1977) ; but in this case, a procedure is developed where the vegetation and soil 
indexes combine in a consistent way to produce an index that is not compromised 
by significant variations of either. A similar two-tiered approach using FIGURE CAPTI 
combinations of Landsat band ratios has been used by Raines et al. (1978) to Sa 
bring out the effects of vegetation (B5/B6) and limonite stained surface 
    
  
  
  
   
  
   
   
     
  
  
  
  
  
  
  
  
   
  
   
  
    
  
  
  
  
  
  
   
  
  
  
  
  
  
  
    
(B4/B6) that combine to indicate areas of uranium mineralization. Figure 1 : A 
In general, Landsat will probably not detect very subtle spectral variations Figure 2 : F 
due to chlorosis ; however, where a geobotanical anomaly is manifested by a 
major change in biomass, the Landsat system can be used to map the anomalous 1 
areas. 
Figure 3 : R 
CONCLUSIONS Figure 4 : F 
4 
It should be emphasized that geobotanical remote sensing is no panacea and must S 
be used in conjunction with other exploration techniques. Not all terrains are 
suitable for geobotanical techniques. For example, the dense Amazon rain Figure 5 : P 
forests may not be suitable because nutrients are transferred directly to the ( 
plants from the decomposing soil litter (Jordan, 1982). Presumably, therefore, 
the root structures of these plants are not developed down into the bedrock Figure 6 : P 
derived soils, and the plants will not feel the effects of heavy metal stress. | 4 
In other forests, however, mineral soils play a more important role in nutrient S 
(and presumably toxic metal) storage and transfer (Jordan, 1982). In these 
latter examples, remote sensing of geobotanical anomalies is very promising. 
While work must be done on many different vegetation types under different | 
kinds of heavy metal stress to determine the specific types and causes of the 
spectral responses, it is clear that remote detection of geobotanical stress is 
one of the methods that should be used in a mineral exploration program. 
Spaceborne remote sensors must be designed to detect geobotanical stress, 
particularly the spectral regions assoicated with chlorophyll absorption in 
the visible part of the spectrum. Narrow spectral bands (10-50 nm) centered in 
the visible part of the spectrum around 475, 560 and 660 nm should be most 
useful. These, of course, correspond to those of the Landsat-D Thematic Mapper. 
These bands coupled with Landsat-D's near infrared bands centered at 850, 
'1600 and 2200 nm should be useful for total biomass determinations where a 
geobotanical anomaly is manifest by plant density. 
ACKNOWLEDGEMENTS 
The work that was done by this author and his colleagues at Dartmouth College 
was primarily supported by NASA Grants NGS 5014 (airborne multispectral scanner 
work) and NAG 5-159. (Landsat Brazil Work) and Dartmouth College. The remote 
| sensing facility at Dartmouth College is also partially supported by NASA 
Cooperative Agreement NCC 5-22 in collaboration with NASA's Goddard Institute 
for Space Studies (GISS) in New York City. The author has benefitted from the 
close collaboration of DR. STEPHEN UNGAR and his colleagues at GISS. The major 
portion of the Brasilian work was accomplished as a Masters Thesis by THOMAS 
A. STONE at Dartmouth College.