Modeling the effects of constructed wetland on nonpoint source pollution control and reservoir water quality improvement

Modeling the effects of constructed wetland on nonpoint source pollution control and reservoir water quality improvement

Journal of Environmental Sciences 2010, 22(6) 834–839 Modeling the effects of constructed wetland on nonpoint source pollution control and reservoir w...

738KB Sizes 2 Downloads 54 Views

Journal of Environmental Sciences 2010, 22(6) 834–839

Modeling the effects of constructed wetland on nonpoint source pollution control and reservoir water quality improvement Jonghwa Ham1 , Chun G. Yoon2,∗, Hyung-Joong Kim1 , Hyung-Chul Kim2 1. Rural Research Institute of KRC, 391 Haean-ro, Sangnok-gu, Ansan-si, Gyeonggi-do 426-908, Korea. E-mail: [email protected] 2. Environmental Science Department, Konkuk University, 1 Hwayang-dong, Gwangjin-gu, Seoul 143-701, Korea Received 07 October 2009; revised 27 December 2009; accepted 02 January 2010

Abstract This article describes the integrated modeling approach for planning the size and the operation of constructed wetlands for maximizing retention of nonpoint source pollutant loads and reservoir water-quality improvement at a catchment scale. The experimental field-scale wetland systems (four sets, 0.88 ha each) have been in operation since 2002, where water depth was maintained at 30–50 cm and hydraulic loading rate was at 6.3–18.8 cm/day. The wetland system was found to be adequate for treating polluted stream water with stable removal efficiency even during the winter. The integrated modeling system (modified-BASINS) was applied to the Seokmoon estuarine reservoir watershed and calibrated with monitoring data from constructed wetland, stream, and reservoir. The calibrated integrated modeling system estimated that constructing wetlands on 0.5% (about 114 ha) of the watershed area at the mouth of reservoir could reduce 11.61% and 13.49% of total external nitrogen and phosphorus loads, respectively. It also might improve the nitrogen and phosphorus concentration of the reservoir by 9.69% and 16.48%, respectively. The study suggested that about 0.1%–1.0% of the watershed area should be allocated for constructed wetland to meet specified water-quality standards for the estuarine reservoir at the polder area where land use planning is relatively less complicated. Key words: constructed wetland; estuarine reservoir; integrated modeling system; nonpoint source pollution control; polder area; water quality improvement DOI: 10.1016/S1001-0742(09)60185-6

Introduction Korea is a land-limited country and large-scale polder projects have been practiced to dike, drain, and reclaim coastal tidelands to develop additional agricultural land and water resources during the last decades in Korean peninsula. After closure of the dikes, much pollutant loads from the watershed has been transported by streams and ditches to the estuarine reservoir, and most estuarine reservoirs are suffering from eutrophication in Korea. Lake and reservoir restoration strategies are primarily aimed at reduction of the external nutrient loads (Cooke et al., 1993). Wetland restoration and construction are considered effective measures to combat the eutrophication of aquatic ecosystems and reduce nutrient loads to lakes and reservoirs. Quantifying nutrient retention in natural and constructed wetlands is important in the effort to enhance waterquality (Lund et al., 2000; Spieles and Mitsch, 2000). Tools to simulate nutrient removal from a wetland on a watershed scale would benefit the process of planning the size and the allocation of wetlands for maximum nutrient removal. To evaluate the effects of a constructed wetland * Corresponding author. E-mail: [email protected]

in controlling nonpoint source (NPS) pollutant loads, a watershed loading model, receiving water-quality model, and wetland model could be employed together. Integrated modeling systems, which can be used to link these models for a comprehensive assessment of a watershed, provide the user with a fully integrated data, analysis, and modeling framework. They are capable of linking models to geographic information system (GIS) databases, and are built from modules that allow the user the flexibility to choose a specialized analysis (He et al., 2001). This article describes the integrated modeling approach for planning the size and allocation of wetlands for maximizing retention of NPS pollutant loads and reservoir water-quality improvement at a catchment scale.

1 Materials and methods 1.1 Study area and experimental wetland systems The Seokmoon watershed is located in Choongnam Province on the west coast of the Korean peninsula (Fig. 1). The watershed area of 22,700 ha is composed of 14% urban, 40% agricultural, and 33% forested lands. The 1600 ha polder area includes an 870 ha estuarine reservoir with average depth of 2.2 m. During 2000–2004, the mean

No. 6

Modeling the effects of constructed wetland on nonpoint source pollution control and reservoir water quality improvement

835

Fig. 1 Seokmoon watershed with monitoring points and experimental wetland systems.

annual precipitation and temperature averaged 1412 mm and 11.8°C, respectively. Four sets of 0.88 ha experimental wetland systems were constructed at the mouth of the Seokmoon estuarine reservoir and have been in operation since 2002 to evaluate their efficiency in controlling NPS pollutant loads from the watershed. Wetland types can be grouped into surface flow (SF) and subsurface flow (SSF) wetlands. In this study, surface flow type wetland was used. Water from the Dangjin Stream flowing into the Seokmoon estuarine reservoir was pumped into the wetlands. Water depth of the wetland was maintained at 30–50 cm, hydraulic loading rate was about 6.3–18.8 cm/day, and retention time was managed at about 3–5 days, during the study period. Emergent plants are allowed to grow in the wetland. After three years, vegetation cover exceeded 90% in all wetlands, and dominant species was common reed (Phragmites australis) and cattails (Typha latifolia). Water samples were taken at inlet and outlet of each wetland cell twice a month during June 2002 to June 2004. The experimental data were divided into two groups: winter (December–February) and growing season (March– November) to analyze seasonal performance. ANOVA and t-tests were used to examine differences between data groups, and statistical analyses were performed using SPSS for windows version 10.0. 1.2 Wetland model and integrated modeling system The NPS-WET, a dynamic compartmental wetland simulation model, was developed to simulate nutrient dynamics in constructed wetlands, to quantify nutrient removal capacity under various conditions, and to examine whether it would be feasible to operate wetlands for optimum nutrient removal (Ham, 2005). The model consists

of nine linked submodels representing the hydrologic, vegetative, periphyton, sediment, phosphorus, microbe, carbon, oxygen, and nitrogen cycles of a wetland (Fig. 2). The model employs a system approach, and various interactions can be simulated among nutrient cycles in a wetland system. It accounts for carbon and nitrogen interactions, as well as for the effect of DO levels upon microbial growth. It also directly links microbial growth and death to the consumption and transformations of nutrients in the wetland system. The NPS-WET model has three layers, a water column layer, an active sediment layer, and a deep sediment layer. The active sediment layer is defined as the top layer with a certain constant depth. This model assumes that sediment and water interaction (sedimentation and leaching) happens only in the top layer. The deep sediment layer compartment acts as a “dead end” in the model. It has no effect on the rest of the model as it was assumed that there was no decomposition in the deep sediments. This model assumes that macrophytes take up nutrients only from deep sediment. It simulates the wetland as a continuously stirred tank reactor, because we consider square-shaped and rectangular-shaped wetland with a low length to width ratio (Persson et al., 1999). The integrated modeling system involves simulating watershed loading, reservoir water-quality, and wetland performance to effectively evaluate the effect of the constructed wetlands on the reservoir water-quality improvement at a watershed scale. HSPF (Hydrological Simulation Program – FORTRAN) is a well-known comprehensive watershed model (Bicknell et al., 2001), and WASP (Water quality Analysis Simulation Program) is widely used to simulate reservoir water-quality (Ambrose et al., 1993). BASINS is a multipurpose environmental

836

Jonghwa Ham et al.

Fig. 2

Vol. 22

Wetland model (NPS-WET) structure and wetland description.

analysis system for watershed and water-quality studies (EPA, 2001). In this study, BASINS was modified to incorporate WASP and NPS-WET in addition to the builtin HSPF. Modules to estimate point source pollutant loads and to link among the add-ons were also added.

2 Results and discussion

being largely a physical (sedimentation) and chemical (adsorption) process, is less likely sensitive to temperature, but may be influenced by the oxygen availability, because the sometimes large role may be played by redox sensitive adsorption to ferrous/ferric oxides (Wittgren and Mæhlum, 1997). The wetland performance demonstrated the effectiveness of nutrient retention and was satisfactory for controlling NPS pollutant loads from the watershed.

2.1 Wetland performance

2.2 Calibration of the integrated modeling system

Average effluent total nitrogen (TN) concentration in the growing season (March to November) and winter (December to February) was 1.5 and 3.7 mg/L, respectively (Table 1). There was nearly no difference in the amount of TN removed between winter (about 1.51 kg/(ha·day)) and the growing season (about 1.46 kg/(ha·day)), whereas the average removal rate in winter (about 30%) was lower than that in the growing season (about 50%). The processes of ammonification, nitrification, and denitrification have all been shown to be temperature dependent in treatment wetlands; therefore, rates of TN reduction will also be temperature dependent (Werker et al., 2002). The removal rate of TN during winter was about 32%, which is in the range of other studies of about 40% at air temperatures around zero (Gumbricht, 1992). Average influent and effluent total phosphorus (TP) concentration was 0.30 and 0.14 mg/L, respectively, showing about 50% removal rate. There was no difference in the effluent and removal rate of TP between the growing season and winter (p > 0.05, n = 172). Phosphorus removal,

Before modeling the effects of constructed wetland on NPS pollution control and reservoir water-quality improvement, integrated modeling system, which incorporated HSPF, NPS-WET, and WASP was calibrated with monitoring data in 2003. To evaluate the wetland model performance, the simulated daily effluent of the wetland model was compared with year-round monitoring data from cell 2 for water depth, and TN and TP concentrations. Figure 3a demonstrates a good agreement between them, and NPS-WET was found to be appropriate to simulate the behaviour of pollutants in the wetland system. When water inflow stopped, the decreasing slope of the simulated water depth was very close to that of the observed data. This implies that the simulation of infiltration and evapotranspiration in the wetland system closely represents the field conditions. Several TP concentrations were underpredicted during the summer monsoon period, which might be attributed to insufficient influent phosphorus data during heavy rainfall for wet-days simulation.

Table 1

Seasonal comparison of concentrations of total nitrogen (TN) and total phosphorous (TP) and removal efficiencies of the wetlands Constituent

Growing season

Winter season

p-Valuea

TN

Influent (mg/L) Effluent (mg/L) Removal (%)

3.3 ± 1.48 1.5 ± 0.96 51.5 ± 26.4

5.5 ± 0.76 3.7 ± 0.92 31.7 ± 13.64

0.000b 0.000b 0.000b

TP

Influent (mg/L) Effluent (mg/L) Removal (%)

0.30 ± 0.149 0.14 ± 0.092 50.6 ± 29.58

0.31 ± 0.089 0.14 ± 0.080 53.0 ± 29.26

0.595 0.998 0.628

Associated probability of t-test; b significantly different at p = 0.05. Data are expressed as mean ± SD, n = 172.

a

No. 6

Fig. 3

Modeling the effects of constructed wetland on nonpoint source pollution control and reservoir water quality improvement

837

Calibration of NPS-WET, HSPF, and WASP in the modified-BASINS. (a) NPS-WET calibration; (b) HSPF calibration; (c) WASP calibration. Table 2 Nutrient retention by a wetland, and its effect on reservoir water-quality Nutrient retention

Loads (ton/yr)

Reduction (%)

Water-quality improvement

Conc. (mg/L)

Improved (%)

0.1% of watershed (24 ha) 0.5% of watershed (114 ha)

12.40 64.39

2.24 11.61

Without wetland 0.1% of watershed (59.31 ha) 0.5% of watershed (114 ha)

3.01 2.94 2.72

– 2.53 9.69

0.1% of watershed (24 ha) 0.5% of watershed (114 ha)

1.41 6.89

2.77 13.49

Without wetland 0.1% of watershed (59.31 ha) 0.5% of watershed (114 ha)

0.175 0.166 0.146

– 4.86 16.48

TN

TP

Figure 3b shows a time series plot of daily simulated and observed data at monitoring point WP3 located at the mouth of estuarine reservoir (Fig. 1). The TN and TP concentrations increased significantly during winter, perhaps due to low flow, while point source pollutant loads were fairly constant throughout the year. The model output simulated the watershed pollutant loads reasonably well, and the seasonal variation was well demonstrated. Figure 3c compares simulated and observed data in the reservoir. The simulated data represents weight-averaged concentration considering water volume and concentration of the reservoir segment, and the observed data are mean concentration with standard deviation derived from monitoring points WP4, WP5, and WP6 (Fig. 1). The model output adequately simulated the observed data, and demonstrated the seasonal variation of the main waterquality parameters. 2.3 Effects of constructed wetland on reservoir water quality with different wetland size The calibrated integrated modeling system was used to estimate the retention capacity of NPS pollutant loads in the wetland system and the effect of nutrient retention on

reservoir water-quality for the condition of 2003. Conversion of a 114 ha paddy rice field to a constructed wetland for NPS pollution control and ecological benefits has been suggested. If it were converted to a wetland system and part of the stream water purified by passage through it, the wetland system would reduce the nutrient loads to Seokmoon reservoir by 64.39 tons TN/yr and 6.89 tons TP/yr at 10 cm/day hydraulic loading rate (Table 2). These amounts account for about 11.61% of TN and 13.49% of TP annual loads from the watershed. The expected waterquality improvement at the Seokmoon reservoir would be 9.69% for nitrogen and 16.48% for phosphorus (Table 2). The simulation results for a wetland system with 0.1% of the watershed area are also summarized in Table 2. The ratio of wetland area to watershed area is often used to explain retention performance (Uusi-K¨ampp¨a et al., 2000; Carleton et al., 2001). Mitsch and Gosselink (2000) suggested that at least 1% of a catchment should be considered for wetlands to trap nutrients. Mitsch and Wang (2000) reported that for low-phosphorus-yield watersheds (< 0.1 g P/(m2 ·yr)), the restoration of 1% of the wetlands could remove 20%–35% of total watershed phosphorus loads. Arheimer and Wittgren (2002) modeled nitrogen

838

Jonghwa Ham et al.

Vol. 22

Fig. 4 Comparison of effluent concentration, removal loading rate, and reservoir water quality improvement rate by 114 ha constructed wetland with hydraulic loading rate. (a) wetland effluent concentration; (b) removal loading rate by wetland; (c) reservoir water quality improvement rate by wetland.

removal in potential wetlands at the catchment scale and showed that the conversion of 0.4% of the catchment to wetlands would reduce annual riverine nitrogen transport by 6%. Ratio of wetland area to watershed area can be increased to meet intended water-quality at model simulation; however, high ratio of wetland area to watershed area is not feasible at many sites. This study suggests that a range of 0.1%–1.0% of the watershed should be allocated to wetlands to protect reservoir water-quality at polder area where land use planning is less complicated. And the combined application of wetland construction and other BMPs is more effective in controlling NPS pollutant loads. 2.4 Determining optimal range of hydraulic loading rate The hydraulic loading rate is often the significant design and operation variable in influencing the performance of a treatment wetland. Hydraulic loading rate affects constituent loading, hydraulic residence time, water velocity, pollutant removal rate, and pollutant removal loading (Kadlec and Knight, 1996). The calibrated integrated modeling system was used to determine the optimal range of the hydraulic loading rate. The relationship between the hydraulic loading rate and effluent concentrations for TN and TP are presented in Fig. 4a. The graph shows a generally increasing effluent concentration with greater hydraulic loading rates. A hydraulic loading rate of between 1.5 and 15 cm/day seems to show a steep increasing of effluent concentration with the increasing hydraulic loading rate. If the objective of a constructed wetland is low effluent concentration, the hydraulic loading rate should be decreased. At over 15 cm/day, although the hydraulic loading rate is decreased, the water quality of wetland effluent can hardly be improved. The relationship between the hydraulic loading rate and removal loading rates for TN and TP are shown in Fig. 4b. In Fig. 4b, a range of nitrogen and phosphorus removal loading rate is 150–750 kg N/(ha·yr) and 5–90 kg P/(ha·day); this result is in agreement with Mitsch and

Gosselink (2000). Mitsch and Gosselink (2000) reviewed several constructed wetlands for freshwater marshes receiving nonpoint source pollution and found a range of nitrogen retention from 30 to 690 kg N/(ha·yr) and phosphorus retention from 4 to 56 kg P/(ha·yr). Over the range of hydraulic loading rates studied here (1.5–30 cm/day) TN and TP exhibit a power relationship, with reservoir water quality improvement rates increasing as the hydraulic loading rate increases (R2 = 0.944 and 0.942). There is only small difference in reservoir water quality improvement rate between 15 cm/day and over 15 cm/day hydraulic loading rate (Fig. 4c). In this study, 5–15 cm/day of hydraulic loading rate was recommended to treat NPS pollutants. Most of the loading of phosphorus to a reservoir in rural areas becomes adsorbed to sediments, so high phosphorus loading has been resulted during storm events. In agricultural watersheds, high pulses of nitrate nitrogen occur during the first storms after fertilizer application to farm fields. A good wetland design should both take advantage of these pulses for system replenishment and provide for excess wet weather storage if nutrient retention is a primary objective (Mitsch and Gosselink, 2000). Constructed wetlands should receive as much storm water as possible, although effluent concentration is increased to maximize nutrient reduction. In this study, we considered two different inflow scenarios. In constant inflow scenario, inflow was maintained constant (10 cm/day) throughout the year. In special operation plan scenario, inflow was maintained constant (10 cm/day) during dry season, and three times as much storm water flow (30 cm/day) was applied during storm events. TN and TP retention loads of 114 ha constructed wetland was increased 6.7% and 13.4%, respectively, and TN and TP concentration of reservoir was decreased 0.07% and 0.68%, respectively, by special operation plan (Table 3). Special operation plan was more effective on TP retention than TN, because adsorbed TP to sediment during storm events can be easily settle down in the wetland.

No. 6

Modeling the effects of constructed wetland on nonpoint source pollution control and reservoir water quality improvement Table 3

839

Nutrient retention and reservoir water-quality by 114 ha constructed wetland with special action plan

Nutrient retention

Loads (ton/yr)

Reduction (%)

Reservoir water-quality Conc. (mg/L)

Loads (ton/yr)

Difference (%)

TN

Constant inflow Special operation plan

64.39 68.70

6.7

Constant inflow Special operation plan

2.720 2.718

–0.07

TP

Constant inflow Special operation plan

6.89 7.81

13.4

Constant inflow Special operation plan

0.146 0.145

–0.68

3 Conclusions The wetland system was found to be an adequate alternative for treating a polluted stream water with stable removal efficiency and is recommended as a practical NPS pollution control measure. The modified-BASINS demonstrated adequate agreement between model output and observed data from incoming stream, wetland, and reservoir, and it was a practical and convenient tool for watershed and reservoir water-quality management. It was estimated that constructed wetland at the mouth of the reservoir on 0.5% (about 114 ha) of the Seokmoon reservoir watershed area would reduce more than 11.61% and 13.49% of external nitrogen and phosphorus loads, respectively. The resulting reservoir water-quality improvement in TN and TP is estimated at 9.69% and 16.48%, respectively. It is suggested that a range of 5– 15 cm/day of a hydraulic loading rate is adequate to provide nutrient retention and water quality values for the landscape. Overall, the modified-BASINS was found to be an appropriate tool for planning the size and operation of wetlands for maximizing retention of NPS pollutant loads and reservoir water-quality improvement at a catchment scale. This study suggests that 0.1%–1% of the watershed should be allocated to wetlands at polder projects to provide reservoir water-quality improvement and ancillary benefits such as wildlife enhancement and flood mitigation.

References Ambrose R B, Wool T A, Martin J L, 1993. The Water Quality Analysis Simulation Program, WASP5 version 5.10. Part A: Model documentation. U.S. Environmental Protection Agency, Office of Research and Development, Environmental Research Laboratory, Athens, GA, USA. Arheimer B, Wittgren H B, 2002. Modeling nitrogen removal in potential wetlands at the catchment scale. Ecological Engineering, 19: 63–80. Bicknell B R, Imhoff J C, Kittle J L, Donigian A S, Johanson R C, 2001. Hydrological Simulation Program-FORTRAN (HSPF): User’s manual for release 12. U.S. Environmental Protection Agency, Environmental Research Laboratory, Athens, GA, USA.

Carleton J N, Grizzard T J, Godrej A N, Post H E, 2001. Factors affecting the performance of stormwater treatment wetlands. Water Research, 35: 1552–1562. Cooke G D, Welch E B, Peterson S A, Newroth P R, 1993. Restoration and Management of Lakes and Reservoirs (2nd ed.). CRC Press, New York. EPA, 2001. Better Assessment Science Integrating Point and Nonpoint Sources (BASINS) Version 3.0 User’s Manual. EPA 823-8-01-001, Office of Water, Washington DC, USA. Gumbricht T, 1992. Tertiary wastewater treatment using the root zone method in temperate climates. Ecological Engineering, 1: 199–212. Ham J H, 2005. Nonpoint Source Pollution Control Using Constructed Wetlands and Integrated Watershed Modeling. Ph.D Thesis. Konkuk University. Seoul, Korea. He C, Shi C, Yang C, Agosti B P, 2001. A windows-based GIS-AGNPS interface. Journal of the American Water Resources Association, 37: 395–406. Kadlec R H, Knight R L, 1996. Treatment Wetlands. CRC Press, FL, USA. Lund L J, Horne A J, Williams A E, 2000. Estimating denitrification in a large constructed wetland using stable nitrogen isotope ratios. Ecological Engineering, 14: 67–76. Mitsch W J, Gosselink J G, 2000. Wetlands. John Wiley & Sons, New York, USA. Mitsch W J, Wang N, 2000. Large-scale coastal wetland restoration on the Laurentian Great Lakes: determining the potential for water quality improvement. Ecological Engineering, 15: 267–282. Persson J, Somes N L G, Wong T H F, 1999. Hydraulics efficiency of constructed wetlands and ponds. Water Science and Technology, 40(3): 291–300. Spieles D J, Mitsch W J, 2000. The effects of season and hydrologic and chemical loading on nitrate retention in constructed wetlands: a comparison of low and high nutrient riverine systems. Ecological Engineering, 14: 77–92. Uusi-Kmpp J, Braskerud B, Jansson H, Syversen N, Uusitalo R, 2000. Buffer zones and constructed wetlands as filters for agricultural phosphorus. Journal of Environmental Quality, 29: 151–158. Werker A G, Dougherty J M, McHenry J L, Van Loon W A, 2002. Treatment variability for wetland wastewater treatment design in cold climates. Ecological Engineering, 19: 1–11. Wittgren H B, Mæhlum T, 1997. Wastewater treatment wetlands in cold climates. Water Science and Technology, 35(5): 45– 53.