The main focus for the Global Forest Survey is information requirements at the national level for forest policy development and national and local implementation of sustainable forest management. Although the general needs for forest and forestry information for policy and planning are well understood, some specific aspects are worth mentioning as a background:
- The overall forestry objectives are well expressed in the Sustainable Forest Management concept, although the specific goal functions may be defined differently throughout the world. Management is to make and implement decisions that, in the long perspective, lead towards the overall goals. This is the classical forestry planning problem. It is usually solved by collecting relevant information and developing scenarios and decision rules. Limitations in the base information are transformed into uncertainties in the planning processes. It is therefore essential to let the decision processes (national and local) guide information requirement analyses;
- National and local forestry planning is seldom dominated by wood production and utilisation. Instead, non-wood products and services, as well as interactions with other land uses are important. Information requirements are therefore multi-disciplinary and include conventional forest inventory parameters (eg volume and growth), other benefits from the forest (eg biodiversity and soil conservation), socio-economic factors (eg land tenure and legislation/policy), as well as pure economic variables (such as market information and substitution effects);
- The information needed for forest management is overwhelmingly local in nature. Although some simplified parameters may be useful on an aggregated national or international level, forest management decisions are dominated by local considerations. Existing national forest inventories have been designed as a large sample of small plots, to ensure that measurements and observations are made in the scale where management decisions are to be made. By aggregating such local observations, estimates for larger areas can be obtained, while at the same time keeping the local scale and variation in the material. Only by preserving this high resolution in data can realistic analyses and scenarios be done for larger geographical units.
International information on forests and forestry usually refers to information that can be compared between countries. This immediately points at the main obstacle in global assessments and monitoring – how can information be made comparable despite the varying climatic, political and socio-economic conditions throughout the world?
It is not difficult to find arguments for standardized collection of forest and forestry information that can be used for international analyses or comparisons (eg Lanly 1996, Lund 1996). The agreements from the United Nations Conference on Environment and Development (UNCED) 1992 is often called upon. The United Nations conventions on biological diversity, climate change and desertification are other milestones. On a more general level, international economic analyses and forecasting require reliable input concerning the forestry sector. The implementations of these international processes all require monitoring of the forest (and other) ecosystems including the production of goods and services, as well as the legal and policy frameworks for the management of land and natural resources.
Besides the formal use of information in the international processes, the requirements from the public at large is important and seems to increase. With the help of more sophisticated information technology, the access to information has improved rapidly over the past decade. Whether communicated through television, printed media or the Internet, information reach more individuals than ever. The demand for public information can be expected to continue to grow rapidly. Public consumption of environmental information would typically be related to the international processes mentioned above. These processes are politically and financially important to many countries and organisations, and the provision of unbiassed and reliable information to the general public is therefore an issue.
It is obvious that international assessments of forests and forestry has developed from a timber-oriented mode some decades ago, to include broad environmental factors, as well as socio-economic aspects. The imaginary land use boundary between forests and agricultural land has become vague as trees are increasingly grown outside the forests, and the values of non-wood forest products are being accounted for. Several environmental parameters, such as carbon cycling and biodiversity, are not confined within land use classifications. It is clear that international assessments of forests and forestry should increasingly be cross-sectoral and include interactions with agriculture and remote benefits from the forest.
Very few developing countries in the world have up-to-date information on their forestry resources and less have national capacity for generating such information. The present status of the baseline information on the resources in the developing world may be deduced from the statistics given in Table 1.
Table 1. Forest inventories at national or local levels in developing countries
|
Continents |
n countries |
n countries |
Forest Cover |
Other Parameters | |||||||
without forest inventory |
partial forest inventory |
National Forest Inventory | |||||||||
repeated1 |
Single shot2 |
state |
change |
volume |
biomass |
biodiv. | |||||
after 1990 |
before 1990 | ||||||||||
Francophone Africa |
31 |
7 |
10 |
0 |
8 |
6 |
|||||
Anglophone Africa |
25 |
6 |
5 |
8 |
4 |
74 |
|||||
Near East |
13 |
11 |
0 |
2 |
0 |
0 |
|||||
Asia3 & Oceania |
22 |
1 |
2 |
3 |
13 |
3 |
10 |
11 |
12 |
||
Latin America |
17 |
3 |
14 |
4 |
6 |
4 |
9 |
1 |
5 |
7 | |
Caribbean |
29 |
2* |
5 |
12 |
17 |
||||||
Totals |
137 |
28 |
33 |
22 |
43 |
34 |
|||||
Notes:
1) "Repeated" is interpreted as systematic monitoring or inventories in a fixed interval.
2) Most countries have more than one single shot inventory, the timeliness (after or before 1990) refers to the most recent one.
3) Asia: here referring to 22 countries in the sub-regions South Asia, Continental SE Asia, Insular SE Asia and East Asia.
The number of countries that have set up monitoring systems for their forest resources through repeated inventories is very limited (22 out of 137). Although the highest number of countries (47) fall under the column of single-shot national inventory conducted more than a decade ago, there is a large number of countries which has done nothing to-date. For clarification, the numbers do not refer merely to studies of forest inventory as such but also include various mapping projects such as land use, soil studies, etc., that provide only limited forestry information - typically the extent of forest cover.
Very few developing countries have conducted thorough national inventories. In the developing world, the African continent has the least information on forest resources, as no country has to-date mounted a monitoring system and few have carried out single-shot national inventories.
Most of the repeated national inventories in the developing countries have been conducted on the basis of different methodologies, which do not produce information on various changes, including forest growth. The national inventories have been prepared for various objectives. Some of the national surveys were undertaken for timber potential determination or general policy development, others for environmental assessment, etc.
The disparity in extent and values of the baseline information at country level within the developing world suggests that international forestry ambitions seem to have failed in addressing the issue of monitoring the forestry resources in the developing countries.
Table 3 provides a time line that, although the figures are not absolutely comparable, is rather depressing. It is a fact that substantial resources have been spent on support to inventories in many countries over the past decades. At some stage most countries therefore have had adequate information about their forest resources, and a capacity to undertake inventories. However, despite remote sensing and modern techniques, the knowledge and capacity often appear to have gone down. In Africa the number in the low category has increased between 1970 and 1990. In Asia and Latin America there are positive trends, but these improvements are partly due to satellite remote sensing projects that do not provide an in-depth assessment of the forest resources. Table 4 illustrates that the field based surveys have declined over the past decades, and that remote sensing and household survey based approaches have increased. As a result, there are less direct observations of forestry processes now compared to ten or twenty years ago.
Table 3. Quality of forest cover information in global assessments 1970-2000 (number of countries, L=low, M=medium and H=high)2
1970 |
1990 |
2000 (change)1 | |||||||
Region |
L |
M |
H |
L |
M |
H |
L |
M |
H |
Africa |
8 |
29 |
2 |
13 |
21 |
5 |
23 |
15 |
5 |
Asia |
7 |
7 |
2 |
1 |
11 |
4 |
2 |
6 |
6 |
Latin America |
6 |
13 |
4 |
4 |
12 |
7 |
4 |
12 |
2 |
1) 2000 figures relate to quality for forest cover change estimations on the national level.
2) Sources: Reidar Persson and FRA 2000 preliminary results.
Table 4. Distribution of deforestation assessment methodologies over time in percent of published studies. Note that no distinction as to the size of the studied area is made. (Rudel et al. 2000)
Remote sensing |
Survey1 |
Field observation |
Secondary sources |
TOTAL | |
Pre-1980 |
8 |
8 |
39 |
45 |
100 |
1980s |
8 |
5 |
30 |
57 |
100 |
1990s |
17 |
15 |
20 |
48 |
100 |
1)Survey refers to household surveys and similar approaches.
There is a strong and a pressing need for a new vision and approach in connection to forest resources assessments in these countries as well as with the national capacity building within the framework of partnership with international organisations. From the mandate point of view, FAO is the international body that should reconsider this co-operation with the developing countries by designing a more fruitful approach for forestry resources monitoring that will benefit more not only the countries themselves but also the international community for global policy adjustments.
National information on forestry resources and related variables in the industrialized countries are often obtained from forest inventories using sampling techniques. The Nordic countries have been reported to be the first ones at planetary level to carry out national forest inventories. The first exercises have been undertaken in 1919 in Norway, 1921-24 in Finland, and 1923-24 in Sweden. The central and western European countries started their national forest surveys after the second world war, while in the United States of America, the first reported work has been conducted in 1930 in Oregon. Initially, the surveys were conducted using rudimentary and tedious techniques although generated first set of valuable information. The techniques have improved by introduction of sound statistical methods and use of new technologies such as computers and GPS.
The sampling designs are based on systematic sampling using permanent and in some cases temporary plots installed in grid nets of different spacing. The random sampling design has been discarded. There are sampling designs, which adopted point-sample plots and others fixed area plots. Both techniques were also used together. Size and shapes of the sample plots differed from country to country. Single sample plots are used in some countries while concentric plots or clusters are used in others. Distribution of plots in clusters also varied. They are located either at the corners of squares, along two lines with right angles or along strips or transects. The spacing between plots varied also from country to country and even within the country itself depending on the forest distribution and topography (Table 5).
Table 5. NFI sampling design for selected industrialized countries and China1.
Country |
Sampling technique |
Grid spacing |
Type of Sample |
Sample form |
Sample Size |
Sample Unit Size |
Sample Unit Shape | |||
USA |
Systematic |
˜ 1x1km |
Permanent |
Fixed area |
120,500 |
1 acre (0.405ha) |
Square? | |||
Sweden |
Systematic |
? |
Permanent Temporary |
Fixed area |
? |
314 m2, 157 m2, |
300m to 1800m Tracts (clusters) | |||
Norway |
Systematic |
3x3 km |
Permanent Temporary |
Fixed area Arealess |
P: 11000 |
1000m2, 250m2, 100m2 |
Concentric circles | |||
Finland |
Systematic |
8x8 km |
Permanent |
Fixed area Arealess |
4100m x 4100m Tracts (clusters) | |||||
Switzerland |
Systematic |
1x1km |
Permanent |
Fixed area |
11,000 field plots out of 41,000 photo plots |
2 (0.8ha) and 5 (2 ha) acres |
Two concentric plots at the intersection of gridlines | |||
Italy |
Systematic |
3x3km |
Permanent |
Fixed area |
0.007% intensity |
600m2 |
Circular | |||
France |
Systematic |
0.548m to 0.632m |
Temporary |
Fixed area |
Aerial photo plots: 2%, Field plots: 0.07% |
707m2, 254.5m2, 113.1m2 |
Circular concentric of 15, 9 and 6 m. independent plots of 2.26m radius for regeneration | |||
Austria |
Systematic |
2.75x 2.75km |
Permanent Temporary |
Arealess |
4 pots/756.25 ha |
Relascope |
200m x 200m tracts (clusters) | |||
Germany |
Systematic |
4x4km, 2.83x2.83km, 2x2km |
Permanent |
Arealess Fixed area |
12,580 tracts (forest) out of 28,978 |
Relascope |
150m x 150m tracts (clusters | |||
Australia |
National inventory does not exist. There are locali surveys whose methods are not yet known to the author s of this paper | |||||||||
China |
Systematic |
6x6km |
Permanent |
Fixed area Arealess |
? |
? |
Rectangular cluster samples | |||
1) Sources: Proceedings of the International IUFRO S. 4.02 and S. 6.04 Symposium, May 14-16, 1990, Birmensdorf, Switzerland; Study on European Forestry Information and Communication System, Vol. 1 & 2, European Commission; Journal of Forestry, Vol. 97, Number 12 December 1999.
The common element in the sampling designs and techniques applied in the countries listed above is the systematic distribution of the sample units over the survey areas using different dot grid designs. All countries relied on permanent plots for monitoring their forestry resources but also on temporary sample units to improve precision and to gather additional data for specific studies when needed.
Two types of sample plots have been used in the industrialised countries namely a). the fixed area plots with different sizes depending on the countries requirements and on the characteristics of the tree population addressed (large DBH, small DBH and regeneration); and b). the point sampling mainly with relascope.
Each country has its own sampling intensity and has selected the sample plot size and shape that thought would achieve the forest inventory operations at the desired precision. For large DBH tree population, plot area varied from 314m2 in Sweden to about 8000m2 in Switzerland. For small tree diameter classes, plots sizes oscillated around 100m2.
Sampling design is also a result of combination of plot size and plot number. Reduced plot sizes are usually used in large number of field plots grouped in clusters of varying extents and shapes. Some clusters are composed of high number of sample units distributed along tracts. other with lesser sample units on 4 corners of a square. Large plot approach is applicable essentially in single plot location and lower plot population.
While many industrialized countries do have a systematic monitoring system for forest resources, it is obvious that the design varies considerably between countries, including sampling, measurement techniques and definitions. It is therefore difficult to make direct comparison at the international level and global assessments must in many cases rely on ad hoc transformation of national results to international classifications.
Furthermore, whereas many parameters related directly to forest management, such as volume, biomass, stand age, stem diameters and volume growth, are often well covered in the inventories, several "new" parameters related to, e.g., environmental issues or benefits from the forests other than timber, are often missing. National or international analyses of forest management options will therefore suffer from poor information quality for several aspects of forestry.
In conclusion, the situation in several industrialized countries (particularly in the CIS area), is less than satisfactory for national and international forest policy development and implementation.
A few years after its foundation in 1945, FAO published its first survey of forest resources. The survey was carried out using a questionnaire approach, and was repeated at five-year intervals during the period 1948 - 1963. Another round was planned for 1968, but the work was interrupted due to erroneous or lacking responses. There was also a lack of funds to complete the project. In the early 70´s, Persson (197x) used a new and more active approach, doing country-by-country research by using all sorts of information as a basis for his expert opinion. This work did not have the same official status as earlier ”World Forest Inventories” since it was not a regular FAO effort.
The next global survey carried out by FAO was published in 1982 (FAO 1982). The methods included (a) country visits and research to collect existing information, (b) interpretation of satellite imagery were existing information was week, (c) re-appraisal and re-classification carried out by a team of specialists and (d) adjustment to the common reference date (end of 1980) by empirical methods.
The Forest Resources Assessment 1990 applied two ”parallel” methods to estimate the state and change of the forest cover. As before, country-level information was gathered through country visits and by correspondence. A deforestation model was developed to take care of the adjustments to the common reference dates 1980 and 1990. The model depended on estimates of vegetation type, initial forest cover and population growth. Results for the tropical countries were published in 1993 (FAO 1993).
The second ”branch” of FRA 1990 was to apply a stratified random sampling scheme to obtain statistically sound estimates of forest cover at regional and global levels (FAO 1995). The sampling units consisted of pairwise Landsat scenes, one from a date close to 1980 and the other from a date close to 1990. The systematic use of satellite imagery covering the whole tropical region was a major step forward in methodology. An important feature was that the fixed location of the dual-date imagery allowed considerable improvements of change estimates.
Since the publication of the FRA 1990 results, FAO has made bi-annual reports called ”State of the World´s Forests” (SOFO) (FAO 1999b). The reports contain updated country information and regional summaries, and can be seen as a first step towards a continuous reporting system. Efforts were made to obtain new information for SOFO by focusing on large countries where new forest surveys had been carried out.
The Global Forest Resources Assessment 2000 (FRA 2000) is a joint endeavour, carried out by FAO in co-operation with partners and member countries. The intent of the assessment is to provide a broad range of up-to-date information on the world’s forest resources. In doing so, FRA 2000 is expected to stimulate critical discussions concerning the management and protection of all forests. The FRA 2000 agenda was developed by a large body of some of the world’s leading forestry experts, during an expert consultation in Kotka, Finland in 1996. In 1997, The FAO Committee on Forestry (COFO 1997) and the Intergovernmental Panel on Forests (IPF IV) approved of the consultation’s findings and recommended that FAO serve as the lead agency for facilitating the assessment and work in partnership with other institutions in its execution.
During the FRA 2000, available information about the extent, composition, protection and utilisation of forests has been compiled and analysed for each country. Special attention has also been given to estimating the rate of change of forest resources and to document the various factors implicated in the changes. The FRA 2000 is a transparent process in the sense that all background material and analyses are published, and each country invited to be involved in the process. In this respect, FRA 2000 provides a participatory process for developing the most comprehensive, reliable and authoritative baseline survey of forest resources for all countries.
In addition to the country-by-country survey, the FRA 2000 also includes an objective pan-tropical remote sensing survey of forest cover change, a set of global maps of forest cover and ecological zones, and special studies on various aspects of the interaction between people and forests.
The results from FRA 2000 are presented mainly through the World Wide Web on the Forestry Department website. An information system is being developed for the compilation and analyses of information, as well as future maintenance and expansion of the dataset. Information is presented in a transparent way, ensuring that the sources of data are possible to find. Results are produced and presented throughout year 2000. Presently (June 2000) about 8,000 web pages with country information have been published (www.fao.org/forestry/FO/COUNTRY/nav_world.jsp). See www.fao.org/FORESTRY/FO/FRA/index.jsp for further information of the Forest Resources Assessment Programme.
Despite the FRA 2000 efforts, one conclusion from is that reliable basic forest and forestry information is not available for a majority of countries.
Several organisations provide international information on forests and forestry. The following lists some significant initiatives.
Within the UN system, the United Nations Environmental Programme (UNEP, www.unep.org) is a provider of primary information, mainly through the World Conservation and Monitoring Centre (WCMC, www.wcmc.org.uk). FAO is collaborating with WCMC in the FRA 2000 on the reporting of protected areas. Within FAO, the Africover project is producing detailed land cover maps for a number of African countries.
The TREES project was set up in 1990 by the European Commission´s Joint Research Centre (JRC, www.jrc.it) and the European Space Agency (ESA). The project aims at (a) monitoring of tropical forests, (b) identifying deforestation hot spots and fires in tropical forests and (c) improving understanding of the tropical forest ecosystem. The main tool is coarse-resolution remote sensing using sensors such as AVHRR and ERS-1 and ERS-2. Development of a World Wide Web interface is high on the agenda. However, the future of the project is currently uncertain.
IUFRO established the Global Forest Information Service Task Force in 1998 (GFIS, http://iufro.boku.ac.at/iufro/taskforce/hptfgfis.htm). The work is mainly carried out by the European Forest Institute (EFI, www.efi.fi) in Joensuu, Finland. The mission is to implement an Internet-based metadata service that will provide world-wide access to forest information. The method is to develop a ”Mother website”, which will include search tools and links to primary providers of information such as FAO and the TREES project. The aim is to have a prototype of GFIS running by August 2000.
The Landsat Pathfinder Humid Tropical Forest Inventory Programme, mainly funded by NASA was designed to study deforestation rates through high-resolution satellite data from three epochs: early 1970s, mid-1980s and mid-1990s (Mayaux et al 1998). Three regions are being mapped: the Amazon basin, Central Africa and continental South-East Asia. A recent web search through www.nasa.gov points to Michigan State University and the Basic Science and Remote Sensing Initiative (BSRSI), a research program on global change. Their web site (www.bsrsi.msu.edu) includes country information regarding forest cover and deforestation as well as map animations showing the historic development of forest cover in selected areas, mainly from the Amazon basin. NASA also supports the interagency initiative Global Observation of Forest Cover (GOFC), which promotes the use of satellite remote sensing for forest monitoring.
The World Resouces Institute (WRI, www.wri.org) recently launched the Global Forest Watch (GFW), a project of WRI´s Forest Frontiers Initiative. The GFW is characterized by combining remote sensing information with on-the ground data collection in target areas. It has a networking approach with NGO partners. The first phase of the program is directed towards Cameroon, Canada, Gabon and Indonesia.
More general information, including forestry, can be found at eg The World Bank (http://www.worldbank.org/data/countrydata/countrydata.html) or The Economist Intelligence Unit (www.eiu.com).
The scientific community provides many publications dealing with forest or forestry development. One study made within the FRA 2000 surveyed the existing scientific literature on tropical deforestation and found that the source documents of forest change processes are strongly biassed to certain regions. Furthermore, the survey approaches vary considerably, including differing definitions and asusmptions as well as variations in scope depending on the implementing agency or organisation.
Conclusively, many initiatives today provide international forestry information, mainly through the world wide web. However, none of the abovementioned organisations provide in-depth and quality controlled information on a global basis. Furthermore, the information is often limited to certain subjects of particular interest to the organisation or project in question. Some information is clearly biassed by promotion of remote sensing technologies. Finally, several organisations quote FAO’s forestry statistics in their presentation of country information.
Knowledge management is to organise and make better use of tangible as well as intangible information (tacit knowledge). Formalized information is often well managed in conventional organisational and information systems. Decisions, however, must often take into account informal (tacit) knowledge that is less organized – if at all.
One way to describe knowledge management in forestry is to view the forestry process as in Figure 1 where the GFS sector is marked with a line. It is assumed that information needs drive the inventory, assessment and that scenarios and policy/planning makes use of the findings, and then provide feedback.
In reality, this flow is cut off at several points. Many inventories are made without adequate investigation of information needs. Scenarios and policy development often use simplified macro-level models. Implementation of policies and plans is sometimes deficient due to institutional imperfections. Feedback mechanisms are typically overlooked. Better knowledge management could potentially reduce the effect of these cut-off points.
Figure 1. Forestry knowledge management. The figure can be applied at local, national or international levels. The outer line indicates the activities of the Global Forest Survey.
Information technology has changed paradigms in many professional fields, including forest inventory. Cheaper and more user-friendly computers have made national forest inventories much more efficient, and the statistical analyses more advanced. Field computers and digital measurement tools have reduced manual errors and facilitated remote digital submissions of data over telephone networks. Computing power and availability is now rarely a limiting factor for forest inventories.
One technology that has gained a lot of attention in forest inventory in recent decades is remote sensing. Although aerial photographs had been around for half a century, the introduction of satellite remote sensing and digital imagery boosted the research and development activities world-wide. Forestry was a field where benefits were expected, especially in light of the relatively expensive field inventories. Despite the hype and substantial financial support, the technology has, however, so far brought little to improve forest inventory or forestry planning (Holmgren & Thuresson 1998). The main reason seems to be that information required for planning (see above) includes many variables not possible to study from above (eg volume growth, biodiversity and socio-economic information). Remote sensing and the related geographic information technology has, however, a great potential to complement a systematic field survey – primarily by ensuring that the field sample is statistically efficient, and by assisting navigation in the forest.
GPS – the Global Positioning System – provides the possibility to locate where you are with high precision, typically within 20 meters, using a small handheld device. The GPS receiver can also be connected to a field PC and track the movement on the ground on top of a map or a remote sensing image. The GPS is revolutionizing forest inventory throughout the world and is one of the prerequisites for a successful implementation of the Global Forest Survey. GPS was successfully used in a pilot study for the GFS (Dalsgaard et al. 2000).
The Internet, and the world wide web in particular, provides new opportunities for large scale forest inventories. In the survey phase, the Internet can be used for data submissions worldwide, using interactive web applications. This technology is already used by FRA 2000, as country information on the Forestry Department website is maintained from allover the world. In the publication phase, very large amounts of data can be published and made dynamically available to users world-wide.
The situation outlined in previous sections is alarming. It is obvious that despite the considerable attention that forestry has on the international agenda, and despite development efforts over past decades, the forestry information situation is inadequate. Whereas the international community calls for extensive measures, including monitoring, to deal with carbon sequestration, biodiversity and sustainable forest management in general, the efforts to survey and monitor the forests in a global and systematic way have not been succesful or sufficient to meet information demands.
Some specific issues are:
- The capacity to assess forests and forestry is insufficient in most countries;
- Consequently, national forest policies and their implementation are in many cases based on insufficient information and analyses;
- The lack of capacity has meant that many countries do not have the initiative in assessing their own resources — instead various international organisation produce non-solicited information with low or unknown accuracy;
- International processes (notably those dealing with carbon and biodiversity) must rely on low quality information for sophisticated monitoring and modelling;
- Efforts in recent years to fill the national and international information gaps has often been dominated by remote sensing and mapping approaches, at the expense of high-resolution and in-depth field surveys;
- Comparability of forest information over time and between countries is in general poor, meaning that estimates of, e.g., forest cover change by country must be assessed through expert opinions.
To address these issues, the Forestry Department of FAO is developing a Global Forest Survey in line with FAO's mandate and responsibilities to ensure capacity building and communicate information for the forestry sector. In the following, a framework for the Global Forest Survey is outlined.
1 Developing countries refers to the regions Africa, Latin America and the Caribbean, Asia excluding former Soviet Union and Japan, and Oceania excluding Australia and New Zealand. The need to reconsider the classification into industrialized and developing countries is acknowledged.
2 Industrialized countries refers to Europe, former Soviet Union, Japan, Australia, New Zealand, Canada and US. The need to reconsider the classification into industrialized and developing countries is acknowledged.
