International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
  
brightness temperatures over these sites are shown in Table 1 
along with the physical temperature at each site. The physical 
temperatures were obtained from the Wakefield weather station 
which is about 40 km away from the Waverly site. 
  
  
  
  
  
July 7 Aug 27 Nov 15 Nov 30 
APR 265.5 252.0 233.0 235.0 
APM 281.2 272.0 263.0 262.0 
T°K 302.4 2 286.3 278.6 
  
  
  
  
  
  
  
Table 1. Brightness and physical temperature at each site. 
An examination of the brightness temperatures in the table 
shows that there is approximately a 30°K change in brightness 
temperature at the APR site while there is a 20°K change in 
brightness temperature at the APM site. Much of this change is 
due to the fact that the physical temperature has changed, 
however the change is larger in the APR site, which has the 
small trees. This represents the influence of changes in soil 
moisture. 
4. FOREST MODEL 
The loblolly pine stands have been modeled using the Peak 
approach, which relates the active and the passive problems 
(Peake, 1959). Basically, the brightness temperature is written 
as the physical temperature times one minus the scattering 
albedo of the forest layer. The scattering albedo of the layer is 
then computed from the bistatic scattering coefficient. The 
passive problem has now been reduced to an active problem. 
The approach has been used before by Chauhan et al (1999) for 
forests, Chauhan et al (1994) for corn canopies, and by Saatchi 
et al (1994) for grass. 
In the work to be presented here, the bistatic scattering 
coefficient of the forest is found by employing a discrete 
scatterer approach wherein the forest is considered to be a 
dielectric layer containing randomly distributed and oriented 
cylindrical scatterers. The trunks, branches and needles are 
modeled as finite length cylinders, which are given prescribed 
orientation statistics obtained from measurements in the forest 
as shown in Figure 6. The individual scattering cross section of 
each type of scatterer is used in the calculation. For more 
details see Chauhan et al (1991). 
«( ESTAR 
  
  
Figure 6. The forest model 
S. GROUND TRUTH 
Ground truth data was collected during and after the 
experiment. Soil moisture measurements were obtained at the 
time of the over-flights while forest stand parameters were 
estimated in February 2001. The forest floor had an organic 
layer of tree litter over a sandy loam soil. The litter layer varied 
in thickness from 0.5 to 5 cm across the forest floor. Samples 
were taken and divided into litter and soil parts. The thickness 
of the litter layer was also recorded. Separate measurements for 
average bulk density (g/cm?) and gravimetric soil moisture, Mg 
(%) of both soil and litter were made. As a first approximation, 
the average of soil and litter were used to compute the bulk 
density and the gravimetric soil moisture. The average bulk 
density and the average gravimetric soil moisture were then 
used to determine the volumetric soil moisture, my , on each 
date. These estimated values of my for the APR and APM sites 
are given in Table 2. Estimates of surface roughness in terms of 
the rms height, & , and correlation length, | , were also made 
but are not given here since the surface scattering effects were 
small. 
  
July 7 Aug. 27 | Nov. 15 | Nov. 30 
  
APR 0.12 0.38 0.46 0.51 
  
APM 0.27 0.33 0.39 0.42 
  
  
  
  
  
  
  
Table 2. Estimated volumetric soil moisture, 7, , for APR and 
APM sites 
The diameter at breast height (DBH), stem density and height 
of the stand were collected for the APR and APM sites. The 
size distribution of the trunks for both sites is given in Figures 7 
and 8 for the APR and APM sites respectively. The branches 
were classified according to their base diameters and the 
average density of each type was estimated. Distributions 
similar to those presented in Figures 7 and 8 were obtained for 
the branches but are not shown here. These distributions were 
incorporated in the forest model to take into account variations 
in trunk and branch sizes. Average stem densities for APR and 
APM sites were 0.23 stems/m. and 0.19 stems/m. respectively. 
The needle dimensions and density were also measured. An 
  
  
  
  
  
  
  
50 
o 40 
Q 
930 
D 
+ 20 
E 
Zz 10 
0 Ï T T 
  
    
  
Diameter (cm) 
Figure 7. DBH distributions for APR site 
Intern 
AY nda oua d d anm E 
averag 
on th 
estimé 
estime 
this n 
montl 
The f 
for thi 
of soi 
mode 
volum 
plotte 
T=3( 
that tl 
moist 
decre: 
indica 
decre: 
more 
chang 
more 
highei 
visible 
condit 
than | 
surfac 
contri 
Apprc 
to ma 
Thea 
given 
tempe 
meast 
Figure 
model 
well. 
Auges 
on Jul 
site is 
believ 
uncert 
tempe