losses 
rivers 
n and 
ig the 
aerial 
e due 
uman 
z ts 
/land- 
river 
in the 
ild be 
risible 
ason). 
> with 
vered 
s and 
m are 
uthern 
ith the 
tor is 
varies 
| river 
ent is 
either 
n the 
Large 
etable 
d and 
nostly 
ttings, 
1 talus 
1S are 
1.0m 
1 only 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 
sediment control and flood protection measures existing in the 
site. Excluding gabions, ditches, installed mainly for irrigation 
purposes, are seen at the site. The completely/partially damaged 
gabions and ditches at the site display the impacts of flood 
events occurring in the monsoon. 
en SA = 
a 
Sardy sreag 
* 
rt 
  
Figure 1. Location of Study Area 
2.1 Geomorphometric Parameters and Land Use/Land 
Cover Classification 
In this study 1988 dated Landsat 5 TM and three Landsat 7 TM 
images taken 1999, 2000 and 2003 were used to investigate 
landuse/land cover changes in the study area. The images were 
geo-referenced and ortho-rectified, using topographic map, 
aerial photograph and field observations. A hierarchical 
classification of four Landuse/land cover classes within the 
watershed was performed by unsupervised multitemporal 
classification by means of the ISODATA (Iterative Self 
Organizing Data Analysis Technique) algorithm in Erdas 
Imaging 8.4 environment. ISODATA, a clustering method, 
classify pixels iteratively, redefine the criteria for each class, 
and classifies again, so that the spectral distance pattern in the 
data gradually emerge. The convergence threshold was set as 
0.95 pixels and the number of iteration was 24. Four land cover 
types, class 1: mix of cultivated land and settlements; class 2: 
mix of sparse forest, vegetable farms and settlements; class 3: 
mix of flood plains, eroded and depleted lands; and class 4: 
dense forest, were classified. Where, classes 1, 2, 3 and 4 will 
be refer as cultivation land, sparse forest, floodplains/eroded 
lands and dense forest, respectively. An example of image 
classification for the year 2003 is given in Figure 2. : 
565 
  
ENSE Cultivation lands 
Sparse forests 
Floodsplains 
Dense forests 
  
2 Kilometers 
Figure 2. Classification Results of Image taken in 2003 
Digital data for deriving geomorphometric parameters: drainage 
area, stream length, altitude difference, numbers of streams, 
drainage density and bifurcation ratio, were prepared from 
topographical map of scale 1:25,000 (Figure 3). Sub-watersheds 
were delineated on the topographical map according to Strahler 
branch ordering method. Delineated topographical map was 
then scanned, digitized and produced DXF (Digital Exchange) 
files by AutoCAD 2000 software, and performed spatial 
analysis in ArcView 3.1 desktop environment. For the 
generation of DEM, contour lines were scanned from 1:25,000 
scale topographical map having contour interval of 20 m. 
Altitude differences, watershed areas and stream lengths within 
the watersheds were measured directly from the topographic 
map. Digital planimeter (X-PLAN 360) was used for area and 
stream length measurement. Areas of the watershed were also 
calculated form the GIS software (ArcView). The difference in 
values obtained from manual and automated measurements are 
shown in Table 1. The total difference in whole Thadokhola 
watershed is approx. 0.10 km?, whereas, total differences in first, 
second and third watersheds are 0.03, 0.07 and 0.11 km? 
respectively. Here, all the geomorphometric parameters: 
watershed area, stream frequency and drainage density those 
implicated in watershed areas, were calculated from the 
automated values. 
The difference values in these two measurements are 
considered due to the errors exerted during. These are; re- 
delineation, re-alignment and re-arrangement of watersheds 
while interpreting and digitizing into the AutoCAD and 
ArcView environment and errors governed by the thickness of 
watershed boundaries drawn on topographic map (watershed 
boundaries were delineated by 0.3 to 0.5 mm ball pens). From 
this, it could be mentioned that doing spatial measurement on 
map, especially in measuring comparatively small watershed 
area of less than 1 km? (1:25,000 scale of standard topographic 
map) needs very fine and precise workmanship. High 
possibilities of occurring human errors, difficulties in re-