ibul 2004
iter the
ed at the
ers were
organic
'er varied
Samples
thickness
ments for
sture, mg
ximation,
the bulk
age bulk
vere then
on each
PM sites
| terms of
Iso made
cts were
ov. 30
0.51
0.42
APR and
1d height
ites. [he
Figures 7
branches
and the
ributions
ained for
ons were
'ariations
APR and
ectively.
|. An
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
15
un
©
9
C 10
Q
5
G E
=. 5 7
= i
Zz ;
Q uu
4-8 27946 20 2428
Diameter (cm)
Figure 8. DBH distributions for APM site
average needle size was used in the model. Furthermore - based
on these measurements - the biomass of the APM stands was
estimated to 160 tons/ha. Although this is close to the biomass,
estimated by using the lumber company's empirical formula,
this number must be reduced somewhat since it was made 18
months after the flights.
6. DATA ANALYSIS
The forest model has been used to reproduce the ESTAR data
for the APR and APM sites. In order to understand the impact
of soil moisture variations on the retrieval of forest biomass,
model generated curves for brightness temperature versus
volumetric soil moisture for both the APM and APR sites are
plotted in Figure 9. In this plot a fixed physical temperature of
T — 300? K has been assumed. As expected the model predicts
that the brightness temperature decreases with increasing soil
moisture This is due to the fact that the emissivity of ground
decreases with increasing soil moisture. The curves also clearly
indicate an increase in sensitivity to soil moisture with a
decrease in foliage density. This is because the foliage becomes
more transparent as its density decreases; hence, the effect of
changes in ground emissivity due to soil moisture becomes
more visible. The combination of these two effects results in
higher sensitivity to biomass when soil is very wet (as is clearly
visible in the figure). This occurs when there are wet
conditions since the contribution of the ground is much less
than from the vegetation canopy. In general, both the soil
surface and the forest canopy (soil moisture and biomass)
contribute together to the observed brightness temperature.
Appropriate corrections for soil moisture must be made in order
to make reliable biomass estimations.
The air temperatures given in Table 1 and the ground truth data
given in Table 2 have been used to calculate the brightness
temperatures using the forest model. The model results and the
measured brightness temperatures (Table 1) are plotted in
Figures 10 and 11 for the APR and APM sites respectively. The
model captures the trends in the data for both sites reasonably
well. The model does very well at the APR site (Figure 10) on
Augest 27 and November 15, but there is a large disagreement
on July 7. The rather complicated trend in the data at the APM
site is reflected reasonably well by the model (Figure 11). It is
believed that some of the error in Figure 10 is related to the
uncertanty in the physical temperature of the canopy. The
temperatures (obtained from the weather station at the Wake-
a re APM Site
280 - mc ei +
APR Site
Brightness Temperature Tg (K)
250 + Tz 3004€ |]
o=20cm
= 140 cm a
N
A
un
T
1 1 1 L L
0 0.05 0.1 0.15 0.2 0.25 0.3
Volumetric Soil Moisture, m, (g/cm?)
0.4
Co
em
e
Figure 9. Model variation of brightness temperature with soil
moisture for the APR and APM sites.
field airport) may not be representative for the APR site. This
may be due to the fact that the APR site consists of small trees
and the canopy and surface temperatures may not be as stable
as the APM site.
290 + T T T
+ ESTAR
Model
nN
co
ce
+
1
J
~~
c
T
1
260 4
Brigthness Temperature, T. (K)
1 L À
July-7 Aug-27 Nov-15 Mov-30
Date
Figure 10. Comparison of ESTAR and model results for APR
site
7. CONCLUSIONS
The measurements at Waverly, Virginia have demonstrated that
L-band brightness temperature is sensitivity to forest biomass.
Although the biomass of the stands observed was between 20
and 200+ tons/ha, it appears that higher values of biomass can
be detected by a passive L-band sensor. A discrete model of
two forest sites at Waverly was used to simulate the
measurements with a reasonable degree of accuracy. They
show that soil moisture variation becomes an increasing
contributor to the brightness temperature as the stand size
becomes smaller. These results also indicate that periods of
high soil moisture have the lowest ground contribution and thus
afford the best opportunity to monitor biomass.
