**3. Results and discussion**

#### **3.1. Model calibration and validation**

Predicted and measured values of cumulative deep percolation (DP) for different soil types are presented in **Figure 1**. Comparing linear relationship between the predicted and measured values of DP with the 1:1 line, the measured values of DP matched well with the predicted values. This indicated that the HYDRUS-1D model is capable to predict DP at different irrigation treatments. The slopes of the linear relationship are statistically equal to 1.0 and the values of NRMSE and "d" are 0.12–0.15, 0.21–0.991, and 0.987–0.976 for sandy loam, loam, and clay loam, respectively. These indicated a high accuracy of the prediction of DP by HYD-RUS-1D model for barley crop.

**Figure 1.** Relationship between predicted and measured values of deep percolation for barley.

Values of measured and predicted leached PO4 for barley crop during the growing season at different soil lysimeters and for different irrigation water are shown in **Figure 2**. The linear Simulation of Phosphorus Transport in Soil Under Municipal Wastewater Application Using Hydrus-1D http://dx.doi.org/10.5772/6621 185

**Figure 2.** Relationship between predicted and measured phosphate leaching for barley.

relationship between the measured and predicted values of leached PO4 were compared with the 1:1 line and the slope and intercept values were calculated. Ideally, the slope and intercept should be one and zero, respectively, indicating a perfect match between predicted and measured values. However, this is a very strict requirement and rarely met in practice. In this study, the slopes of the linear relationship for PO4 is statistically equal to 1.0 and intercept values were 0. 216, 0.870, and 0.036 for sandy loam, loam, and clay loam, respectively. The close similarity between the measured and predicted PO4 content at different soil profile depths over


1 AE, the average error; RMSE, the root mean square error; NRMSE, normalized root mean square error; and d, the index of agreement.

\*W1: freshwater, W2: mixture of and effluent, W3: effluent, W4: wastewater.

\*\*S: sandy loam, L: loam, C: clay loam.

**Model parameter Soil sample\***

Soil bulk density, g cm−3 1.51 1.43 1.35 Longitudinal dispersivity, cm 1 1.15 1.23 Equilibrium constant-adsorption isotherm coefficient, cm3 mg−1 1 1.25 1.35 Shape fitting parameter-adsorption isotherm coefficient, – 1.35 1.45 1.6

Predicted and measured values of cumulative deep percolation (DP) for different soil types are presented in **Figure 1**. Comparing linear relationship between the predicted and measured values of DP with the 1:1 line, the measured values of DP matched well with the predicted values. This indicated that the HYDRUS-1D model is capable to predict DP at different irrigation treatments. The slopes of the linear relationship are statistically equal to 1.0 and the values of NRMSE and "d" are 0.12–0.15, 0.21–0.991, and 0.987–0.976 for sandy loam, loam, and clay loam, respectively. These indicated a high accuracy of the prediction of DP by HYD-

**Figure 1.** Relationship between predicted and measured values of deep percolation for barley.

Values of measured and predicted leached PO4 for barley crop during the growing season at different soil lysimeters and for different irrigation water are shown in **Figure 2**. The linear

\*

S: sandy loam, L: loam, C: clay loam.

**3. Results and discussion**

RUS-1D model for barley crop.

**3.1. Model calibration and validation**

**Table 7.** Transport and reaction parameters for different soil samples.

184 Soil Contamination - Current Consequences and Further Solutions

**S L C**

**Table 8.** Statistical indexes for calibration and validation of HYDRUS-1D1 . time resulted in a high correlation coefficient (0.991), high index of agreement (0.984), low average error (0.077), low root mean square error (0.312 mg l−1), and low normalized root mean square error (9%), demonstrating a very good calibration of the model (**Table 8**). These indicated a high accuracy of the prediction of leached PO4 by HYDRUS-1D model for barley crop in different soil types. The model overestimated the measured phosphate leaching in all soil types used in the model simulation. Correlation coefficient values were at around 0.914, index of agreement at around 0.907, average error at around 0.305, root mean square error values at around (0.0298 mg−1), and normalized root mean square error at around 11% for all lysimeter soil. Overall, the values calculated for phosphate leaching demonstrate a good correlation of the model to field data.

#### **3.2. PO4 1 leaching to depth**

The findings of phosphor concentration in different kinds of irrigation and drainage water are displayed in **Figure 3**. The percentage of phosphate removal was high in all treatments (between 91 and 99%), which revealed the good potential of crop and soil system in phosphate removal. In **Table 9**, the averages of phosphate in drained water in different treatments during growing season are displayed. The effects of soil and irrigation water on transfer of phosphor to root zone are described below:

**Figure 3.** Mean phosphate leaching during the growing season.

Simulation of Phosphorus Transport in Soil Under Municipal Wastewater Application Using Hydrus-1D http://dx.doi.org/10.5772/6621 187


\*W1: freshwater, W2: mixture of and effluent, W3: effluent, W4: wastewater.

(1)Total phosphorus inputs in terms of milligrams per liter, from irrigation water.

(2)Total phosphorus output in milligrams per liter, measured in lysimeter drainage water.

(3)Total phosphorus output in milligrams per liter, simulated in lysimeter drainage water.

(4)Percent transfer, represents the amount of total phosphorus observed in drainage water drains compared with the

input values of irrigation water at each sampling time.

\*\*S: sandy loam, L: loam, C: clay loam.

time resulted in a high correlation coefficient (0.991), high index of agreement (0.984), low average error (0.077), low root mean square error (0.312 mg l−1), and low normalized root mean square error (9%), demonstrating a very good calibration of the model (**Table 8**). These indicated a high accuracy of the prediction of leached PO4 by HYDRUS-1D model for barley crop in different soil types. The model overestimated the measured phosphate leaching in all soil types used in the model simulation. Correlation coefficient values were at around 0.914, index of agreement at around 0.907, average error at around 0.305, root mean square error values at around (0.0298 mg−1), and normalized root mean square error at around 11% for all lysimeter soil. Overall, the values calculated for phosphate leaching demonstrate a good

The findings of phosphor concentration in different kinds of irrigation and drainage water are displayed in **Figure 3**. The percentage of phosphate removal was high in all treatments (between 91 and 99%), which revealed the good potential of crop and soil system in phosphate removal. In **Table 9**, the averages of phosphate in drained water in different treatments during growing season are displayed. The effects of soil and irrigation water on transfer of phosphor

correlation of the model to field data.

186 Soil Contamination - Current Consequences and Further Solutions

**3.2. PO4 1 leaching to depth**

to root zone are described below:

**Figure 3.** Mean phosphate leaching during the growing season.

**Table 9.** Mean phosphate input, output, and transfers percentage.

**The effect of soil:** Types of soil had significant effect (*p* < 0.05) on phosphate concentration in lysimeters drained water. LSD test showed that the amount of phosphate transferred to root zone in sandy loam lysimeters was significantly higher than in loam lysimeters. Also, the amount of phosphate transferred to root zone in loam lysimeters was lower (except in control treatment) than clay lysimeters. One possible reason for this difference is considerable growth of crop in loam soil and also different permeability of different soil types. Low permeability of clay soil and phosphate absorption by soil particle are the factors influencing less transfer of phosphate to the depth. Of the loam lysimeters irrigated by effluent, wastewater, and mixture of freshwater and effluent, only about 0.97–6.2% of influent phosphor was drained. Also, in clay and sandy loam lysimeters about 1–6.7% and 1.2–8.1% of influent phosphor was drained, respectively. Since in sandy loam soil the amount of phosphor uptake by crop was not high (because of nonconsiderable growth of crop), the removal of more than 90% of phosphor in sandy loam soil suggested the ability of soil in the removal of phosphor available in wastewater and effluent. The findings are consistent with Kardos and Hook [16], who reported in their study that in loam and clay loam, the amount of phosphor leaching in the depth of 120 cm were 1 and 0.1% lower than influent phosphor, respectively. About 97–99% of phosphor removal in crop and soil system was reported by Hasan Oghli et al. [17].

**The effect of irrigation water**: Simulation results showed that the effect of type of irrigation water on phosphate concentration in drainage water of lysimetrs was significant at *p* < 0.05. There was no significant difference among the amount of phosphate in drained water of lysimeters irrigated with wastewater, effluent, and mixture of freshwater and effluent. However, there were significant differences between the amount of phosphate in drained water of freshwater treatments and the other treatments. According to the findings, we can say that the amount of phosphate output from lysimeters was dependent on the growth of crop and type of soil compared to type of irrigation water.

**The effect of sampling time**: The findings showed that sampling time had no significant effect on the amount of transferred phosphate; however, in the middle of growing season, the amount of transferred phosphate to the depth was at the maximum level.

Once the discharge of drainage water from underground drains to surface water and groundwater is considered, the amount of phosphate phosphor should not be more than the determined standards. In our research, in the worst situations, the amount of phosphate in lysimeters drained water did not exceed 0.11 mg l−1, which was lower than the standard level [10].
