CAS Workshop on Ecosystem Succession Theory and Practice of Ecological Restoration

The use of submerged aquatic vegetation treatment wetlands for phosphorus removal from the everglades agricultural area runoff: an overview of research 

Gu Binhe

(Southern District Restoration Department, South Florida Water Management District, 3301 Gun Club Road, West Palm Beach, Florida 33406, USA) 

Abstract: The Everglades is an internationally recognized ecosystem that covers approximately two million acres in South Florida and is the largest subtropical wetland in the United States.  The biotic integrity of the Everglades ecosystem has been endangered by alterations of hydrologic and nutrient regimes due to urban and agricultural development.  Reduction of total phosphorus (TP) from the Everglades Agriculture Area (EAA) runoff is a prerequisite to restoring and protecting the remaining Everglades natural resources. The 1994 Everglades Forever Act requires that water released from the EAA into the Everglades Protection Area (EPA) meet a threshold discharge limit for TP.  Preliminary studies within the Everglades have suggested that the ultimate protective threshold TP could be a low as 10 ΅g/L.  The EFA and the 404 US Corps permit for the Everglades Construction Project mandate the South Florida Water Management District to evaluate a series of treatment technologies to achieve this goal.

Submerged Aquatic Vegetation and Limerock (SAV/LR) Treatment system is one of the green technologies evaluated by the South Florida Water Management District and Florida Department of Environmental Protection. Prior to this research, SAV has been used only for nutrient removal from wastewater systems at a relatively small scale. The SAV/Limerock (LR) technology uses indigenous submerged plants to remove P from the water column, along with a LR filter positioned at the end of the system. Removal of P is believed to be accomplished by plant uptake, as well as by adsorption to (or co-precipitation with) calcium carbonate (CaCO₃) that precipitates from the water column due to photosynthesis-related pH elevations. The LR further removes a small amount of particulate P (PP) and dissolved organic P (DOP). The SAV research programs consisted of a number of research projects conducted on various platforms ranging from small mesocosms to full-scale constructed wetlands. Additionally, data and samples from natural SAV ecosystems were analyzed in order to obtain information on the long-term P removal performance.

Field experiments were conducted in the Everglades Nutrient Removal (ENR) Project site which is a 15.45 km² treatment wetland built by the District on former agricultural land which is located in Palm Beach County, Florida, USA (26°38’N and 80°25’W). The ENR research site is composed of four large treatment cells with surface area between 1.47 and 5.79 km² per cell and two banks of 20 small test cells (2,000 m²/each). Mesocosm (tank) experiments were conducted at the north and south research sites which received flows with low and high TP concentrations, respectively. The ENR research site has been fully operated since August 1994.

Research on Mesocosms

Controlled experiments designed to evaluate the effects of water depth, hydraulic residence time, flow rate, pulse loading, harvesting and substrate type were conducted in the mesocosms (~4 m²). Prior to the experiments, SAV (Najas quadalupensis, Ceratophyllum demersum, Chara spp. and Potamogeton sp.) collected from the ENR project site was stocked into the mesocosms.

Results from HLR experiments indicated that SAV mesocosms operated for 3 years at HLRs of 11, 22 and 53 cm/day reduced mean TP inflows of 97 mg/L to 25, 31 and 51 mg/L, respectively. Results from water depth experiments showed that under static conditions, P removal performance by SAV was comparable over a water depth range of 0.4 to 1.2 m. SAV mesocosms subjected to a fluctuating depth regime exhibited a slight reduction in P removal performance. Pulse hydraulic loadings, where long periods of stagnation were interspersed with high flows, reduced overall P removal effectiveness of the SAV mesocosms relative to constant hydraulic loading rate (HLR) conditions. Detrimental effects of pulsing were most pronounced with mesocosms that received a high average HLR (i.e., 53 cm/day), and were minimal in SAV wetlands that received a lower HLR (11 cm/day).

Partial harvesting of SAV from mesocosms resulted in short-term (~7 weeks) impairment of treatment performance. Because of performance impairment and high costs, SAV harvest is unlikely to be performed in full-scale STAs.

Over two years, the standing crop of SAV in mesocosms cultured on muck was more robust than SAV cultured on more inert (limerock or sand) substrates. SAV cultured on muck provided comparable P removal performance to SAV cultured on limerock.

Research on Small Constructed Wetlands (Test Cells)

Phosphorus removal performance was also conducted in small-scale constructed wetlands (test cells) with a surface area of ~2,000 m² per cell. Similar to the mesocosm experiments, several SAV species were stocked into the test cells prior to the experiments. During the entire monitoring period, the two north test cells received inflow TP concentrations ranging from 25 to 189 mg/L with a mean of 74 mg/L. Both test cells reduced TP concentrations relative to inflow with mean outflow TP concentrations of 24 mg/L, providing 68% TP reduction. Total P reduction at the north test cells was also estimated at different hydraulic loading rates over time.

The inflow TP concentrations in the two south test cells ranged from 12 to 55 mg/L with a mean of 23 mg/L. The mean outflow TP concentrations in the two south test cells were 19 and 18 mg/L, respectively. This reduced TP removal (23% TP reduction) at the south test cells was also observed in the STA Optimization cattail dominated systems and was most likely due to the low concentration of soluble reactive P (SRP) in the south test cell inflow water.

Research on a Full Scale Constructed Wetland (Cell 4)

The four large treatment cells were constructed and arranged as two flow-ways with inflow water moving from Cell 1 to Cell 3 and Cell 2 to Cell 4. Treatment Cells 1 to 3 were dominated by cattails (Typha latifolia and T. domingensis). Treatment Cell 4 has been actively maintained as a SAV/periphyton system dominated by southern naiad (N. quadalupensis), with lesser quantities of coontail (C. demersum L.) and pondweed (Potamogeton sp.).

Water depth in these cells varied from 0.2 to 0.9 meter. Hydraulic flow rate averaged approximately 95 cubic feet per second (cfs). Nominal hydraulic retention time for Cell 4 ranged from 11 to 32 days with a medium value of 21 days. Hydraulic loading rate ranged from 12.0 to 16.5 cm/day with an average of 14.8 cm/day.  Inflow TP concentration ranged from 30 to 57 mg/L and mass loading rate for total P range from 1.48 to 3.69 g/m²/yr with an average of 2.48 g/m²/yr. From February 1995 through September 2001, Cell 4 reduced TP concentrations from an average of 52 to 22 mg L^-1, at a mass removal rate of 1.4 g P m^-2-yr^-1. Removal efficiency averaged 55%. Total P settling rate averaged 42 m/yr, compared to world average of 12 m/yr for treatment wetlands.

Sediment cores collected from the inflow and outflow region of Cell 4 were incubated under amended pH, calcium/alkalinity and anoxic conditions. Phosphorus fractionations revealed that much of the P sequestered in Cell 4 SAV communities was associated with a fairly stable, Ca-bound sediment fraction. Relative to inflow region sediments, Cell 4 outflow region sediments were extremely stable, and exhibited little P release in response to anoxia, desiccation and low pH conditions.

Long-Term P Removal in Natural SAV Systems

Water quality data from several Florida lakes and rivers dominated by SAV were analyzed for P removal rates using an input-output model. The overall conclusion of this analysis was that SAV-dominated lakes and rivers typically removed P from the water column and the likely long-term sink for this P was the newly accreted sediment.  These calculated long-term removal rates were higher than those for full-scale wetlands dominated by emergent vegetation. Mass removal rates estimated for SAV-dominated lakes and rivers overlapped those from the SAV-dominated constructed wetland and mesocosm studies, but on average were generally lower.  These removals were clearly influenced by inlet P loading rates as a function of both P inlet concentration and hydraulic loading rate. Based on this analysis, caution should be taken when extrapolating P removal results from relatively short-term, small-scale studies to the design of full-scale, long-term operating SAV-dominated wetlands.

Sedimentation Composition and P Accrual Rate in a SAV-dominated Lake

Two sediment cores from Lake Panasoffkee, a large and shallow SAV-dominated lake located at southwest Florida, were analyzed for nutrient contents, P sedimentation rates, plant fragments, pollen, ²¹⁰Pb and stable carbon isotopes. The purposes of this study were to understand major organic sources and P accrual rates in Lake’s sediment. Results indicated that SAV was not the primary source of organic matter in sediment. Pollen analysis showed that SAV had become the dominated plant community in the lake since the turn of last century. Phosphorus removed from the water column was likely associated with calcium carbonate in the lake’s sediment.

Water Quality Conditions in Shallow SAV Lakes

Florida has over 7,700 lakes that range in size from 40,000 m² to over 1,800 km² ha. Many of these are shallow lakes vegetated with SAV. The objectives of this analysis were to (1) reveal community structure (species composition and richness) and (2) assess relationships between SAV biodiversity, biomass and environmental conditions in Florida SAV dominated lakes. We examined the physical, chemical and biological conditions of selected SAV lakes using existing data. The shallow nature of these lakes with high water temperature and long growth period were among the several key variables that favor SAV growth in this subtropical region. Our analysis revealed that eight genera with approximately 15 species of SAV inhabited these shallow lakes, which range in size from less than 20,000 m² ha to 23 km² ha. The SAV community within a lake was generally occupied predominantly by a single or a few species. Utricularia and Hydrilla were the most common SAV found in the study lakes. Ceratophyllum, Najas and Vallisneria often dominated lakes with high nutrient concentrations (mean TP = 34 to 53 mg/L) while Chara, Utricularia, Potamogeton and Myriophyllum preferred to inhabit lakes with relatively low nutrient concentrations (mean TP = 8 to 13 mg/L). Many SAV species grew well in a wide range of water quality conditions although biomass tended to increase with increasing lake size, pH, alkalinity and calcium concentration.

Key words: constructed wetlands; Everglades; phosphorus; stormwater management areas; submerged aquatic vegetation; treatment systems

Affiliation Links
©1988-2002, All rights reserved by Sino-Eco.  Last modified at 11/20/02  Web Statistics