trees appear as differences of spectral 
age. The comparison would give 
information about forest structures. 
Following features were interpreted 
visually from the four age class images. 
a) Dense mature spruce stands showed 
older spectral ages in summer images 
than those in winter images. 
b) Sparse spruce stands showed opposite 
tendency of dense stands. Namely, 
spectral ages appeared younger in 
summer images than those in winter 
images. 
Those results suggested that it would be 
possible to monitor successional changes 
in’ spruce stands ‘quite well. Thus the 
successional spectral trajectory is 
useful as the training data of forest age 
classification with less seasonal 
differences. This method would be 
utilized for growth monitoring 
practically in the forest managemant. 
5. CONCLUSION 
The exponential relationships made it 
possible to estimate averages and 
standard deviations of each stand age for 
the minimum distance classifier. They 
reduced effects of seasonal spectral 
variation in training areas. Thus, quite 
stable age class classification for 
spruce became possible using winter and 
summer imagery. 
However, the exponential curve suggest 
that there is no clear relationship 
between the stand age and DN after about 
30 years  old.- This caused that the 
estimated age class appeared rather 
younger in old stands. older than that age. 
Though this may be the limitation of 
forest age class classification 
spectrally, the usefulness of 
successional spectral trajectory was 
confirmed from the four. classification 
results. 
Comparison of summer and winter imagery 
would make clear structural difference 
between stands, namely dense or sparse 
with: spruce. These results should be 
analyzed further to make confirm what 
they mean. 
ACKNOWLEDGMENT 
This study was carried out as part of the 
'Japanese Experimental Study in - the 
Arctic Area' supported by the Science and 
Technology Agency of Japan. Landsat TM 
data over Tomakomai area were supplied by 
the National: Space Development‘ Agency 
(NASDA) of Japan. We appreciated for 
their contribution on this study. 
30 
References 
Awaya,Y. and Tanaka, T., 1996. 
Successional and seasonal pattern of 
spruce spectra: as a basis of boreal 
forest monitoring. The 26th International 
Symposium on Remote Sensing, Vancouver, 
B.C., Canada, pp.142-146. 
Awaya,Y., Tanaka,N., Moriyama,T., 
Maesato,S. and Oguma,H., 1996. The 
successional spectral trajectory of Ezo 
spruce’ during ‘the “four seasons. J. Jpn. 
For. Soc. forthcoming. in Japanese 
Nilson, T. and Peterson, U., 1994. Age 
dependence of forest reflectance: 
Analysis of main driving factors. Remote 
Sens’. ‘Environ. ,- 48, ‘pp.319-331. 
Peterson, U. and 
Successional reflectance trajectories in 
northern: temperate forests. INT: de 
Remote Sensing, 14, pp.609-613. 
Takagi, M° ‘and  Shimoda,H. (ed.), TOO. 
Gazou Kaiseki  Handobukku (Handbook of 
Image Analysis), Tokyo, ‘Tokyo: Daigaku 
Shuppankai, 775pp. in Japanese 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996 
Nilson, 7:5 12950,