D SOIL
sling /
ement of
SAR data
n, testsite
logy and
ency and
(Shuttle
sets from
‘borne C-
on with
larization
) analyze
ly and to
| maps of
n paper,
|. (1996)
and TM-
r of an
in April
n, with à
ce angle.
ving pre:
djustment
f a MAP
e without
blurring. Due to the good quality of the scene, the visual
analysis was made directly on the computer screen. A
field and an aerial survey were made in this region, to
control the findings of interpretation.
3 OVERALL DESCRIPTION OF THE AREA
UNDER STUDY
The area under study is localized at the border between
Amazonas and Acre States, west of the city Sena
Madureira, with approximate geographical coordinates of
S 837 - W 68717.8$:8934;- W^ 68?55' t0 S 915 -
Ww 68°09’, S 9°17” - W 68°27’. The geology of this
region is made up essentially by the Cenozoic Solimôes
Formation (mainly very friable sandstones and siltstones
and recent sediments) that covers entirely the area under
study. The climate is humid, presenting two distinct
seasons (humid / dry) along the year. The annual rainfall
is around 2,000-2,500 mm (PMACI I, 1990). The main
vegetation cover is of two types: Tropical Lowland
Forest with Palms and Bamboo (open forest) and Dense
Tropical Forest, on higher sections, such as on the
Plateau (I). Along the floodplain of the main river, Rio
Purüs, a wet Forest dominates, locally known as "Igapó
forest".
4. MAIN GEOMORPHOLOGICAL UNITS
The criteria used to . discriminate. 8 main
geomorphological units (Figure 1) were: texture, tone,
context and previous experiences from aerial and ground
surveys in this region. The higher section (I) is part of the
Lower West-Amazonian Plateau, defined in Brazil (1976)
as part of a much larger relief unit, that has been strongly
eroded at its’ edges. The average height of (I) is around
250-300 m, only about 100 m above the Rio Purus
floodplain (III). The plateau is covered, at least partially,
by Al-, Fe-, Mn holding laterites, known in Brazil as
Canga. The main soil types of this section are Red
Yellow Latosols and Podzols (PMACI I).
At NW of the plateau (units II, VII and VIII) there are
highly dissected areas of relatively small hills, covered by
Open Tropical Rainforest, dominated by Bamboo. Nelson
(1994) presented the first map showing the extension of
bamboo-dominated forests in Acre, indicating also areas
of death of entire bamboo formations during the second
half of the 80's. Taking into account the completely
different physiognomy of bamboo (specially of the
canopy due to a different architecture of trunks, twigs and
leaves) as compared to the canopy of the Tropical
Rainforest in Acre, the relief units VII and VIII
(dominated by bamboo) present very distinct smoother
texture features and tones, as compared to unit VI (no
bamboo dominance), although the relief height and
geological substratum are the same at these 3 units. This
395
indication was confirmed by an independent analysis of
TM-Landsat data.
The floodplain of Rio Purüs is apparently the best
defined relief unit of the area under study. This is
certainly due to differences of soils and vegetation with
high (III) and lower (IV) moisture content. Furthermore,
the Upper Rio Purüs Terrace (IV), 10-15 m above (III),
comes out sharply, due to subtle changes in the geometry
of the channels of the tributaries, after leaving units (VI)
and (V) to (IV). The scene under study would also allow
to establish a classification of fluvial features such as,
abandoned meanders, oxbows (open and vegetation
covered, etc.) that are important for environmental
planning purposes.
S. DISCUSSION
The excellent experiences with X-band SAR data
obtained worldwide, and specially in the tropics, as
mentioned at the beginning of this paper, will certainly be
a strong argument towards an operational spaceborne
SAR system at this frequency, which is missing
presently. Such a system would allow both to map the
geomorphology and soils of unknown areas or upgrade
already existing ones. If optical data (TM-Landsat) is
available, merging of SAR and TM data and RGB/IHS
transformations become feasible, as it was shown by Kux
et al (1995) in Acre with airborne C-Band SAR, thus
allowing further informations related to vegetation, land
use and hydrology.
The floodplain of all the major rivers in Amazonia is very
dynamic, and changes can occur in a timeframe of a few
years. An X-SAR, together with other sensors, would
contribute to show the changes that are taking place.
Furthermore it should be mentioned that floodplains, like
that of Rio Purüs shown here, are regions of extremely
high primary production, and due to that, an area of
potential high dispute among humans and animals, for
living grounds. Again, X-SAR data could help to detect
those areas that are being occupied and those that deserve
environmental protection (e.g. breeding areas for
wildlife).
This study, together with its’ companion paper of Santos
et al. (this Symposium), present the first results obtained
from the analysis of SIR-C/X-SAR testsite in Acre.
Further studies, using the polarimetric and multi-
frequency information to be obtained from data of this
Mission, will certainly give new insights to problems
like, inasmuch canopy moisture can be better detected at
shorter wavelengths (X-band) as compared to longer
wavelengths (L-band) or on how to improve information
related to soil erosion, from surface roughness, ratioing
images from different frequencies and polarizations. Soil
erosion problems are found almost anywhere in Acre,
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996
