Geography Reference
In-Depth Information
The mixing between the natural vegetation class and the rain-fed based crops
areas was lesser than it was in the LANDSAT-TM-May-2007 coverage. Also, the
total water area was lesser than it was in the 2007-coverage, because of the
recently constructed dams. There were areas located in the five agricultural sta-
bilization zones, that classified in 1987-coverage as cultivated, but were reclas-
sified in 2007-coverage as natural vegetation and/or bare areas. This was related to
the prevention of agriculture after 1990 in the fifth agricultural stabilization zone.
There was also (using the all available remote sensing data for this study) a
problem in separation and misclassification between fallow and bare areas with
dark soils colours.
The ability to separate the inactive volcanoes area from the dark colored fallow
fields and classify them, is better rather than with the MSS, because the higher
spatial and spectral resolution of TM.
Another problem detected while using TM-data was correlation between the
Euphrates River waters in the shallow and less wide areas because of the number
of the training samples concentrating on cultivated lands at the expense of the
water areas. However, this problem was overcome by increasing the number of the
representative training samples of the water, which were distributed suitably over
the whole borders of Euphrates River.
TERRA-ASTER-May-2005 data fused with LANDSAT-ETM+-May-2005 data.
These data were found to be optimal in classifying the ERB to the 5 general LULC-
classes (Fig.
6.4
), especially for classifying the artificial surfaces which had been
classified with poor accuracy using the 3 other data-coverages. So, their outcomes of
classification can be considered as a base when more accurate statistical information
is required about the distribution of the general classes within the ERB.
LANDSAT-TM-May-2007 data. There was a problem in spectral separation
between the inactive volcanoes area with dark colour and the fallow lands. Yet,
after drawing a large ROI in the volcano area, the spectral separability increased
from 1.61 to 1.78. This accentuated the high significance of the process in
selecting the training-samples and the bases it included (number of the experi-
mental areas that represented each class, area or total areas of the samples for each
class, the shape of the samples whether pixel or polygon, and the geographical
spatial distribution of these training samples within the study area).
There was a limited spectral correlation between natural vegetation and the
cultivated areas, particularly those lands of rain-fed plantations, because of their
similarity in the green vegetation bio-mass (Fig.
6.5
).
6.1.2 The Temporal Development Mapping of the Irrigated
Areas
In Syria, the majority of irrigation projects were devised for two main crops, wheat
in winter and cotton in summer. However, since the 1990s other crops have
emerged, such as yellow corn, sunflower, peanuts and watermelon. These new
Search WWH ::
Custom Search