Geography Reference
In-Depth Information
The classification of natural vegetation outside the irrigation areas in arid and
semi-arid regions in Syria, especially in the fifth agricultural stabilization zone,
was made difficult because of the dominant and variable soil signal (spectral
response) (Huete et al.
1994
). This was also true for the agricultural crops and trees
(because of the relatively wide dimensions between the trees in relation to the
spatial and spectral resolution of the used remotely sensed data), that were culti-
vated particularly in the third and fourth agricultural stabilization zones.
In this study, two approaches of change detection techniques were applied to
almost all agricultural areas in the arid and semi-arid ERB-environment in Syria,
to test the effectiveness of the two techniques in mapping the changes. The pre-
classification change detection approach that was based on image differencing was
very effective in mapping the change from bare areas to cultivated areas (the new
irrigation projects) over the time period 1975-2007. The post-classification
approach detected, mapped and defined 21 type of change. However, it offered a
lower accuracy (83 %) rather than the first method (86 %), because it depended on
the quality of already achieved classification and dealt with more types of change
(21) rather than the first method (3). Therefore, it contradicted the assumption that
this was the most accurate change detection approach (Mas
1999
). The two
approaches were easy to interpret.
Based on the pre-classification change detection approach (see
Chap. 6.3.1
)
,
there were three major trends of activities of land use/land cover: The first trend
(no change) was stable and the most dominant with about 88.62 %. The second
was negative, where as in most arid and semi-arid regions, the major cause of
natural vegetation change to bare areas (4.74 %) was related to the climatic factor
of precipitation, which is unstable and changes from one year to another. In
addition the human factor of overgrazing must be taken into consideration. The
largest area of natural vegetation exists in the fifth Agricultural Stable Zone (ASZ)
which is made up of natural pastures. The third major driving force was positive
because it accounted to a decrease in bare areas (7.19 %), where as in most
developed countries, it was related to the activities of cultivation agriculture.
These results were confirmed by applying the post-classification change detection
approach (see
Chap. 6.3.2
)
, where the change value of natural vegetation was
68.33 %, in which 42.95 % was transformed to cultivated areas and 24 % trans-
formed to bare areas, and the loss was at 43.22 %; the change value of cultivated
areas was 17.86 %, and the gain was at 35.49 %. This was on account of the
natural vegetation with 304,983 ha and bare areas with 263,863 ha. The change in
bare areas was about 13.89 %, where 10.01 % transformed into cultivated areas,
and the gain was only 0.23 %.
The major limitations of this study were the MSS-data of 1975 that were
characterized by low spatial resolution of 60 9 60 m and low spectral resolution
of four bands. The corrected LANDSAT-ETM+/SLC-OFF/data of 2005 that were
fused with ASTER-data to increase their spectral resolution from three bands to six
bands, were obtained after the applying a correction method from USGS. The time
period lag between the remotely sensed data of the years 1975 and 1987, and the
field-work in 2007 and 2009 limited the full usefulness of using the remotely
Search WWH ::
Custom Search