Methodology
This study followed a three-step methodology. First, groundwater quality was simulated with a focus on the parameters of NO3 and chloride. In the second step, different scenarios were testified based on the quality status of aquifers and standard limits to reach a cost-effective WLA policy among PS and NPS. These steps provided a framework for evaluating the economic benefits of implementing the discharge permit market in the study area. Consequently, in the third step, the highest beneficial trading market, and WLA strategy were finally selected.
Study area
The methodology is testified on the aquifer of Varamin plain, located in south eastern of Tehran province, Iran between the latitude of 35.39°N and 35.07°N and longitude of 51.26°E and 51.55°E. The surface area of this plain is approximately 957 km2.
In the recent decade, the droughts gradually decreased groundwater recharge in the study area, while its extraction was increasing. Since this water resource was rapidly depleting, its status has changed as “prohibited” by the government for additional water extraction. As a consequence, its quality management has become more critical (Noghreyan et al., 2022).
Varamin plain supplies the agricultural and drinking water of two counties, Pakdasht and Varamin, and simultaneously receives the raw sewage of their urban residential areas in association with the drainage of farmlands. The farming area covers about 86% of the overall plain. Approximately 70% of farmlands in this area are covered with wheat and barley crops, 15% with corn, and the remaining 15% includes vegetables. In this area, the secondary treated wastewater of Eastern Tehran is also available for artificial aquifer recharge (Ministry of power, 2013; Nouri et al., 2020). The main point sources in this area are the discharges of Pakdasht, Varamin, and Eastern Tehran, while the major non-point emission source is the leaching of farmlands. Table 1 outlines the general specifications of these point sources. Here, the blueprint of WW3 is the approximate location for discharging this treated effluent to the plain.
Table 1
Point emission sources with their blue prints and specifications
WWTP | Blueprint | Population covered (Capita) | Average flow (m3/s) |
WW1 | 35°24’0” N, 51°53’0” E | 284000 | 0.52 |
WW2 | 35°28’0” N, 51°48’0” E | 351000 | 0.65 |
WW3 | 35°31’0” N, 51°37’0” E | 1350000 | 2.5 |
Accordingly, this study searches for a cost-effective and environmentally sound WLA that responds to the questions: 1) the construction and operation of which wastewater treatment plant (WWTP), Varamin (WW1) or Pakdasht (WW2), should be prioritized with respect to groundwater quality variations and total costs, 2) Whether treated wastewater reuse of Tehran WWTP (WW3) is necessary for the groundwater rehabilitation or not.
The study area with its geological structure is illustrated in Fig. 2. Here, the approximate locations of point sources, groundwater level and flow directions, with the location of piezometer wells are marked. Since the area is rather large, it is divided into six zones based on land-uses and available observation wells with water quality data. This zoning provides more elaborate framework to show the sensitivity of groundwater to WLA scenarios in short term (3 years) and long term (10 years). In the study area, eight observation wells with available temporal water quality data were highlighted to monitor the quality and groundwater level of the plains during the study period. In regions with two observation wells, the results were based on their average data.
Simulation
Varamin aquifer was simulated by GMS software (Version 10.1) in which the groundwater level and quality were simulated by MODFLOW and MT3DMs modules, respectively. Here, the study area was defined and gridded by 250000m2 squares. The locations of observation wells, their temporal variations on groundwater level, and the aquifer recharge were initially determined based on the available data. In addition, water quality data were assigned to the nodes including the concentrations of pollutants in groundwater and the discharges of emission sources (Ministry of Energy, 2013). Groundwater quality modelling was carried out in three time steps. Calibration was based on available data from 2009–2011 in which T0 represents 2011 when the model is calibrated. T3 and T10 show the predicted groundwater quality in each WLA scenario for a 3-year and 10-year period after the simulation period, respectively. Table 2 summarizes the calibration status of groundwater level in 42 piezometers regarding R2 and Nash-Sutcliffe (NSE) indices based on observed data.
Table 2
R-square and NSE values in simulated piezometers at T0
Piezometer No. | R2 | NSE | Piezometer No. | R2 | NSE | Piezometer No. | R2 | NSE |
2 | 0.88 | 0.91 | 34 | 0.81 | 0.98 | 64 | 0.77 | 0.42 |
6 | 0.81 | 0.85 | 37 | 0.78 | 0.87 | 67 | 0.74 | 0.24 |
8 | 0.86 | 0.97 | 39 | 0.84 | 0.86 | 68 | 0.79 | 0.98 |
12 | 0.69 | 0.48 | 41 | 0.89 | 0.96 | 72 | 0.81 | 0.49 |
17 | 0.82 | 0.95 | 43 | 0.9 | 0.91 | 73 | 0.86 | 0.85 |
18 | 0.78 | 0.86 | 44 | 0.88 | 0.85 | 76 | 0.84 | 0.94 |
20 | 0.77 | 0.79 | 48 | 0.77 | 0.98 | 77 | 0.82 | 0.98 |
21 | 0.63 | 0.57 | 50 | 0.79 | 0.66 | 79 | 0.85 | 0.84 |
25 | 0.67 | 0.58 | 52 | 0.78 | 0.76 | 85 | 0.72 | 0.7 |
26 | 0.88 | 0.95 | 54 | 0.83 | 0.9 | 87 | 0.75 | 0.73 |
28 | 0.88 | 0.96 | 55 | 0.75 | 0.53 | 91 | 0.69 | 0.37 |
29 | 0.77 | 0.61 | 56 | 0.85 | 0.81 | 97 | 0.84 | 0.83 |
31 | 0.84 | 0.92 | 57 | 0.82 | 0.24 | 100 | 0.83 | 0.91 |
32 | 0.83 | 0.63 | 58 | 0.81 | 0.98 | 103 | 0.9 | 0.76 |
WLA scenario
13 WLA scenarios were designed based on the calibrated model in which their impacts on the quality of aquifer were examined and compared. These scenarios can be classified in two main WLA scenarios as shown Table 3. The first class includes WLA policies that emphasize point emission sources constructing and operating municipal WWTPs. Their differences are upon their locations and nitrogen (N) removal efficiencies (S1.1-S1.9). The second class recommends controlling agricultural discharges (NPS) by changing the main crop patterns in the study area (S2.1-S2.3). In this study, S0 represents the basic status of the simulated aquifer in the study area without projecting any WLA strategy.
Table 3
The definition of different WLA scenarios in this study
Targeted emission sources | Scenario | Explanation |
- | S0 | Basic scenario |
Point sources | S1.1 | WW2 with 25% N removal + aquifer recharge (A2) |
S1.2 | WW2 with 50% N removal + aquifer recharge (A2) |
S1.3 | WW1 with 25% N removal + aquifer recharge (A4) |
S1.4 | WW1 with 50% N removal + aquifer recharge (A4) |
S1.5 | S1.1 + S1.3 |
S1.6 | S1.2 + S1.4 |
S1.7 | WW3 with 50% N removal + aquifer recharge (A1) |
S1.8 | S1.1 + S1.3 + S1.7 |
S1.9 | S1.2 + S1.4 + S1.7 |
Non-point sources | S2.1 | Wheat and Barley plants: 15% Corn plant: 70% Vegetable plant: 15% |
S2.2 | Wheat and Barley plants: 15% Corn plant: 15% Vegetable plant: 70% |
S2.3 | Wheat and Barley plants: 50% Corn plant: 25% Vegetable plant: 25% |
In scenarios S1.1-S1.9, the abated N was allocated to the target residential areas as reduced NO3 concentration (%). This approach is due to the fact that Pakdasht and Varamin cities do not have wastewater collection and treatment systems and domestic wastewater is traditionally disposed underground. N abatement in S1.1-S1.9 means that NO3 discharge and infiltration to groundwater would be reduced from the residential areas by constructing and operating WWTPs with different N removal efficiencies. The impacts of these scenarios were then analyzed and compared in the observation wells of both the targeted zone and the study area.
For NPS in S2.1-S2.3, the direct water footprint (WF) of each crop was referred in addition to their average yield and cultivation area. Therefore, the impacts of changing crop patterns on the study area in form of water extraction and pollution reduction were estimated based on the average WFs of targeted zones in comparison with S0. Here, crop yields were estimated according to the recent works of literature as 0.52 kg/m2 for wheat and barley, 3.23 kg/m2 for corn, and 3.89 kg/m2 for vegetables (Vogel et al., 2019; Yousefi, 2018). Accordingly, grey WF would be reduced 30%, 23%, and 27% by applying S2.1, S2.2 and S2.3, respectively.
It should be noted that in the study area, chloride is mainly influenced by quantitative variation of groundwater due to annual recharges. It is assumed that wastewater treatment or changing crop patterns can barely reduce chloride concentration in domestic and agricultural discharges. In a nutshell, chloride is a pollutant that is mainly affected by water volume rather than treatment systems.
WLA scenarios were inputted to the calibrated model separately, and the outputs from GMS software were obtained for NO3 and chloride concentrations in eight observation wells. The impact of each scenario with pollutants concentrations is estimated according to the average concentration of all observation wells for each time span. Afterward, WLA scenarios were qualitatively ordered according to their impacts on groundwater quality. Here, the most environmentally friendly WLA scenario is a scenario with the lowest pollution concentration in the whole plain on average with respect to both NO3 and chloride concentrations in T10. Yet, a sustainable WLA scenario requires more analysis with economic perspective.
Economic analysis
For economic analysis of WLA scenarios with the evaluation of discharge permit markets and their benefits, it is necessary to calculate related total costs (TC). For PS, TC is attributed to both the construction and operation of WWTPs during their lifetime. This can be defined per required biochemical oxidation demand (BOD) and N removal efficiencies in WWTPs regarding WLA scenarios. Eq. 1 has been recently developed with this purpose (Jamshidi & Niksokhan, 2016) and used in this study similarly.
Where C W is the annual capital and operation costs (M$/yr) in which M notes as a million in units, Q is the annually average wastewater inflow (m3/s), and T is the annual capital and operating cost of WWTPs per unit volume (M$/m3) which is calculated by Eq. 2.
$$T={T}_{BOD}+{T}_{NO3}$$
2
Here, T BOD and T NO3 are the costs of reducing BOD and NO3 pollutants, respectively (M$/m3). These values depend on the required efficiency as calculated by Equations 3 and 4, respectively.
$${T}_{BOD}=1.7{X}^{2}+0.9X+0.11$$
3
X denotes the abatement of BOD concentration in the treatment plant ranges between 0 and 1. In this study, BOD concentration reduction for all treatment plants is considered as 0.9 which means that WWTPs should at least remove 90% of BOD concentration of wastewater in any cases and scenarios. However, it is noteworthy that BOD is only used for cost evaluations and is not included in environmental and water quality assessment in simulation and WLA.
$${T}_{NO3}=-2.8{Z}^{3}+4.1{Z}^{2}-0.3Z$$
4
Where Z represents NO3 removal efficiency of WWTPs and it ranges between 0 and 1. For example, in S1.1 with 25% N removal, Z equals 0.25.
In addition to economic evaluation of different WLA scenarios, this study emphasizes on calculating the economic benefits of possible discharge permit markets in the study area. For this purpose, the average marginal costs (MC) of PS should be calculated for determining the costs of permits and penalties. This was calculated by Eq. 5. It is noteworthy that for PS, there is no need to subtract the total cost of PS-based WLA scenarios with S0 because the whole abatements in S1.1-S1.9 in these cases are additional to the basic scenario of S0.
$${MC}_{W}=\frac{{C}_{W} }{R}$$
5
Where MC W is the average marginal cost of each PS-based WLA scenario (M$.L/yr.mg), R is the reduction of pollutants concentration for each WLA scenario with S0 (mg/L). C W is defined earlier.
For NPS, cost analysis was carried out differently. The crop MC was determined by the amount of production of each crop (Eq. 6), in association with all costs (Cf) and benefits (Bn) of producing each product (Equations 7 and 8). Based on the calculated benefits, the net benefit of each crop was calculated per m2.
P is the annual farm productions (kg/yr), A is the cultivated area (m2), and Y is the average crop yield in the study area (kg/m2yr).
$${C}_{f}=P\times {C}_{a}+{C}_{t}$$
7
Where C f is the annual farming costs (M$/yr) and C a is the average cost of cultivating a crop unit (M$/kg). C a is the estimated costs according to annual payments required for cultivation including labour, seeds purchase, the costs of fertilizer, pesticide, installation and irrigation, agricultural machineries, and other services. In addition, C t shows the transaction costs of switching crop pattern in each NPS-based WLA scenario. For example, changing from wheat to corn or vegetables in a farmland requires costs related to training workers and farmers, replacing equipments, preparing lands, education and official documents (M$/yr).
$${B}_{n}={S}_{P}\times P-{C}_{f}$$
8
B n is the annual net benefit (M$/yr), S P is the product’s local selling price (M$/kg), and P and C f were defined earlier. This equation calculates the net benefit for agricultural scenarios depending on the type of crops and land use. On the contrary to MC W , the MC of farming WLAs (MC L ) should contain the subtraction with S0. Otherwise, there would be duplications in calculating costs for lands without changing crop patterns. In other words, the reduction in overall benefits of farmers, or their total economic losses (C L ) was calculated and equalized the total costs of these diffuse pollution sources. It should also be considered that in NPS-WLA scenarios (S2.1-S2.3), WWTPs are not required to be equipped with tertiary treatment units for nitrogen removal. Therefore, their related costs are limited to the BOD abatement of domestic wastewater (TNO3 = 0). It means that WWTPs should at least remove BOD to the required level regardless of PS or NPS WLA scenarios. These are shown in Equations 9 and 10.
$${C}_{L}={(B}_{i}-{B}_{n})+{T}_{BOD}\times Q$$
9
$${MC}_{L}=\frac{{C}_{L} }{R}$$
10
Where, MC L is the marginal cost of NPS-based WLA scenarios (M$.L/yr.mg), B i is the net benefit of S0 (M$/yr) without implementing groundwater quality conservation programs and environmental punishments, C L is the annual economic loss of farmers for WLA scenarios without considering punishments (M$/yr), T BOD , Q, R (mg/L) and B n (M$/yr) were defined earlier.
WQT assessment
Prior to the benefit assessment of WQT strategies, the possible seller and buyer of the tradable discharge permits should be determined according to their MCs which can be calculated by MCW or MCL. Nevertheless, in order to induce some economic incentives for polluters for attending a water quality conservation program, it is required to define a price for monetary punishments. Therefore, according to the estimated MCs of discharge permits, the penalty price was calculated by Eq. 11.
$$F=2\times {MC}_{A}\times E$$
11
Where F is the total cost of environmental fines (M$/yr) and MC A is the average MC of all scenarios (M$.L/mg). E is the violations (mg/L) remains in WLA scenario which is the negative difference of simulated and predicted pollutants from the allowable groundwater quality limit (AQL). S0's total cost would equal the calculated penalties, multiplied by the amount of violations.
This punishment would be allocated per violations to polluters in scenarios which cannot satisfy the threshold of AQL. An environmental violation equals the difference between the average water quality of observation wells of each scenario with AQL concentrations of NO3 and Cl−. Scenarios with concentrations below AQL were highlighted as “Negative violation” (N), whereas others were classified as “Positive violation” (V). It should be noted that AQL was estimated based on simulation tool according to water quality remediation potential of aquifer, and practical reductions of PS and NPS. Accordingly, the benefits of WQT are the economic savings (%) of scenarios determined as TC in comparison with the total costs of S0's scenario (Eq. 12).
$$TC={C}_{W} +{C}_{L}+F$$
12
Here, TC is the aggregate of C W and C L (M$/yr), in addition to probable penalties for scenarios with violations (V). However, F equals 0 in market-based WLA scenarios as they should be selected from strategies without violations (NV).
In WQT, since the directive of market-based environmental protection is founded on economic incentives, both discharge permit sellers and buyers should find some economic motivation for this policy. Therefore, this study also evaluates the total costs of sellers and buyers separately.
Here, TCs is the total costs of sellers which is relatively compensated with money gained out of selling permits (M$/yr) while K is the total outcome of sold permits (M$/yr). Obviously, buyers are those stakeholders who do not pay for WWTPs or changing crop patterns but only for buying permits from sellers as Eq. 14.
Where MC denotes the marginal costs of sellers and equals MC W or MC A depending on WLA scenario (M$.L/mg). D represents the differences of estimated groundwater quality with AQL (mg/L).