JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Liikanen, A.
Right arrow Articles by Hartikainen, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liikanen, A.
Right arrow Articles by Hartikainen, H.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Liikanen, A.
Right arrow Articles by Hartikainen, H.
Related Collections
Right arrow Water Quality
Right arrow Wetlands and Aquatic Processes
Right arrow Phosphorus
Right arrow Nutrients
Right arrow Nutrient Cycling
Published in J. Environ. Qual. 33:1124-1132 (2004).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Wetlands and Aquatic Processes

Phosphorus Removal in a Wetland Constructed on Former Arable Land

Anu Liikanena, Markku Puustinen*,b, Jari Koskiahob, Tero Väisänenc, Pertti Martikainena and Helinä Hartikainend

a Department of Environmental Sciences, Research and Development Unit of Environmental Health, University of Kuopio, P.O. Box 1627, FIN 70211, Kuopio, Finland
b Finnish Environment Institute, P.O. Box 140, FIN 00251, Helsinki, Finland
c North Ostrobothnia Regional Environment Centre, P.O. Box 124, FIN 90101, Oulu, Finland
d University of Helsinki, Department of Applied Chemistry and Microbiology, FIN 00014, University of Helsinki, Finland

* Corresponding author (markku.puustinen{at}ymparisto.fi).

Received for publication April 15, 2003.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Phosphorus in surface runoff water may cause eutrophication of recipient water. This study clarifies the mechanisms of P removal in the wetland of Hovi, Finland, constructed on arable land in 1998. Before the construction, the surface soil (removed in the construction) and subsoil (the current wetland bottom) were analyzed for Al and Fe oxides (Alox and Feox) reactive in P sorption, and for the distribution of P between various pools as well as for P exchange properties. Retention of P from runoff water within the wetland was studied from 1999 to 2001 in situ and factors affecting the P removal (O2 availability and P concentration in water) were investigated in a laboratory microcosm. The processes taking place in the wetland diminished by 68% the total P load and by 49% the dissolved reactive P load. Desorption–sorption tests indicated that without removal of the surface soil, there would have been a risk of the wetland being a source of P, since the equilibrium P concentration of the soil removed was high compared with the mean P concentration of the inflowing water. The subsoil contained less P and high amounts of reactive oxides, which could bind P. Evidently, the P sorption by Alox played an important role in a first phase removal of P, since the wetland retained P efficiently even under anoxic conditions, where Fe tends to be reduced. Fine-textured, mineral soil on the bottom of the wetland (subsoil of the former arable land) seemed to be very efficient in retaining P from agricultural runoff.

Abbreviations: Alox, oxalate-soluble aluminum • CW, constructed wetland • DRP, dissolved reactive phosphorus • EPC0, equilibrium phosphorus concentration • Feox, oxalate soluble iron • TP, total phosphorus


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
CONSTRUCTION OF WETLANDS to decrease pollution of surface waters has become more common in recent years. Constructed wetlands (CWs) have been successfully used to retain nutrients from treated wastewaters (Gale et al., 1994; White et al., 2000), and recently to purify nonpoint source pollutants from urban and agricultural runoff (Nairn and Mitsch, 2000; Uusi-Kämppä et al., 2000; Woltemade, 2000). Runoff water from agricultural soil is the most important nonpoint source of nutrients to rivers and lakes provoking their eutrophication (DePinto et al., 1981; Ekholm, 1994). Thus, constructed wetlands may offer a promising tool for the treatment of agricultural runoff, which is difficult and expensive to treat with conventional wastewater purification techniques.

The main objective of the CWs established in agricultural areas is to reduce the load of suspended solids, P, and N. In particular, the removal of P is of importance, since in most cases, P is known to be the major limiting nutrient for algal growth in freshwater ecosystems (Wetzel, 2001). Wetlands remove P through biological, chemical, and physical processes. Sedimentation is an important mechanism for removal of particulate inorganic and organic P. Sediment burial is considered to be the major long-term P storage in wetlands (Howard-Williams, 1985; Reddy et al., 1999; White et al., 2000). Some wetlands store large amounts of organic matter and therefore sequester also a large quantity of P bound to organic molecules (Craft and Richardson, 1993). However, part of the sedimented organic P can be mineralized to dissolved inorganic P, which can subsequently be partially released back to water (D'Angelo and Reddy, 1994; McLatchey and Reddy, 1998). Dissolved inorganic P is either immobilized via assimilation by macrophytes (Brix, 1997; Tanner, 1996), algae (Wetzel, 1990), or microorganisms (Gächter and Meyer, 1993) or it can be retained by amorphous and crystalline Fe or Al oxides, or by Ca (Harter, 1969; Khalid et al., 1977; Richardson, 1985). Part of the P immobilized in biomass is remineralized back to a dissolved inorganic form. Thus, biomass only represents short-term P storage; the turnover time for P in aboveground parts of macrophytes varies from months to the annual growing season (Reddy et al., 1999). In contrast, P bound to inorganic sorption components may be a more long-term storage for P. Iron-bound P is sensitive to changes in the redox potential, and during anoxic conditions it can be released (Boström et al., 1988; Ann et al., 2000). However, part of the mobilized P can be resorbed by Al oxides (Koski-Vähälä and Hartikainen, 2001). Therefore, a large amount of sorption components in the wetland soil is essential to achieve a long-term and stable P retention. In fact, wetlands constructed on mineral soils may be more efficient in retaining P than natural wetlands, where a high organic matter content often diminishes the amount of active sorption components (Johnston, 1991; Gale et al., 1994).

In the construction of wetlands, it is important to create good physical, chemical, and biological environments for nutrient removal. Our aim was to study how former mineral, agricultural soil is functioning as a wetland bottom matrix to remove P from agricultural runoff water. When a former agricultural soil is converted to wetland there is a risk that soil is a source of nutrients rater than a sink (Pant et al., 2002a, 2002b). We will clarify how properties of soil matrix and environmental conditions, such as O2 availability and P content, in the wetland affect nutrient removal. This information is needed for the correct construction of the wetlands to reduce the nutrient load from agriculture.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The factors affecting the efficiency of a CW to retain P from agricultural runoff water were investigated using different approaches. Before the construction work, we investigated the P status and P exchange characteristics of topsoil and subsoil of the planned water basin. Subsequently, the function of the wetland was monitored in situ and factors affecting P removal were clarified in laboratory microcosm studies.

Site Description
The Hovi CW (Finland; 60°25'N, 24°22'E) of 0.6 ha was constructed on arable land in 1998. The 12-ha catchment of the CW consists of arable land (5% wetland to catchment ratio) and water entering the CW is high in nutrients (Table 1). The characteristics of the CW and its catchment and runoff water are presented in Table 1.


View this table:
[in this window]
[in a new window]
  		Table 1. Characteristics of the Hovi wetland and its catchment.

 
During the construction of the wetland, the topsoil (0–30 cm) was removed to avoid the release of P accumulated in the plow layer as a result of the previous fertilization. Additionally, a deeper region was excavated at the inlet of the CW to enhance sedimentation (Fig. 1) . Two spits of land and an islet were created from the removed soil to diversify the landscape, increase the hydraulic efficiency, and enrich the biodiversity by generating a wider littoral ecosystem. Since the wetland area remained bare after the construction, the development of vegetation was enhanced by planting and seeding (Table 1). However, within the first 3 yr, these activities yielded far less biomass than the spontaneous propagation of vegetation, mostly cattail (Typha latifolia L.). Cattail was the dominant species in the wetland; other species are presented in Table 1. The aboveground biomass in autumn 2001 was estimated to be 44 g m–2 (Koskiaho et al., 2003).



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 1. Schematic image of the Hovi wetland. Surface soil and subsoil samples were taken before construction of the wetland from idle meadow area (1 and 2) and actively cultivated area (3, 4, and 5). The term A represents deep water sediment samples, while B is littoral sediment samples.

 
Soil Physicochemical Analysis
The properties of the soil were studied before the construction of the wetland to evaluate the capacity of the soil to retain P. Samples from the surface soil to be removed during the construction and from the corresponding subsoil were taken from five different sites on the basin of the current wetland (Fig. 1). Two of the sampling sites were in the meadow (not tilled or fertilized with commercial fertilizers) and three in the actively cultivated area (fertilized annually with P). Soil samples were analyzed for hydrated oxides of Al and Fe (Alox and Feox) representing the main sorption components in Finnish soils (Hartikainen, 1979), P status, and P exchange characteristics.

Reactive oxides (Feox and Alox) were extracted in the dark with 0.05 M NH4–oxalate (pH 2.9) (1:20 w/v) and determined by atomic absorption spectrophotometry (AAS) (Niskanen, 1989). The soil P reserves were characterized by determining inorganic P fractions according to the Chang and Jackson (1957) sequential fractionation method modified by Hartikainen (1979): NH4Cl-extractable P was taken to represent an easily soluble fraction, NH4F-extractable P as a Alox–bound fraction, NaOH-extractable P as a Feox–bound fraction, and H2SO4–extractable P as an apatitic Ca-bound fraction. Degree of phosphorus saturation (DPS) of Alox and Feox oxides was calculated as 100 x the molar ratio of NH4F-P to Alox and NaOH-P to Feox, respectively (Hartikainen, 1982). The organic P in the soil samples was estimated by an indirect ignition method where the samples were ignited at 600°C for 1 h. Both ignited and corresponding unignited samples were extracted with 0.1 M H2SO4. The organic P content of the soil was calculated as the difference between P extracted from the unignited sample and P extracted from the ignited sample (Kuo, 1996). Phosphorus in ignited samples was taken to represent total P in soil excluding P in the microbial matrix, which is very difficult to dissolve.

The P exchange characteristics of the surface soil and subsoils were depicted using quantity/intensity (Q/I) plots prepared for samples taken from one point on cultivated area and from one point on meadow area. For the graphs, 1 g of air-dried or corresponding amount of moist soil (three replicates) was shaken for 1 h with 50 mL of a series of standard KH2PO4 solutions containing 0 to 3.0 mg P L–1, allowed to stand for 23 h, reshaken for 10 min, and centrifuged. The supernatant filtered through a 0.2-µm filter (Nuclepore polycarbonate membranes; Whatman, Maidstone, UK) was analyzed for soluble P colorimetrically after phosphomolybdate reduction with ascorbic acid. Transfer of P from or to soil was calculated from changes in the P concentration of the solution and plotted as a function of the P concentration in the equilibrium solution. The Freundlich isotherm modified by Fitter and Sutton (1975) to the form:

was used to compute the equilibrium phosphorus concentration (EPC0). In the equation, Y represents P sorbed or desorbed (mg kg–1), X is the P concentration measured in the equilibrium solution (mg L–1), and a (a slope of the equation) and b are adjustable positive valued constants. The constant b can vary between 0 and 1 and is related to sorption sites available. In theory, the graph can be extrapolated to the y axis at the equilibrium P concentration X = 0. This intercept, the constant c in the equation, is an estimate of instantly labile P. The intercept on the x axis (Y = 0) gives the EPC0 where no net sorption to or desorption from the solid phase occurs. The parameters of the equation were determined using iteration by the NLIN procedure of SAS (SAS Institute, 1989).

Phosphorus Removal In Situ
Removal of P in the Hovi CW in situ was studied during a total of 11 mo in the period 1999–2001 (April 1999–May 1999, October 1999–April 2000, April 2001, and October 2001). Between the sampling gaps (summer 1999, 2000, 2001, and winter 2000–2001) the wetland was occasionally dry and not sampled. The daily mean inflows and outflows of the CW were derived from the curves drawn by continuously recording water-stage gauges at control weirs (with pre-known stage–discharge relationships) installed at the inlet and outlet. Water samples for P analysis were taken automatically from the inflow and outflow at 8-h intervals and stored in a refrigerator. From these samples, daily composite samples were made by mixing equal portions of three 8-h samples. Within a week, from two to five daily composite samples were analyzed for total phosphorus (TP) and dissolved reactive phosphorus (DRP). The choice of samples analyzed was made according to the hydrograph to obtain samples representative of major runoff peaks. A total of 89 pairs of samples formed the basic dataset of P concentrations for the 11 mo studied. The DRP was analyzed from filtered (Nuclepore polycarbonate membranes, 0.4-µm pore size) and TP from unfiltered samples digested with K2S2O8 using the molybdate blue method (Murphy and Riley, 1962).

Daily TP and DRP fluxes (mg P m–2 d–1) to (a positive flux) and from (a negative flux) the CW were calculated by multiplying the daily concentration (concentrations for the days between the sampling periods were obtained by linear interpolation) by the daily mean flow of the corresponding day using a wetland area of 6000 m2 (a typical wet area during a high water situation). Daily P fluxes were obtained by subtracting output P loading from input P loading. From this daily dataset, average P fluxes (mg P m–2 d–1) and P removals (%) in each month were calculated for the CW. December 1999 and February 2000 were combined due to the low number of samples taken manually because of freezing damage in the automatic samplers. The monthly DRP removal was compared with monthly average of input DRP concentration and runoff volume. The use of measured daily values instead of using the interpolation and monthly averaging would have been misleading. Due to the long retention time (Table 1) in the Hovi CW the P concentration in the outflow may reflect the incoming water P concentration with a delay of one to several weeks. Removal percentage for TP and DRP was also calculated for the period of all 11 mo.

Phosphorus Removal in Laboratory Microcosms Sediments
To determine the effect of DRP and O2 availability on P removal, intact wetland sediment cores were incubated in a laboratory microcosm (at 15°C) using a continuous water flow technique (Liikanen et al., 2002). The sediments were taken in June 2001 from the deep water area (water depth = 1.5 m) and in August 2001 from the littoral area (water depth = 0–10 cm) directly to the incubation cores (transparent acrylic tube, i.d. = 94 mm, core height = 650 mm, sediment height = approximately 150–300 mm) with a piston core sampler.

The deep water sediments (eight replicates) without plants were incubated in the dark according to Liikanen et al. (2002). Water was pumped from the reservoirs over the cores by a peristaltic pump at a rate of 50 mL h–1; overlying water volume was 670 mL and water retention time in the cores was 13 h. Four sediments were incubated with anoxic water (O2 concentration in a water reservoir being <1 mg O2 L–1, deoxygenized with N2) and four with oxic water (O2 concentration in the water reservoir being 8.6–9.4 mg O2 L–1, in equilibrium with air) The sediments were first incubated with water from the CW (81 µg TP L–1, 60 µg dissolved total phosphorus [DTP] L–1, 23 µg DRP L–1), and then with artificial runoff water having 16, 47, or 155 µg DRP L–1 derived from KH2PO4. The concentrations of P used in the artificial runoff water represented minimum, mean, and maximum DRP concentrations in the inflow of the CW in 1999–2000. In addition, artificial runoff water contained also 140 µg NH4+–N L–1, 6.8 mg NO3––N L–1, 8 mg Ca L–1, and 2.4 mg Mg L–1 (derived from NH4Cl, NaNO3, CaCl2, and MgCl2, respectively). Each incubation lasted one week.

The littoral sediments (nine replicates) with living plants (cattail) were incubated under artificial lighting. The incubation was performed only with oxic water, since littoral sediments are not exposed to anoxic water in situ. Water flowed into the cores through a tube placed above the overlying water and flowed out through a tube placed 3.5 cm above the sediment surface keeping water level at this height throughout the experiment. Water flow was 26 mL h–1, volume of overlying water was 240 mL, and water retention time in the core was 9.3 h. During the first week, all sediments (nine replicates) received water from the CW (41 µg TP L–1, 26 µg DTP L–1, 6 µg DRP L–1) and during the second week artificial runoff water with 16, 47, or 155 µg DRP L–1 (three replicates).

The flux of DRP (mg P m–2 d–1) was determined from concentration differences between in- and out-flowing water using flow rates and sediment surface area (69 cm2). When the sediments were incubated with water from the CW also the flux of TP was determined. Removal (%) of TP and DRP were calculated on the basis of the P concentration in the in- and out-flowing water. Analyses were performed according to Finnish standards SFS 3026 for TP (not filtrated) and SFS 3025 for DRP (SFS Standardization 1986a, 1986b). At the end of each incubation period O2 concentration in the overlying water (1 cm above sediment surface) was measured with an O2 electrode (Dissolved Oxygen Meter Oxi 330 with Dissolved Oxygen Probe CellOx 325; WTW, Weilheim, Germany).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Characteristics of the Soils before the Construction of the Wetland
The amount of total soil P (H2SO4–extractable P in ignited samples) (Fig. 2A) , both organic (Fig. 2B) and inorganic (Fig. 2C), tended to be higher in the surface soil than in the subsoil at the depth below 30 cm (not statistically significant difference, t test, P > 0.005). The only exception was Sample 2 (meadow site), which had a higher content of organic P at a deeper depth. Total P content in meadow surface soil was lower than in the actively cultivated soil (Fig. 2A). On average, about half of the P was in the inorganic and half in the organic fraction in both soil layers (Table 2).



View larger version (35K):
[in this window]
[in a new window]
  		Fig. 2. Amount of P in various fractions in the soil samples (1–5) before the construction in various parts of the constructed wetland (CW) (see Fig. 1). (A) Total P (H2SO4–P in ignited samples), (B) organic P, (C) inorganic P, (D) NH4F-soluble "Al bound P," (E) NaOH-soluble "Fe bound P," and (F) acid-soluble "Ca bound P." Closed symbols (1, 2) present idle meadow and open symbols (3, 4, 5) present the actively cultivated area.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Proportion of organic and inorganic P of the total P, the relative distribution of inorganic P in various fractions, and the degree of phosphorus saturation (DPS) of Al and Fe oxides in surface soil layer (0–20 cm) and subsoil layer (20–60 cm) samples, with ranges in parentheses.

 
Most soils were dominated by Alox over Feox (Fig. 3) . However, the NaOH-P fraction (i.e., Fe-bound P) was greater than the NH4F-P fraction (i.e., Al-bound P) (Fig. 2, Table 2) (P < 0.001, t test). As a result, Feox was more saturated with P than Alox (Table 2) (P < 0.001, t test). However, only negligible amounts (less than 7%) of Alox and Feox were saturated with inorganic P (Table 2). The amount of Alox and Feox generally decreased with soil depth (Fig. 3) as did also the oxide-bound P (Fig. 2D, 2E) (not statistically significant differences, t test, P > 0.005). Phosphorus bound by Alox (Fig. 2D) and Feox (Fig. 2E) as well as total P (Fig. 2A) in the surface layers of meadow soils were one-half to one-third of those in the cultivated field containing accumulated fertilizer P. Most of the inorganic P was bound to Ca and Feox (Table 2), whereas only a negligible amount, 1.3 mg P kg–1 (i.e., 0.3% of the inorganic P pool), was in an easily soluble form (data not shown). The distribution of Ca-bound P between surface and subsoil showed no clear trends (Fig. 2F).



View larger version (19K):
[in this window]
[in a new window]
  		Fig. 3. Amount of (A) Alox and (B) Feox in the soil samples before the construction from various parts of the constructed wetland (CW) (see Fig. 1). Closed symbols (1, 2) present idle meadow and open symbols (3, 4, 5) present the actively cultivated area.

 
The EPC0 value of the surface soil from meadow was 8 µg P L–1 for the moist sample and 68 µg P L–1 for the air-dried one. In the cultivated soil, surface soil EPC0 values were higher, 55 and 112 µg P L–1, respectively. Thus, the air-drying markedly reduced the ability of soil to retain P from water. For the subsoils, all samples studied yielded EPC0 values of 3 µg P L–1, which is much lower than the DRP concentration of water flowing to the CW (Fig. 4) . However, in the original surface soil that was removed, the EPC0 was similar to or higher than the DRP concentrations in the inflowing water (Fig. 4). Thus, without removal of the surface soil, there would have been a risk that inflowing water P concentration had exceeded the EPC0 of the soil resulting in desorption of P from the soil to water.



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 4. Inflowing water dissolved reactive phosphorus (DRP) concentration (dots) and equilibrium phosphorus concentration (EPC0) of the soil samples (horizontal lines) taken from the wetland area to be constructed. Reprinted with permission from Elsevier from Ecological Engineering, 20:91–105.

 
Phosphorus Removal In Situ
During the in situ measurement period, a total of 68% of the TP and 49% of the DRP load was retained in the CW (Table 3). The in situ removal of P by sediment varied from 0.03 to 2.6 mg P m–2 d–1 (Fig. 5A, 5B) for DRP and from 0.03 to 66.7 mg P m–2 d–1 for TP (data not shown). Phosphorus retention by the CW increased with increasing input runoff (Fig. 5B), but the relative retention, that is, removal percentage, decreased as DRP concentration (Fig. 5C) or runoff (Fig. 5D) increased. The retention of P was not related to inflowing water DRP concentration (Fig. 5A). At the time of snow melt in April (1999, 2000, and 2001), the wetland retained TP more effectively than DRP (Table 3). In April 2000 the CW did not retain any DRP (Table 3); this sample represents an exception when DRP retention is compared with the inflowing water DRP concentration (76 µg P L–1) and runoff (30 L s–1 km–2) (Fig. 5). In October 2001 the removal of DRP was good, whereas only a minor amount of TP was retained (Table 3). Otherwise, the relative removals of DRP and TP were similar.


View this table:
[in this window]
[in a new window]
  		Table 3. Removal of total phosphorus (TP) and dissolved reactive phosphorus (DRP) in situ and in the laboratory microcosm sediments from the constructed wetland (CW) water.

 


View larger version (30K):
[in this window]
[in a new window]
  		Fig. 5. Monthly averages of (A) flux of dissolved reactive phosphorus (DRP) vs. input DRP concentration, (B) flux of DRP vs. runoff, (C) DRP removal percentage vs. input DRP concentration, and (D) DRP removal percentage vs. runoff in the wetland in situ.

 
Oxygen Conditions and Phosphorus Removal in the Microcosm Sediments
The O2 concentration in overlying water (1 cm above sediment surface) in the sediment cores with the anoxic flow was on average 0.3 ± 0.1 (SE) mg O2 L–1, close to the detection limit of the O2 electrode. With the oxic water flow, overlying water O2 concentration was higher in the littoral sediments, on average 7.8 ± 0.2 mg O2 L–1, than in the deep water sediments, on average 3.5 ± 0.1 mg O2 L–1. The deep water sediments were incubated under a sealed lid, which did not allow O2 to diffuse directly from the atmosphere to the overlying water. Oxygen was derived only from the inflowing water. However, the littoral sediment was exposed to open air, and O2 could diffuse to the microcosm directly from the atmosphere.

The sediment retained DRP from overlying water in all conditions (Fig. 6) . As expected, with increasing DRP concentration in the inflowing water, the flux of DRP from water to sediments increased (Fig. 6A). However, the DRP removal percentage decreased with increasing P load in deep water sediments but was almost constant in littoral sediments at all DRP concentrations (Fig. 6B). Fluxes of DRP to the sediment observed in the laboratory were higher, from 0.15 to 10.5 mg P m–2 d–1, than measured for the whole CW in situ. However, the relative DRP removal from artificial water measured in the microcosm was similar, on average 56%, compared with the average of 49% in situ. The DRP removal percentage was highest when the inflowing water DRP concentration was 47 µg P L–1, the mean DRP concentration of the inflowing water in the Hovi CW.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 6. (A) Flux of dissolved reactive phosphorus (DRP) and (B) DRP removal percentage in various wetland sediments with oxic and anoxic water flow studied in the laboratory microcosm.

 
The sediments incubated under different conditions did not significantly differ in their P retention when P concentrations in the inflowing water were low (Fig. 6A). At the highest P concentration (155 µg P L–1), however, the oxic littoral sediment retained more P than did the deep water sediments (P = 0.003, one-way ANOVA, Tukey's b). The deep water sediments had similar retention abilities. The DRP removal percentage from the artificial runoff water was higher in the littoral sediment than in the deep water sediment (Fig. 6B) (P < 0.001, one-way ANOVA, Tukey's b). Removal percentages of various P fractions from the CW water in the microcosm sediments are presented in Table 3. The removal of DRP (26–46%) was higher than that of TP (8–18%).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Removal of Phosphorus in the Hovi Constructed Wetland
Both in situ and laboratory measurements showed that the CW was efficient in retaining P. The total removals of 49 and 68% for DRP and TP, respectively, measured for the Hovi CW in situ are similar to the average P removal of 58% for 83 various wetlands reviewed by Reddy et al. (1999). In the laboratory incubations, the removal of DRP was equivalent to the in situ retention, but for TP the removal in laboratory experiments was lower than that observed in situ. This discrepancy is probably attributable to the continuous stirring in the microcosm limiting the sedimentation of particulate P, an important component of the TP. When negligible retention for DRP was detected in April 2000 (Table 3), some P was probably released to dilute snowmelt water from newly sedimented matter rich in P (Rekolainen, 1989). However, the removal for TP was always good, including during the snowmelt period in spring.

The P removal percentage decreased with the highest P load as found earlier by Sakadevan and Bavor (1999) and Pant et al. (2001). This response indicates that the P retention was mainly attributable to chemical sorption. The quantity/intensity plots, generally used to depict the dynamic equilibrium between the solid and solution phase P, show that the relative P sorption by solids will decrease with increasing P additions as a result of the increasing degree of P saturation on oxide surfaces (see, for example, Hartikainen, 1991). In fact, retention of P in wetland soil is reported to correlate with amorphous and poorly crystalline forms of Al, Fe, and Mn (Reddy et al., 1998; Pant et al., 2001). The bottom soil of the Hovi wetland, which had been exposed to P-rich agricultural runoff waters for rather short time, had excellent capacity to retain P even at low concentrations. Owing to the low P saturation of Alox and Feox, the subsoil was able to maintain in the solution phase a very low EPC0 value (Fig. 4) above which the soil acts as a sink for P. However, with time, those sorption sites will become more saturated, which may reduce the retention capacity of the soil. Although the P removal percentage decreased with increasing DRP concentrations, the relative retention in the laboratory tests was at its highest when the inflowing water DRP concentration was 47 µg P L–1, the mean DRP concentration the Hovi CW receives.

At low DRP concentrations, the retention of dissolved P was efficient irrespective of O2 conditions (Fig. 6A). At the highest DRP load, however, the retention markedly diminished under anoxic flow, when Fe3+ obviously was at least partly reduced to Fe2+. Here we did not determine the flux of Fe between sediment and water, which could have indicated a reduction-induced dissolution of Fe2+. However, in our previous experiments with lake sediments under similar anoxic conditions (approximately 0.3 mg O2 L–1), sediments have released Fe and P (Varjo et al., 2003), indicating that under the prevailed conditions reduction of Fe3+ was obvious. Because the wetland sediment was able to retain P also under anoxic flow conditions, the high amount of Alox has had an important role as a sorption component. Under reduced conditions, the Al oxides are found to be of importance in binding P since they do not dissolve at low redox values as do Fe oxides (Koski-Vähälä and Hartikainen, 2001; Uusitalo and Turtola, 2003). However, P bound to Alox, in turn, can be partly replaced by OH– ions produced as a result of increase in pH with lowering redox potential (Bartlett, 1999). Phosphorus can also be bound by Ca (Reddy et al., 1999). However, in Finnish soils, which are typically noncalcareous and very young, the retention of P by Ca compounds is shown to be immaterial (Hartikainen, 1979). The high Ca-bound P reserves found in the soils studied are attributable to the apatitic P present in weakly weathered soils.

The decrease of P removal percentage with increasing runoff volume was probably attributed to decreased water retention times as the runoff increased. When water retention time decreases, the contact time of water with soil decreases and binding of P to solids is not so efficient (Tanner et al., 1995; Moustafa, 1999; Sakadevan and Bavor, 1999).

The removal of TP was good during the in situ measurements but not in the laboratory microcosms, which suggests that sedimentation and vegetation in the CW contributed to retention of TP. According to the laboratory incubations, vegetation may have some importance in taking up dissolved inorganic P at high DRP loads. Many studies have found that planted wetland sediments retain P more efficiently than unplanted ones (Tanner et al., 1995; Tanner, 1996). In contrast to our sediments, plant efficiency to remove P has been reported to be greatest at low P concentrations (Reddy et al., 1987). Mineral wetland soils with high amounts of Alox and Feox, like the soil in the Hovi CW, have been reported to have an optimal P retention capacity without vegetation (Yang et al., 2001). However, in the Hovi CW, vegetation probably had a significant physical role by stabilizing the soil, increasing the sorption area and distributing and reducing the velocity of the flowing water (Brix, 1997). These mechanisms are found to favor sedimentation of suspended solids and to reduce the risk of erosion and resuspension (Brix, 1997). Nutrient assimilation by plants follows their seasonal growth pattern being maximal during the growing season in the summer. In the boreal region, the maximal load of P and growth rate of plants do not match, the highest nutrient load occurs in spring when the growth of plants is negligible, and during summer months when the biomass production is highest and the flow of water and nutrients is lowest (Braskerud, 2002; Hyvärinen, 2003). Therefore, P retention by vegetation bound to the seasonal growth pattern is not as effective as P retention by chemical sorption reactions, which are fast and occur throughout the year. For example, the vegetated laboratory microcosm sediments taken in August were not in a state of growth and therefore vegetation probably did not contribute to any significant improvement in P uptake.

Construction Factors Contributing to the Phosphorus Retention
The removal of the topsoil was essential to achieve a good P removal efficiency of the CW, since the topsoil contained high amounts of P (Fig. 2A). The EPC0 of the surface soil was similar to the DRP concentrations in the inflowing water. This means that if the topsoil had been left at the bottom of the CW, there would have been the possibility that the topsoil would release P into the flowing water. When a former agricultural soil is flooded to convert the soil to wetland, an agricultural soil with a high P content would not retain P, and may even be a potential source for P (Pant et al., 2002a, 2002b). For example, in the Hovi CW, release of P would have been more crucial from the actively cultivated soil having higher P content and EPC0 than from the meadow soil, which had not been under P fertilization. However, although the topsoil rich in P is removed, the characteristics of the subsoil to be exposed have to be suitable for P retention. The EPC0 of the Hovi subsoil, the current bottom of the CW, was lower than the DRP concentration of inflowing water and therefore the wetland was able to retain P. In addition to removal of P, the peeling of the topsoil removed organic matter, which, at the bottom of the CW, would have degraded and released nutrients and, thus, contributed to O2 consumption and development of anoxia (D'Angelo and Reddy, 1994). In anaerobic conditions, P retention by wetland soil is reduced (Reddy et al., 1998) or P can be even mobilized from soils or sediments to the overlying water (Boström et al., 1988; D'Angelo and Reddy, 1994).

The Hovi CW, which contains both deep and shallow water areas, provided good environmental conditions for the removal of both particulate and dissolved P. The deep water area immediately at the inlet of the CW acted as a sedimentation pond for particulate matter and the shallow water area with its tortuous flow path provided good contact surface for the removal of dissolved P. Moustafa (1999) reported that P was efficiently retained only when the water depth in the wetland was less than 20 cm; in deeper waters dissolved P is not in contact with sediment and cannot be removed.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
The Hovi wetland constructed on a former agricultural soil was efficient in retaining P from agricultural runoff waters, due to the careful design and implementation of the CW. The removal of topsoil with a history of fertilization, excavation of a deep sedimentation pond, and creation of a tortuous shallow flow path and the large area enabling long water retention time were essential to obtain good purification efficiency for both particulate and dissolved P. Sedimentation reduced the concentration of TP, and chemical sorption reactions of P onto oxides of Al and Fe were the key mechanisms accounting for the DRP retention. Thus, CWs established on mineral soils with a low degree of P saturation on oxide surfaces may be more efficient in retaining P than those constructed on natural, organic wetland soils (Johnston, 1991; Gale et al., 1994). Before the construction of a wetland on arable land, proper soil analyses for characterization of P resources and P exchange properties are needed to evaluate the applicability of the soil for effective P removal. The Hovi CW had excellent properties for both sedimentation and sorption, both of which are considered as long-term storage mechanisms for P.


   ACKNOWLEDGMENTS
 
The study was financed by the project "PRocess based Integrated Management of constructed and Riverine wetlands for Optimal control of wastewater at catchment ScalE" (PRIMROSE; EVK1-CT-2000-00065) of the EU 5th Framework Programme. Mari Räty is acknowledged for the help with soil analyses, Päivi Noponen for conducting the microcosm studies, and the personnel of North Ostrobothnia and North Savo Regional Environment Centres for water chemical analyses.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2004 33: 799-804. [Full Text]  



This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
A. K. Sovik, J. Augustin, K. Heikkinen, J. T. Huttunen, J. M. Necki, S. M. Karjalainen, B. Klove, A. Liikanen, U. Mander, M. Puustinen, et al.
Emission of the Greenhouse Gases Nitrous Oxide and Methane from Constructed Wetlands in Europe
J. Environ. Qual., October 27, 2006; 35(6): 2360 - 2373.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
B. C. Braskerud, K. S. Tonderski, B. Wedding, R. Bakke, A.-G. B. Blankenberg, B. Ulen, and J. Koskiaho
Can Constructed Wetlands Reduce the Diffuse Phosphorus Loads to Eutrophic Water in Cold Temperate Regions?
J. Environ. Qual., November 7, 2005; 34(6): 2145 - 2155.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Liikanen, A.
Right arrow Articles by Hartikainen, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liikanen, A.
Right arrow Articles by Hartikainen, H.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Liikanen, A.
Right arrow Articles by Hartikainen, H.
Related Collections
Right arrow Water Quality
Right arrow Wetlands and Aquatic Processes
Right arrow Phosphorus
Right arrow Nutrients
Right arrow Nutrient Cycling

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome