engaged in preparing an assessment for 1990, due to be released in 1992 or 1993 [Singh 
1989]. For these assessments, three levels of data products are used. The primary resource 
is through questionnaires and logging permit records. These are backed up by analysis of 
the normalized global vegetation index derived from AVHRR data. (Neither of these 
sources provides the accuracy, reliability, or specificity required for the Assessment.) The 
third resource is high resolution satellite images such as SPOT and Thematic Mapper data. 
These are used relatively seldom by FAO, however, due to cloud cover for key regions, and 
due to data costs. The United Nations has not funded FAO to the extent required for 
extensive use of these data sources. It is generally agreed that the accuracy of the FAO 
Assessments would be markedly improved if full access to high resolution global remote 
sensing data were available. 
A major intervention initiative is the Tropical Forestry Action Plan (TFAP) [FAO 
1988], for which the UN FAO is the Secretariat. That Plan seeks to provide a framework 
for stimulating coordinated response from diverse groups, with an emphasis on changes in 
policies and practices leading towards sustainable use. At the present time, national forestry 
reviews are being conducted in more than 60 developing countries to assess the extent of 
deforestation, and to plan strategies to reduce deforestation trends [Hazlewood 1989]. 
However, the advisory group for the TFAP has expressed concern over its inability to 
monitor and evaluate the level of progress achieved as the Plan is implemented. 
Institutional activity related to tropical forest resources is expanding, including 
multilateral and unilateral funding agencies, private sector groups, and the governments of 
tropical nations. Several aid agencies, including the World Bank, are including resource 
assessment using remote sensing as a part of their grant programs, in response to 
accelerated resource depletion caused in part by their development projects. Again, the 
need for active environmental monitoring has been recognized, but its implementation in 
this context is just beginning, and there seems to be no mechanism for its large scale 
adoption. 
There are similar situations in other key environmental areas. However, the situation 
is substantially the same as for tropical forests. 
Technical Considerations 
The data volume and technical implications of active monitoring are substantial. 
Globally, there are on the order of 10 billion hectares of critical environmental regimes, 
including portions of tropical and boreal forests undergoing change, polar and glacial ice 
margins, desert boundaries, and oceanic coastal zones. Most of these areas are subject to 
change induced by human activities that require observation at reasonably fine resolution 
for detailed monitoring. If these regions are fully imaged by 100 km square segments, 
approximately 10,000 frames of data would be required for each monitoring cycle, assuming 
no margins for cloud cover nor allowance for data overlap or loss. Even if one assumes