ISPRS, Vol.34, Part 2W2, “Dynamic and MultiPimensional GIS’’, Bangkok, May 23-25, 2001 
• Network definition 
using an approach similar to that proposed by Foster and 
Wischmeier: 
The stream and watershed network was determined so that 
there was a single stream segment for each watershed that had 
been modeled. The DEM cells that form the streams are 
defined as the union of two sets of grid cells. The first set 
consists of all cells whose flow accumulation is greater than a 
user-defined threshold value. This set identifies the streams 
with the largest drainage area, but not necessarily with the 
largest flow because flow depends on other variables that are 
not related exclusively to topography. The second set is defined 
interactively by the user by clicking a certain point on the map, 
which results in an automatic selection of all downstream cells. 
To include these streams using the threshold criterion, it would 
be necessary to lower the threshold value for the entire system, 
thus defining unnecessarily a much more dense stream 
network. 
Y = EL..K { C ,P,S, 
Aj (A, /22.13)" - A H (^.,/22.13)" 
A,-A, 
Where Yj, Kj, Cj, Pj, and Sj are the sheet and rill erosion, soil 
erodibility factor, cover and management factor, practice factor 
and slope factor, respectively, in the jth celJ Aj and Aj.i are the 
upslope contributing areas where flow exits and enters each 
cell, respectively. Ehois the rainfall energy-intensity. 
The LS factor derived from unit stream power theory used with 
GIS enhanced pesticides model appeared to perform 
satisfactorily. It has an advantage over the traditional method in 
that the inputs to the relationship are derived easily from the 
terrain analysis and physically can better account for the effects 
of flow convergence and divergence on erosion. 
Sub-basin outlets were also defined as the union of two sets of 
grid cells. The first set, based on the stream network, consists 
of all cells located just upstream of the junctions. Consequently, 
at a junction, two outlet cells are identified, one for each of the 
upstream branches. The system outlet is also identified as an 
outlet. The second set is defined interactively by the user by 
clicking on any cell on the stream network such as those 
associated with gages or other water control points. After the 
sub-basin outlets have been defined, a unique identification 
code is assigned to each stream segment connecting a 
headwater cell with a sub-basin outlet, or two sub-basin outlets. 
• Length-slope factor 
A minimum cell area of 280mx400m was applied with 
Pesticides Losses Runoff Model in this paper to allow a 
representative field slope length to be used in estimating the 
length-slope factor (LS) in the Universal Soil Losses Equation 
(USLE). The USLE is used to calculate the sheet and rill 
erosion in each cell. A theoretical equation derived from unit 
stream power theory is used to estimate the lengthslope factor 
in the USLE (Cialella et. al. 1997). This equation also better 
represents the effects of flow convergence aid divergence on 
erosion: 
LS = 
A, T 
[ s ' n P, 
.With 
A, 
" and S = 
sin /?, 
_22.13J 
[o.0896j 
|_22.13J 
[(10896J 
Where A s is the specific catchment area (=A/b), defined as the 
upslope contributing area (A) per unit width normal to the flow 
direction (b); p is the slope gradient in degrees; n=0.4 and 
m=1.3. The within-cell sheet and rill erosion is then estimated 
RESULTS AND DISCUSSION 
For this paper is based on an undergoing project, there are little 
results that can be showed here. Through the application of the 
methods mentioned in previous section, the slope and flow 
direction of the study basin were carried out (Fig.5 and Fig.6). 
To evaluate the reliability of terrain analysis, the results 
generated by the integration of GIS and remote sensing and the 
field measured slope and direction were compared. It was 
showed there were some disagreement though satisfying in 
general. 
Two reasons have been identified for the mismatc h between 
true topographic surface form, and its representation as a DEM 
within a GIS. Firstly, the methods themselves provide some 
conceptual limitations. It is not possible to represent fully, a 
continuous, undifferentiatable surface with a discrete, finite 
resolution elevation model. Secondly, the process of elevation 
interpolation required for DEM generation can lead to model 
error (Garbrecht and Starks, 1995; Gong, et al., 2000). 
The lack of agreement implies further refinement of thisterrain 
analysis methodology may provide more insight intopesticides 
losses runoff modeling. It suggests that GIS data, no matter 
what specific methodology is employed, may need to be 
augmented with site-specific sampling data to facilitate 
pesticides control decision making. However, the important 
point should be confirmed is the GIS and remote sensing 
integrated modeling is a useful tool for the pesticides losses 
control and the results are receivable.