; such 
], the 
with 
nd to 
nere - 
elling 
sition 
vhose 
>y are 
arious 
nd to 
ically 
  
yy the 
re the 
or the 
Idrisi; 
ershed 
raphic 
of the 
region 
erived: 
:d from 
| of the 
> of our 
create a 
Yeshed 
allow 
). To 
| -local 
eration 
consists of fulfilling the "holes" until the lowest level of 
peripheral meshes so as to find an exit out for waters of the 
central mesh. 
The method the most commonly used consists of establishing 
four (04) main data levels (matrix): the numerical model of 
the terrain image, the M.N.T without "holes", direction of 
flows and the drainage system. 
The chronological elaboration of the four stages has been 
realized using G.I.S Arc Info station: 
-the "raw" M N T: is considered as a data file of basis for 
the outcome to the reasoning and as compulsory point in this 
work. 
- the filtered M.N.T and without "holes" (processed): 
techniques of elaboration of M.N.T produce models that are 
not hydrologically consistant, it is to tell exempt minima- 
premises. These last are more frequent in zones and in the 
talwegs. The study of the minima - premises proves necessary. 
The mathematical morphology allows to envisage a reply to 
this problem with the creation of intermediate curves by 
squelletisation of surfaces between two successive curves. We 
note that in practice the utilization of a filtering of data 
associated with an increase of resolution eliminates already a 
good part of the minima - premises. Once the minima - 
premises processed, we will have a filtered M.N.T and 
without "hole" and a hydrologic model consisting of a 
draining M.N.T 
- the direction of flows: The direction of flow for a mesh is the 
direction of the exit of the water of the former. The objective 
consists in making stream of all meshs of the numerical model 
of the terrain processed until the outlet of the watershed and 
recording their direction. Among the eight possibilities of 
directions, we will retain the direction of the maximal slope: 
the maximum variation rate of the altitude and the exhibition 
(orientation or aspect) that corresponds to the bearing of the 
direction of the gradient. According to the morphology of the 
terrain, three classified slopes have been determined : 0-10 
%, 10 - 20%, > 20%.The distribution of surfaces by class of 
slope for Arzew watershed is given in the table hereafter 
(Table 1): 
  
  
  
  
  
  
  
Class %Class 
Class 1 0-10% 58.04% 
Class2 10-25% 33.33% 
Class3 >25% 8.63% 
  
  
Table 1: The distribution of surfaces by class of slope for 
Arzew watershed 
- The system of drainage: it represents, for each target mesh, 
the number of stitches that pour in the former as well as the 
progress of water. Several methods have been used . These 
methods consist in creating a file of similar format to the 
M.N.T where the inscription of a path in the file increment 
values of passage of the path of (01) . In this manner, one will 
be able to know for each mesh of the watershed the number of 
meshs that are drained. Meshs mean that a flow of a river 
type appears with the specification of the surface to drain. It 1s 
necessary therefore to search a threshold comprised between 
the size of the resolution and the surface of the watershed 
allowing to restore the hairy hydrographic . For a threshold 
equal to the resolution of the mesh, one will obtain the limit 
of the watershed. On the other hand, the search of an 
appropriate threshold is necessary to have the correspondence 
between cartographied rivers and those obtained by automatic 
extraction. The obtained results are satisfatory: 
- The hydrographic system extracted from the M.N.T with the 
choice of an adequat threshold, obtained from the satellite 
imagery that can improve its hierarchisation and digitalised 
from the topographic map are very close. 
- For a resolution of 100 x 100 m the surface of the 
watershed extracted automatically is 61,30 Km2. and the 
digital surface manually on a digitizing table is 59.9 km2. 
4.1.2 Land cover by remote sensing: Modifications 
affecting the vegetation cover of a given watershed,having 
anthropic or natural origins, have a direct influence on the 
relationship existing between precipitations and flows of 
surface. For rains with similar frequency, the evolution of the 
hydrogrammes and inondables zones are dependent on the 
change of the occupation of the ground. Classes of occupation 
of the necessary ground for the modelisation have been 
obtained from the satellite image LANDSAT TM of 15 
March 1993 of the watershed of Arzew. Classes have been 
chosen as compared to the production of the streams and 
characterized by necessary attributes for the modelisation: 
Forest, Crop , Nude Ground, Building. They have been 
obtained by a supervised classification based on the method 
of maximum likelihood. 
The distribution of surfaces by class of occupation of the 
ground after windowing on the limit of the watershed is 
given by the table hereafter (Table 2): 
  
  
  
  
  
  
classes area (km?) %class 
Building 9.47 15.46 
Crop 16.03 26.17 
Forest 26.01 42.46 
Nude Ground 975 15.92 
  
  
  
  
Table 2: The distribution of surfaces by class of occupation of 
the ground 
4.2 Meteorological data 
It is important to specify that in the meshed Hydrologic 
Model, each rainy episode is considered separately. In our 
work, we have determined a hyetogramme of the rain of 
project with the quantitative characteristic contribution of the 
most usual of rains namely parameters of Intensity curves - 
Duration - Frequency (IDF). This process is used currently in 
most of hydrologic studies as enteries of all form of 
hydrologic model. 
We have based our study on data of pluviometric stations on 
the region of Arzew for the construction of our hyetogramme 
with the adjustment of the law of Talbot and the application 
of the formula of CHICAGO for period of return of 5, 10, 20 
,50 years. Parameters of adjustment of the law of Talbot for 
different periods of return are found on the table 3. 
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 93 
ÓOOÜIÓÓ—MÁ——MÁÜÁ—3