373 
TABLE 2 
Usefulness of Thematic Mapper Bands and Transformations 
for Discerning Metal Stress in Coniferous Tree Canopies 
Rank 
Spain TM 
Scenes 
Austria TM Scenes 
Spain-Austria 
TM Scenes 
January 
August 
June August 
All Dates 
Group I 
TM4 
- 
BD1 
ND1 
PC1 
TMB 
ND1 
R41 
BD1 
TM1 
ND3 
TMB 
BD2 
TM5 
TM7 
Group II 
TM3 
TMW 
BD3 
BD 1 
TM4 
PC1 
TM7 
R41 
TMG 
BD2 
TM2 
BD7 
BD2 
BD3 
BD3 
TM5 
ND7 
PC1 
PC1 
TMG 
BD5 
TM4 
R43 
BD7 
TM5 
TM5 
R31 
PC1 
ND5 
Group III 
TMG 
R45 
TMB 
TM4 
BD1 
R47 
TMG 
TMB 
TM7 
R23 
BD7 
TM5 
BD3 
BD7 
Group IV 
R2 3 
R57 
TM7 
TM2 
TMW 
BD1 
ND3 
TM7 
ND2 
TMB 
R43 
alues" of 
alues re- 
ups is 
orrespond- 
p-ranked 
ions for 
TABLE 2, as 
n bands and 
scene 
and trans- 
August scene 
respondence 
:ase with 
: Spanish 
iormations 
;d Spanish 
:ed also 
1 transfor- 
/■idual scene 
ire repre- 
7 three of 
p-ranked 
md BD1) are 
Lscriminat- 
Eourth ranked 
4. DISCUSSION AND CONCLUSIONS 
The application of Thematic Mapper bands 
and transformations for geobotanical pro 
specting requires that a stress discrimina 
tor be both robust and simple to employ. It 
needs to be able to work equally well with 
scene data acquired from clear as well as 
hazy or cloudy (hence varying scene illumi 
nation conditions) scene dates, and with low 
as well as high solar elevation angles. Both 
the first principal component (PC1) and the 
band difference BD1 appear to meet these 
criteria, with the band difference being the 
simpler of the two to apply. A potential 
problem inherent in utilising the first 
principal component is the scene dependency 
of its derived eigenvector coefficients, 
which may restrict its use with data from 
scenes differing significantly from the 
one(s) employed in its formulation. The The 
matic Mapper brightness (TMB) index, how 
ever, is based on the intrinsic physical 
characteristics of the scene features, there 
by allowing it to be applied universally to 
data sets from different regions and acqui 
sition dates. 
The dominant representation of Thematic 
Mapper bands 5 and 7 in the higher ranking 
bands and transformations of the August 
Spanish scene implies a strong presence of 
water stress in the pine tree canopy. This 
is to be expected, as July and August are 
the hottest and driest months of the year 
in this part of Spain and the effects of 
water stress in the trees would be most 
acute at this time. 
The lack of any correspondence between 
the top-ranked TM bands and transformations 
of the January and August Spanish scenes is 
likely the consequence of the marked differ 
ences in solar elevation angles (and hence 
canopy shadow) between the two scene dates, 
the presence of numerous cumulus clouds in 
the immediate vicinity of the test site in 
the August scene (resulting in varying can 
opy illumination conditions), and the like 
lihood of severe water stress conditions 
being present in the pine tree canopy in the 
August scene. 
ACKNOWLEDGEMENTS: NASA Ames Research Center, 
ZGF in Munich, and the European Space Angency 
kindly supplied the Landsat TM CCT's of the 
Spanish and Austrian test areas used in this 
study. 
REFERENCES 
Banninger, C., 1984. Detection of Heavy Metal 
Stressed Vegetation Using Landsat Digital 
Data; Proc. 18th Int. Symp. Rem. Sen. Envi 
ron., Paris, France, pp. 1101-1105. 
Banninger, C., 1985a. Geobotanical Remote 
Sensing of Heavy Metal Stressed Vegetation 
Using Landsat MSS Data; Proc. Fourth The 
matic Conf.: Rem. Sen. for Explor. Geology, 
San Francisco, Environ. Research Inst. 
Michigan, pp. 339-345. 
Banninger, C., 1985b. Comparison Between 
Landsat MSS and Thematic Mapper Data for 
Geobotanical Prospecting in the Spanish- 
Portuguese Pyrite Belt; Proc. Int. Geo 
science and Rem. Sen. Symp. (IGARSS'85), 
Univ. Massachusetts, Amherst, pp. 949-956. 
Banninger, C., 1985c. Spectral Analysis of 
a Heavy Metal Stressed Forest Canopy Using 
Landsat Thematic Mapper Data; 3rd Int. 
Colloquium Spectral Signatures of Objects