IT. INSTRUMENT DESCRIPTION 
The GSFC Mark-II Three Band-Held Radiometer consists of two modules, which are 
connected by an electrical cable (see Fig.1). The hand held module contains 
three channels, which approximate to Landsat 4's thematic mapper bands TM3 i; 
(0.63-0.69 pm), TM4 (0.76-0.90 um) and TM5 (1.55-1.75 um). The read-out module 
contains the power supply, 3 liquid crystal displays, the digitizer, display 
driver circuits, range controls, battery check, a "sample and hold" switch, and 
"off-on" switch.. Radiometric accuracy is 1 part in 1,000 (^-10 bits) with a 
total range of 1, 000,000 (20 bits). This represents a dynamical range from 
1 milliwatt/m? to 1 kilowatt/m2 within the optical bandpass. Radiometric 
sampling occurs at the rate of 4 times per second. For detailed description 
and operation of the radiometer, see Tucker et al (1980) 
III. FIELD MEASUREMENTS 
In the field operation of the radiometer, the hand-held module.is mounted ôn a 
holder, which is attached to the top of a 3 meter stainless steel pole. A rectan- 
gular plate with 5 holes on it is fixed above the ground by the way of a. 
sharpened stainless pipe, which is attached with the plate. The five holes on 
the plate will make the sensor making the angles of 0°, - 22.59, and t Are with 
respect to the zenith, when the Hole at the other end of the three meter pole 
matches to those five holes. The whole set-up is shown in Fig. 2a and 2b. 
A 2m x 2m aluminum plate coated by "Halon" power (trade name for. powdered 
fluorocarbon) is used for calibration purpose. Such kind of calibration board 
has been used in many occasions with great success (see Mclean 19793. deerit is 
close to a lambertian surface with high reflectivity. With no preyious experien- 
ce, our calibration board is less satisfactory. The relative variations is 
. reflectance, looking in different angles accounts for 109 - 15%, depending on 
the wavelength. 
Four test sites of the rice paddy are located just outside of the campus. Imme- 
diatly next to the test sites, there is a campus meteorological obseryatory, 
where meteorological data are routinely taken. In a typical field operation, two 
assistants, one handling the pole with sensor while the other reading the radio- 
meter read out, take 5 measurements over the calibration board, corresponding 
to five different looking angles. They, then, move the wholé.set-up.to the 
near-by test sites and take 5 measurements for each of the test sites. 
Normally, it takes about one hour to take a whole set of. data. Fig. 3 illustra- 
tes a typical data sheet for a given day of measurements. The time and the 
major meteorological data are shown on the top-left portion of the data sheet. 
The relative orientation of the four test sites is indicated on the extreme 
left side. Fl, F2, Z, B2 and B1 represent 5 different view angles, corresponding 
to 45°, 22.5°, 0°, -22.5° and -45° with respect to the sun. (The angle is : 
positive when the sun and the sensor are on each side of the zenith). For a 
given view angle at a specific site, 3 measurements are taken. Test site No.Q 
represents the calibration board. The reflectance ratio is computed by diyiding 
“ the average of 3 measurements over that taken at the calibration board. 
We started taking our field measürements on May 17, 1982, which was about 5Qdays 
after the planting of the first crop season. Our data ended on August 20, 1982, 
which was about ten days after the planting of the second Crop season. Because 
of weather condition, we took 16 clear-day over the whole period. 
IV. EXPERI 
A casual ri 
of measurer 
referred ii 
(19703. "Tt 
by the scei 
As mention: 
we take thi 
for comput 
Fig. 4 sho 
as functioi 
‘May 17, 19: 
August 13 : 
crop seaso 
a general : 
curves, wh 
band. The 
changes fo 
stage. The 
The ratios 
and fluctu 
. variations 
To see how 
we depict 
of 130 day 
small that 
wet soil. 
basically 
seeding gr 
getting doi 
brown grai 
TM3 is chi 
band. Thus 
what we wo 
Admittely, 
vegatative 
those take 
system con 
meter are 
one radiom 
designed d 
automatica 
data taken 
October 30 
for 0° vie 
correspond 
between ba 
that the v 
in general 
and furthe 
To further 
we plot th 
between th