Artificial Neural Network - Water Quality Management: Assessment and Interpretation - page 81

5.4 Application of Arti cial Neural Network

5.4.1 A Case Study: Zaribar Water Quality (Iran)

Zaribar Lake is located near the North West of Marivan (North West of Iran), in
46 latitude and at an altitude of 1,250 m from the sea

surface. The lake area varies due to the variation in water volume during different
seasons of the year. The minimum depth is about 2 m and the maximum is 6 m. The
perimeter of the lake is about 22.5 km. The average rate of rainfall is 786 mm per
year and relative humidity is 58.4 %; average vaporization is 1,900 mm annually.
Figure 5.4 and Table 5.1 indicates the location and a statistical summary of water
quality data in Zaribar Lake, respectively (Irani 1991 ).
Sewages from Marivan city and its surrounding villages are discharged into
Zaribar Lake. This resulted in high amounts of Phosphor and Nitrogen and phos-
phate and nitrate chemicals which are washed up by rain water from the adjacent
farmland. Unfortunately, this causes a decrease of DO of water by growth of algae.
This may cause an eutrifcation phenomena in the lake. Figure 5.4 presents the
situation of Mari van (Zaribar Lake) in the map of Iran.
The aim of this case study was to

° ,6 ′ longitude and 35 ° ,31 ′

find effects of TP and TN parameters in
increasing and predicating BOD parameter (eutrophication) in Zaribar Lake using
the MLP technique and indicating application of MLP neural network.
Twelve year data of Zaribar Lake (Table 5.1 ) with 1 month time delay was used
to the modeling process. Eight years of the data were used for network training and

Fig. 5.4 Situation of Marivan
(Zaribar Lake) in map of Iran