INTRODUCTION
According to the definition of international Ramsar convention wetland is natural
or artificial marshy, permanent or ad interim with fresh water, marginal water,
or brackish water zones. Wetlands have area equal 856 million hectares, the
international wetlands have 75 million ha and the number of them is 1118. The
wetlands areas of Iran are 22.5 million ha and cover 1.5% of the total area
of Iran. The numbers of international wetlands of Iran are 20 with 1.3 million
ha areas.The most important factors that causes environmental crisis for wetlands
in Iran are drainage, discharge wastewater, human encroachment, immethodical
hunting, etc. The results of these factors are decreasing of aquatics production,
decreasing groundwater resources, decreasing quality of human environmental
and death of wetland. Wetlands have more advantage such as; direct recovery
of water with people, securement and storage water in aquifer, regularization
water flow especially flood flow, debarment of infiltrating saline ground and
surface water, conservation articles of food, gene bank, sources for natural
production of wetland. So, knowledge about water quality helps for management
wetland. Several methods for evaluation quality of surface water are existed
in the world. Between those, using of water quality index is the most application
and the easiest.
References to important water quality, some index for evaluation water quality
have represented. Some chemical and physical index in papers and researches
have been represented that can be divided to general index, special purpose
index, design index, statistical index and biological index. These main groups
have some sub index such as, Horton quality index, USA National Sanitation Foundation
index, Prati index, Mc Duffie index, Dinius index, Dojlido index, Walski and
Parker index, Nemerow and Sumitomo index, Oregon index, National Canada index,
Harkins index, Beta index and etc. (Ott, 1978). By evaluation
mathematical and quality structure of indexes are determined advantages and
limitation. In statistical index cannot be compared some data to another. Design
index use for management and are not use for classification surface water. Some
parameter that requirement for general index are not measured in Iran so, cannot
be used. Also, some indexes such as, Oregon index is local index. Between indexes
that have introduced, USA National Sanitation Foundation (NSFWQI) index is the
most application and easiest index. Nevertheless, although, USA National Sanitation
Foundation index has many advantages, has limitation. In this index some parameter
such as, color and oily material do not use. These parameters are important
for sportive purpose. So, if use this index for evaluation water quality for
sportive purpose, result will show that water is good but maybe has perceivable
color and oily material. Water quality index is an important factor for evaluated
surface water. Sixty two stations are selected for determination WQI for ecology^{'}s
stream. Results showed that the change of water quality index increased. Donizetti
and Galizia (2003) measured temperature, pH, electrical conductivity, dissolve
oxygen, PO_{4}, NO_{3}, coli form and biochemical dissolve oxygen
in watershed of Jaboata River from March 1998 to February 1999 for evaluation
water in eight points of river monthly. Results showed that the amount of dissolve
oxygen, PO_{4} and coli form were critical. Lin
et al. (2003) measured on Clyde frith from 1990 to 2000 in the Southwest
of Scotland water quality variables on seven stations regularly. In addition,
of data of input into water, data of ebb and flow and air temperature collected.
Then, by using artificial neural networks were applied of these data for control
and surveillance this zone and control of pathogens. Vandenberghe
et al. (2002) did a research for calibration the model of water quality
in Dender River in Belgium. Water quality model that used was ESWAT Model. In
this study, the percentage of safety model as a function of measurement data
has defined. Najafi (1994) evaluated self purification
of Jajrood River in north of Iran by using water quality and amount of self
purification by physical, chemical, biological and input wastewater were determined.
Gharae (1994) evaluated during four years quality variation
of GharehGhach River. The analysis showed that electrical conductivity was
increased and pH did not have any variation. Nasrolahzadeh
and Varedi (2003) evaluated quality of Tajan River on north of Iran with
water quality index in four stations. Results showed that in two stations of
upstream water was pure and in two stations of downstream start of severe variation
in quality was beginning. Generally, water quality of this river was in good
classification. Jaafarzadeh et al. (1998a) did
a research for evaluation effects of wastewater on quality of Dez River from
North of Dezful to BanGhir. Jaafarzadeh et al. (1998b)
did a research for evaluation variation of Karoon River on raw water of refinery
of drinking water in Ahvaz. For that, reason by using ten years data, variation
of water quality index have evaluated and seasonal variation was calculated.
Then, the effects of water quality in each steps of water purification have
analyzed. Morovati et al. (2000) evaluated effects
of industrial wastewater on water quality of Karoon River. In this study the
sources of pollution have recognized, hydrology and morphology of river and
the proportion of industries on pollution Karoon River have evaluated.
Because of important Bamdezh wetland and water management of it in this study were obtained the relation between water quality index and physical and chemical parameter of water in this wetland. After that, the optimum equations were obtained by using statistical index.
MATERIALS AND METHODS
Declared wetlands in Khuzestan Province in Southwest of Iran are; Shadegan, HorAlazim, Bamdezh and Miangaran.
Bamdezh wetland was located on 40 km of Northwest of AhvazDezful road. Kharkheh river located on West and Dez river located in east of Bamdezh wetland. This wetland located between 48^{0} 27^{' }to 48^{0} 42^{'} Eastern longitude and 31^{0} 38^{'} ^{'} to 31^{0} 55^{'} ^{'} Northern latitude. The area of this wetland is equal to 40 km^{2} with 11 km of length and 4 km of width. Bamdezh wetland is a natural wetland, permanent fresh water marsh and recharge with Shavoor River as surface water. Vegetation of this wetland consists of; hydrophytes, halophyte, xerophytes and wildlife consist of; water, beside water and xerophile fowl, fishes, amphibian and reptilian. This area has an average annual precipitation of 260 mm, temperature of 24 degree of centigrade and evaporation of 1900 mm.
Bamdezh wetland is suitable ecosystem for birds, suitable places for growing
plant, suitable place for fish of fresh water, flood control, creation microclimate
that cause increasing relative humidity and decreasing temperature, natural
landscape, etc (Bostanzadeh, 2003).
The input of Bamdezh wetland is Shavoor River and the outlet of this is Kharoor
that after pass of Tavana canal discharge to Dez River. According to data of
2002 the mean, minimum and maximum of input discharge of wetland respectively
was 16.58, 14.60 and 18.90 m^{3} sec^{1} without to take into
account the different use such as agricultural use before enter to wetland.
The mean, minimum and maximum of output discharge of wetland that has been obtain
of balance equation, respectively was 12.03, 6.54 and 15.05 m^{3} sec^{1}
(Afkhami, 2004). Figure 1 shows the
satellite image of Bamdezh wetland.
In this study for determination WQI has been used temperature, dissolve oxygen,
biochemical dissolve oxygen, chemical dissolve oxygen, NO_{3}, PO_{4},
pH,Turbidity, TSS and Coliform of Bamdezh wetland in four measurement stations
of 20012002 and 20072008.

Fig. 1: 
Satellite image of Bamdezh wetland 
By using measurement data, water quality index of each parameter has been calculated
then by function 1 water quality index of each station has been determined.
where, W_{i} is sub index of each variable and Q_{i} is weight factor of each variable.
To obtain mathematical models were used SPSS12.0 software. Water quality parameter
as independent variable (X) and WQI as dependent variable (Y) were used as input
of SPSS. Then 10 mathematical models were obtained. For evaluated mathematical
formulas performance were calculated Maximum Error (ME), Root Mean Square Error
(RMSE), Coefficient of Determination (CD or R^{2}), Modeling Efficiency
(EF) and Coefficient of Residual Mass (CRM). The mathematical expressions of
these statistics are as follows (Homaee et al., 2002):
where, P_{i} is the predicted data, O_{i} the measured data and n is the number of samples. By statistical index, determine sum of degree and was classified 10 equations.
RESULTS AND DISCUSSION
For determination optimum function between WQI and BOD, NO_{3}, PO_{4},
pH, Turbidity, TSS, Coliform, COD and DO 10 function (Linear, Logarithmic, Inverse,
Quadratic, Cubic, Power, Compound, S, Growth, Exponential) were obtained by
using SPSS12.0. The results of SPSS 12.0 and statistical index that were calculated
have been represented on Table 19.
For determination optimum equation was used statistical index and was calculated ME, RMSE, CD (R^{2}), EF and CRM between observation and measurement. The ME indicate the worstcase performance of the model. The RMSE value shows how mush the simulation overestimate or underestimate the measurements. The CD or R^{2} gives the ratio between the scatter of the simulated and measurement. Compares between the simulated data to the average measured values determine with EF.The CRM is a measure of the tendency of the model to overestimate or underestimate the measurements. For determination optimum equation should be arranged ME, RMSE, EF and CRM of each equation from minimum to maximum and R^{2 }of each equation should be arranged from maximum to minimum. Then, average of degree was calculated for equations and each equation that was obtained maximum number was selected as optimum equation. According to Table 1, cubic equation is the best imperial mathematical model between WQI and DO. Total ranking of statistical indexes indicate that cubic equation the best equation.
Table 2 indicates that cubic equation is the best imperial
mathematical model and by measurement COD and using this equation can be determined
WQI with coefficient of determination equal to 0.913. Also, the maximum error
between measurement and calculated when using this equation is 1.008 that is
suitable error in natural phenomena.
Statistical indexes in Table 3 show that cubic equation is
the best imperial mathematical model between WQI and BOD.
According to Table 4, total ranking of statistical indexes indicates that cubic equation is the best imperial mathematical model between WQI and PO_{4}. By using this equation can be determined WQI with coefficient of determination equal to 0.910.
Table 5 shows that exponential equation is the best imperial mathematical model and by measurement temperature and using this equation can be determined WQI with coefficient of determination equal to 0.745.
Results in Table 6 shows that total ranking of statistical indexes of compound equation is number one and chosen as best mathematical model between WQI and NO_{3}. By using this equation can be determined WQI with coefficient of determination equal to 0.600.
According to Table 7, equation is the best imperial mathematical model between WQI and EC. Total ranking of statistical indexes indicate that s equation the best equation.
Table 8 and 9 shows that exponential formula is the optimum equation between WQI and pH and TSS.
Generally, all tables that have been represented, show the relationship between WQI and water quality parameter is nonlinear and with application these equations and measurement BOD, NO_{3}, PO_{4}, pH, Turbidity, TSS, Coliform, COD and DO can be determined WQI in Bamdezh wetland.
CONCLUSION
Generally, for determination optimum equation in researches often is used coefficient of determination (R^{2}). This study was shown that R^{2 }is not sufficient and should be used other statistical index such as; Maximum Error (ME), Root Mean Square Error (RMSE), Modelling Efficiency (EF) and Coefficient of Residual Mass (CRM). Because some equation had high coefficient of determination but according to mention statistical index is not optimum equation. The results of this study show that the best relation between water quality index and DO, COD, BOD, PO_{4} is cubic and the optimum equation between water quality index and temperature and pH and TSS is exponential. S and compound equations are the best imperial mathematical models for water quality index, EC and NO_{3}. Generally, the results were indicated that the relationship between water quality index and physical and chemical parameter is nonlinear. By using obtaining equations and measurement BOD, NO_{3}, PO_{4}, pH, Turbidity, TSS, Coliform, COD and DO can be determined WQI with suitable coefficient of determination in Bamdezh wetland.
ACKNOWLEDGMENT
Khuzestan Water and Power Authority Company and Office of Research and Irrigation and Drainage Networks Standard supported this research. The authors are grateful to their support.