Monitoring of Landuse using 3S'S, A case study for Palung Sub-Watershed
Figure 2. Detail map of Palung sub-watershed
Methodology
Monitoring land use began with a visual interpretation of the 1992 orthophotomap. Land use was
classified into three major categories: forest, agriculture and shrub. Since orthophotomaps
replicate the ground and show its features clearly, this classification was easily done. Then these
three classes of land use were traced on tracing paper and scanned in TIFF format using an A0
scanner. The TIFF image was duly registered in modified universal transverse mercator (MUTM).
To reduce error in digitisation, software which converts raster data into vectors was used. After
the conversion, polygons of land use were built in Arc/Info. The attributes of the three land-use
classes were added to the corresponding polygons of land use.
The IRS1C geocoded photo is gray and topographical features are not clearly identifiable. A
visual interpretation was made, assuming that black areas were forests and whitish ones,
agricultural land. It is virtually impossible to distinguish shrub land, so only two categories were
classified. Agriculture and forestland uses were classified on the map and traced on tracing
paper. The above mentioned process is used to covert the traced land-use map to the GIS
database.
Once both maps are available in the GIS database, GIS algorithms can be used to reveal the
land-use changes which took place between 1992 and 2001. The final land-use map has five
categories: forest, agriculture, agriculture changed to forest, forest changed to agriculture and
shrub changed to forest. A flowchart of all the steps in this method is presented in Figure 3.
Similar concept is used in (Mariamni, 1997) and (Chien 2000).
Figure 3. Flow chart of the Analysis Procedure
Since the land use shown on IRS1C is not as clear as that shown on the orthophotomap, GPS
was used to provide intensive field verification. Rangers and villagers were asked to identify land
use. Sample plots were decided upon and Differential Global Positioning System (DGPS) was
carried out for those plots. After post processing the DGPS for sample plots, the results were
plotted on the land-use change map. It was found that the interpretations of land use on most of
the plots were correct and the few inaccuracies were adjusted.
