DIGITAL SITE CLASSIFICATION 
presented at ISPRS Commission VII, September 17-21, 1990 
in Victoria, B.C. Canada. 
Peter Holmgren, M.Sc forestry 
Faculty of Forestry, Remote Sensing Laboratory 
Swedish University of Agricultural Sciences 
S-901 83 UMEA, SWEDEN 
phone: int+46 90 166554 fax: 165925 
ABSTRACT 
Forest site variables were determined digitally by analysing a Digital Elevation Model, a soil map and a 
cadastral map, using definitions in the existing Swedish Site Classification System. The variables were 
estimated for each pixel (10x10m) in a 5x5 km square of land located in northern Sweden. A site index 
image was then produced for the test area. 
The estimates of sub-surface waterflow and ground moisture were evaluated against a field reference of 
1000 circular plots. In no case could the computer estimates be labelled invalid. Instead it was 
discovered that a large number of the field reference plots were doubtfully classified due to subjective 
aspects in the existing classification system. 
It appears that site variables can be estimated digitally with good reliability. Subjective measurements 
in the field can give unreliable input to the otherwise objective and unbiased Site Classification 
System. 
INTRODUCTION 
Site classification, an overview of the Swedish system 
Site classification is a major component in forest management. All planning, operational as well as 
long-term strategic, has to consider the site quality as one of the most important factors in 
determining the proper action and timing of actions within a forest estate. The single most critical 
decision, when attempting economically efficient forestry, is at what age a stand should be clearcut. 
For Swedish conditions, the optimal cutting age can range between 65 and 150 years due to variations in 
site quality, including climate. 
The site does not only influence decisions made on economic and biological facts; also the silvicultural 
law of Sweden has several of its paragraphs based on the site conditions. As an example, for each site 
quality class a minimum cutting age is required by the law. 
It is thus evident that a rather sophisticated and objective Site Classification System (SCS) is 
required for Swedish forestry. Such a system has been developed and established over the last 15 years 
(HAGGLUND & LUNDMARK 1977,1982; HAGGLUND 1979). Today it is used by all interested parties, from law 
authorities to farmers. 
The base unit in the system is the Site Index, SI, which is defined as the height of a dominant tree at 
a reference age, normally 100 years, in a well managed stand of a certain species. Thus the SI T28 means 
that a Scots pine (T) stand grows to a height of 28 meters in 100 years; G32 that a Norway spruce (G) 
stand grows to 32 meters in 100 years. The SI can be converted to volume growth (m^/ha*year'^). 
870