.
For this purpose, the Forestry Department of FAO is developing a Global Forest Survey (GFS) (FAO, 2000). GFS have two main development objectives: 1: National Capacity building for sustainable forest management and 2) Forestry knowledge management for international processes. In addition to improving forest management and forestry knowledge management, the first objective is ‘to support the member countries in establishing continuous forest resource assessment and monitoring systems and in developing their national capacities for the purpose’, and ‘to provide technical assistance to member countries in building national capacities in the areas of methodology development, use of new technologies and establishment of forest monitoring systems.’ The second objective is ‘to provide assessment variables required by international forestry processes, and establish feedback mechanism to evolve the collection of information and reporting over time’, ‘to develop a practical approach for the assessment and monitoring of the forestry resources at country and global levels that meets the concurrence of the member countries and donor community’, and ‘to establish, under the leadership of the FAO, partnership and co-operation linkages with well-established national and international institutions for the carrying out world wide forest survey in continuous basis’. According to the Working Paper (FAO, 2000), GFS ‘should be regarded as a component of and development of FAO's global resource assessments, for which FAO has received a strong mandate from its member countries’.
The objective of this paper is to discuss opportunities to carry out a remote sensing aided forest resource survey (RSFS), independent of the countries' own inventories for the whole globe. The emphasis is in the global level change estimates. For technical purposes, large region level estimates are also considered, e.g., estimates for Europe, North America, South America, with a possible differentiation into Temperate, Boreal and Tropical zones. The global level estimates are constructed from the large region level estimates.
The different opportunities for remote sensing data acquisition are presented. The properties of the sensors are presented in Appendix 2. Parameters that require quality control most urgently and for which the available resources give possibilities in RSFS are areas of forest land (FL), other wooded land (OWL) and other land (OL) as well as their changes. Tree stem volume and biomass are also key variables in assessing the status of worlds forests. The estimation of these variables requires thorough field measurements. The parameter estimation methods are discussed. We also demonstrate how the sampling designs and related errors could be evaluated from existing data sources, e.g., global land cover maps. This analysis is one of the main emphases of this paper and should be carried out with sufficient resources for a possible RSFS. The aim is neither to present real final sampling designs nor error estimates but rather outline methods which could be used in evaluating different designs when enough resources are available. Our study is presented in Appendix 1. The FRA 2000 global forest map of FAO (Zhu and Walter 2001) is applied in our study. Other land cover maps will probably be available when planning and realising a possible RSFS.
One of the basic questions in RSFS will be the availability of field data. Field sampling is needed because remote sensing based forest resource surveys should to be supported by field observations/measurements. Field sampling intensity depends on the available budget, variability of the target parameters in the field and the applied remote sensing material. Often, a minimum number of field plots is required for each image scene. A successful relative calibration of images may reduce this need.
Field measurements and designs are out of the focus of this paper. Possible rough cost estimates are presented if an independent sample will be collected. A possible field sampling design could be carried out utilising global land cover maps and same principles as presented in Appendix 1. Other possibilities for the field data include the GFS data or countries own data, provided by country correspondents and possibly evaluated by means of a sampling approach. GFS could also be used for a local evaluation of RSFS estimates.
The alternatives of RSFS vary depending on the ground sampling and remote sensing sampling densities. The variation can be illustrated in the space of 'Remote sensing intensity' and 'Field work intensity' dimensions (Figure 1). Both sampling densities are affected by the applied remote sensing material. The sampling intensities may vary also by regions, see Chapter 2.
Figure 1: Examples of different sampling designs (adopted from Peter Holmgren)
