Seasonal and storm event nutrient removal by a created wetland in an agricultural watershed

This study examines the effectiveness of a 1.2-ha created/restored emergent marsh at reducing nutrients from a 17.0 ha agricultural and forested watershed in the Ohio River Basin in west central Ohio, USA, during base flow and storm flow conditions. The primary source of water to the wetland was surface inflow, estimated in water year 2000 (October 1999–September 2000) to be 646 cm/year. The wetland also received a significant amount of groundwater discharge at multiple locations within the site that was almost the same in quantity as the surface flow. The surface inflow had 2-year averages concentrations of 0.79, 0.033, and 0.16 mg L⁻¹ for nitrate + nitrite (as N), soluble reactive phosphorus (SRP), and total phosphorus (TP), respectively. Concentrations of nitrate–nitrite, SRP, and TP were 40, 56, and 59% lower, respectively, at the outflow than at the inflow to the wetland over the 2 years of the study. Concentrations of SRP and TP exported from the wetland increased significantly (α = 0.05) during precipitation events in 2000 compared to dry weather flows, but concentrations of nitrate–nitrite did not increase significantly. During these precipitation events the wetland retained 41% of the nitrate–nitrite, 74% of the SRP, and 28% of the TP (by mass). The wetland received an average of 50 g N m⁻² per year of nitrate–nitrite and 7.1 g m⁻² per year of TP in 2000. Retention rates for the wetland were 39 g N m⁻² per year of nitrates and 6.2 g P m⁻² per year. These are close to rates suggested in the literature for sustainable non-point source retention by wetlands. The design of this wetland appears to be suitable as it retained a significant portion of the influent nutrient load and did not lose much of its retention capacity during heavy precipitation events. Some suggestions are given for further design improvements.     

Mass balance on water column trace metals in a free-surface-flow-constructed wetlands in Sacramento, California

A mass balance has been performed on trace metals concentrations and hydrology observed between 1994 and 1996 at the Sacramento Demonstration Constructed Wetlands using a first-order areal plug flow model. Water losses to infiltration and evapotranspiration from a typical cell are estimated to average 35 and 7% of influent flow, respectively. The wetlands effluent metals concentrations consistently meet proposed discharge criteria. Annual total mass loadings for all trace metals average 14.0 kg ha⁻¹ yr⁻¹, 88% of which consists of zinc, copper, and nickel. Effluent metals leaving the wetland average 3.1 kg ha⁻¹ yr⁻¹, 79% of which consists of the same three metals. Annual vegetation harvest events do not appear to account for more than 5% of annual trace metals mass removal, although harvest does appear to represent a significant loss pathway for some metals like mercury, lead, nickel, and chromium. Metals mass removals resulting from first-order removal interactions within the wetland range from 27 to 81%, with the exception of arsenic and nickel which display poor mass removals in part due to their high dissolved concentrations. An average of 7.6 kg ha⁻¹ yr⁻¹, or 54% of influent metals loadings, is sequestered within the internal wetland compartments.     

A simulation model for nitrogen retention in a papyrus wetland near Lake Victoria,Uganda (East Africa)

Papyrus wetlands around Lake Victoria, East Africa play an important role in the nutrient flows from the catchment to the lake. A dynamic model for nitrogen cycling was constructed to understand the processes contributing to nitrogen retention in the wetland and to evaluate the effects of papyrus harvesting on the nitrogen absorption capacity of the wetlands. The model had four layers: papyrus mat, water, sludge and sediment. Papyrus growth was modelled as the difference between nitrogen uptake and loss. Nitrogen uptake was modelled with a logistic equation combined with a Monod-type nitrogen limitation. Nitrogen compartments were papyrus plants, organic material in the floating mat; and total ammonia, nitrate and organic nitrogen in the water, sludge and sediment. Apart from the uptake and decay rates of the papyrus, the model included sloughing and settling of mat material into the water, mineralization of organic matter, and nitrification and diffusion of dissolved inorganic nitrogen. Literature data and field measurements were used for parameterization. The model was calibrated with data from Kirinya wetland in Jinja, Uganda which receives effluent from a municipal wastewater treatment plant. The model simulated realistic concentrations of dissolved nitrogen with a stable biomass density of papyrus and predicted accumulation of organic sludge in the wetland. Assuming that this sludge is not washed out of the wetland, the overall nitrogen retention of the wetland over a three-year period was 21.5 g N m⁻² year⁻¹ or about 25% of input. Harvesting 10, 20 and 30% of the papyrus biomass per year increased nitrogen retention capacity of the wetland to 32.3, 36.8 and 38.1 g m⁻² year⁻¹, respectively. Although the nutrient flows estimated by the model are within the ranges found in other papyrus wetlands, the model could be improved with regard to the dynamics of detrital nitrogen. Actual net retention of nitrogen in the sludge is likely to be lower than 21.5 g N m⁻² year⁻¹ because of flushing out of the sludge to the lake during the rainy season.     

A comparison of soil carbon pools and profiles in wetlands in Costa Rica and Ohio

Wetlands are large carbon pools and play important roles in global carbon cycles as natural carbon sinks. This study analyzes the variation of total soil carbon with depth in two temperate (Ohio) and three tropical (humid and dry) wetlands in Costa Rica and compares their total soil C pool to determine C accumulation in wetland soils. The temperate wetlands had significantly greater (P < 0.01) C pools (17.6 kg C m⁻²) than did the wetlands located in tropical climates (9.7 kg C m⁻²) in the top 24 cm of soil. Carbon profiles showed a rapid decrease of concentrations with soil depth in the tropical sites, whereas in the temperate wetlands they tended to increase with depth, up to a maximum at 18–24 cm, after which they started decreasing. The two wetlands in Ohio had about ten times the mean total C concentration of adjacent upland soils (e.g., 161 g C kg⁻¹ were measured in a central Ohio isolated forested wetland, and 17 g C kg⁻¹ in an adjacent upland site), and their soil C pools were significantly higher (P < 0.01). Among the five wetland study sites, three main wetland types were identified – isolated forested, riverine flow-through, and slow-flow slough. In the top 24 cm of soil, isolated forested wetlands had the greatest pool (10.8 kg C m⁻²), significantly higher (P < 0.05) than the other two types (7.9 kg C m⁻² in the riverine flow-though wetlands and 8.0 kg C m⁻² in a slowly flowing slough), indicating that the type of organic matter entering into the system and the type of wetland may be key factors in defining its soil C pool. A riverine flow-through wetland in Ohio showed a significantly higher C pool (P < 0.05) in the permanently flooded location (18.5 kg C m⁻²) than in the edge location with fluctuating hydrology, where the soil is intermittently flooded (14.6 kg C m⁻²).     

Evaluation of nutrient retention in four restored Danish riparian wetlands

During the last 15–20 years, re-establishment of freshwater riparian wetlands and remeandering of streams and rivers have been used as a tool to mitigate nutrient load in downstream recipients in Denmark. The results obtained on monitoring four different streams and wetland restoration projects are compared with respect to hydrology, i.e. flow pattern and discharge of ground or surface water, retention of phosphorus (P), and removal of nitrogen (N). Furthermore, the monitoring strategies applied for quantifying the post-restoration nutrient retention are evaluated. The four wetland restoration projects are the Brede River restoration (including river valley groundwater flow, remeandering and inundation), Lyngbygaards River restoration (groundwater flow, irrigation with drainage water, inundation with river water and remeandering), Egeskov fen (fen re-establishment and stream remeandering) and Egebjerg Meadows (fen restoration and hydrological reconnection to Store Hansted River). Retention of phosphorus varied between 0.13 and 10 kg P ha⁻¹ year⁻¹, while the removal of nitrogen varied between 52 and 337 kg N ha⁻¹ year⁻¹. The monitoring strategy chosen was not optimal at all sites and would have benefitted from a knowledge on local hydrology and water balances in the area to be restored before planning for the final monitoring design. Furthermore, the outcome concerning P retention would have benefitted from a more frequent sampling strategy.     

The role of sediments for phosphorus retention in the Kirinya wetland (Uganda)

Phosphorus dynamics were examined and modelled in a Cyperus papyrus and Phragmites mauritanus wetland on the Ugandan coast of Lake Victoria receiving secondary treated wastewater. Using a series of transversal transects, concentrations of nitrogen (N) and phosphorus (P) were found to decrease gradually as water moved downstream, giving nutrient retention capacities which ranged between 40% and 60%. Near-zero oxygen and nitrate concentrations were observed as well. To investigate the phosphorus retention characteristics in more detail, laboratory experiments were carried out on sediment samples and sediment cores retrieved from points along the wetland. Following a P shock load to cores of the wetland sediment, it was possible to determine a sediment P uptake rate of 0.016 day⁻¹. Sediment P adsorption studies were also performed, showing significant Freundlich and Langmuir isotherm behaviour. With these data a maximum P adsorption capacity of 4 mg P/g for the wetland sediment could be estimated. A plug-flow model was used to evaluate the phosphorus retention dynamics of the Kirinya wetland. A good correspondence between the actual and simulated P retention was observed. Comparing the daily P uptake (g/m³day) in the Kirinya wetland with the maximum sediment P uptake capacity, it can be concluded that the total P retention capacity of the wetland will only be sufficient for 30 more years under the present P loading and wetland management.     

Flow velocity retardation and sediment retention in two constructed wetland–ponds

Hydraulic properties of two constructed wetland–ponds in agricultural watersheds in southern Finland were examined by simulations with two-dimensional hydrodynamic and water quality transport models. Hydraulic efficiency was determined for the existing and hypothetical layouts of both wetlands to find out the effects of different design options. Suspended sediment retention in the wetlands was simulated with a two-dimensional model for sediment transport. Hydraulic efficiency was found to be highly improved by baffles that direct the main flow to optimally exploit the wetland acreage. Also, an elongated shape of wetland appeared beneficial; hydraulic efficiency was high regardless of the size or position of the inlet. As for suspended sediment retention, the wetland area in relation to its watershed proved to be an essential factor. According to water quality observations, the wetland which occupied 5% of its watershed area was capable of reducing the inflow of total suspended solids (TSS) concentrations during flood events by 43–72%, whereas the reduction varied between −7 and 5% in the other wetland with a corresponding ratio of 0.5%. Model simulations of the same flood events plausibly predicted the output TSS concentrations, even though with wider ranges of the reductions (26 to 74 and −13 to 43%, respectively).     

Aquatic metabolism as an indicator of the ecological effects of hydrologic pulsing in flow-through wetlands

Water column metabolism is a major component in the functioning of wetland ecosystems and can be used as an indicator of ecosystem health. The effect of hydrologic pulsing on water column metabolism was studied with 2 year's field data and a validated model for two 1-ha created riparian wetlands in Columbus, OH, USA. Aquatic gross primary productivity (GPP) was measured during hydrologic pulses the first week of April, May and June of 2004 and compared to GPP during steady flow-conditions in April, May and June 2005. Pulses reduced diurnal variation of water temperature, pH and dissolved oxygen, and negatively affected GPP rates. Mean GPP measured during hydrologic flood pulses was 5.4 ± 2.6 kcal m⁻² day⁻¹, significantly lower than that measured for comparable months with steady-flow hydrology (10.8 ± 3.3 kcal m⁻² day⁻¹). Solar-normalized productivity values of 0.08 ± 0.01% of solar energy during pulses and 0.2 ± 0.02% for steady-flow conditions were also significantly different. Different hyperbolic curves of optimum productivity with water temperature were seen for pulsing and steady-flow conditions. A simulation model with hydrology, metabolism, and dissolved oxygen sub models was calibrated with 2005 steady-flow year data and validated with 2004 pulse year data. Results from both the field study and model simulations suggest that there was a threshold hydraulic inflow rate between 30 and 50 cm day⁻¹ where aquatic metabolism became negatively affected by flow.     

The impacts of re-introducing Mississippi River water on the hydrologic budget and nutrient inputs of a deltaic estuary

Most wetlands of the Mississippi deltaic plain are isolated from riverine input due to flood control levees along the Mississippi River. These levees have altered hydrology and ecology and are a primary cause of massive wetland loss in the delta. River water is being re-introduced into coastal basins as part of a large-scale ecological engineering effort to restore the delta. We quantified freshwater, nitrogen, and phosphorus inputs to the Breton Sound Estuary for three climatically different years (2000, 2001, and 2002). Water budgets included precipitation, potential evapotranspiration, the diversion, stormwater pumps, and groundwater. Precipitation contributed 48–57% of freshwater input, while the diversion accounted for 33–48%. Net groundwater input accounted for less than 0.05% of freshwater inputs. Inputs of ammonium (NH₄-N), nitrate (NO₃-N), total nitrogen (TN), and total phosphorus (TP) were determined for each of the water sources. Atmospheric deposition was the most important input of NH₄-N (57–62% or 1.44 × 10⁵–2.32 × 10⁵ kg yr⁻¹) followed by the diversion. The diversion was the greatest source of NO₃-N (67–83%, 7.78 × 10⁵–1.64 × 10⁶ kg yr⁻¹) and TN (60–71%). The diversion contributed 41–60% of TP input (1.17 × 10⁵–2.32 × 10⁵ kg yr⁻¹). Annual loading rates of NH₄-N and NO₃-N were 0.17–0.27 and 1.2–2.3 g N m⁻² yr⁻¹, respectively, for the total basin indicating strong retention of nitrogen in the basin. Nitrogen retention through denitrification and burial was estimated for the upper basin.     

Macrophyte productivity and community development in created freshwater wetlands under experimental hydrological conditions

Macrophyte biomass production and species richness were monitored from 1988 through 1991 in four freshwater wetlands constructed on the floodpain of the Des Plaines River, Lake County, Illinois, USA. The wetlands were constructed in 1988 and pumping of river water began in 1989 under two differentd hydrologic regimes: two wetlands received high water inflow (equivalent to 40 cm wk⁻¹ of water depth) and two received low flow (11 cm wk⁻¹). Biomass production showed no relationship to the hydrologic inflows after two years of experimentation, with both the highest and lowest production occuring in low flow wetlands. Rates of primary production increased between 1990 and 1991 under low flow conditions and decreased under high flow conditions, primarily as a result of the initial composition of the plant community. The change from dry conditions in 1988 to flooded conditions in 1989 altered the species composition in each wetland to include almost 100% wetland-adapted species. Similarity in species composition among the four wetlands diverged from 1988 to 1989 as the plant community adjusted to flooded conditions and then converged in both 1990 and 1991 as the wetlands developed.     

Integrated stream and wetland restoration: A watershed approach to improved water quality on the landscape

Water quality in Upper Sandy Creek, a headwater stream for the Cape Fear River in the North Carolina Piedmont, is impaired due to high N and P concentrations, sediment load, and coliform bacteria. The creek and floodplain ecosystem had become dysfunctional due to the effects of altered storm water delivery following urban watershed development where the impervious surface reached nearly 30% in some sub-watersheds. At Duke University, an 8-ha Stream and Wetland Assessment Management Park (SWAMP) was created in the lower portion of the watershed to assess the cumulative effect of restoring multiple portions of stream and former adjacent wetlands, with specific goals of quantifying water quality improvements. To accomplish these goals, a three-phase stream/riparian floodplain restoration (600 m), storm water reservoir/wetland complex (1.6 ha) along with a surface flow treatment wetland (0.5 ha) was ecologically designed to increase the stream wetland connection, and restore groundwater wetland hydrology. The multi-phased restoration of Sandy Creek and adjacent wetlands resulted in functioning riparian hydrology, which reduced downstream water pulses, nutrients, coliform bacteria, sediment, and stream erosion. Storm water event nutrient budgets indicated a substantial attenuation of N and P within the SWAMP project. Most notably, (NO₂⁻ + NO₃⁻)-N loads were reduced by 64% and P loads were reduced by 28%. Sediment retention in the stormwater reservoir and riparian wetlands showed accretion rates of 1.8 cm year⁻¹ and 1.1 cm year⁻¹, respectively. Sediment retention totaled nearly 500 MT year⁻¹.     

Hydrological and nutrient budgets of freshwater and estuarine wetlands of Taylor Slough in Southern Everglades, Florida (U.S.A.)

Hydrological restoration of the Southern Everglades will result in increased freshwater flow to the freshwater and estuarine wetlands bordering Florida Bay. We evaluated the contribution of surface freshwater runoff versus atmospheric deposition and ground water on the water and nutrient budgets of these wetlands. These estimates were used to assess the importance of hydrologic inputs and losses relative to sediment burial, denitrification, and nitrogen fixation. We calculated seasonal inputs and outputs of water, total phosphorus (TP) and total nitrogen (TN) from surface water, precipitation, and evapotranspiration in the Taylor Slough/C-111 basin wetlands for 1.5 years. Atmospheric deposition was the dominant source of water and TP for these oligotrophic, phosphorus-limited wetlands. Surface water was the major TN source of during the wet season, but on an annual basis was equal to the atmospheric TN deposition. We calculated a net annual import of 31.4 mg m^–2 yr^–1 P and 694 mg m^–2 yr^–1N into the wetland from hydrologic sources. Hydrologic import of P was within range of estimates of sediment P burial (33–70 mg m^–2 yr^–1 P), while sediment burial of N (1890–4027 mg m^–2 yr^–1 N) greatly exceeded estimated hydrologic N import. High nitrogen fixation rates or an underestimation of groundwater N flux may explain the discrepancy between estimates of hydrologic N import and sediment N burial rates.     

Effects of vegetation and hydrologic load on sedimentation patterns in experimental wetland ecosystems

Vegetation and water velocity effects on patterns of sediment deposition were tested by monitoring sedimentation rates in dense cattail, open water, and transitional vegetation zones at distances of 5, 10, and 20 m from the inflows of two experimental wetland basins at the Des Plaines River Wetlands Demonstration Project, northeastern Illinois, USA. One basin received a high hydrologic load (up to 50 cm/wk) and one basin received a low load (up to 6 cm/wk). Sediment deposition rates within 20 m of the inflows reached 3300 g dry wt m⁻² day⁻¹ in the high-load basin and 700 g dry wt m⁻² day⁻¹ in the low-load basin. Vegetation patterns did not have a significant effect (P > 0.05) on sediment deposition rates in the high-load basin, whereas water velocity effects on rates of sedimentation were significant (P < 0.01) in three of four periods of monitoring. In the low-load basin, vegetation effects were significant (P < 0.01) during the entire period of investigation. Experimental research at this scale aids in the assessment of design criteria for constructed wetlands.     

Nutrient and metal removal in a constructed wetland for wastewater treatment from a metallurgic industry

This contribution summarizes the nutrient and metal removal of a free water surface constructed wetland, compares it with the previous small-scale prototype and discusses the observed differences. Several locally available macrophyte species were transplanted into the wetland. Eichhornia crassipes (water hyacinth) showed a fast growth and it soon became dominant, attaining 80% cover of the wetland surface. Typha domingensis (cattail) and Panicum elephantipes (elephant panicgrass) developed as accompanying species attaining 14 and 4% cover. The wetland removed 86% of Cr and 67% of Ni. Zn concentrations were below 50 μg l⁻¹ in most samplings. The FeS precipitation probably caused the high retention of Fe (95%). The outcoming water was anoxic in most samplings. Phosphate and ammonium were not retained within the wetland while 70% and 60% of the incoming nitrate and nitrite were removed. Large denitrification losses are suggested. Cr, Ni and Zn were retained by the macrophytes in the larger wetland and in sediment in the small-scale one. Differences in the retention mechanism of the two wetlands are discussed.     

Dynamics of nitrogen and phosphorus retention during wetland ecosystem succession

We compared the mechanisms of nitrogen (N) and phosphorus (P) removal in four young (<15 years old) constructed estuarine marshes with paired mature natural marshes to determine how nutrient retention changes during wetland ecosystem succession. In constructed wetlands, N retention begins as soon as emergent vegetation becomes established and soil organic matter starts to accumulate, which is usually within the first 1–3 years. Accumulation of organic carbon in the soil sets the stage for denitrification which, after 5–10 years, removes approximately the same amount of N as accumulating organic matter, 5–10 g/m²/yr each, under conditions of low N loadings. Under high N loadings, the amount of N stored in accumulating organic matter doubles while N removal from denitrification may increase by an order of magnitude or more. Both organic N accumulation and denitrification provide for long-term reliable N removal regardless of N loading rates. Phosphorus removal, on the other hand, is greatest during the first 1–3 years of succession when sediment deposition and sorption/precipitation of P are greatest. During this time, constructed marshes may retain from 3 g P/m²/yr under low P loadings to as much as 30 g P/m²/yr under high loadings. However, as sedimentation decreases and sorption sites become saturated, P retention decreases to levels supported by organic P accumulation (1–2 g P/m²/yr) and sorption/precipitation with incoming aqueous and particulate Fe, Al and Ca. Phosphorus cycling in wetlands differs from forest and other terrestrial ecosystems in that conservation of P is greatest during the early years of succession, not during the middle or late stages. Conservation of P by wetlands is largely regulated by geochemical processes (sorption, precipitation) which operate independently of succession. In contrast, the conservation of N is controlled by biological processes (organic matter accumulation, denitrification) that change as succession proceeds.     

Nitrogen and phosphorus removal by wetland mesocosms subjected to different hydroperiods

The effect of hydroperiod on nutrient removal efficiency from simulated wastewater was investigated in replicate wetland mesocosms (area, 2 m², planted with Scirpus californicus). Alternate draining and flooding of sediments (pulsed discharge) increased nutrient removal efficiency compared to the continuous-flow “control”. Average PO₄³⁻ removal efficiency was 20–30% higher in wetland mesocosms that drained twice daily compared to the control. Inorganic N removal efficiency was less affected than phosphate removal by hydroperiod variation. At the higher NH₄⁺ loading rate (1.83 g N m⁻² day⁻¹), inorganic N removal efficiency was consistently 5–20% higher in pulsed-discharge wetland mesocosms than in the control. At the lower NH₄⁺ loading rate (0.9 g N m ⁻² day ⁻¹), pulsed-discharge hydrology had no effect on inorganic N removal efficiency. Twice-daily drainage exhibited average inorganic N removal efficiencies of 96% (lower N loading rate) and 87% (higher N loading) and average phosphate removal efficiencies of 81% (lower P loading) and 90% (higher P loading). Mass balance data from the continuous-flow treatment revealed that the aquatic macrophyte Scirpus californicus was the most important nutrient sink, assimilating 50% of the NH₄⁺ and PO₄³⁻ supply. The high plant productivity in the mesocosms (15.6 kg m⁻² year⁻¹) occurred under conditions of high light (high edge per mesocosm area) and high root contact with nutrient-rich influent (shallow, sandy substrate) and may overestimate plant uptake in larger wetlands. The addition of a nitrification-inhibitor (N-Serve) indicated that 34% of the NH₄⁺ supply was transformed to NO₃⁻ by nitrifying bacteria.     

Nitrate-nitrogen retention in wetlands in the Mississippi River Basin

Nitrate-nitrogen retention as a result of river water diversions is compared in experimental wetland basins in Ohio for 18 wetland-years (9 years × 2 wetland basins) and a large wetland complex in Louisiana (1 wetland basin × 4 years). The Ohio wetlands had an average nitrate-nitrogen retention of 39 g-N m⁻² year⁻¹, while the Louisiana wetland had a slightly higher retention of 46 g-N m⁻² year⁻¹ for a similar loading rate area. When annual nitrate retention data from these sites are combined with 26 additional wetland-years of data from other wetland sites in the Basin Mississippi River (Ohio, Illinois, and Louisiana), a robust regression model of nitrate retention versus nitrate loading is developed. The model provides an estimate of 22,000 km² of wetland creation and restoration needed in the Mississippi River Basin to remove 40% of the nitrogen estimated to discharge into the Gulf of Mexico from the river basin. This estimated wetland restoration is 65 times the published net gain of wetlands in the entire USA over the past 10 years as enforced by the Clean Water Act and is four times the cumulative total of the USDA Wetland Reserve Program wetland protection and restoration activity for the entire USA.     

Predictive models for phosphorus retention in wetlands

The potential of wetlands to efficiently remove (i.e., act as a nutrient sink) or to transform nutrients like phosphorus under high nutrient loading has resulted in their consideration as a cost-effective means of treating wastewater on the landscape. Few predictive models exist which can accurately assess P retention capacity. An analysis of the north American data base (NADB) allowed us to develop a mass loading model that can be used to predict P storage and effluent concentrations from wetlands. Phosphorus storage in wetlands is proportional to P loadings but the output total phosphorus (TP) concentrations increase exponentially after a P loading threshold is reached. The threshold P assimilative capacity based on the NADB and a test site in the Everglades is approximately 1 g m^–2 yr^–1. We hypothesize that once loadings exceed 1 g m^–2 yr^–1 and short-term mechanisms are saturated, that the mechanisms controlling the uptake and storage of P in wetlands are exceeded and effluent concentrations of TP rise exponentially. We propose a One Gram Rule for freshwater wetlands and contend that this loading is near the assimilative capacity of wetlands. Our analysis further suggests that P loadings must be reduced to 1 g m^–2 yr^–1 or lower within the wetland if maintaining long-term low P output concentrations from the wetlands is the central goal. A carbon based phosphorus retention model developed for peatlands and tested in the Everglades of Florida provided further evidence of the proposed One Gram Rule for wetlands. This model is based on data from the Everglades areas impacted by agricultural runoff during the past 30 years. Preliminary estimates indicate that these wetlands store P primarily as humic organic-P, insoluble P, and Ca bound P at 0.44 g m^–2 yr^–1 on average. Areas loaded with 4.0 g m^–2 yr^–1 (at water concentrations>150 g·L^–1 TP) stored 0.8 to 0.6 g m^–2 yr^–1 P, areas loaded with 3.3 g m^–2 yr^–1 P retained 0.6 to 0.4 g m^–2 yr^–1 P, and areas receiving 0.6 g m^–2 yr^–1 P retained 0.3 to 0.2 g m^–2 yr^–1. The TP water concentrations in the wetland did not drop below 50 g·L^–1 until loadings were below 1 g m² yr^–1 P.     

Eight years of internal phosphorus loading and changes in the sediment phosphorus profile of Lake Søbygaard,Denmark

During each of the first 8 years following an 80–90% reduction in external phosphorus loading of shallow, hypertrophic Lake Søbygaard, Denmark in 1982, phosphorus retention was found to be negative. Phosphorus release mainly occurred from April to October, net retention being close to zero during winter. Net internal phosphorus loading was 8 g P m^–2 y^–1 in 1983 and slowly decreased to 2 g P m^–2 y^–1 in 1990, mainly because of decreasing sediment phosphorus release during late summer and autumn. The high net release of phosphorus from Lake Søbygaard sediment is attributable to a very high phosphorus concentration and to a high transport rate in the sediment caused by bioturbation and gas ebullition. Sediment phosphorus concentration mainly decreased at a depth of 5 to 20 cm, involving sediment layers down to 23 cm. Maximum sediment phosphorus concentration, which was 11.3 mg P g^–1 dw at a depth of 14–16 cm in 1985, decreased to 8.6 mg P g^–1 dw at a depth of 16–18 cm in 1991. Phosphorus fractionation revealed that phosphorus release was accompanied by a decrease in NH₄Cl-P + NaOH-P and organic phosphorus fractions. HCl-P increased at all sediment depths. The Fe:P ratio in the superficial layer stabilized at approximately 10. Net phosphorus release can be expected to continue for another decade at the present release rate, before an Fe:P ratio of 10 will be reached in the sediment layers from which phosphorus is now being released.     

A hydrologic assessment of the Everglades Nutrient Removal Project, a subtropical constructed wetland in South Florida (USA)

The Everglades Nutrient Removal Project (ENRP), a 1544-ha constructed wetland in south Florida, was intensively monitored throughout its five-year operational history. Water budgets for the ENRP and each of its interior treatment cells were dominated by surface flows (≥85% of inflows; ≥68% of outflows) with smaller contributions from precipitation, evapotranspiration, groundwater flux, and change in storage. The mean water depth, hydraulic loading rate for surface water, and nominal hydraulic retention time for the entire wetland were 0.6 m, 3.1 cm d⁻¹ and 17.7 d, respectively, and were comparable to values anticipated in design. The east flow-way was slightly shallower (0.2 m) and received proportionately more flow (61%) than the west flow-way. The hydrology of other treatment wetlands is often driven by surface flows. All treatment cells in the ENRP were to some extent hydraulically short-circuited. There was net groundwater inflow to the ENRP from Water Conservation Area 1 (WCA-1) resulting from significant head differences between these wetlands. Groundwater outflow to the adjacent farmlands was greatest in Cell 2 and substantially exceeded groundwater inflow. All hydrologic parameters exhibited seasonality to some extent; fluctuation in water depth and groundwater inflows corresponded with the seasonal change in stage in WCA-1. Errors in the ENRP and individual cell water budgets were generally less than 10% and within the range of errors for water budgets from other wetlands.