A Pathfinder Global Soils Research Initiative 
Pathfinder Imagery for Global Soils Identification 
Robert C. Lozar 
Construction Engineering Research Laboratory 
USA 
(NB: This plus other global presentations are available on the Internet at: 
http://www.cecer.army.mil/WWWDEMO/global_apps/global_menu.html ) 
Objective 
To develop a method which will improve the global soils map. The method will incorporate the high 
confidence statistical characteristics derived from NASA's latest Mission To Planet Earth (MTPE) 
satellite imagery from the Pathfinder Data Sets with currently available digital mapped information. 
The Pathfinder Data is being generated to test the Earth Observing System Data and Information 
System (EOSDIS). 
Problem 
The only existing global soils map is the one compiled by the United Nations Food and Agriculture 
Organization (FAO) from individual countries using a categorization based on a combination of 
systems. Though this is a vast step forward, it is difficult to use to: 
* generate soil interpretations, 
* assign statistical quality evaluations, or 
* update. 
Why are Soils Identification Important? 
Soils are a key indicator of many environmental concerns. They indicate (among many other 
concerns): 
* the sensitivity to environmental impacts, 
* the parent materials and the mineralogy, 
* many subsurface conditions, 
* the potential agricultural productivity, 
* expected soil moisture, 
* the recent climatic conditions, and 
* the susceptibility to erosion. 
Basic Technique 
We will use the Pathfinder data sets (such as the Normalized Vegetation Index) to extract statistical 
characteristics of the data, correlate that with the central sections of the FAO soils units to derive 
significant correlations. How the statistical characteristics change over time will also be investigated. 
These, in combination with other information types (such as soils maps, topography, terrain types, 
water flow concentration, or climate character) will be used to improve the reliability of the current 
soils map and provide a basis for continued improvement as more data becomes available. In 
addition, using remotely sensed data can provide a map from which seasonally varying interpretations 
can be 
applie 
In ord 
resultii 
10511 
of the 
charac 
locatic 
genera 
Previ 
CERL 
USDA 
data de 
digital