1 Background, aim, and scope

Chemical fertilizers have been playing a very important role in agricultural production in the modern society. Because of crops’ quick response to chemical fertilizers, to many farmers, fertilizer application seems to be the only guarantee of high crop yield. But the ever increasing use of fertilizer results in significant buildup of nutrients, such as N and P, in the soils (Smith et al. 2007). When the soils are saturated, these nutrients are subjected to losses by leaching and surface runoff. Water quality is impaired and water availability is reduced because of accelerated eutrophication (Carpenter et al. 1998).

Water quality throughout south Florida has been a major concern for many years. Nutrient enrichment has been considered to impact ecological functions of the Everglades National Park, Lake Okeechobee, and Indian River Lagoon (Capece et al. 2007; Ritter et al. 2007). Various water quality problems affect the Indian River Lagoon (IRL), most of which are associated with the development of an intricate network of the canals that drain the surrounding urban and agricultural lands. Canals C-23, C-24, and C-44 in the St. Lucie Basin, which are connected to the IRL, are estimated to collectively deliver at least 8.6 × l05 kg of N, 9.1 × 105 kg of P, and 3.6 × l08 kg of suspended solids (SS) to the estuary annually (Graves and Strom 1992). Overall IRL total N load is projected (year 2010) to increase by 32% (Woodward-Clyde Consultants 1994).

The St. Lucie Estuary is facing challenges of eutrophication due to increased inputs of nutrients, especially N and P from nonpoint sources. Results from a recent monitoring study by He et al. (2006) indicate that more than 50% of the surface runoff water samples contained a total N of 1 to 5 mg l−1 and total P above 1.0 mg l−1. Mean concentrations of total N and total P in the runoff were 4.1 and 1.6 mg l−1, respectively, which are much greater than the USEPA critical levels for surface waters (1.5 mg l−1 for total N and 0.1 mg l−1 for total P) (U.S. Environmental Protection Agency 1976).

Best management practices have been used to reduce N and P export from urban areas and agricultural fields, approximately 10–15% reduction may be realized based on our previous BMPs project (He et al. 2005). This reduction is still far below the goals (30–70% reduction in N and P) established in the Surface Water Improvement and Management Plan (SWIM plan) (SFWMD and SJRWMD 1994) for the St. Lucie Estuary watershed. The stormwater needs to be further cleaned before it is dischargeable to the St. Lucie Estuary.

Physical and chemical treatments to remediate eutrophication in waters are not cost effective, less flexible in terms of design modifications, and are targeted primarily to remove BOD and, to a lesser extent, to reduce N and P levels. Phytoremediation has been increasingly used to clean up contaminated soil and water systems because of its lower costs and fewer negative effects than physical or chemical engineering approaches (Gumbricht 1993; Kowalik et al. 1998; Mahujchariyawong and Ikeda 2001). The principles of phytoremediation systems for cleaning up stormwater include: (a) identification and implementation of efficient aquatic plant systems; (b) uptake of dissolved nutrients including N, P, and metals by the growing plant; and (c) harvest and beneficial use of the plant biomass produced from the remediation system.

Large constructed wetlands or stormwater treatment areas have been operating since the early 1990s to filter nutrients in eutrophic stormwater from Everglades Agricultural Area before they are drained into a water conservation area in the Everglades National Park. Similar wetland systems are also under construction in the Indian River area to reduce nutrients (N and P) before the stormwater from the agricultural areas is discharged into the IRL. Key to the performance of wetlands in reducing nutrient and metal loads is the establishment and sustainability of desired vegetation communities.

In many cases, especially in tropical or subtropical areas, invasive plants such as the water hyacinth (Eichhornia crassipes) and water lettuce (P. stratiotes L.) are used in these phytoremediation water systems (Karpiscak et al. 1994; El-Gendy et al. 2005). This is because, compared to native plants, these invasive plants show a much higher nutrient removal efficiency with their high nutrient uptake capacity, fast growth rate, and big biomass production (Reddy and Sutton 1984). In the active growth season, for instance, water hyacinth plants can double in number and biomass in 6 to 15 days (Lindsey and Hirt 1999). Thus, one of the large-leaved floating invasive plants, water lettuce was chosen in this study. And the primary objectives of this study were to evaluate the effectiveness of water lettuce (P. stratiotes L.) in removing nutrients, including N and P, from stormwater in the constructed water detention systems before it is discharged into the St. Lucie Estuary using phytoremediation technologies and to quantify the potential of this plant in improving stormwater quality in detention pond systems.

2 Materials and methods

2.1 Experimental design

Two detention ponds (East and West Ponds) in the St. Lucie Estuary watershed, each with a control and a treatment plot, were selected. Water lettuce (P. stratiotes) was grown in the treatment plots, while no plant was maintained in the control plots.

Water samples were collected weekly from the control and the treatment plots and analyzed for water quality parameters, including total N and P, NO3–N, NH4–N, ortho-P, pH, electrical conductivity (EC), suspended solids, and turbidity.

Water lettuce was sampled monthly from the treatment plots. After being rinsed thoroughly with D.I. water and blotted dry, root and shoot were separated and their fresh weights were recorded. Plant parts were oven dried at 70°C for 3 days and ground to <1 mm using a stainless ball mill prior to analysis for total N and P.

Besides monthly sampling, plants were also periodically harvested to maintain an optimum plant density. For each harvest, the total fresh weight of the lettuce plant was recorded, plant moisture was determined, and total quantity of dry plant biomass yield was calculated for each plot. Harvested plant materials were applied to the field as organic amendments. Total amounts of N and P removed from the water by the harvested plant were quantified by multiplying the amounts of plant biomass by the concentrations of N and P in the plant.

2.2 Chemical analysis

Prior to filtration, pH and EC of the water samples were determined using a pH/ion/conductivity meter (pH/Conductivity Meter, Model 220, Denver Instrument, Denver, CO, USA) following EPA method 150.1 and 120.1, respectively. Turbidity of water samples was measured using a turbidity meter (DRT-100B, HF Scientific Inc., Fort Myers, FL, USA). Total P in the unfiltered water sample was determined by the molybdenum-blue method after digestion with acidified ammonium persulfate (EPA method 365.1). Sub-samples of the water were filtered through Whatman 42 filter paper. Portions of the sub-samples were filtered further through a 0.45-μm membrane for measuring total dissolved P and PO4–P. Concentrations of NO3–N and PO4–P were measured within 24 h after sample collection using an ion chromatograph (DX 500; Dionex Corporation, Sunnyvale, CA, USA) following EPA method 300. The concentrations of NH4–N and total Kjeldahl N (TKN) in the water sample were measured using a discrete autoanalyzer (EasyChem, Systea Scientific Inc., Italy) following EPA method 353.2. Total N in the water samples was calculated as the sum of TKN and NO3–N. Total dissolved P in water was determined using inductively coupled plasma atomic emission spectrometry (ICP-AES, Ultima, JY Horiba Inc. Edison, NJ, USA) following EPA method 200.7.

Plant N content was determined using a CN analyzer (vario Max CN, Elemental Analysensystem GmbH, Hanau, Germany). Sub-samples (each 0.400 g) of plant material were digested with 5 ml of concentrated HNO3 in a digestion tube using a block digestion system (AIM 500-C, A.I. Scientific Inc., Australia), and P concentration in the digester was determined using the ICP-AES.

3 Results and discussion

3.1 General water quality improvement

Total suspended solids, turbidity, EC, and pH in the waters of both plots changed seasonally; increasing during the rainy season of May to November and decreasing during the dry season of November to May with values in the rainy season several times, or up to three units of pH higher than for those in the dry season (Figs. 1, 2, 3, and 4). The increase in these parameters during the rainy season was likely to be due to the large input of stormwaters, which bring soil particles and solutes, including nutrients. The growth of water lettuce improved water quality. Total suspended solids in the water column were decreased by an average of approximately 10% in the treatment plots as compared with those in the control plots (see Fig. 1 and Table 1) due to sedimentation in a more favorable environment provided by the plants (Brix 1997). On average, water turbidity was reduced by 65.5% and 63.3% in treatment plots as compared with the controls in East and West Ponds, respectively, in the period from August 2005 to August 2007 (see Fig. 2 and Table 1). Water lettuce growth decreased water EC in both ponds (see Fig. 3), due to salt removal from the waters by plant uptake or root adsorption. In a similar trend, water lettuce growth decreased water pH (see Fig. 4), which was not expected for it is well known that water pH rises with plant photosynthesis. One explanation is that nearly complete coverage of the water surface by the floating lettuce effectively blocked out sunlight for the growth of other plants (such as submerged plants and algae) which carry out photosynthesis in the water and contribute to the pH rise. On the contrary, some algae might grow in the control plot due to higher N and P concentrations and thus caused a pH increase. It is also well known that oxygen oversaturation happens concurrently with pH rise, but dissolved oxygen (DO) monitoring results did not show an oxygen oversaturation scenario in the water during the day with DO concentration <1.5 and 0.7 mg/l in East and West Ponds, respectively. Thus, pH decrease here was probably due to reduced or eliminated growth of algae or other submerged vegetation by the floating plants.

Fig. 1
figure 1

Suspended solids in control and treatment plots of East and West Ponds

Full size image
Fig. 2
figure 2

Water turbidity in control and treatment plots of East and West Ponds

Full size image
Fig. 3
figure 3

Water EC in control and treatment plots of East and West Ponds

Full size image
Fig. 4
figure 4

Water pH in control and treatment plots of East and West Ponds

Full size image
Table 1 Water quality improvement in treatment plots of East and West Ponds
Full size table

3.2 N and P concentration reduction

Changes of inorganic N (NH4–N plus NO3–N), total N, PO4 3−, and total P concentrations in water for the period from August 2005 to August 2007 are shown in Figs. 5, 6, 7, and 8. Like total suspended solids and water turbidity, nutrient concentrations in the waters showed seasonal changes during the year, which were affected by external input from stormwaters during the rainy season.

Fig. 5
figure 5

Inorganic N in the waters of control and treatment plots of East and West Ponds

Full size image
Fig. 6
figure 6

Total N in the waters of control and treatment plots of East and West Ponds

Full size image
Fig. 7
figure 7

Water PO4 3− in control and treatment plots of East and West Ponds

Full size image
Fig. 8
figure 8

Total P in the waters of control and treatment plots of East and West Ponds

Full size image

Although there are many reports showing that aquatic plants, such as Salvinia molesta and Elodea densa, preferred NH4–N to NO3–N (Reddy et al. 1987; Shimada et al. 1988) and theoretically NH4 + uptake is energetically more efficient than that of NO3 −, there were no differences in concentration reductions between NH4–N and NO3–N in both ponds with reduction rates of approximately 50–60% (see Table 1). Besides plant uptake, denitrification may also contribute to the decreased NO3–N concentration in the treatment plots as a more anaerobic condition (dissolved oxygen <1.5 and 0.7 mg/l in East and West Ponds, respectively) was created by the growing plants at the water’s surface and other anaerobic micro-sites (Gumbricht 1993; Reddy 1983).

Inorganic P (PO4 3−) removal (14% and 23% in East and West Ponds, respectively) was not as efficient as inorganic N (NH4–N+NO3–N) in both remediation systems (see Table 1), which was also the case in Sheffield’s research with a reduction rate of 40–55% in ortho-P compared to a reduction rate of 94% in inorganic N in a water hyacinth system (Sheffield 1967). Total P had a higher reduction than inorganic P (see Table 1), which indicates that the role aquatic plants play in such a remediation system is far more than uptake. Instead, nutrient uptake is only of quantitative importance in low-loaded systems (surface flow systems). More importantly, the aquatic plants play a crucial role by creating a favorable environment for a variety of complex chemical, biological, and physical processes that contribute to the removal and degradation of nutrients, which was thought by Brix (1997) to be the most important functions of aquatic plants. A higher removal rate in total P than in dissolved total P can come from the additional sedimentation effect on particulate P.

3.3 Plant N and P removal potential

Total N and P concentrations in the plant were approximately 17 and 3 g kg−1, respectively, with minimal differences between root and shoot (Figs. 9 and 10). Nitrogen and P content typically average 15–40 g N and 4–10 g P kg−1 for such large-leaved floating plants as water lettuce and water hyacinth (E. crassipes) (Aoi and Hayashi 1996).

Fig. 9
figure 9

Nitrogen contents in plant roots and shoots from East and West Ponds

Full size image
Fig. 10
figure 10

Phosphorus contents in plant roots and shoots from East and West Ponds

Full size image

Annual removal of N and P by plants were 190 and 24.6 kg ha−1, respectively, in East Pond and 329 and 34.1 kg ha−1, respectively, in West Pond, with dry matter being approximately 9 Mg ha−1 (East Pond) and 15 Mg ha−1 (West Pond). Much research has been performed on another invasive, large-leaf floating aquatic plant, water hyacinths (E. crassipes). Very high uptake rates have been reported in this research, for instance, 1,980 kg N and 322 kg P ha−1 year−1 by Boyd (1970), 2,500 kg N and 700 kg P ha−1 year−1 by Rogers and Davis (1972), and up to 5,350 kg N ha−1 year−1 and 1,260 kg P ha−1 year−1 by Reddy and Tucker (1983). The reasons behind this big difference in nutrient uptake rate between our research and the above research can be: (1) it is known that water hyacinth has a higher nutrient uptake and biomass yield potential than water lettuce; (2) this research was done using nutrient medium whose nutrient contents were much higher than that in the stormwater retention ponds; and (3) these high reported values were based on short-term experiments and extrapolated to 1 year, which often overestimates the nutrient uptake rate of the plant. On the other hand, the low nutrient uptake values from this research also indicated that the water lettuce was far from reaching their maximum nutrient uptake potential in these stormwater retention ponds.

3.4 Physiological limits

Plant growth is influenced by many environmental factors such as solar radiation, rainfall, and temperature, so is nutrient removal efficiency, as reflected in both nutrient concentrations in the plant (see Figs. 9 and 10) and plant yield of water lettuce (data not shown), showed strong seasonal dependence. This seasonal variability in plant growth and nutrient removal capacity was also discussed by Reddy and Sutton (1984).

West Pond worked better than East Pond in removing total N and total P from the waters (see Table 1), which could be related to the differences in total organic carbon (NPOC, averages of 30 and 12 mg l−1 in East and West Ponds, respectively) and EC of waters (180–2,000 and 100–400 μS cm−1 in East and West Ponds, respectively; see Fig. 3) between these two ponds. It was reported that an EC of 2,683 μS cm−1 was toxic to water lettuce (Haller et al. 1974). High EC in East Pond negatively affected water lettuce’s growth, leading to low efficiency in nutrient removal from the water.

Water lettuce is an invasive species, which means that it grows very well in nutrient-rich waters, but may not work well to our purpose in low nutrient waters.

4 Conclusions

Phytoremediation can be an important approach for cleaning eutrophicated stormwaters from agriculture and urban areas via man-made wetlands such as STAs and water detention/retention systems. Water lettuce has a great potential for removing N and P, reducing water suspended solids and turbidity from stormwaters, and improving water quality.

5 Recommendations and perspectives

For efficient water purification, grown-up biomass of aquatic macrophytes must be removed from water bodies to keep an optimum plant density. If not harvested, the vast majority of the nutrients that have been incorporated into the plant tissue will be returned to the water by decomposition processes (Brix 1997). Harvested plant biomass can be used as soil amendment or processed into livestock feed.

As water lettuce is an invasive species, it is important that the plant be strictly confined in the remediation system so that we can make full use of its nutrient scavenging ability without bringing unnecessary damage to the ecosystem.

More studies on how a variety of aquatic plants perform in different waters (with different nutrient ranges, pH, EC, or OC) under different environments (temperature, solar radiation, etc.) are needed for applying the right plant to the right water to achieve a maximum purification of the water.