While it is recognized that computer assisted classification methods 
could be used to estimate irrigated lands, this study relies entirely 
upon photo interpretation of CIR transparencies coupled with a variable 
probability sampling design. Visual interpretation methods require less 
training and less expensive equipment than methods using computer assisted 
techniques. Also, simpler methods may be easier to apply for developing 
countries which may be constricted by low budgets and lack of access to 
computer technology. 
  
There is some precedence in using photo interpretation techniques 
on satellite images and photos to inventory irrigated and forest lands. 
For example, Draeger (1976) reported on an image interpretation method 
used on irrigated lands in Oregon whereby interpreters placed a dot grid 
with 400 dots per square inch (6.45 sq. cm) over a LANDSAT CIR print 
enlarged to 1:250,000. Sample errors at 95 percent confidence interval 
for two interpreters were 7 and 13 percent respectively. Langley et al. 
(1969) described a technique using a multistage probability sampling 
design to estimate timber volume within a grid composed of four mile 
square blocks (6.4 km) overlaying an Apollo 9 CIR transparency. In one 
area, including parts of Mississippi, Louisiana and Arkansas, the error 
of estimate was about 13 percent with a ground sample size of one to a 
million in terms of area. 
Armed with the knowledge that irrigated lands exhibit a unique 
signature on the color composites and a sampling design that was applicable 
to the estimation problem, we selected a pilot test area, 27 by 60 miles 
(43.2 by 96 km), in southern Idaho to determine how closely we could 
estimate irrigated lands by visual interpretation methods. 
PROCEDURES: AND METHODOLOGY 
Description of Study Areas 
The location of the pilot inventory test site is in southwestern 
Idaho (Fig. 1). The site is a rectangular area 18 by 60 miles (28.8 by 
96 km). The area is that of a plain bisected by the Snake River within 
a canyon averaging 400 feet (13.0 m) in depth. The region consists of a 
series of plateaus, separated by the canyons of the Snake River and its 
tributaries, with elevations from 2800 to 3200 feet (850 to 970 m) above 
sea level. Within the canyons, irrigated agriculture has existed for 
some time. On the plateaus newly irrigated lands have been developed 
recently by the use of high lift pumping from the Snake River. The 
climate of the test site is dry during the growing season with precipitation 
averaging less than 10 inches (25.4 cm) annually. Summers are hot and 
winters are mild. 
Figu