Selection of Water Quality Monitoring Stations - Water Quality Management: Assessment and Interpretation - page 5

was introduced by Bellman ( 1957 ). Letternmaier et al. ( 1984 ) suggested an optimi-
zation method, which they used the DPA, for inspection of water quality station. They
applied the technique for reduction of the number of stations in the urban water quality
monitoring network. The results showed a reduction of monitoring stations from 81 to
47. The DPA was studied and extended by some researchers such as Harmancioglu
et al. ( 1994 , 2004 ) and Icaga ( 2005 ). Harmancioglu et al. ( 2004 ) applied the DPA on
Gediz River in the West of Turkey for reduction of water quality monitoring stations.
Icaga ( 2005 ) assessed existing water quality monitoring stations of the Gediz River
applying the DPA for different water usage and allocated a different weight for indices
of each water quality monitoring site. Cetinkaya and Harmancioglu ( 2012 ), applied
the DPA for assessment of water quality monitoring stations. The results showed that
the DPA was a suitable tool for optimization of the number of monitoring stations
which are going to be remaining. Asadollahfardi et al. ( 2014 ) used the DPA for
assessment of existing water quality of Sef

d River in North of Iran. The results
described that the DPA reduced the number of stations of the network.
As mentioned previously, there are numbers of methods for selection of stations.
In this chapter Sanders, MCDM and dynamic approach is brie

ı ̄ d-Ru

y described.

2.2 Sanders Method

The method proposed by Sanders et al. ( 1983 ) involves the identi cation of sam-

pling reaches in a river basin (Macro location) when the intent is to allocate
monitoring sites along the entire basin. According to Sanders et al. ( 1983 ), the
objectives of the sampling must be de

ned prior to the actual design process.
The emphasis on water quality management efforts has recently been shifted from
detection of stream standard violations to the assessment of overall trends in water
quality because of various complications in compliance monitoring, such as inter-
mittent or random sampling practices and incorrectly selected sampling locations. As
a result, restrictions on ef cant than those on

stream quality. In this case, a network developed for the assessment of trends must
cover sampling points which will yield information characteristic of reaches of the
river and in composite with other stations will yield information characteristic of the
condition of the river system in general (Sanders et al. 1983 ). Sanders et al. ( 1983 )
proposed their method for site selection in a water quality monitoring network with
the primary objective of detecting, isolating and identifying a source of pollution.
Sanders et al. ( 1983 ) describe three approaches for macro location:

uent quality have become more signi

￿ Allocation by the number of contributing tributaries;

￿ Allocation by the number of pollutant discharges;

￿ Allocation by measures of BOD loadings.

These approaches, although they may produce a rather different system of stations,
work pretty well in initiating a network when no data or very limited amounts of data
are available. It must be noted that, by applying these methods, one may roughly