Modeling Water Quality of Rivers Using QUAL2Kw Model (Case Study: Shahroud River)

Modeling water quality of rivers can be used as one of the most effective tools for water quality management in rivers and reducing the environmental impacts of entering pollutants. The purpose of this paper is to use the valid QUAL2Kw 5.1 model to model water quality in Shahrood River.

In this paper, seven parameters of water quality have been used including dissolved oxygen (DO), biochemical demand oxygen (BOD), pH, total dissolved solids (TDS), total phosphorus, and total nitrogen four times in Shahrood River. Data from October of 2007 and July of 2008 were used to calibrate and data from September and October of 2008 were used to verify the model. Auto-calibration of model coefficients was done using genetic algorithm of the model. In order to compare simulated results with the observed data, determination coefficient and mean absolute error were used.

The most important calibration coefficients of the model were related to TSS, BOD, total nitrogen and phosphorus. This model in simulation of pH and EC with mean absolute error of 0.19 and 163.89 during verification stage showed the most and the least accuracy, respectively. On average the minimum and maximum DO were measured 6.93 and 9.99 mg/L in September and October of 2008 respectively in Shahroud River. Also the highest and lowest accuracy of the model in simulating these parameters were related to July and October of 2008 with mean absolute error of 0.86 and 1.29, respectively. In addition the results showed accurate hydraulic modeling of hydraulic parameters changes of the river along the river had a great influence on modeling of the water river quality.

The results of this paper show the accuracy of the QUAL2Kw model in simulating water quality parameters of Shahrood River. On the other hand, the accuracy of the simulation of each parameter varies with the amount of its variation along the river so that the less the changes in a parameter along the river and at different intervals, the higher the accuracy of the model in simulating this parameter will be.