Environ. Sci. Technol. 2005, 39, 4621-4627
Recovery and Fractionation of Phosphorus Retained by Lightweight Expanded Shale and Masonry Sand Used as Media in Subsurface Flow Treatment Wetlands MARGARET G. FORBES,* KENNETH L. DICKSON, FARIDA SALEH, AND WILLIAM T. WALLER Institute of Applied Science, University of North Texas, P.O. Box 310559, Denton, Texas 76203-0559 ROBERT D. DOYLE Department of Biology, Baylor University, 1311 South 5th Street, Waco, Texas 76798-0003 PAUL HUDAK Department of Geography, University of North Texas, P.O. Box 305279, Denton, Texas 76203-5279
Most subsurface flow treatment wetlands, also known as reed bed or root zone systems, use sand or gravel substrates to reduce organics, solids, and nutrients in septic tank effluents. Phosphorus (P) retention in these systems is highly variable and few studies have identified the fate of retained P. In this study, two substrates, expanded shale and masonry sand, were used as filter media in five subsurface flow pilot-scale wetlands (2.7 m3). After 1 year of operation, we estimated the annual rate of P sorption by taking the difference between total P (TP) of substrate in the pilot cells and TP of substrate not exposed to wastewater (control). Means and standard deviations of TP retained by expanded shale were 349 ( 171 mg kg-1, respectively. For a substrate depth of 0.9 m, aerial P retention by shale was 201 ( 98.6 g of P m-2 year-1, respectively. Masonry sand retained an insignificant quantity of wastewater P (11.9 ( 21.8 mg kg-1) and on occasion exported P. Substrate samples were also sequentially fractionated into labile P, microbial P, (Fe + Al) P, humic P, (Ca + Mg) P, and residual P. In expanded shale samples, the greatest increase in P was in the relatively permanent form of (Fe + Al) P (108 mg kg-1), followed by labile P (46.7 mg kg-1) and humic P (39.8 mg kg-1). In masonry sand, there was an increase in labile P (9.71 mg kg-1). Results suggest that sand is a poor candidate for long-term P storage, but its efficiency is similar to that reported for many sand, gravel, and rock systems. By contrast, expanded shale and similar products with high hydraulic conductivity and P sorption capacity
* Corresponding author phone: (361)749-7601; fax: (361)749-6786; e-mail:
[email protected]. Present address: University of Texas Marine Science Institute, 750 Channel View Dr., Port Arkansas, TX 78373. 10.1021/es048149o CCC: $30.25 Published on Web 05/11/2005
2005 American Chemical Society
could greatly improve performance of P retention in constructed wetlands.
Introduction An increasing number of permitting authorities in the United States are requiring tertiary phosphorus (P) removal from secondarily treated effluents. Many free surface water wetlands are not effective for long-term P removal due to relatively rapid saturation of P storage compartments, followed by periodic P export (1-5). Exceptions are systems with mineral soils (6) or inputs of dissolved calcium, iron, or aluminum that have the ability to precipitate or sorb P and retain it in long-term storage (7, 8). The primary mechanisms of dissolved P storage in natural wetlands are microbial uptake, plant uptake, incorporation of organic P into soil peat, and soil adsorption (9, 10). The sediment-litter compartment is the major P pool (>95%) in natural wetlands (10) with plant-bound P and microbial P playing a smaller, seasonal role in P cycling (1, 2). Wetlands sediments exposed to high P loading may store P in easily desorbed inorganic or organic forms; in more recalcitrant Ca, Al, and Fe compounds; or in humic and fulvic compounds (11). Phosphorus retention is most often related to soil characteristics such as hydraulic conductivity (12, 13) and Fe, Al, Ca, and Mg content (9, 14, 15). In free surface water wetlands, opportunities for dissolved P to associate with sediments are limited because only a fraction of the overlying water comes into contact with sediment surfaces. Subsurface flow systems promote opportunities for P sorption by forcing wastewater through the filter media. In European countries, subsurface flow “reed bed” systems are widely utilized for wastewater treatment in rural communities (16). Most use gravel or a sand and gravel mix to minimize clogging, and have very low hydraulic loading rates (5-20 cm day-1). Many of these systems report high P retention, and although P removal is often attributed to biological processes (13) and physical settling, the fate of stored P is rarely determined. A variety of substrates such as shale, lightweight aggregates, iron-rich sands, zeolite, and blast furnace slags have been proposed for use as sorbents for phosphorus removal (17-22). Sorption isotherm and column flow-through experiments have been used as screening tools for comparing different filter media and soils (18, 19, 21-24). A few researchers have correlated P removal to Al + Fe or Ca + Mg content of their media (15, 18, 19, 21), while others (20, 23, 25-28), determined chemical forms of P retained by their materials. Sorption isotherms are useful for initial screening of materials, and flow-through column studies have been used to predict long-term P-removal. Pilot cells provide a more realistic estimate of potential P retention by wetlands because they are subject to both biotic and abiotic influences, and they reflect hydraulic inefficiencies such as clogging and short-circuiting. A primary factor that discourages use of treatment wetlands in the United States is the large acreage required to achieve discharge standards (2, 29). The global average of P storage through peat accretion alone is 0.5 g of P m-2 year-1 (1), and at this rate over 500 ha of wetland would be required to reduce 1 million gallons of wastewater per day (1 MGD) from 2.0 to 0.1 mg of P L-1. Storage via particulate settling is approximately an order of magnitude higher (8). If alternative substrates that maximize P storage can be found, VOL. 39, NO. 12, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4621
the land requirements for treatment wetlands may be substantially reduced. Materials with high hydraulic conductivity are preferred because they permit higher loading rates and, therefore, greater economic efficiency (30). The goal of our research was to demonstrate the feasibility of using a subsurface flow wetland to remove dissolved P. We used two locally available filter materials in pilot-scale systems, and at the end of 1 year of operation, recovered the added P to calculate an aerial rate of P retention. The two materials, expanded shale and masonry sand, were used as the substrate because they each showed moderate to high P sorption in previous batch isotherm and flow-through column experiments (5). To evaluate the long-term potential of our pilot cells, we sequentially fractionated substrate samples to determine quantities of labile P, microbially bound P, (Fe + Al)-bound P, humic P, (Ca + Mg)-bound P, and residual P. We also examined the spatial distribution of P fractions within the pilot cells. Finally, we compared P retention in our substrates to published results of similar wetlands. Given the increases in anthropogenic P in the environment and the resulting widespread eutrophication of natural waters, it is important to expand treatment options for removal of P from domestic wastewater as well as stormwaters, agricultural runoff, and industrial discharges.
TABLE 1. Substrate and Pilot Cell Characteristics substrate characteristics
masonry sand
expanded shale
bulk density (kg/m3) porosity effective size (D10) (µm) uniformity coefficient (D60/D10) hydraulic conductivity (m/day) Langmuir phosphorus sorption maxima (mg/kg)
1670 0.304 110 1.4 17.3 58.8
728 0.594 720 2.3 92.2 971
pilot cell characteristics (per cell)
A, B, and C
X, Y, and Z
volume (excluding pea gravel) (L) surface area (excluding pea gravel) (m2) mass substrate (kg) days operated volume wastewater treated (m3) average hydraulic loading rate (cm/day)
1930 2.16 3220 392 430 46
1930 2.16 1400 416 490 55
Materials and Methods Filter Material. Expanded lightweight shale utilized in this study is manufactured by Texas Industries from shallow shale deposits mined near Streetman, TX. The shale is sorted by size and then fired in a rotary kiln to approximately 1093 °C (2000 °F). The result is a stable, porous material with high hydraulic conductivity, increased surface area, and chemistry similar to the parent material. A partial list of components of expanded shale reported by the manufacturer includes SiO2 62.06%, Al2O3 15.86%, Fe2O3 5.80%, CaO 1.44%, and MgO 1.68%. Note that Si and Al are major components of this material, while Ca and Mg are minor components. Previous experiments indicate that this product has high P sorption capacity without the poor hydraulic conductivity associated with unaltered clays and shales (5). Our masonry sand was washed, fine-grained construction sand commonly used as an ingredient in mortar and plaster and as fill during various construction activities. Pilot Cells. Three replicate pilot cells (A, B, and C) containing masonry sand and three cells (X, Y, and Z) containing expanded shale were constructed for this study. Each cell was planted with 30 plugs of Schoenoplectus acutus (Muhl. Ex Bigelow) A. Love & D. Love. All cells received secondarily treated municipal effluent from the Pecan Creek Reclamation Facility in Denton, TX, from September 2000 through January 2002. Substrate and pilot cell characteristics are described in Table 1 and in Forbes et al. (5). Each cell received wastewater through PVC pipes discharging onto pea gravel, followed by horizontal subsurface flow through substrate (shale or sand) to an outlet pipe surrounded by pea gravel (Figure 1). Water levels were kept just below the surface of the substrate with swivel standpipes. Cell A developed a leak early in the study and was taken off line. Hydraulic Efficiency. Inlet flow rates were varied to determine the effects of flow rate on actual hydraulic retention times (τ). High flow rates were determined by increasing the flow to each cell until ponding occurred and then decreasing it slightly. Low flow rate for sand was 3 L h-1, and the high flow rates were 40 L h-1 for cell B and 93 L h-1 for cell C. Each shale cells was loaded at 10 L h-1 (low flow) and 87 L h-1 (high flow). Rhodamine FWT dye (20 mL) was poured onto the surface of pea gravel located at the inlet zone of each cell and samples were collected from the outlet pipe approximately every 20 min. Dye concentrations were measured with a Turner Designs 10-AU fluorometer and plotted against 4622
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 12, 2005
FIGURE 1. Cross-sectional view of one pilot cell. Cells are approximately 90 cm high, 120 cm wide, and 240 cm long. time with τ taken as the centroid of the exit tracer concentration distribution (31). The product of τ and Q represents the active volume of the unit (Va), which is the volume of pore space occupied by wastewater (31). The potential volume of the unit is nV, where n is the porosity and V is the pilot cell volume (Table 1). The hydraulic efficiency of the unit at a given flow rate is the ratio of active volume to potential volume (Va/nV). Water Quality. Biweekly samples of pilot cell inflow and outflows were analyzed for soluble reactive phosphorus (SRP), pH, temperature, and electrical conductivity. SRP was determined colorometrically by the ascorbic acid method (32). Total Phosphorus and Total Inorganic Phosphorus. After approximately 1 year, sand, shale, and pea gravel samples were collected and chemically extracted for total inorganic P (TIP) and total P (TP). To examine possible differences in spatial distribution of P retained by the pilot cells, each cell was divided into five horizontal zones parallel to the flow paths (Figure 2). Three sand and shale cores were collected within each zone of each cell by driving a 1-in. diameter thin-walled PVC pipe down to the bottom of the substrate. Each core was split into a top (top 30 cm) and bottom (bottom 30 cm) sample. One composite surface sample of pea gravel was taken from both the inlet (zone 1) and the outlet (zone 5). Substrate Analysis: Total Inorganic Phosphorus and Total Phosphorus. All P analyses of substrates were performed on field moist samples. Controls were defined as material delivered during construction of the pilot cells and stored in plastic bags at room temperature. For TIP, 2.5 g aliquots of substrate were placed in 50 mL centrifuge tubes with 20 mL of 1.0 M HCl. Tubes were shaken on a mechanical reciprocating shaker for 3 h, centrifuged at 3500 rpm for 10
FIGURE 2. Schematic of sampling design for phosphorus extraction from pea gravel and substrate. Zones 1 and 5 contain pea gravel; zones 2-4 contain sand or shale. Arrows indicate direction of wastewater flow. min, and filtered through Whatman no. 41 filter paper. TP was determined on duplicate samples that were ashed at 550 °C for 4 h and extracted with hot 6 M HCl. This extract was analyzed for total phosphorus following sulfuric acidpotassium persulfate digestion in an autoclave or on a hot plate (32). All forms of P were determined by the ascorbic acid colorimetric method (32) after the specified treatments. We calculated TP retained by each pilot cell on the basis of the difference between mean TP of samples from the pilot cells and mean TP of control samples. Phosphorus Fractionation. After determining that TP and TIP from replicate cells were not significantly different, fresh substrate samples were again collected from one sand pilot cell (C) and one expanded shale pilot cell (Y) for fractionation analyses. Extractions, centrifugation, and filtration were accomplished as described above and in the sequential extraction scheme depicted in Figure 3. Extraction procedures using alkali and acid reagents were originally proposed by Chang and Jackson (33) and modified by Qualls and Richardson (11), Reddy et al. (34), and Ivanoff et al. (35) for use with organic wetland soils. Total Phosphorus Retained in Pilot Cells. Estimates of total P retained per cell were calculated by taking the difference between mean TP extracted from the substrate (milligrams of P per kilogram of dry weight) and mean TP of the control material and then multiplying the difference by the mass of material in the cell (kilograms). Annual aerial retention rate (grams of P per square meter per year) was determined by dividing the above value by the area containing the material (2.19 m2) and the length of time the systems received wastewater.
Results and Discussion Water Quality. Incoming wastewater P ranged from 0.37 to 4.17 mg L-1, with the lowest values occurring in the rainy season (Figure 4). Soluble reactive P (SRP) concentrations of outflow from shale cells were consistently lower than those from sand cells (Tukeys, R ) 0.05). Mean (( standard deviation) removal efficiency of shale and sand cells was 51% ( 26% and 14% ( 20%, respectively (5). Variability in P removal may be attributed to season, pH, saturation of easily accessible sorption sites, and fluctuations
in influent P concentrations. Both materials retained P during the first 60 days of loading (Figure 4), followed by a period of approximately 120 days where retention decreased (winter). As temperatures increased and the growing season resumed, retention by both materials improved, suggesting P uptake by the plants influenced P removal. We hypothesized that initial P retention by sand was predominantly via nonspecific physical adsorption (ion exchange), where soluble anions such as PO43- sorb to ionically bonded cations on the solid surface (10). This was further supported by periodic release of phosphate in sand cells, particularly following periods where incoming phosphate concentrations were low. This P desorption by our sand was observed in earlier sorption-desorption isotherm and column experiments (5). In shale cells, initial high P retention was likely due to ion exchange. Presumably, as easily accessible surface sorption sites are occupied, P removal in shale would decrease; however, expanded shale contains abundant micropores and sorption to those internal surfaces may continue at a slower rate. Phosphate sorbed on the surfaces via ion exchange is initially reversible. Phosphate may, however, replace an -OH2 or an -OH group in the surface structure of Al or Fe hydrous oxide minerals (ligand exchange), which is less reversible; and over time, the exchanged phosphate ion may further penetrate the mineral surface by forming a stable binuclear bridge (36). Under continuous loading, sorbed phosphate ions may undergo an entire sequence of these reactions, becoming increasingly unavailable (36, 37). Hydraulic Efficiency. Results of tracer dye studies indicated that masonry sand had substantially more dead volume (short-circuiting) than expanded shale. At low flow rates (3 L h-1), both sand cells had a retention time of approximately 20 h, for a hydraulic efficiency of only 8%. With high flow rates, cell B (40 L h-1), had a retention time of 6 h, for a hydraulic efficiency of 30%. In cell C (93 L h-1), the high flow retention time was 4.75 h for a hydraulic efficiency of 55%. When expanded shale had low flow rates of 10 L h-1, cells yielded an observed retention time of approximately 22 h, for a hydraulic efficiency of 22%. Increasing the flow rates to 87 L h-1 yielded a retention time of 10.8 h, increasing the hydraulic efficiency to 89%. As is commonly observed, hydraulic efficiency was improved by increasing flow rates through the materials (31). In masonry sand, however, the poor hydraulic efficiency undoubtedly contributed to the lower P retention rates. Total Inorganic Phosphorus and Total Phosphorus. Mean TIPs of substrate from each pilot cell were grouped as follows: Z ) Y > X > shale control > B ) C ) sand control (Student-Newman-Keuls on log-transformed data, R ) 0.05). TP means were grouped as follows: X ) Y ) Z > shale control > C ) SaC ) B (Student-Newman-Keuls, R ) 0.05). Although water quality monitoring indicated that sand pilot cells retained P, the small amounts of TIP and TP extracted from the sand (Figure 5) led us to conclude that P removal was not attributable to the sand. Rather, plant uptake was the likely pathway of P retention in the sand pilot cells (38). By contrast, TIP increases in shale pilot cells demonstrate that wastewater P associated with inorganic components in the shale. Spatial Differences in P. The rate and amount of sorption is greatly influenced by sorbate concentration. Therefore, a saturation front is likely to occur closest to the inlet area, where P concentrations are highest (1, 31). TIP and TP gradients are evident in the expanded shale (Figure 6); however, means were not statistically different. This may be a result of the relatively short length of our pilot cells (approximately 2 m). We also expected that organic P would be higher in top samples than in bottom samples because, VOL. 39, NO. 12, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4623
FIGURE 3. Sequential phosphorus fractionation procedure. Pi ) inorganic phosphorus; Po ) organic phosphorus. The number of hours shown is the extraction time. in natural wetlands, soil accretion can account for the majority of newly stored phosphorus (1). The rate of soil accretion is slow, however. For example, soil accretion in enriched cattail marshes in the Everglades accounted for 0.63 g of P m-2 year-1(1). Pea-sized gravel, another common filter material for subsurface flow wetlands, contained relatively small amounts of P after exposure to wastewater (Figure 6). Sequential Phosphorus Fractionation: Comparison of the Sum of Inorganic Fractions to Total Inorganic Phosphorus. The sum of individual P fractions (sum P) recovered from 63% to 84% of the value obtained from TP analyses. We confirmed during trial fractionations that the organic component of the HCl extracts (HCl Po) represented approximately 27% of total P extracted. Our fractionation scheme did not include quantification of this organic P pool; therefore, it was discarded. It is doubtful that this fraction represents a major storage compartment for added P in either of our materials; however, the fractionation procedure should be amended to include analyses of the HCl Po component. Masonry Sand. Although total P increases in sand were not statistically significant, seven of the individual P fractions had small but statistically significant increases (Table 2). The largest increase was in labile Pi (7.5 mg kg-1). Labile Pi represents weakly adsorbed inorganic P that acts as a buffer regulating solution concentrations of P, thus providing temporary and intermediate storage of P (11). Labile Pi has been shown to increase significantly in systems with anthropogenic inputs of soluble P (11, 34, 39). Models suggest that P sorption is governed by the formation of metal phosphate precipitates on the surfaces of metal oxide particles, followed by diffusion of P through the metal phosphate precipitate to the unreacted metal oxide surface (40). Diffusion, which controls the sorption rate, is a function of the P concentration gradient between the solution and the interface between the metal oxide and the metal phosphate (40). Initial adsorption induces a diffusion gradient 4624
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 12, 2005
toward the interior of the particle, which begins the solidstate diffusion process (41). Thus, labile Pi may be converted from short-term to long-term storage by precipitation or incorporation into the interior structure of the substrate (37, 42). It appears unlikely that interior diffusion occurred in our sand. Instead, most P remained in the reversible, labile Pi form and was periodically desorbed when incoming wastewater P concentrations were very low. Microbial P represents both inorganic P (Pi) and organic P (Po) held in live microorganisms associated with the substrates. This pool turns over rapidly to supply Pi for plant use (43). Microbial P accounted for 22% (3.76 mg kg-1) of the increase in sand P. Microbial P has been found to represent 16-37% of total P in Everglades peat soils (11) with higher relative proportions in unenriched soils. Fe + Al P accounted for only 23% of the P increase in sand. Humic P and labile Po also accounted for small increases in sand P. The humic P pool was expected to be small in both treatments due to the young age of the wetlands and relatively small amounts of plant biomass. Phosphorus associated with Ca and Mg accounted for the largest single P fraction in both pilot cell sand (40.6 mg kg-1) and control sand (40.9 mg kg-1); however, this pool did not act as a storage fraction for added P. We concluded that sand is a poor candidate for retaining dissolved P in a subsurface flow system. Del Bubba et al. (22) reported Langmuir sorption maxima for 13 Danish sands ranged from 20 to 129 mg kg-1, with all but three of the sands below 100 mg kg-1. Thus, the Langmuir sorption maximum of our sand [59 mg kg-1 (5)] is within the range of sands likely utilized in many European treatment wetlands. Del Bubba et al. (22) also fractionated the 13 sand samples after exposure to added P and found residual P was the largest P pool in all but one sand; labile P ranged from 2.6 to 17.3 mg kg-1. The authors did not, however, compare P content to a control, so it is not possible to draw conclusions about how the added P was stored.
FIGURE 4. Soluble reactive phosphorus (SRP) in inflow (SRPin) and in the effluent (SRPout) from each pilot cell. Top panel, sand (A, B, C); bottom panel, shale (X, Y, Z). Expanded Shale. In our shale, three fractions, labile Pi, (Fe + Al) P, and humic P, accounted for most of the retained P, while two fractions, (Ca + Mg) P and residual P, decreased by 27.4 and 32.0 mg kg-1, respectively. Labile Pi increased from 8.03 to 54.7 mg kg-1. This fraction may have been larger in the shale than in sand due to abundant micropores, which provide a substantially greater surface area (44). In addition, the higher hydraulic loading rates increased opportunity for reactions between wastewater and substrate. Due to its high Al content, we hypothesized that (Fe + Al) P would be a major storage pool in expanded shale. Mean (Fe + Al) P increased 107.6 mg kg-1, a 9-fold increase over the control. The (Fe + Al) P fraction may also contain phosphate that has penetrated the mineral surface, replacing Al-bound oxygens with phosphate-bound oxygens. Humic P in pilot cell shale increased from 17.0 to 56.8 mg kg-1, a substantially larger increase than in sand. Increases in labile Po and microbial P were similar to those in sand. As with the sand, these pools are probably limited by the overall low organic carbon content and associated microbial communities in the neophyte systems. This does suggest, however, that biofilms are not a significant P storage pool in these systems. The (Ca + Mg) P fraction was the largest component in both substrates and their controls; however, this pool did not act as a storage compartment for added P in the shale. Clearly, P interactions with Ca and Mg were not a viable P retention process in either expanded shale or masonry sand.
FIGURE 5. Substrate total inorganic phosphorus (TIP) and total phosphorus (TP) for pilot cells and respective controls. B and C contained masonry sand; X, Y, and Z contained expanded shale. ShC ) shale control, SaC ) sand control. For each pilot cell, n ) 18; ShC, n ) 8; SaC, n ) 7. Phosphorus retained (not shown) was calculated as the difference between mean P of pilot cell material and mean P of control material. Spatial Differences in P Pools. Although we expected to see higher P concentrations in material located closest to incoming wastewater, there were no statistically significant differences in the horizontal distribution of any P fractions. However, labile Pi and (Fe + Al) P were significantly higher in samples collected from the top 12 in. than in samples taken from the bottom 12 in. (two-sample t-test, p < 0.05). Reddy et al. (34) found that (Fe + Al) P was higher in surface layers of Everglades soils. It is unclear whether higher (Fe + Al) P in top layers of our pilot cells was associated with increased biological phenomenon such as litter fall or mineralization, uneven hydraulic loading resulting in greater exposure to wastewater in the top layer, or redox conditions in deeper pore spaces that would facilitate release of Febound P. Estimates of Total Phosphorus Retained. Mean TP (( SD) retained by expanded shale was 349 ( 171 mg of P (kg dry weight)-1, and the aerial P retention was 201 ( 98.6 g of P m-2 year-1. Thus P retention in shale was many times higher than masonry sand and 3 orders of magnitude higher than the average global P retention via peat accretion (0.5 g m-2 year-1). The higher hydraulic loadings to shale cells conVOL. 39, NO. 12, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4625
TABLE 2. Comparisons of Means or Medians from P Fractions in Pilot Cells and Control Materialsa labile Pi pilot cell control P increase top layer bottom layer pilot cell control P increase top layer bottom layer
8.45 0.92 7.53b 9.17 5.89b 54.7 8.0 46.7b 64.7 44.6b
labile Po 4.08 1.90 2.18b 5.28 2.89 11.1 13.1 -2.0 9.87 2.39
microbial P
(Fe + Al) P
Masonry Sand (mg kg-1) 3.76 5.18 0.0 1.30 b 3.76 3.88b 5.39 6.57 1.63 3.78b Expanded Shale (mg kg-1) 8.5 121 6.2 13.4 2.3 107.6c 3.31 130 6.09 113c
humic P 2.68 1.02 1.66c 3.72 0.84b 56.8 17.0 39.8b 72.8 40.8
(Ca + Mg) P 40.6 40.9 -0.31 44.2 37.1c 151.2 178.6 -27.4 159 143c
residual P 13.5 11.9 1.52 14.5 10.8b 55.4 87.4 -32.0 59.5 51.3
sum P 74.5 58.0 16.5 78.0 59.7 450 318 133 493 389
a Pilot cell P fractions from top and bottom layers of cells were compared. Means are shown in roman type; medians are shown in italic type. Sum P is the sum of all P fractions exclusive of microbial P, which is a subset of labile Po. Nonnormally distributed data were compared by Kruskal-Wallis on ranked data. Normally distributed data were compared by one-tailed, two-sample t-test. b p < 0.01. c p < 0.05.
FIGURE 7. Comparison of P loads in and out for studies with sand and gravel (O), soil (9), light expanded clay aggregate (4), and shale (1). We calculated loads from referenced studies, detailed in Table 3 in Supporting Information.
FIGURE 6. Substrate total inorganic phosphorus (TIP) and total phosphorus (TP) by zone. G1,G5 ) pea gravel; Sa2-Sa4 ) zones with sand; Sh2-Sh4 ) zones with shale. Zone 1 was closest to the inlet. tributed to greater aerial retention; however, the Langmuir sorption maxima for expanded shale (971 mg kg-1) was 16 times higher than the maxima for masonry sand (5), and therefore the shale’s superior P retention is not entirely attributable to higher loading. We compared our results to those we calculated from other subsurface flow wetland studies (Figure 7) that used expanded clay (23), shale (17), soil (23), sand (5, 25), gravel (45, 46), or a sand and gravel mixture (3). Despite high P retention by sands reported by others, the relative efficiency of P retention is low, even when compared to lightweight clays (23) and shales (5, 17). Drizo et al. (17) used shale in mesocosms and achieved P retention that we extrapolated to 407 g of P m-2 year-1. Their mesocosms, however, were 4626
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 12, 2005
subjected to much higher P concentrations (10-120 mg L-1) than any used in this study. Most of the studies shown in Figure 7 have relatively low hydraulic loading rates in order to provide detention times for removal of organics. It is not known whether expanded shale would be effective in treating such organic-rich waste, but for tertiary phosphorus removal, expanded shale and similar materials show far more promise than the more commonly used sands and gravels. Important properties of our shale include excellent hydraulic efficiency, high sorption capacity, and ability to store P in recalcitrant forms. The ultimate P storage potential of a shale wetland system is also unknown; however, mechanisms associated with long-term P storage such as complexation with Fe, Al, and humic materials were demonstrated in this study. By contrast, the primary mechanism of P retention by sand and gravel appears to be ion exchange, which is easily reversed and thus has limited capacity for long-term P storage. Further research is recommended to determine the life of such systems and further optimize their design and operation.
Acknowledgments We thank the City of Denton for their assistance with construction and operation of pilot cells, Texas Industries, Inc., for donation of lightweight expanded shale, and the University of North Texas for financial support.
Supporting Information Available Table 3, supplemental information for Figure 7, substrate treatment wetlands for P removal.
Literature Cited (1) Richardson, C. J.; Craft, C. B. Efficient phosphorus retention in wetlands: Fact or fiction? In Constructed Wetlands for Water Quality Improvement; Moshiri, G. A., Ed.; CRC Press: Boca Raton, FL, 1993; pp 271-282. (2) Kadlec, R. H.; Knight, R. L. Treatment wetlands; CRC Press: Boca Raton, FL, 1996; pp 443-480. (3) Scho¨nerklee, M.; Koch, F.; Perfler, R.; Haberl, R.; Laber, J. Tertiary treatment in a vertical flow reed bed systemsa full scale pilot plant for 200-600 p.e. Water Sci. Technol. 1997, 35 (5), 223230. (4) Greenway, M.; Woolley, A. Constructed wetlands in Queensland: Performance efficiency and nutrient bioaccumulation. Ecol. Eng. 1998, 12, 39-55. (5) Forbes, M. G.; Dickson, K. R.; Golden, T. D.; Doyle, R. D.; Hudak, P. Dissolved phosphorus retention of lightweight expanded shale and masonry sand used in subsurface flow treatment wetlands. Environ. Sci. Technol. 2004, 38, 892-898. (6) Masscheleyn, P. H.; Pardue, J. H.; DeLaune, R. D.; Patrick, W. H., Jr. Phosphorus release and assimilatory capacity of two lower Mississippi Valley freshwater wetland soils. Water Res. Bull. 1992, 28 (4), 763-773. (7) Cooke, J. G.; Stub, L.; Mora, N. Fractionation of phosphorus in the sediment of a wetland after a decade of receiving sewage effluent. J. Environ. Qual. 1992, 21, 726-732. (8) Mitsch, W. J.; Cronk, J. K.; Wu, X.; Nairn, R. W. Phosphorus retention in constructed freshwater riparian marshes. Ecol. Appl. 1995, 5 (3), 830-845. (9) Richardson, C. J. Mechanisms controlling phosphorus retention capacity in freshwater wetlands. Science 1985, 228, 1424-1426. (10) Faulkner, S. P.; Richardson, C. J. Physical and chemical characteristics of freshwater wetlands soils. In Constructed Wetlands for Wastewater Treatment; Hammer, D. A., Ed.; Lewis: Chelsea, MI, 1989; pp 41-72. (11) Qualls, R. G.; Richardson, C. J. Forms of soil phosphorus along a nutrient enrichment gradient in the northern Everglades. Soil Sci. 1995, 160 (3), 183-198. (12) Reed, S. C.; Brown, D. S. Constructed wetland designsthe first generation. Water Environ. Res. 1992, 64, 776-781. (13) Perfler, R.; Haberl, R. Actual experiences with the use of reed bed systems for wastewater treatment in single households. Water Sci. Technol. 1993, 28 (10), 141-148. (14) Nichols, D. S. Capacity of natural wetlands to remove nutrients from wastewater. J. Water Pollut. Control Fed. 1983, 55 (5), 405505. (15) Reddy, K. R.; D’Angelo, E. M. Biogeochemical indicators to evaluate pollutant removal efficiency in constructed wetlands. Water Sci. Technol. 1997, 35 (5), 1-10. (16) Vymazal, J., Brix, H., Cooper, P. F., Green, M. B., Haberl, R., Eds. Constructed wetlands for wastewater treatment in Europe; Backhys Publishing: Leiden, The Netherlands, 1998. (17) Drizo, A.; Frost, C. A.; Smith, K. A.; Grace, J. Phosphate and ammonium removal by constructed wetlands with horizontal subsurface flow, using shale as a substrate. Water Sci. Technol. 1997, 35 (5), 95-102. (18) Zhu, T.; Jenssen, P. D.; Mæhlum, T.; Krogstad, T. Phosphorus sorption and chemical characteristics of lightweight aggregates (LWA)spotential filter media in treatment wetlands. Water Sci. Technol. 1997, 35 (5), 103-108. (19) Sakadevan, K.; Bavor, H. J. Phosphate adsorption characteristics of soils, slags and zeolite to be used as substrates in constructed wetland systems. Water Res. 1998, 32 (2), 393-399. (20) Johansson, L. Blast furnace slag as phosphorus sorbentss column studies. Sci. Total Environ. 1999, 229, 89-97. (21) Arias, C. A.; Del Bubba, M.; Brix, H. Phosphorus removal by sands for use as media in subsurface flow constructed reed beds. Water Res. 2001. 35(5), 1159-1168. (22) Del Bubba, M.; Arias, C. A.; Brix, H. Phosphorus adsorption maximum of sands for use as media in subsurface flow constructed reed beds as measured by the Langmuir isotherm. Water Res. 2003, 37, 3390-3400. (23) Jenssen, P. D.; Mæhlum, T.; Krogstad, T. Potential use of constructed wetlands for wastewater treatment in northern environments. Water Sci. Technol. 1993, 28 (10), 149-157. (24) Johansson, L. The use of Leca (Light Expanded Clay Aggregates) for the removal of phosphorus from wastewater. Water Sci. Technol. 1997, 35 (5), 87-93.
(25) Geller, G. Horizontal subsurface flow systems in the German speaking countries: Summary of long-term scientific and practical experiences; recommendations. Water Sci. Technol. 1997, 35 (5), 157-166. (26) Luederitz, V.; Eckert, E.; Lange-Weber, M.; Lange, A.; Gersberg, R. M. Nutrient removal efficiency and resource economics of vertical flow and horizontal flow constructed wetlands. Ecol. Eng. 2001, 18, 157-171. (27) Meuleman, A. F.; van Logtestijn, M. R.; Rijs, G. B. J.; Verhoeven, J. T. A. Water and mass budgets of a vertical-flow constructed wetland used for wastewater treatment. Ecol. Eng. 2003, 20 (1), 31-44. (28) Drizo, A.; Comeau, Y.; Forget, C.; Chapuis, R. P. Phosphorus saturation potential: A parameter for estimating the longevity of constructed wetland systems. Environ. Sci. Technol. 2002, 36 (21), 4642-4648. (29) Cueto, A. J. Development of criteria for the design and construction of engineered aquatic treatment units in Texas. In Constructed Wetlands for Water Quality Improvement; Moshiri, G. A., Ed.; CRC Press: Boca Raton, FL, 1993; pp 99-105. (30) Jenssen, P. D.; Siegrist, R. L. Technology assessment of wastewater treatment by soil infiltration systems. Water Sci. Technol. 1990, 22 (3/4), 83-92. (31) Tchobanoglous, G.; Schroeder, E. D. Water quality: characteristics modeling modification; Addison-Wesley: Reading, MA, 1987; pp 289 and 423. (32) APHA. Standard methods for the examination of water and wastewater, 19th ed.; American Public Health Association, American Water Works Association, and Water Environment Federation: Washington, DC, 1995. (33) Chang, S. C.; Jackson, M. L. Fractionation of soil phosphorus. Soil Sci. 1957, 84, 133-144. (34) Reddy, K. R.; Wand, Y.; DeBusk, W. F.; Fisher, M. M.; Newman, S. Forms of soil phosphorus in selected hydrologic units of the Florida Everglades. Soil Sci. Soc. Am. J. 1998, 62, 1134-1147. (35) Ivanoff, D. B.; Reddy, K. R.; Robinson, S. Chemical fractionation of organic phosphorus in selected histosols. Soil Sci. 1998, 163, 36-44. (36) Brady, N. C.; Weil, R. R. Elements of the Nature and Properties of Soils; Prentice Hall: Englewood Cliffs, NJ, 2000; pp 400-409. (37) Stumm, W.; Morgan, J. Chemical equilibria and rates in natural waters, 3rd ed.; Wiley: New York, 1996; pp 404-414. (38) Forbes, M. G. Phosphorus retention and fractionation in masonry sand and lightweight expanded shale used as substrate in a subsurface flow wetland. Dissertation, University of North Texas, Denton, TX, 2002; 211 pp. (39) Koch, M. S.; Reddy, K. R. Distribution of soil and plant nutrients along a trophic gradient in the Florida Everglades. Soil Sci. Soc. Am. J. 1992, 56, 1492-1499. (40) Pierzynski, G. M. The chemistry and mineralogy of phosphorus in excessively fertilized soils. Crit. Rev. Environ. Control, 1991, 21 (3/4), 265-295. (41) Barrow, N. J. A mechanistic model for describing the sorption and desorption of phosphate by soil. J. Soil Sci. 1983, 34, 733750. (42) Weber, W. J., Jr.; McGinley, P. M.; Katz, L. E. Sorption phenomena in subsurface systems: concepts, models and effects on contaminant fate and transport. Water Res. 1991, 25 (5), 499528. (43) Page, A. L.; Miller, R. H.; Keeney, D. R. Methods of soil analysis part 2schemical and microbiological properties, 2nd ed.; American Society of Agronomy and Soil Science Society of America: Madison, WI, 1982. (44) Sloan, J. J.; George, S. W.; Mackay, W. A. Moisture and nutrient dynamics in expanded shale. Research Update; Texas Agricultural Experiment Station, Texas A&M: College Station, TX, 1999. (45) Neralla, S.; Weaver, R. W.; Lesikar, B. J.; Persyn, R. A. Improvement of domestic wastewater quality by subsurface flow constructed wetlands. Biores. Technol. 2000, 75, 19-25. (46) Vymazal, J. The use of subsurface constructed wetlands for wastewater treatment in the Czech Republic: 10 years experience. Ecol. Eng. 2002, 18, 633-646.
Received for review November 24, 2004. Revised manuscript received April 7, 2005. Accepted April 8, 2005. ES048149O
VOL. 39, NO. 12, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4627
