[ing 
emote 
essive 
Image 
| and 
Raey, 
along 
Delta, 
No. 8, 
tum to 
| ^78, 
1g and 
ational 
insing, 
venier, 
Optical 
Water 
, No.3, 
$ 
, and 
source 
g and 
3. 
COMPARISONS OF THREE REMOTELY SENSED DATA ON FOREST 
CROWN CLOSURE AND TREE VOLUME ESTIMATIONS 
Kuo-mu Chiao 
Professor,Dept. of Forestry, National Taiwan University , 
Taipei, Taiwan, The Republic of China 
Commission VII, Working Group 3 
KEY WORDS : Forestry, Landsat, SPOT, Multispectral, Remote Sensing, Crown- 
closure, Tree-volume 
ABSTRACT 
The aim of this study isto examine the airborne multispectral scanning data, the SPOT HRV data, and 
the Landsat-TM data on forest crown closure and stand volume estimation. The airborne multispectral 
scanning data used in this study were MSS] to MSS6 and MSSg to MSS]. The SPOT data were XS] ^ 
XS2 and XS3, and the Landsat-TM data were TM] to TM5 and TM7. AII the digital data were collected 
from each sample plots which were respectively located on geometric corrected and radiomatic corrected 
airborne multispectral scanning, the SPOT and the Landsat-TM imageries. The stepwise regression 
approach was used to establish the regression equations in which the spectral bands, and their 
combinations were used as the independent variables and the forest crown closure and the stand volume 
were as the dependent variable. The spectral bands was the individual bands of the multispectral scanning 
data, the SPOT data and Landsat-TM data. The combinations were (1) band difference(2) normalization 
difference (3) tree ages and crown closure. The optimum equation was sclected by their significant 
statistical F value and higher coefficient of determination(R ). The findings of this paper were as follows : (1) 
The crown closure equations that were estabished by three remotely sensed data were eligible. Among the 
three equations, the equation which — was developed by Landsat-TM data was the best, The airborne 
multispectral scanning data the second, the SPOT HRV data the third. (2) The stand volume equation that 
were estabished by three remotely sensed data were not eligible. Another parameters such as tree ages and 
crown closures were added. Especially the age factor was much more effectively. (3) On the crown closure 
and stand volume equation establishment, thrce independent variables were enough. No more variables were 
needed. 
INTRODUCTION 
Forest stand volume estimation is one of the main in 1980's. Several new sensors such as Landsat-TM, 
problems of forest mensuration. Foresters have SPOT HRV have.bcen available. More spectral bands 
examined the potentialities and limitations of and fine resolution of these sensors made the detailed 
estimations of tree volume and its rclated factors in the forest measurements possible . Some of the research 
forests or on the aerial photograph for many years. Since interest were focus on forest canopy determination, and 
the Landsat MSS data have been available in 1970's, the green biomass and tree volume estimation. 
digital data and derived imagery have been used for 
monitoring vegetation change in extensive area but little The aim of the work reported in this paper was to 
utilized on forest stand estimation. The important reason investigate the effects of different remotely sensed data 
for this was the relatively coarse resolution (80m) of the on forest crown closure and tree volume estimation. 
Landsat MSS data. However, this situation has altered These remotely sensed data were the airborne 
123 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996