ANNUAL COUNTRY WIDE FOREST MONITORING IN SWEDEN: A PRELIMINARY 
ASSESSMENT OF REQUIREMENTS FOR THE WIDE-SCALE INTRODUCTION OF 
REMOTE SENSING IN SWEDEN’S NATIONAL FOREST INVENTORY 
Randall Thomas, Visiting Professor 
and 
Bo Ranneby, Professor 
Swedish University of Agricultural Sciences 
Faculty of Forestry 
Department of Forest Survey 
S-90183 UmcS, Sweden 
Abstract 
Sweden annually performs an inventory of its 450,000 square kilometer area 
to obtain information on existing forest and other natural resources and 
their change through time. The current inventory procedure is based on a 
systematically-located sample of square traverses or "tracts", on which are 
located either temporary or permanent sample plots. Approximately 18,500 
of these sample plots are measured every year in the Swedish National Forest 
Inventory (NFI) in order provide information for national and regional 
policy making and planning. 
In order to respond to the need for more precise forest state, change, and 
environmental impact statistics, and to provide information for smaller 
administrative areas, the Swedish NFI will be adapted over the next several 
years to use remote sensing data on a wide-scale. This process will involve a 
phased implementation of satellite- and aircraft-acquired remote sensing 
data, taking advantage of the extensive NFI ground plot system. Key findings 
and recommendations from a preliminary assessment of what it will take to 
accomplish this goal are presented here. 
Key Words: forest inventory, remote sensing, sampling, estimation, 
cooperation 
CONTEXT 
The Department of Forest Survey of the Swedish Faculty of Forestry has 
determined that a program will be initiated to integrate the use of remote 
sensing techniques in their annual, nationwide forest inventory conducted 
on behalf of the government. The Swedish National Forest Inventory (NFI) 
is currently based on data obtained from approximately 18,500 ground plots 
occurring along permanent and temporary rectangular sample tracts 
(Ranneby and others, 1987). These tracts, forming sample plot clusters, are 
in turn systematically located across the country. While conducted 
efficiently and on annual basis, growing national and local requirements for 
more precise estimates of forest change and environmental impact, and for 
location-specificity, require that the system be improved. 
To obtain this improved capability, the Department of Forest Survey wishes to 
combine the wealth of information available from its current ground plot 
system, modified and/or supplemented as necessary, with satellite and 
possibly also aircraft-acquired remote sensing data. Specific parameters for