How to Minimize the Nitrogen Pollution Risk of Applying Reclaimed Sewage for Urban Turfgrass Irrigation

Abstract

Reclamation of treated sewage is an important way to alleviate urban water scarcity and optimize ecological layout, especially in irrigating urban turfgrass. Nevertheless, the irrational use of reclaimed sewage could result in risk of excessive nitrogen (N) pollution, which requires a scientific understanding and assessment. This study examined the water-N transport process of the turfgrass system with a HYDRUS-2D model that was accurately calibrated and validated using a set of field experimental data in North China. By integrating 15 scenarios with different irrigation levels and N applications into the model, the turfgrass water flow and N fate characteristics were estimated. The results showed that the adjusted HYDRUS-2D model effectively simulated the volumetric soil water content, drainage water, N leaching, and soil N residual. The temporal variation in turfgrass water loss and N leaching consistently followed that of precipitation and irrigation, with more than 60% of the total drainage water occurring from June to August. The N leaching was at its peak during April and August, and total ammonium-N and nitrate-N leaching was 2.86 and 2.02 kg/hm², respectively. In simulated scenarios, the turfgrass drainage water was significantly reduced by 26.82% under I_60%S_1/3-I_60%S₃ scenarios (I was 100%, 80%, or 60% of total irrigation and S was 1/3, 1/2, 1, 2, or 3 times the experimental sewage concentration), while root water uptake only decreased by 0.85%. Meanwhile, N leaching and soil N residual were significantly reduced by 3.94% and 26.56% under I_60%S_1/2, respectively. Furthermore, by the TOPSIS entropy weight method, I_60%S_1/2 was identified as an optimal turfgrass irrigation strategy for the semi-arid region of North China. These results provide a guiding basis for sewage green treatment and urban sustainable irrigation on turfgrass.

1. Introduction

Water is a crucial strategic resource for supporting social and ecological development, but ensuring its sustainability presents huge challenges. Due to the highly uneven distribution of global freshwater resources, combined with rapid population growth and accelerated urbanization, some countries and regions have already experienced varying degrees of water scarcity. It is predicted that at least 5 billion people will face water insecurity by 2050 worldwide [1]. Although China’s total water resources (2876.12 billion m³) are relatively abundant, accounting for approximately 5.1% of the world’s total water resources, per capita water resources are only one-quarter of the global average [2]. Meanwhile, China’s sewage discharges have been increasing year by year, with urban sewage annual discharge rising sharply from 46.27 billion tons in 2012 to 63.93 billion tons in 2022 [3].

Urban turfgrass is an essential component of urban green land, playing important roles in regulating water balance, improving microclimate conditions, and preventing soil erosion [4]. In recent years, the rapid development of urbanization has led to the continuous expansion of urban turfgrass space. Currently, the area of turfgrass in the United States is between 1640 and 2020 hm², more than three times the area of any irrigated crop [5]. In 2022, China’s urban green land coverage rate achieved 42.69%, but the establishment and maintenance of turfgrass lacked matching water resources. Rational reclamation of treated sewage (also known as reclaimed sewage) for irrigating turfgrass has become a major strategy to ensure the sustainable development of water-shortage cities. For instance, the state of Florida in the United States is a major consumer of reclaimed sewage, with more than 50% of reclaimed sewage used for irrigation purposes in golf courses, parks, sports fields, and other recreational landscapes [6]. Australia uses 300 million m³ of reclaimed sewage annually, of which about 76% is allocated for landscape environment and agricultural irrigation [7]. China has achieved the treated sewage reuse of 6 billion m³ annually, with approximately 25% of urban green land adopting reclaimed sewage for irrigation in Beijing City [8]. Recycling of treated sewage provides an effective approach to address global water scarcity.

Due to the advantages of stable supply, abundant quantity, and easy accessibility, reclaimed sewage has become an important water source for turfgrass irrigation in arid and semi-arid regions [9]. On the one hand, large amounts of nitrogen (N) in sewage provide important nutrients for crop growth and development, thereby reducing fertilizer application [10]. On the other hand, excessive N input into the plant–soil–groundwater system poses threats to environmental safety and human health [11]. Determining the optimal concentration of N in reclaimed sewage to meet the basic requirements of plants while preventing groundwater contamination and excessive N accumulation in the soil poses a major challenge [12]. Some studies on irrigating turfgrass with reclaimed sewage have primarily focused on the effects on soil physicochemical properties (e.g., soil water, salt, and nutrient status) [13,14], turfgrass growth and quality [15,16], microbial community characteristics [17], and environmental pollution risks [18,19,20]. However, there is very limited knowledge of combining different reclaimed sewage irrigation strategies to evaluate the turfgrass water and N dynamics.

For measuring water flow and N transport processes in turfgrass, the complexity and uncertainty of the various influences within the soil environment render field investigations expensive and time- and labor-consuming. The HYDRUS-2D model developed by the U.S. Salinity Laboratory can address these limitations and effectively assess the impact of different approaches to field management on the two-dimensional movement of water and solutes in the variably saturated zone [21]. This model has been successfully used to simulate water flow and N leaching of turfgrass [22,23,24], demonstrating that HYDRUS-2D is an effective tool for conducting such studies and the simulation results are convincing. Despite some relevant studies, there have been no investigations into developing an optimal reclaimed sewage irrigation strategy for turfgrass using HYDRUS-2D.

Therefore, this study aimed to address the following objectives based on turfgrass field experiments and the application of the HYDRUS-2D model: (1) to calibrate and validate the processes of soil water, ammonium-N (NH₄⁺-N), and nitrate-N (NO₃⁻-N) transport, (2) to clarify the dynamic response of turfgrass water flow and N leaching under reclaimed sewage irrigation, and (3) to assess the impact of different irrigation strategies on soil water and N fate.

2. Materials and Methods

2.1. Field Experiments

2.1.1. Site Description

The field study was carried out during the 2020 growing season of turfgrass at the Hebei University of Architecture in Zhangjiakou, China. The site is located at 40°45′ N latitude, 114°53′ E longitude, located in an arid and semi-arid region, with an average altitude of 724.2 m above sea level. Weather conditions during the experimental period (5 April to 5 November) (Figure 1) showed that the daily maximum temperature ranged from 6.48 °C to 37.56 °C, the daily minimum temperature ranged from −4.26 °C to 24.44 °C, with a total precipitation of 305.21 mm, which occurred mostly from June to August. The soil profile at the experimental field was 120 cm with the texture of sandy loam, a pH of 8.51, organic matter of 7.8 g/kg, and conductivity of 0.13 ms/cm [2]. The initial soil N concentrations were 0.49 g/kg (total N), 38.28 mg/kg (alkali-hydrolyzable N), 3.45 mg/L (NH₄⁺-N), and 1.59 mg/L (NO₃⁻-N).

2.1.2. Experimental Design

The turfgrass species selected for the experiment was Poa annua L., which was planted in June 2019. The experimental device was 80 cm × 80 cm × 120 cm with a filter layer of 10 cm gravel at the bottom, drainage hole was installed to collect regularly the leachate of free drainage using a plastic container (Figure 2). The irrigated water experienced primary and secondary treatment by the campus sewage treatment station, in which the average concentrations of total N, NH₄⁺-N, and NO₃⁻-N were 29.62, 23.09, and 4.65 mg/L, respectively. The irrigation amount per time and irrigation frequency of the turfgrass were formulated according to the reference evapotranspiration (ET₀) subtracted from precipitation (P) in this month of the previous year, where ET₀ was determined by the FAO Penman–Monteith equation [25]. Irrigation was performed thirty times in total during the experimental period, with the irrigation amount per time ranging from 9.38 to 15.63 mm. Turfgrass was mowed every half month at a height of 5 cm, while weeds were pulled in the experimental field.

2.1.3. Field Measurements and Analysis

In order to obtain experimental data for model calibration and validation, the volumetric soil water content (VSWC, cm³/cm³) of the turfgrass was constantly monitored every thirty minutes using soil moisture monitoring sensors (5TE) at soil depths of 30, 50, 70, and 90 cm in 2022. At the same time, the amount of drainage water (mm) from the bottom of the soil (120 cm) at the experimental plot was observed using a rain gauge from 21 June to 6 August 2022. After each irrigation and precipitation event, the volume of collected drainage water was measured and stored immediately in a refrigerator at 4 °C. After the water samples were filtered through a 0.45 µm filter membrane using the vacuum pump, NH₄⁺-N and NO₃⁻-N concentrations of the water samples were determined in less than 24 h by Nassler’s reagent spectrophotometry and ultraviolet spectrophotometry with national standards [26,27], respectively. On the last day of the experimental period (5 November), soil samples were collected by layer at the depths of 0–30, 30–50, 50–70, 70–90, and 90–120 cm with an auger, soil NH₄⁺-N and NO₃⁻-N contents were determined by a UV–visible spectrophotometer. Basic data such as temperature, precipitation, relative humidity, wind speed, and sunshine hours were constantly monitored relying on the HOBO-U30 weather station. The turfgrass evapotranspiration (ET) was calculated and corrected based on the ET₀ and the water balance method with a large-scale weighing lysimeter.

2.2. Modeling Approach

2.2.1. Model Description

The HYDRUS-2D model is based on a numerical approach to calculate two-dimensional finite elements of water flow and solute transport processes in variably saturated porous media [28]. Neglecting the effects of gas phase and temperature in soil water flow movement, the isothermal Darcy two-dimensional flow control equation is obtained by solving the Richards equation:

$$\frac{\partial \theta }{\partial t}=\frac{\partial }{\partial x_i}\left[K\left(h\right)\left(K_{ij}\frac{\partial h}{\partial x_j}+K_{iz}\right)\right]-S$$

(1)

where θ is the volumetric soil water content (cm³/cm³); t is time (d); h is the pressure head (cm); K(h) is the unsaturated hydraulic conductivity function (cm/d); K_ij and K_iz are components of dimensionless anisotropic tensor; x_i and x_j are the spatial coordinates (cm); S is the sink for water uptake by crop roots (cm/d), which is defined by Faddes et al. [29] as

$$S\left(h\right)=\alpha \left(h\right)S_p$$

(2)

where α(h) is the water stress response function, which ranges from 0 to 1; S_p is the potential root water uptake rate (d⁻¹). The two-dimensional spatial distribution function of the potential root water uptake is implemented in HYDRUS-2D [30,31,32] by the following equation:

$$b\left(x,z\right)=\left(1-\frac{z}{Z_m}\right)\left(1-\frac{x}{X_m}\right)e^{-\left(\frac{p_z}{Z_m}\left|z^{*}-z\right|+\frac{p_r}{X_m}\left|x^{*}-x\right|\right)}$$

(3)

where Z_m and X_m are the maximum rooting lengths in the spatial coordinate directions (cm); z* and x* are the depths or radius of maximum intensity (cm); p_z and p_r are the empirical parameters. The unsaturated soil hydraulic properties parameters were expressed using the Van Genuchten model [33] as follows:

$$\theta \left(h\right)=\{\begin{array}{r}{\theta }_r+\frac{{\theta }_s-{\theta }_r}{{\left(1+{\left|\alpha h\right|}^n\right)}^m}\begin{array}{cc}& h<0\end{array}\\ {\theta }_s\begin{array}{ccc}\begin{array}{cc}& \end{array}& & h\ge 0\end{array}\end{array}$$

(4)

$$K\left(h\right)=K_s{S}_e^l{\left[1-{\left(1-S_e^{1/m}\right)}^m\right]}^2$$

(5)

$$S_e=\frac{\theta -{\theta }_r}{{\theta }_s-{\theta }_r}$$

(6)

where θ_s and θ_r are saturated and residual water content (cm³/cm³), respectively; α is the inverse of the air entry value; n is the pore size distribution index, m = 1 − 1/n (n > 1); l is the pore connectivity parameter; S_e is the effective saturation. Transformations and reactions of soil solute are described using convection–dispersion equations:

$$\frac{\partial \theta c_k}{\partial t}=\frac{\partial }{\partial x_i}\left(\theta D_{ij,k}\frac{\partial c_k}{\partial x_j}\right)-\frac{\partial q_i{c}_k}{\partial x_i}-\left({\mu }_1+{\mu }_k^{\prime }\right)\theta c_k+{\mu }_{k-1}^{\prime }\theta c_{k-1}+{\gamma }_k\theta -S{c}_{r,k}$$

(7)

where c is the solute concentration in the liquid phase (mg/cm³); k is the number of reaction chains; D_ij is the dispersion coefficient tensor (cm²/d); q_i is the volumetric flux density (cm/d); µ is the first-order reaction rate constant (d⁻¹); µ^’ is the similar first-order reaction rate constant (d⁻¹); γ is the zero-order reaction rate constant (d⁻¹); c_r is the concentration of the sink term (mg/cm³).

2.2.2. Domain Initial and Boundary Conditions

Measured soil water and N content at five depths (0–30 cm, 30–50 cm, 50–70 cm, 70–90 cm, and 90–120 cm) before the experiment was used as the initial conditions, and missing data were supplemented by linear interpolation, while observation points were set up at the corresponding locations in the profile of the study area. To simplify the transformation process of soil N, due to the organic N concentration in the irrigation water used for the turfgrass being small, the solute reactions were only considered for the adsorption, volatilization, and nitrification of NH₄⁺-N and denitrification of NO₃⁻-N. The upper boundary of the domain was set as the atmospheric boundary, the lower boundary as free drainage, and the left and right boundaries as no flux (Figure 2). At the same time, the flux boundary condition (third type) was selected as the upper and lower boundaries for solute transport. The measured ET in the experimental field was divided into E and T as follows:

$$\begin{array}{c}E=ET\exp \left(-k\times LAI\right)\\ T=ET-E\end{array}$$

(8)

where k is the extinction coefficient, which is 0.463 [34]; LAI is the leaf area index, which is 4.11 [35].

2.2.3. Input Parameters

The initial values of soil hydraulic parameters were obtained using neural network predictions (Rosetta) in the HYDRUS-2D model and optimized by continuous inverse solution results (

Table 1. Soil hydraulic parameters of the water flow model.

Soil Depths (cm)	θr (cm−3·cm−3)	θs (cm−3·cm−3)	α (cm−1)	n	Ks (cm·d−1)
0–30	0.047	0.241	0.0016	1.149	79.3
30–50	0.059	0.296	0.0012	1.114	56.5
50–70	0.043	0.246	0.0011	1.109	47.2
70–90	0.062	0.307	0.0018	1.069	39.7
90–120	0.076	0.297	0.0031	1.209	34.4

), in which the l value was set to 0.5. Initial values of solute transport parameters were determined based on previous studies [36,37,38,39] and calibrated by inverse analysis (

Table 2. Soil solute parameters of N transport model.

Soil Depths (cm)	Ks,1 (cm−3·mg−1)	μw,1 (d−1)	μ’w,1 (d−1)	Ks,2 (cm−3·mg−1)	μw,2 (d−1)
0–30	0	0.005	0.01	0	0.01
30–50	0.04	0	0.001	0.0001	0
50–70	0.49	0	0	0.0001	0
70–90	0.49	0	0	0.0004	0
90–120	0.49	0	0	0.0004	0

Note: K_s,1 is the NH₄⁺-N adsorption coefficient; μ_w,1 is the NH₄⁺-N volatilization rate constant; μ^’_w,1 is the nitrification rate constant; K_s,2 is the NO₃⁻-N adsorption coefficient; μ_w,2 is the denitrification rate constant.

). The longitudinal dispersivity (D_L) was considered to be one-tenth of the soil profile depth (1.2 cm) and transverse dispersivity (D_T) was set to one-tenth of the D_L (0.12 cm) [37]. Molecular diffusion in the soil air (D_a) was neglected [38] and the molecular diffusion coefficient in free water (D_w) for NH₄⁺-N and NO₃⁻-N was 1.52 and 1.64 cm²/d, respectively [39]. Root water uptake was described using water stress parameter values (model defaults) proposed for turfgrass by Feddes et al. [29]. The root solutes were considered as passive uptake by setting the C_max to 10 mg/L [23], which represented allowing N uptake to occur without limitation.

2.2.4. Evaluation of the Model

In this study, measured water and N data from 5 April to 5 November 2022 were used to optimize the parameters of HYDRUS-2D. For the hydraulic parameters, the time step was set to 2 days with odd days selected for the calibration period (total 108 days) and even days selected for the validation period (total 107 days). For solute parameters, the N leaching concentrations during the experimental period (5 April to 5 November) were used for the calibration period and the soil N residual on the last day (5 November) was used for the validation period. The root mean square error (RMSE) and index of agreement (d) were calculated to assess the simulation results by

$$RMSE=\sqrt{{\displaystyle \sum _{i=1}^n\frac{{\left(P_i-O_i\right)}^2}{n}}}$$

(9)

$$d=1-\frac{{\displaystyle \sum _{i=1}^n{\left(P_i-O_i\right)}^2}}{{\displaystyle \sum _{i=1}^n{\left(\left|P_i-O_{avg}\right|+\left|O_i-O_{avg}\right|\right)}^2}}$$

(10)

where P_i is the simulated values; O_i is the measured values; n is the amount of data; O_avg is the average of the measured values. Close to zero value of RMSE and to one value of d indicate a better efficiency of the model.

2.3. Optimization of Irrigation Strategies

2.3.1. Setup of Irrigation Scenarios

After calibration and validation of the HYDRUS-2D model, it was used to simulate water and N dynamics under different irrigation scenarios. The two factors (the amount of irrigation water and the concentration of N application) were considered in the scenario setting. Based on the cumulative reference evapotranspiration (ET₀) and cumulative precipitation (P) of the turfgrass from April to October, three different levels of irrigation water were set as 100% (ET₀-P) (full irrigation), 80% (ET₀-P) (mild deficit irrigation), and 60% (ET₀-P) (moderate deficit irrigation). The amount of water for each irrigation was set at 15.63 mm and the frequency of irrigation was equally distributed according to the precipitation situation of the month. Five concentration levels of N application (1/3, 1/2, 1, 2, and 3 times) were set based on the NH₄⁺-N and NO₃⁻-N concentrations of reclaimed sewage in the experiment. Fifteen irrigation scenarios were obtained for each combination of irrigation levels and N concentration levels (

Table 3. Scenario setting for irrigating turfgrass with reclaimed sewage.

Irrigation Scenarios	Volume (mm)/ Number of Irrigation	NH4+-N Concentration of Reclaimed Sewage (mg·L−1)	NO3−-N Concentration of Reclaimed Sewage (mg·L−1)
I100%S1/3	414.28/27	7.70	1.55
I100%S1/2		11.54	2.33
I100%S1		23.09	4.65
I100%S2		46.17	9.31
I100%S3		69.26	13.96
I80%S1/3	331.43/22	7.70	1.55
I80%S1/2		11.54	2.33
I80%S1		23.09	4.65
I80%S2		46.17	9.31
I80%S3		69.26	13.96
I60%S1/3	248.57/16	7.70	1.55
I60%S1/2		11.54	2.33
I60%S1		23.09	4.65
I60%S2		46.17	9.31
I60%S3		69.26	13.96

), enabling the comparison of simulated changes in water and N leaching, root water and N uptake, and soil N distribution (30, 50, 70, 90, 120 cm depths).

2.3.2. Evaluation of Irrigation Scenarios: TOPSIS Entropy Weight Method

Currently, commonly used comprehensive indicator methods are mainly the analytic hierarchy process (AHP) and fuzzy comprehensive evaluation (FCE). However, the subjective factors to be considered are numerous and complex, which do not apply to the assessment of irrigation scenarios in this study. The TOPSIS entropy weight method, an integration of TOPSIS (technique for order of preference by similarity to ideal solution) and entropy, calculates the objective weights for each indicator and evaluates the advantages and disadvantages of the alternatives based on weighted distances close to the ideal solution [40]. This methodology not only alleviates the shortcomings of conventional subjective weighting but also enables a comprehensive evaluation of the original data across multiple levels, making results more objective and reliable. Consequently, the TOPSIS entropy weight method was selected to evaluate the effects of different irrigation scenarios on turfgrass water and N dynamics. The specific calculation steps are as follows:

(1) Use the entropy method to determine evaluation index weight

The initial matrix is X = {x_ij}_m×n, where m and n are the number of programs and the number of indicators, respectively. Construction of normalized decision matrix:

$$\{\begin{array}{l}{x}_{ij}^{\prime }=\frac{x_{ij}-\min \left(x_j\right)}{\max \left(x_j\right)-\min \left(x_j\right)}\\ x_{ij}^{\prime }=\frac{\max \left(x_j\right)-x_{ij}}{\max \left(x_j\right)-\min \left(x_j\right)}\end{array}$$

(11)

Calculation of the information entropy:

$$r_{ij}=\frac{x_{ij}^{\prime }}{{\displaystyle \sum _{i=1}^n{x}_{ij}^{\prime }}}$$

(12)

$$E_j=-\frac{1}{\ln m}{\displaystyle \sum _{i=1}^m{r}_{ij}\ln r_{ij}}$$

(13)

Calculation of index weight:

$$w_j=\frac{1-E_j}{{\displaystyle \sum _{j=1}^n\left(1-E_j\right)}}$$

(14)

(2) Use the TOPSIS model to evaluate the scheme

Construction of weight decision matrix:

$$z_{ij}=w_j\times y_{ij}$$

(15)

Determination of the positive and negative ideal solution:

$$\{\begin{array}{l}{z}^{+}=\max \left(z_1,z_2,\cdot \cdot \cdot ,z_n\right)\\ z^{-}=\min \left(z_1,z_2,\cdot \cdot \cdot ,z_n\right)\end{array}$$

(16)

Calculation of distance to the positive and negative ideal solution:

$$\{\begin{array}{l}{d}^{+}=\sqrt{{\displaystyle \sum _{j=1}^n{\left(z_{ij}-z^{+}\right)}^2}}\\ d^{-}=\sqrt{{\displaystyle \sum _{j=1}^n{\left(z_{ij}-z^{-}\right)}^2}}\end{array}$$

(17)

Calculation of closeness coefficient:

$$C=\frac{d^{-}}{d^{+}+d^{-}}$$

(18)

3. Results and Discussion

3.1. Model Calibration and Validation

3.1.1. Simulation of Water Flow

The simulation efficiency of N transport is closely related to the water flow simulation. The hydraulic parameters of HYDRUS-2D were tested with the measured VSWC at depths of 30, 50, 70, and 90 cm and deep drainage water at a depth of 120 cm in the experimental field. Figure 3 and Figure 4 illustrate the difference between the measured and simulated values of turfgrass water flow movement during the calibration (odd days in experimental period) and validation (even days in experimental period) periods. The results indicated that HYDRUS-2D was capable of effectively simulating the variations in VSWC, with the RMSE ranging from 0.004 to 0.005 cm³/cm³ and the d ranging from 0.60 to 0.75 (Figure 3). During the 90 to 110 days, the measured values showed a significant fluctuating downward trend. Due to the peak growth period for turfgrass, which had a large demand for water. However, the precipitation and irrigation were not sufficient. Additionally, the high temperature resulted in rapid depletion of soil moisture and upward movement of water in deep soil, which reduced the VSWC to an unsaturated state. The simulated values of VSWC were overall lower than the measured values at 90 cm soil depth in the later phases of the experiment, which may be attributed to the model’s inability to incorporate the influence of excessive precipitation and irrigation on the water retention properties of heavily textured soil.

Regarding the simulation results for drainage water during the calibration and validation periods, the RMSE was 3.50 and 3.78, respectively, and d was 0.95 and 0.92 during the calibration and validation periods (Figure 4). Although the measured values of drainage water were slightly lower than the simulated values due to measurement errors of the experimental equipment, there were the same variations with the simulated and measured drainage water. In summary, the HYDRUS-2D model had a reasonable agreement on simulating turfgrass water flow under the reclaimed sewage irrigation based on the selected evaluation indicators (RMSE and d).

3.1.2. Simulation of N Transport

Using the optimized soil hydraulic parameters, the solute transport parameters of HYDRUS-2D were further tested based on the N field data (N leaching and N residual in different soil layers) that were collected during the turfgrass growing season in 2020. Figure 5 shows the turfgrass NH₄⁺-N and NO₃⁻-N leaching from 120 cm soil depth during the calibration period. The RMSE and d between the simulated and measured NH₄⁺-N leaching were 1.49 mg/m² and 0.87, while the RMSE and d of NO₃⁻-N leaching were 0.99 mg/cm² and 0.92, respectively, which indicated that the simulated N leaching had a good agreement with the measured values. Although the overall trends of N leaching were the same, there were still differences between the simulated and measured values. The main reason was that the model failed to consider that the high salt concentration in the reclaimed sewage may lead to a decrease in soil porosity [41], which affected soil hydraulic conductivity and resulted in a deviation between the measured N leaching and the simulated values.

Figure 6 shows differences between simulated and measured N residual at varying soil depths during the validation period. The results showed that the RMSE and d for soil NH₄⁺-N residual were 0.17 mg/kg and 0.75, and for NO₃⁻-N residual, they were 0.40 mg/kg and 0.73, respectively. The evaluation indicators remained within a reasonable range, indicating a satisfactory performance of the model. However, there was still acceptable deviation in the spatial trend between the simulated and measured values for N residual, probably because the model did not consider more complex soil N transformation processes (e.g., organic N mineralization and ammonium immobilization). Additionally, long-term irrigation and precipitation may cause non-linear solute concentration gradients in the soil, whereas the model assumed a homogeneous soil environment [42]. Generally, the optimized set of solute transport parameters in the model was reasonable and suitable for simulating the migration and transformation of N in the turfgrass of the study area.

3.2. Characterization of Water and N Dynamics

3.2.1. Volumetric Water Content and Deep Seepage

Based on the VSWC measured in the field and the deep drainage water simulated by HYDRUS-2D, the temporal dynamic variations in turfgrass water flow were analyzed during the experimental period in 2022 (Figure 7). For the entire soil profile (0–120 cm), the VSWC of turfgrass ranged from 0.209 to 0.308 cm³/cm³ with a similar temporal trend. The VSWC at different depths decreased in the order of 90, 50, 70, and 30 cm, in which the VSWC at 30 cm depth was affected by surface evaporation and plant uptake, resulting in a strong response to precipitation and irrigation infiltration with the greatest fluctuation. The average VSWC at 50 cm depth was 0.286 cm³/cm³, which was 20.09% higher than the VSWC at 70 cm soil depth. This difference may be attributed to the fact that the 50–70 cm soil layer was the root tip meristematic zone of the turfgrass [43], where a significant amount of soil water and inorganic salts could be absorbed, and thus, the VSWC at 70 cm depth was relatively lower. The trend of VSWC at 90 cm depth fluctuated steadily, with an average value of 0.301 cm³/cm³.

Deep soil drainage is the primary pathway for turfgrass water loss. Similar to the distribution characteristics of VSWC, there was a significant positive correlation between drainage water and precipitation as well as irrigation, with a cumulative drainage water of 719.79 mm in the turfgrass growing season. The drainage water showed an uneven seasonal distribution predominantly concentrated in June and August, which accounted for 27.60% and 20.96% of the cumulative drainage water with daily average drainage water of 7.19 and 5.43 mm in these months, respectively. A heavy precipitation event occurred from 26 to 28 June with a cumulative precipitation of 95.2 mm, resulting in the maximum drainage water of 41.95 mm on 28 June in the turfgrass system.

3.2.2. N Leaching

N leaching is the main pathway for soil N loss from a turfgrass system, and its variations are determined by N application, precipitation, irrigation management, and plant uptake [44]. Figure 8 illustrates the daily changes in simulated N leaching at 120 cm soil depth during the experimental period. The results showed that NH₄⁺-N and NO₃⁻-N leaching varied significantly with the fluctuations in precipitation and irrigation, which was consistent with the transport process of soil water flow. The cumulative NH₄⁺-N leaching was 2.86 kg/hm², slightly higher than the cumulative NO₃⁻-N leaching (2.02 kg/hm²). The N leaching dynamics were consistent with the turfgrass water flow characteristics, and the NH₄⁺-N and NO₃⁻-N leaching accounted for 49.36% and 49.48% of the cumulative leaching in June and August, respectively. Influenced by the heavy precipitation, the NH₄⁺-N and NO₃⁻-N leaching reached the maximum values of 0.15 and 0.10 kg/hm² on 28 June, respectively. Therefore, precipitation played an important role in increasing N loss from turfgrass fields, which was consistent with the findings of Mehrabi et al. [45] on winter wheat fields.

There were seasonal differences in N leaching variations during the turfgrass growth period. The NH₄⁺-N and NO₃⁻-N leaching accounted for 81.38% and 73.12% of the cumulative N leaching from 5 April to 31 August, respectively, which indicated that turfgrass N leaching mainly occurred in spring and summer. In the autumn (1 September to 5 November), the cumulative NO₃⁻-N leaching (0.54 kg/hm²) was 1.95% higher than the cumulative NH₄⁺-N leaching, indicating that NO₃⁻-N leaching was dominant during the period. The main reason for this was that the soil NH₄⁺-N initial concentration in turfgrass was relatively high, and using reclaimed sewage irrigation may promote increased soil organic matter, which affected the nitrification reaction rate, resulting in a gradual increase in soil NO₃⁻-N content [46]. Under long-term precipitation and irrigation, NO₃⁻-N is prone to leaching deeper soil layers and into groundwater. Therefore, it is important to adopt appropriate irrigation management to enhance the efficient utilization of N and reduce the background concentrations of soil N.

3.3. Assessing Different Irrigation Scenarios

3.3.1. Cumulative Drainage Water and Root Water Uptake

Variations in irrigation amount and frequency will affect soil water loss through the pathways of deep drainage water and root water uptake (RWU). The turfgrass cumulative drainage water and cumulative RWU were simulated under three irrigation scenarios (I_100%, I_80%, and I_60%) using the HYDRUS-2D model (Figure 9). The results revealed significant decreases in both drainage water and RWU with decreases in irrigation amount. Under the premise of maintaining proper turfgrass growth, the cumulative drainage water of the I_80% and I_60% scenarios decreased by 12.01% and 26.82% in comparison to that of the I_100% scenario (587.46 mm), respectively, which indicated that the soil water loss was significantly lower in the I_60% scenario than in the I_80% and I_100% scenarios. However, the cumulative RWU varied slightly among the three irrigation scenarios. Specifically, the cumulative RWU in the I_80% and I_60% scenarios only exhibited reductions of 0.43% and 0.85% in comparison to the I_100% scenario (553.67 mm), respectively. Similar results were reported by Karandish et al. [47] on the positive effects of reducing irrigation levels in maintaining ET_a (actual crop evapotranspiration). In summary, the I_60% scenario exhibited significantly reduced drainage water while maintaining a favorable level of RWU, indicating that the soil water use efficiency of the turfgrass was highest.

3.3.2. Cumulative Root N Uptake and N Leaching

Figure 10 displays the cumulative root N uptake and cumulative deep soil (120 cm) N leaching under different reclaimed sewage irrigation scenarios. The results indicated that when the N application was fixed, a decrease in irrigation amount (from I_100% to I_60%) would lead to a slight reduction in root N uptake by 1.22% to 7.60%, which was attributed to the potential negative impact on root N uptake due to a decrease in RUW. However, the N leaching significantly decreased by 12.55% to 25.86%, suggesting that the negative impact on root N uptake was overshadowed by the positive effect of N leaching. Therefore, considering the previous water movement results (Figure 9), adopting the I_60% irrigation scenario was effective in reducing turfgrass water loss and N leaching.

In the I_60% irrigation scenario, the root N uptake increased significantly with increasing N application concentration (from S_1/3 to S₃) by 11.24% to 37.67%. However, NH₄⁺-N leaching (1.70 kg/hm²) remained unchanged, and excess NH₄⁺-N was retained in the soils. The increase in N application concentration resulted in an obvious upward trend in NO₃⁻-N leaching, rising by 4.38% to 22.40%, which was consistent with the results of Groenveld et al. [48], with the N emission being higher as the N application increased. This demonstrated that high levels of N application concentration may result in NO₃⁻-N leaching into deeper soil layers, potentially reducing the efficiency of root NO₃⁻-N uptake [49].

3.3.3. Soil N Residual

The soil N residual in the turfgrass was analyzed for different irrigation scenarios on the last day of the simulation period (31 October, Figure 11). When the N application was fixed, the NH₄⁺-N residual at 0–30 cm soil depth under the I_80% and I_60% irrigation scenarios was reduced by 11.75% to 16.80% and 16.47% to 25.44%, respectively, compared to the I_100% scenario (0.74 to 5.37 mg/kg). For the 30–120 cm soil depth, the NH₄⁺-N residual under the I_80% and I_60% scenarios were decreased by 1.46% to 4.54% and 1.80% to 8.15%, respectively, compared to the I_100% scenario (0.64 to 0.75 mg/kg). This indicated that the response of the NH₄⁺-N residual for the 0–30 cm soil depth was significantly greater than for the 30–120 cm, which was attributed to the fact that NH₄⁺-N was easily adsorbed by negatively charged soil colloids, making it difficult to leach downward with water transport, thus mainly aggregating in the surface soil. However, it was easily leached into deeper soil along with water movement for NO₃⁻-N, resulting in the NO₃⁻-N residual at different soil depths all responding significantly to irrigation amount variations. Under the I_80% and I_60% scenarios, the NO₃⁻-N residual at the 0–120 cm soil depth was reduced by 1.04% to 10.28% and 2.79% to 23.75%, respectively, compared to the I_100% scenario (0.40 to 2.23 mg/kg). Overall, the N residual showed a decreasing trend with a decrease in irrigation amount, which indicated that moderate water deficit irrigation (I_60%) was beneficial for promoting turfgrass root N uptake and reducing N loss.

Under the I_60% irrigation scenario, the rise in N application concentration caused a significant increase in NH₄⁺-N residual at the topsoil (0–30 cm) by 34.11% to 46.45%, whereas NH₄⁺-N residual at the 30–120 cm soil depth only increased by 0.56% to 3.22%. As for the NO₃⁻-N residual, the entire soil profile (0–120 cm) showed a significant increase of 21.26% to 40.92% with the rise in N application. The excessive N that was not absorbed by crops accumulated in the soil during the winter posed a potential pollution risk as it was susceptible to leaching into water bodies by precipitation and irrigation during spring and summer.

3.3.4. TOPSIS Entropy Weight Method

According to integrating five evaluation indicators (drainage water, RWU, root N uptake, N leaching, and soil N residual), the TOPSIS entropy weight method was utilized to identify the optimal reclaimed sewage irrigation scenarios (

Table 4. Comprehensive evaluation and ranking of irrigation scenarios.

Irrigation Scenarios	d+	d−	C	Rank
I100%S1/3	1.210	0.325	0.212	11
I100%S1/2	1.210	0.305	0.201	12
I100%S1	1.214	0.246	0.168	13
I100%S2	1.232	0.134	0.098	14
I100%S3	1.262	0.065	0.049	15
I80%S1/3	0.671	0.641	0.489	6
I80%S1/2	0.670	0.633	0.486	7
I80%S1	0.673	0.609	0.475	8
I80%S2	0.692	0.573	0.453	9
I80%S3	0.725	0.552	0.432	10
I60%S1/3	0.065	1.262	0.951	2
I60%S1/2	0.058	1.259	0.956	1
I60%S1	0.066	1.249	0.950	3
I60%S2	0.135	1.234	0.902	4
I60%S3	0.214	1.223	0.851	5

). The calculation results for fifteen irrigation scenarios indicated that I_60%S_1/2 achieved the minimum value (0.058) in the distance to the positive ideal solution (d⁺) and a relatively large value (1.259) in the distance to the negative ideal solution (d⁻). Additionally, the closeness coefficient (C, 0.956) was very close to 1. Compared to the I_60%S₁ scenario, although cumulative root N uptake decreased by 27.36%, the cumulative N leaching and soil N residual (0–120 cm) of the I_60%S_1/2 scenario decreased by 3.94% and 26.56%, respectively. Therefore, I_60%S_1/2 is selected as the optimal reclaimed sewage irrigation strategy for turfgrass in the study area.

4. Conclusions

In order to evaluate the effects of different reclaimed sewage irrigation strategies on water and N dynamics in urban turfgrass system, this study conducted field soil column leaching experiments. After calibration and validation, the HYDRUS-2D model demonstrated a good agreement on simulated values and measured values of variations in volumetric soil water content, drainage water, N leaching, and N residual. By analyzing the temporal variations in water and N during the turfgrass growing period in 2022 using the simulated values, the volumetric water contents at different depths were ranked as 90 cm > 50 cm > 70 cm > 30 cm, ranging from 0.209 to 0.308 cm³/cm³. The drainage water mainly occurred from June to August, which accounted for 62.53% of the cumulative drainage water (719.79 mm). For the soil NH₄⁺-N and NO₃⁻-N, leaching was mainly concentrated in spring and summer (April to August), accounting for 81.38% and 73.12% of the cumulative NH₄⁺-N (2.86 kg/hm²) and cumulative NO₃⁻-N (2.02 kg/hm²) leaching, respectively. Furthermore, the N leaching was primarily dominated by NO₃⁻-N in autumn (September to October). The evaluation of the irrigation scenarios (I_100%S_1/4-I_60%S₄) indicated that I_60%S_1/2 (irrigation amount of 248.57 mm, and NH₄⁺-N and NO₃⁻-N application concentrations of 11.54 and 2.33 mg/L, respectively) was the optimal irrigation strategy with the highest water use efficiency and the minimum pollution risk. In conclusion, the application of these simulation results will help formulate appropriate water and N management decisions for the study area, enabling more environmentally friendly turfgrass irrigation practices.