212
Regional Environmental Changes: Application to Soil Erosion Modeling
7
DISCUSSION
From the analysis of the results, it can be deduced that maritime Syria experienced
important changes in its overall land cover types during the 1977-1993 period
(Table 2).
Statistical analysis of the differentiation of the two classified NDVI images showed
that change occurred in 47% of the total terrestrial part of the image. In the 19771993 period, lowering of the NDVI values is registered by 31.08% of the total
terrestrial pixels: 24.74% by one, 5.50% by two and 0.84% by three slices
respectively. Meanwhile, increase of the NDVI values is recorded by 15.81 % of
the total terrestrial pixels distributed as: 13.14% by one, 2.30% by two and 0.37%
by three slices. The major change is observed for one degree of variation of NDVI
classes (-24.74% and +13.14%).
Surface distribution of each class for the two dates is depicted as follows: 14.23%
of the terrestrial part of the scene was occupied in 1977 by class 1 (non-vegetable
elements) which increased to 21.69% in 1993; class 2 (low vegetation density)
registered no significant change (23.07% in 1977 and 23.41 % in 1993); class 3
and 4 clearly record an important decrease in vegetation. In fact, these two classes
which correspond to a high and medium density of vegetation are affected by the
previously noted decrease of NDVI by one slice. The surface of class 3 occupied
32.72% of the total terrestrial surface in 1977 and in 1993 29.15% (78.5 km 2 ).
Class 4 was 29.98% in 1977 and became 25.75% in 1993 (93 km 2 ). These results
have been compared to those relative to the 1986 scene. It is deduced that no
major change occurred between 1977 and 1986, i.e. during an eleven year time
span. What caused such an important change between 1986 and 1993? The factors
responsible for this change can be subsumed under five main themes:
1 - Forest fires in summer 1992 in the North-Western part of the region;
2 - The building of the Al Kebir dam after 1984;
3 - Development of communication structures (roads, railways, yards ... );
4 - General urban growth particularly appreciable in Lattaquie city;
5 - Deforestation practices.
However, the real impact of each kind of factor needs to be all the more controlled
by field observations as other factors may be involved in change processes.
Additional investigations being conducted include the analysis of seasonal
variations especially during and after the dry season when the main degradation
processes in vegetation cover were shown to occur in the study region.
8
CONCLUSION
The multi-temporal NDVI analysis applied to maritime Syria made it possible to
detect land cover differences during the 1977-1993 period. The RGB-NDVI
composite image technique contributed to visualize and quantify vegetation
change but was unable to qualify the nature of change. This procedure was
combined with the post-classification comparison method. Classified NDVI values
were assigned to four classes referring to vegetation density. The interpretation of
the resulting image demonstrates the possibility of evaluating change both
quantitatively and qualitatively.
Regional Environmental Changes: Application to Soil Erosion Modeling
7
DISCUSSION
From the analysis of the results, it can be deduced that maritime Syria experienced
important changes in its overall land cover types during the 1977-1993 period
(Table 2).
Statistical analysis of the differentiation of the two classified NDVI images showed
that change occurred in 47% of the total terrestrial part of the image. In the 19771993 period, lowering of the NDVI values is registered by 31.08% of the total
terrestrial pixels: 24.74% by one, 5.50% by two and 0.84% by three slices
respectively. Meanwhile, increase of the NDVI values is recorded by 15.81 % of
the total terrestrial pixels distributed as: 13.14% by one, 2.30% by two and 0.37%
by three slices. The major change is observed for one degree of variation of NDVI
classes (-24.74% and +13.14%).
Surface distribution of each class for the two dates is depicted as follows: 14.23%
of the terrestrial part of the scene was occupied in 1977 by class 1 (non-vegetable
elements) which increased to 21.69% in 1993; class 2 (low vegetation density)
registered no significant change (23.07% in 1977 and 23.41 % in 1993); class 3
and 4 clearly record an important decrease in vegetation. In fact, these two classes
which correspond to a high and medium density of vegetation are affected by the
previously noted decrease of NDVI by one slice. The surface of class 3 occupied
32.72% of the total terrestrial surface in 1977 and in 1993 29.15% (78.5 km 2 ).
Class 4 was 29.98% in 1977 and became 25.75% in 1993 (93 km 2 ). These results
have been compared to those relative to the 1986 scene. It is deduced that no
major change occurred between 1977 and 1986, i.e. during an eleven year time
span. What caused such an important change between 1986 and 1993? The factors
responsible for this change can be subsumed under five main themes:
1 - Forest fires in summer 1992 in the North-Western part of the region;
2 - The building of the Al Kebir dam after 1984;
3 - Development of communication structures (roads, railways, yards ... );
4 - General urban growth particularly appreciable in Lattaquie city;
5 - Deforestation practices.
However, the real impact of each kind of factor needs to be all the more controlled
by field observations as other factors may be involved in change processes.
Additional investigations being conducted include the analysis of seasonal
variations especially during and after the dry season when the main degradation
processes in vegetation cover were shown to occur in the study region.
8
CONCLUSION
The multi-temporal NDVI analysis applied to maritime Syria made it possible to
detect land cover differences during the 1977-1993 period. The RGB-NDVI
composite image technique contributed to visualize and quantify vegetation
change but was unable to qualify the nature of change. This procedure was
combined with the post-classification comparison method. Classified NDVI values
were assigned to four classes referring to vegetation density. The interpretation of
the resulting image demonstrates the possibility of evaluating change both
quantitatively and qualitatively.