multispectral, permitting an assessment of comparative infrared reflect- 
ance values, unobscured by the effects of reflectance in other ranges 
of the spectrum. Finally, discrimination of forest types by photo 
interpretation depends largely on the judgement of the individual 
ng interpreter. On the other hand, forest typing using digital analysis 
: of Landsat data is an automatic selection procedure, once meaningful 
selection criteria have been introduced. 
ed 
ed Since the digital analysis methodology does not permit distinctions 
, to be made between similar reflectance values on the basis of the 
investigator's familiarity with the forest conditions, the application 
of the methodology to forest regions where many species and complex 
io associations of species are present would be problematical. Therefore, 
, the approach discussed in this paper is proposed for application to the 
boreal forest of Ontario, where pure stands are frequent, only four 
S major and four minor species occur, and extensive homogeneous forests 
es are found. 
id 
The major species in the boreal forest of Ontario are black spruce 
| (Picea mariana (Mill.) BSP), jack pine (Pinus banksiana Lamb), the : 
tea poplars (Populus) and white birch (Betula papyrifera Marsh.). The minor 
species are white spruce (Picea glauca (Moench) Voss), balsam fir 
a (Abies balsamea (L.) Mill.), tamarack (Larix laricina (Du Roi) K.Koch), 
nal and white cedar (Thuja occidentalis Le). 
When the forest is in full leaf, the difference in the reflectance 
values for these species result from variations in the surface area 
y of the foliage which is exposed to the sun, crown width, crown density 
ei and crown closure. Other factors affecting absolute reflectance values 
are sun angle, atmospheric conditions and site conditions. 
gases Because, for specific stand comparisons, there may be similarities 
lar in the above-mentioned stand features or compensations of factors, it is 
not always possible to identify the component species of any stand. As 
examples, although poplar and white birch can be distinguished from one 
the another on aerial photography, their deciduous foliage makes their 
d ral infrared reflectance indistinguishable on satellite data: it appears 
a arnative that they are identifiable on Landsat only as a poplar-birch stand. If 
in this jack pine, white spruce or balsam fir occur as less than 30% of a hard- | 
wood stand it is difficult or impossible to distinguish which one of | 
them is present. | 
I 
sat d If black spruce or jack pine occur as 50% or more of a stand, 
used i species identification by characteristic reflectance values is possible. 
The low reflectance from the narrow crown and dense stand of black 
spruce is particularly indicative of the species, while jack pine 
ie reflects a greater proportion of incident infrared because the crown 
is broader and the stand is more open. 
iffic- 
nent, 
values 
al to 
> 
is 
Io