Journal of Plant Ecology Advance Access originally published online on February 14, 2008
Journal of Plant Ecology 2008 1(1):67-74; doi:10.1093/jpe/rtm002
| ||||||||||||||||||||||||||||||||||||||||||||||||
Phosphorus removal in small constructed wetlands dominated by submersed aquatic vegetation in South Florida, USA
Everglades Division, South Florida Water Management District, West Palm Beach, FL 33406, USA
* Correspondence address. Everglades Division, South Florida Water Management District, 3301 Gun Club Road, West Palm Beach, FL 33406, USA. E-mail: bgu{at}sfwmd.gov
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Aims: Free-surface flow-constructed wetland is a powerful means for the reduction of contaminants from agricultural runoff. Wetlands dominated by submerged aquatic vegetations (SAVs) may take up nutrients, particularly phosphorus (P), from surface flow with high efficiency. The objective of this study was to assess P removal performance by the SAV community under high and low P concentrations.
Methods: Weekly or biweekly inflow and outflow water samples were collected from four small constructed wetlands (test cells) planted with SAV in South Florida, USA, between September 1999 and September 2001. These test cells were divided into two groups, with the north test cells receiving a higher inflow total phosphorus (TP) concentration (average = 75 µg l–1) than the south test cells receiving a lower TP concentration (average = 23 µg l–1). Limerock (LR) berms were installed in two of these test cells to allow an evaluation of the efficiency of this physical barrier to enhance wetland performance.
Important findings: North test cells displayed high TP removal of 
60% while the removal efficiency of the south test cells was only 20%. Soluble reactive phosphorus concentrations in both north and south test cells were sequestered down to near-detection limit. High removal efficiencies for particulate phosphorus were also observed in the north test cells. The LR berms at the two test cells were found to be associated with decreases of an average TP removal of 2 µg l–1. Outflow TP concentration did not increase with inflow TP concentration, but increased with nominal hydraulic loading rates. Findings from this study demonstrated high P removal from inflow water containing high TP concentration by the SAV wetland and the importance of hydraulic regime to wetland performance.
Keywords: constructed wetlands • hydraulic loading rates • limerock berm • removal efficiency • submerged aquatic vegetation (SAV) • total phosphorus (TP)
Received: 3 July 2007 Revised: 16 August 2007 Accepted: 19 October 2007
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Free-surface flow-constructed wetlands have increasingly been used to treat storm water runoff (Kadlec 2005; Kadlec and Knight 1996; Mitsch et al. 1998). The vegetation community in most of these systems is dominated by emergent plants, including the large constructed wetlands in South Florida, i.e., the Stormwater Treatment Areas (STAs), that were built for the restoration of the Everglades ecosystem (Chimney and Goforth 2001). Recent research indicates that the emergent plants-based STAs are capable of reducing the total phosphorus (TP) concentration from
150 µg l–1 at inflow down to 40–50 µg l–1 at outflow (Nungesser and Chimney 2001; Pietro et al. 2006). However, even greater P reduction may be needed to fully protect the Everglades (Sklar et al. 2005). Recent studies have documented high TP removal rates in wetlands dominated by submerged aquatic vegetation (SAV) (Dierberg et al. 2001; Gu et al. 2001). Direct uptake by SAV and its associated periphyton, gravity settling and co-precipitation of calcium carbonate (CaCO3) with soluble reactive phosphorus (SRP) are the mechanisms thought to regulate P removal in these systems (Dierberg et al. 2001, 2002). Because the TP removal efficiency (RE) of SAV communities can be greater than that of emergent macrophytes, SAV currently is established in portions of the STAs to enhance their treatment performance (Pietro et al. 2006).
Herein we report on several experiments that examined the treatment performance of small constructed SAV wetlands, referred to as 'test cells'. The test cells were used as analogs of the STAs and had biological and hydrological characteristics that were similar to the larger systems (Chimney et al. 2000). This research has provided a greater understanding of how SAV wetlands function, which has helped to optimize the design and operation of the STAs. The objective of this paper was to evaluate TP removal performance under various hydraulic loading rates (HLRs). In addition, the effect of limerock (LR) berms on TP removal was also addressed.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Test cell description
The test cells were constructed on the agriculture land in Palm Beach County, Florida (26°38'N and 80°25'W), in 1998. They are small (2000 m2) rectangular wetlands (
91 x 21 m) arranged into two banks of 15 cells (Fig. 1). The research described in this paper was conducted in two north test cells (N1 and N15) and two south test cells (S4 and S9). Data for TP removal from two north and two south test cells vegetated by cattail (Typha sp.) were also used for performance comparison with the SAV test cells. All test cells had 0.6 m of peat substrate over 0.6 m of sand. Each test cell is hydrologically isolated by a full liner to prevent seepage from the adjacent test cells. Inflow water came from a canal downstream of the Everglades Agricultural Area and the surrounding treatment wetland. Water was first pumped into an elevated storage cell and then delivered in parallel fashion to the test cells via a pipe distribution system. Inflow rate was regulated by changing calibrated orifice caps fitted to the end of the distribution pipes. Distribution manifolds were installed to deliver water at multiple points across the top of the test cells to improve their hydraulic efficiency compared to point-source inflows (Chimney et al. 2000). The north test cells had inflow with relatively high TP concentrations of 60–150 µg l–1 while the south test cells received lower TP concentrations of 30–50 µg l–1 (Newman and Lynch 2001).
|
There had been several changes in vegetation types in the test cells. The test cells were initially dominated by cattail (Typha sp.). Prior to this study in 1998, the wetlands were drained and the emergent plants were removed. Upon reflooding, all test cells became dominated by Chara sp. By early 2000, however, the plant community in the south test cells had shifted largely to emergent macrophytes and Hydrilla verticillata, an undesirable exotic species. In June 2000, the south test cells were again drained, herbicided and stocked at
0.9 kg m–2 with Najas guadalupensis harvested from an adjacent wetland. In addition, 1 ton of Ceratophyllum demersum was added to the inflow region of the north test cells in July 2000. By mid-2001, the north test cells were dominated by N. guadalupensis with lesser amounts of C. demersum and Chara sp., while the SAV community in south test cells were dominated by Chara sp.
Experimental design
Hydrologic assessment
During the study period, the hydrological regime of the SAV test cells changed several times to meet operation, management and experiment requirements. The alterations of water depth, HLR and hydraulic retention time (HRT) are detailed in Table 1.
|
LR berm evaluation
One goal of this research was to assess the ability of LR, i.e., limestone, to adsorb P onto its surface and enhance the treatment performance of SAV wetlands. Berms of crushed LR (
2.5 cm diameter) were constructed across the width (and perpendicular to the direction of flow) at a distance 88% of the flow path length downstream of the distribution manifolds in N15 and S9 in April 2000. The berms were constructed directly on the peat substrate; they measured 4.9 m at their base and tapered to 1.2 m at their top. The berms were designed to allow water to pass through them. The estimated travel time of water through the berm was 5–10% of the test cell's HRT. Head loss through the berm was negligible (
0.3 cm) even after 17 months of continuous operation.
Water quality monitoring
TP in the test cells was monitored by collecting weekly or biweekly grab samples at the inflow and outflow, from September 1999 to September 2001. Sampling was suspended in the south test cells from March 2000 until August 2000 and in N15 from April 2000 until June 2000 due to experimental manipulations and/or vegetation management activities. In addition to TP, weekly grab samples were collected at the inflow and outflow from October 2000 to April 2001 and from July to September 2001 and analyzed for SRP, total dissolved phosphorus (TDP), total suspended solids (TSSs), calcium (Ca) and alkalinity following standard methods (APHA 1998). Water temperature, pH and conductivity were measured in the field with a Hydrolab® (Hydrolab-Hach Co., Loveland, CO) on each sampling date. Particulate phosphorus (PP) was calculated by the difference between TP and TDP; dissolved organic phosphorus (DOP) was calculated by the difference between TDP and SRP. Dissolved Ca concentration, alkalinity, conductivity and pH in the north and south test cells after the installation of the LR berms at N15 (June 2000 to September 2001) and at S9 (August 2000 to September 2001) were measured to assess the potential effects of LR berms on water chemistry.
RE for various forms of P was calculated as:
 | (1) |
Mass removal rate (g m–2 year–1) was calculated as:
 | (2) |
Settling rate constant (k) was calculated as:
 | (3) |
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Hydraulic loading and inflow water quality
Hydraulic regime changed considerably during the study period (Table 1). All test cells were shallow with water depths ranging from 0.3 to 0.96 m with the north test cells deeper than the south test cells. Nominal HLRs varied from 5.5 (S4) to 22.5 cm day–1 (N1). Nominal hydraulic residence time was from 2.7 (N1) to 7.7 days (N15).
Table 2 summarizes several ambient water quality variables from the inflow except nutrients. The inflow water from the EAA runoff was high in alkalinity, Ca, conductivity and total organic carbon content and low in DO and TSS and had above-neutral pH. There were no significant differences in these variables between the north and south test cells (all at P > 0.05).
|
P concentration and removal
TP removal performances between the north and south test cells differed considerably (Fig. 2). Inflow TP concentration at the north test cells ranged from 25 to 188 µg l–1 with an average of 75 µg l–1. TP removal occurred in most sampling events with average outflow concentrations of 27 and 24 µg l–1 for N1 and N15, representing 67 and 73% removal, respectively (Table 3). Reductions for all the P species contributed to the overall reduction in TP concentration in the test cells. The average removal rate for SRP at north test cells was >90% and virtually all of the SRP was sequestered (Table 3). The removal rate for PP was also high (>60%) followed by DOP removal (>40%).
|
|
TP removal for the south test cells was low over the entire monitoring period (Fig. 2). Average inflow TP concentration was 23 µg l–1 compared to average outflow concentrations of 21 and 20 µg l–1 for S6 and S9, respectively (Table 3). Average TP removal rates were only 15 and 35%, respectively. Removal of SRP was the highest (>60%) compared to the removal of PP and DOP, which were <20% (Table 4).
|
Changes in plant community and TP removal
Inflow TP concentration between the initial period dominated by Chara and the second period dominated by Najas at both north and south sites did not show significant differences (P > 0.05). Except for the N1, outflow TP concentrations in the Najas period were significantly lower than those in the Chara period (Fig. 2), indicating that the majority of the test cells improved their TP removal. However, the relationship between plant species and TP removal is complicated by the timing of experiments and hydraulic loading.
TP removal by LR berm
Water quality data were split into two periods, representing pre-LR period and post-LR berm period. For the period prior to the placement of the LR berm (September 1999 to April 2000), the average TP concentration was higher in the outflow of N1 by 6 µg l–1 than in the outflow of N15 (Table 3). Differences in the outflow TP concentrations probably reflect the considerable changes in the HLR to the cells during this period. For example, the HLR was increased from 10–13 to 40 cm day–1 in N1 for 20 days and then was reduced to 2–3 cm day–1 for the following 6 months. The HLR for N15 was never set higher than the initial 10–13 cm day–1 and was maintained at the lower rate of 2–3 cm day–1 for only 53 days. Thus, N1 was subjected to the extremes in HLR more frequently and for longer periods than was N15, which may have affected the TP removal performance of the wetland.
We observed marked performance differences between N1 and N15 after placement of the LR berm in N15 and resumption of P monitoring in that test cell in June 2000. The mean TP outflow concentration from N15 was lower than that of N1 by 4 µg l–1 during both post-LR periods (Table 3). Since the HLR (and thus the P loading) to N1 was more than double that of N15 (12.1 versus 5.1 cm day–1) for one-third of the post-LR berm period, the mass removals achieved in the two test cells (3.25 g P m–2 year–1 for N1 and 2.20 g P m–2 year–1 for N15) cannot be compared based only on the presence or absence of the LR berm.
The HLRs for S4 and S9 were comparable during the monitoring periods, which enabled us to compare nutrient removal performance. The outflow TP concentration of S9 was higher by 6 µg l–1 than that of S4 during the pre-LR berm period in S9 (Table 3). Since TP concentrations in the outflow of S9 were higher than the inflow concentrations, P was being exported from the test cell prior to the installation of the LR berm. Most of the export occurred during the beginning of the monitoring period (Fig. 2) and may have been due to high P leaching from the native muck that was deposited into the cell during construction. Since a cattail community inhabited the cell prior to being converted into an SAV-dominated system, release from decomposing cattail tissues or cattail-derived organic detritus may also have contributed to a flush of P at the beginning of the monitoring period. After LR berm installation at S9 and the resumption of monitoring on 4 August 2000, the outflow concentration at S9 was on average 5 µg l–1 lower than that of S4 for the post-LR period (Table 3).
Samples were collected for TP analysis from in front of the LR berm (‘pre-berm’ samples), as well as from the test cell outflows. This enabled us to quantify the performance of the LR berm and the associated downstream ‘polishing wetland’. On average, both cells removed 2 µg l–1 of TP (excluding an extreme value from S9). However, consistently net removal did not take place until 6 months later for S15 and 12 months later for S9 (Fig. 3). Dissolved Ca, total alkalinity and specific conductivity decreased from inflow to outflow stations in all the test cells during the post-LR berm period (Table 4). The south test cells exhibited a greater reduction in these constituents than did the north test cells, probably because of their shallower depths and lower HLRs (Table 1).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Aquatic systems dominated by SAV exhibit higher P RE and mass accumulation rates than systems dominated by emergent macrophytes (Dierberg et al. 2001). Our findings from test cells are in agreement with the results reported from other SAV systems including mesocosms (Dierberg et al. 2001), large-scale constructed wetlands (Gu et al. 2001) and natural lakes (Knight et al. 2003). These studies indicate that the average settling rate constant for SAV systems range from 99 m year–1 for mesocosms, to 30 m year–1 for test cells and to 40 m year–1 for a full-scale constructed wetland and 48 m year–1 for natural Florida lakes. It is likely that the settling rate constant for mesocosms and test cells reflect the start-up condition when SAVs experience unsustainable growth. However, rate constants for an STA and the natural lakes are calculated based on data from 8 to 20 years of records, demonstrating the high long-term P removal by the SAV systems. Similarly, a recent study indicates that the littoral zone of lakes with 50% of the sediment surface colonized by SAV was responsible for one-third of the whole lake P and two-thirds of the whole lake bulk sediment accumulation annually (Rooney et al. 2003).
On the other hand, the average rate constant for emergent wetlands is 12 m year–1 (Kadlec and Knight 2001). A comparison of TP removal between the SAV and cattail test cells, which had similar HLR for the same time period, reveals that the north SAV test cells performed much better than their counterparts while both south SAV and cattail test cells performed similarly at low removal rates (Table 5). The SAV treatment systems have several advantages on P removal over wetlands dominated by emergent plants. First, in contrast to emergent plants, which only use roots for nutrient uptake, the entire vegetation surface and the associated periphyton are capable of assimilating nutrients from the water column. Our data indicate that the SAV community in the test cells removed essentially all of the SRP from the inflow. Recent experiments have demonstrated high rates of SRP uptake by N. guadalupensis (Dierberg et al. 2002) and C. demersum (Pietro et al. 2006) in South Florida. Second, the dense SAV community served as a physical filter for the removal of suspended PP that made up 40% of the TP at the test cell inflow. High percentages (45–66%) of PP were removed from the test cells. However, wetlands with emergent plants are also efficient in removing particulate matter (Coveney et al. 2002). Third, intensive photosynthesis by SAV removes CO2 from the water column, leading to an increase in pH and CaCO3 co-precipitation with SRP (Kleiner 1988; Murphy et al. 1983). Recent studies have shown consistent CaCO3 saturation in SAV mesocosms and a full-scale wetland (Dierberg et al. 2002; Gu et al. 2006). The relative importance of TP removal through CaCO3 co-precipitation and diel changes in pH and CaCO3 concentration in the SAV wetlands requires further research.
|
Several factors may influence the P removal performance in the SAV wetlands. The first is the quantity of P from the inflow waters, which includes TP concentration and mass loading. In our study, high removal efficiencies were found in the north test cells with average inflow TP concentration of 75 µg l–1. At an inflow TP concentration of
40 µg l–1 in an STA dominated by SAV, treatment efficiency was also considerably high (Gu et al. 2001). South test cells with inflow TP concentrations ranging from 19 to 31 µg l–1 exhibited low removal efficiencies. This is because a majority of the SRP had been removed from the inflow downstream of the STAs. Further removal from waters with low concentrations is increasingly difficult. This is supported by the similar long-term outflow TP concentration in the full-scale STAs to the results from this study (Juston and DeBusk 2006). It appears that the SAV treatment system is more suitable to treat inflow water with TP concentration >30 µg l–1. HLR is another factor affecting wetland performance. HLRs for the test cells were lower than those in an STA (15 cm day–1) and natural Florida lake and river systems dominated by SAV (24 and 48 cm day–1 respectively) reported by Knight et al. (2003). The time-adjusted average P removal rate was 3.72 g m–2 year–1 for the north test cells and 0.11 m–2 year–1 for the south test cells (Table 5) and falls in the range for Florida lakes (Knight et al. 2003). The removal rate for the north test cells is higher than that (1.83 m–2 year–1) of STA dominated by SAV (Nungesser and Chimney 2001) and the long-term rate of 1 m–2 year–1 proposed by Richardson et al. (1997) for emergent wetlands. The P loading and removal rates increased with HLR (Fig. 5). Therefore, it is not surprising that the removal rate was considerably lower in the south test cells than in the north test cells because the former had low HLR and P mass loading rate. Rate constant (k value) for north and south test cells ranged from 3.6 to 55.4 m year–1 and did not always increase with HLR, with the highest K value at moderate HLRs (Table 5). However, a constructed wetland dominated by SAV in South Florida had a HLR of 15 cm day–1 and a k value of 40 m year–1 (Nungesser and Chimney 2001). Knight et al. (2003) reported an average HLR of 23.7 cm day–1 and a k value of 15 m year–1 for several SAV-dominant natural systems. It appears that constructed systems performed better than the natural systems in terms of RE of TP. However, P in the sediment from recent deposits is still subject to biogeochemical manipulation, one cannot judge the system performance by simply comparing k values with different time horizons.
Outflow TP concentration is the key nutrient criterion for the successful Everglades restoration mandated by the 1994 Florida Everglades Forever Act. Typically, inflow TP concentration and hydraulic loading play important roles in controlling outflow TP concentration (Dierberg et al. 2001; Gu et al. 2001). However, we found a weak or no correlation (all P > 0.05) between inflow and outflow TP concentrations (Fig. 4) with an exception for N1 (r = 0.51, P < 0.05). Especially, during the final 4–5 months when inflow TP concentrations increased dramatically, the outflow TP concentrations remained similar to those of previous months for 3 out of 4 test cells (Fig. 2). This finding has two important implications for STA management. First, it suggests that the test cells under the given HLR and other environmental conditions were capable of removing TP from inflow concentration as high as 140 µg l–1 down to 20 µg l–1. Second, it suggests that outflow TP concentration and TP removal rate at these test cells were a function of HLR (Fig. 5). High HLR implies less time for the plants to interact with the nutrients and a potentially high probability for hydraulic short-circuiting, which may reduce nutrient RE in the test cells.
|
|
Data collected for the LR evaluation indicated that test cells with LR removed more TP than the test cells without LR with consistent performance delayed for several months to up to 1 year. This is largely in agreement with the LR berm data from previous mesocosm experiments, which indicated that a several month conditioning period is required before PP removal will occur within an LR bed (Dierberg et al. 2001). The prolonged delay for net removal in S9 is unknown, but may be associated with low nutrient status in the inflow. Possible removal mechanisms that took place within the LR include co-precipitation of CaCO3 with SRP and filtration of PP. A more accurate evaluation of LR performance is needed since no samples were collected at the locations immediately post-LR. It is also likely that LR may serve as a physical barrier for better flow distribution, especially in the full-scale STAs, which will require additional research.
|
Conclusions |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Our experiments using small test cells demonstrated that SAV-dominated wetlands were very efficient in removing P, even at high hydraulic (20–22 cm day–1) and TP (7.5–8.3 g m–2 year–1) loading rates. SRP was the most readily sequestered P species at the north and south test cells. Regardless of the inflow SRP concentrations and HLR, essentially all SRP was removed within each test cell. Outflow TP concentration increased with increasing HLRs, but generally had no relation to inflow TP concentration. Internal hydraulic distribution and biogeochemical processes also likely controlled outflow TP concentration and overall wetland performance. The role of LR berms on TP removal needs more research with better sampling design and longer time scale.
|
Acknowledgements |
|---|
I thank D.B. Environmental Inc. for fieldwork and sample analyses and Mike Chimney, Seán Sculley, Thomas DeBusk and Thomas Dreschel for comments.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
-
APHA. Standard Methods for the Examination of Water and Wastewater (1998) 20th edn. Washington, DC: American Public Health Association, American Water Works Association and Water Pollution Control Federation.
Chimney MJ, Goforth G. Environmental impacts to the Everglades ecosystem: a historical perspective and restoration strategies. Water Sci Technol (2001) 44:93–100.[ISI][Medline]
Chimney MJ, Nungesser M, Newman J, et al. Chapter 6: Stormwater treatment areas—status of research and monitoring to optimize effectiveness of nutrient removal and annual report on operational compliance. In: 2000 Everglades Consolidated Report (2000) West Palm Beach, FL: South Florida Water Management District. 6–1–6–127.
Coveney MF, Stites DL, Lowe EF, et al. Nutrient removal from eutrophic lake water by wetland filtration. Ecol Eng (2002) 19:141–59.[CrossRef][ISI]
Dierberg FE, DeBusk TA, Jackson SD, et al. Submerged aquatic vegetation-based treatment wetlands for removing phosphorus from agricultural runoff: response to hydraulic and nutrient loading. Water Res (2001) 36:1409–22.
Dierberg F, DeBusk TA, Potts J, et al. Biological uptake vs. coprecipitation of soluble reactive phosphorus by ‘P-enriched’ and ‘P-deficient’ Najas guadalupensis in hard and soft waters. Verh Int Ver Theor Angew Limnol (2002) 28:1865–70.
Gu B, Chimney MJ, Newman J, et al. Limnological characteristics of a constructed wetland in south Florida. Ecol Eng (2006) 27:345–60.[CrossRef][ISI]
Gu B, DeBusk TA, Dierberg FE, et al. Phosphorus removal from Everglades agricultural area runoff by submerged aquatic vegetation/limerock treatment technology: an overview of research. Water Sci Technol (2001) 44:101–8.[ISI][Medline]
Juston J, DeBusk TA. Phosphorus mass load and outflow concentration relationships in stormwater treatment areas for Everglades restoration. Ecol Eng (2006) 26:206–23.[CrossRef][ISI]
Kadlec RH. Phosphorus removal in emergent free surface wetlands. J Environ Sci Health (2005) 40:1293–306.[CrossRef]
Kadlec RH, Knight RL. Treatment Wetlands (1998) Boca Raton, FL: CRC Press.
Kleiner J. Coprecipitation of phosphorus with calcite in lake water: a laboratory experiment modeling phosphorus removal with calcite in Lake Constance. Water Res (1988) 22:1259–65.
Knight RL, Gu B, Clarke RA, et al. Long-term phosphorus removal in Florida aquatic systems dominated by submerged aquatic vegetation. Ecol Eng (2003) 20:45–63.[CrossRef][ISI]
Mitsch WJ, Wu X, Nairn RW, et al. Creating and restoring wetlands. Bioscience (1998) 48:1019–30.[CrossRef][ISI]
Murphy TP, Hall KJ, Yesaki I. Coprecipitation of phosphate with calcite in a naturally eutrophic lake. Limnol Oceanogr (1983) 28:58–69.
Newman JM, Lynch T. The Everglades Nutrient Removal Project test cells: STA optimization—status of the research at the north site. Water Sci Technol (2001) 44:117–22.[ISI][Medline]
Nungesser MK, Chimney MJ. Evaluation of phosphorus reduction in a south Florida treatment wetland. Water Sci Technol (2001) 44:109–15.[ISI][Medline]
Pietro K, Bearzotti R, Chimney M, et al. Chapter 4: STA performance, compliance and optimization. In: South Florida Environment Report (2006) West Palm Beach, FL: South Florida Water Management District. 1–114.
Pietro KC, Chimney MJ, Steinman AD. Phosphorus removal by the Ceratophyllum/periphyton complex in a south Florida (USA) freshwater marsh. Ecol Eng (2006) 27:290–300.[CrossRef][ISI]
Richardson CJ, Qian S, Craft CB, et al. Predictive models for phosphorus retention in wetlands. Wetlands Ecol Manage (1997) 4:159–75.
Rooney N, Kalff J, Habel C. The role of submerged macrophyte beds in phosphorus and sediment accumulation in Lake Memphremagog, Quebec, Canada. Limnol Oceanogr (2003) 48:1927–37.
Sklar FH, Chimney MJ, Newman S, et al. The ecological-societal underpinnings of Everglades restoration. Front Ecol Environ (2005) 3:161–9.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||