e c o l o g i c a l e n g i n e e r i n g 2 7 ( 2 0 0 6 ) 279–289

avai lab le at www.sc iencedi rec t .com

journa l homepage: www.e lsev ier .com/ locate /eco leng

Periphyton as a potential phosphorus sinkin the Everglades Nutrient Removal Project

Paul V. McCormicka,∗, Robert B.E. Shuford III a, Michael J. Chimneyb

a Everglades Division (MSC-4440), South Florida Water Management District, 3301 Gun Club Road, West Palm Beach, FL 33406, USAb STA Management Division (MSC-4470), South Florida Water Management District, 3301 Gun Club Road, West Palm Beach,FL 33406, USA

a r t i c l e i n f o

Article history:

Received 8 May 2006

Accepted 30 May 2006

Keywords:

Algae

Constructed wetland

Everglades Nutrient Removal Project

Microcosm

Periphyton

Phosphorus

a b s t r a c t

Phosphorus uptake and release by periphyton mats were quantified in the Everglades

Nutrient Removal Project (ENRP) to evaluate the potential for periphyton P removal.

Short-term P uptake rates were determined by incubating cyanobacteria (Oscillatoria

princeps and Shizothrix calcicola) and Chlorophycean (primarily Rhizoclonium spp.) algal mat

samples for 0.5–2 h under ambient conditions in BOD bottles spiked with soluble reactive

P (SRP). Cyanobacterial mats removed P more than twice as fast (80–164 �g P h−1 g−1 AFDM)

as Chlorophycean mats (33–61 �g P h−1 g−1 AFDM) during these incubations. In a longer

term study, fiberglass cylinders were used to enclose 1.8 m2 plots within the wetland and

were dosed weekly for 7 weeks with: (1) no nutrients; (2) SRP (0.25 g P m−2 week−1); or (3)

SRP plus nitrate (0.42 g N m−2 week−1) and ammonium (0.83 g N m−2 week−1). Phosphorus

uptake rates by this periphyton assemblage, which was dominated by the chlorophytes

Stigeoclonium spp. and Oedogonium spp., were measured weekly and were similar among

nutrient treatments on most dates, indicating that the algal storage compartment for P was

not saturated despite repeated P additions. Decomposition rates and P loss by cyanobacteria

and Chlorophycean mats were determined by measuring biomass loss and SRP release

in darkened BOD bottles over 28–42 day periods under anaerobic and aerobic conditions.

First-order aerobic and anaerobic decomposition rates for cyanobacterial mats (k = 0.1095

and 0.1408 day−1, respectively) were 4–20-fold higher than rates for Chlorophycean mats

(k = 0.0066 and 0.0250 day−1, respectively) and cyanobacteria released considerably more

P back to the water column. Our findings suggest that periphyton can be an important

n treatment wetlands and that retention is strongly affected by the

n o

has been acknowledged as a potentially important short-

short-term sink for P i

taxonomic compositio

1. Introduction

Nutrient retention in freshwater wetlands is attributed to sed-iment accretion and chemical sorption processes (Nichols,

1983; Reed et al., 1995; Richardson et al., 1997; Reddy et al.,1999; Verhoeven and Meuleman, 1999). The principal bio-logical mechanism for long-term phosphorus (P) removal is

∗ Corresponding author. Present address: USGS Leetown Science CenterTel.: +1 304 724 4478; fax: +1 304 724 4465.

E-mail address: pmccormick@usgs.gov (P.V. McCormick).0925-8574/$ – see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.ecoleng.2006.05.018

f the periphyton assemblage.

© 2006 Elsevier B.V. All rights reserved.

the accretion of macrophyte detritus in anoxic sedimentswith low mineralization rates (Richardson and Craft, 1993;Kadlec and Knight, 1996). While algal and microbial biomass

, 11649 Leetown Road, Kearneysville, WV 25430, USA.

term sink for P in wetlands, the working assumption is thatthese biological compartments play a minor role in long-term P retention due to high turnover (e.g., decomposition)

r i n g

280 e c o l o g i c a l e n g i n e e

rates (Richardson and Craft, 1993; Richardson et al.,1997).

The South Florida Water Management District (District)has converted large tracts of Everglades peatland previouslydrained and fertilized for agricultural use into treatment wet-lands referred to as stormwater treatment areas (STAs). Theprimary purpose of these wetlands is to remove P from agri-cultural runoff that enters the Everglades. When flooded, theSTA soils were predicted to be a source of rather than asink for, P at the anticipated inflow P concentrations (Reddyand Graetz, 1991; Reddy et al., 1999). Consequently, biologi-cal mechanisms must account for the bulk of P removal inthese systems. Although algal and cyanobacterial mats (i.e.,periphyton) occur in most wetlands, their role in P uptake andretention has not been widely studied (but see Knight, 2003;Scinto and Reddy, 2003). Unlike most emergent aquatic macro-phytes, algae obtain the bulk of their nutritional demandsdirectly from the water column and exhibit rapid growthresponses to increased availability of limiting nutrients. Inaddition, the rate of algal decomposition, and subsequentlyP return to the water column, is generally regarded as beingrelatively fast compared to macrophyte decay (e.g., Enriquezet al., 1993). This view, however, is based largely on studiesof phytoplankton decomposition and ignores the possibilitythat a substantial fraction of the P associated with periphytonmay be in recalcitrant forms. For example, in carbonate-richsystems such as the Everglades, some P precipitates onto peri-phyton mats as insoluble inorganic complexes (Otsuki andWetzel, 1972; Scinto and Reddy, 2003).

The objective of this study was to assess the potential sig-nificance of periphyton-P fluxes in the Everglades NutrientRemoval Project (ENRP), a prototype STA. Short-term incuba-tions and enclosure experiments were designed to measurethe capacity of periphyton mats to remove P from surfacewater and the potential for periphyton P to be released back tothe water column during mat decomposition. Different typesof periphyton mat were used to evaluate the influence of tax-onomic composition on these P fluxes.

2. Methods

2.1. Study site

All experiments were conducted in ENRP Cell 4 (147 ha) (Fig. 3in Chimney and Goforth, 2006) during its initial start-up phase(1994–1995). The District began recirculating water betweeninterior cells of the ENRP on a limited basis in spring 1994and started flow-through operations in August 1994. The his-tory, operation and layout of the ENRP are described in Guardoet al. (1995) and Chimney and Goforth (2006). Cell 4 was anopen-water system that contained scattered floating vegeta-tion (primarily alligatorweed [Alternanthera philoxeroides]) butalmost no emergent macrophytes. Sparse periphyton growthoccurred as free-floating mats or tangled in the floating veg-etation. Attached periphyton was largely absent due to thelack of submerged surfaces in this newly created wetland.

Light attenuation likely limited benthic periphyton as almostall photosynthetically active radiation (PAR) was lost withinthe top 30 cm of the water column and all but the margins ofthis treatment cell were flooded to depths >30 cm. Relatively

2 7 ( 2 0 0 6 ) 279–289

low chlorophyll a and soluble reactive P [SRP] concentrations(median values = 10 and 4 �g L−1, respectively) and a historyof modest changes in diel dissolved oxygen (DO) levels (max-imum change ∼4 mg L−1; Newman and Pietro, 2001) at a loca-tion in Cell 4 near our experiments suggested that our studysite had low to moderate water-column net productivity dur-ing this study.

2.2. Experiment dates

Short-term periphyton P uptake experiments were conductedin March 1994 and January 1995. A longer term study usingenclosures (i.e., mesocosms) was initiated in October 1994to quantify the response of periphyton decomposition andP cycling to repeated nutrient additions. Both cyanobacterialand Chlorophycean periphyton mats were used in the short-term trials, while the periphyton assemblage that developedin the enclosures was dominated by Chlorophytes.

2.3. Short-term experiments

2.3.1. Characterization of periphyton matsCyanobacterial and Chlorophycean periphyton mats were col-lected from Cell 4 in March 1994 and examined microscopicallyto determine the dominant taxa and qualitatively assess theextent of mineral deposits (e.g., carbonates) on the surface ofalgal filaments. Approximately 1-g wet weight (WW) samplesof each mat type were placed in individual light and dark 300-mL BOD bottles, filled with ambient water and incubated in situjust below the water surface for 0.5–2 h to determine gross pri-mary productivity (GPP) and respiration (R) using the oxygenevolution method (Wetzel and Likens, 2000). Following incuba-tion, all periphyton samples were returned to the laboratory,dried to a constant dry mass (DM) at 70 ◦C and ashed at 500 ◦Cfor 1 h to determine the ash-free dry mass (AFDM) and per-cent ash content (SM 10300C; APHA, 1989). Respiration rateswere normalized for AFDM and GPP rates were normalized forboth AFDM and the amount of PAR received during the incu-bation. Percent ash content was used as an indicator of matcarbonate content. Diatom abundance in both mats was low,so we assumed that relatively little of the ash was attributableto silica.

Additional samples of each mat type were frozen at −10 ◦Cupon return from the field and analyzed later for total P(TP), total carbon (TC), total nitrogen (TN) and total calcium(TCa) content. Total Ca was used as a second indicator of car-bonate deposits. Samples were thawed, dried for 48–72 h at85–90 ◦C to a constant DM, finely ground in a Wiley mill andredried to remove any moisture. A homogenized aliquot ofeach sample was then assayed for TC and TN content using aCarbon–Nitrogen elemental analyzer (Fisons NA1500). Total Pcontent was determined by digesting a second aliquot withperchloric acid and analyzing the filtrate for P using stan-dard methods (EPA 365.2; USEPA, 1983). Total Ca content wasmeasured by flame atomic absorption spectrophotometry (SW7140; USEPA, 1986).

2.3.2. Phosphorus uptakeSamples of both periphyton mats were incubated in individ-ual BOD bottles spiked with known amounts of SRP to measure

i n g

siaacapbBdtawmtt2tisetpl

i

U

wtBbtictoto

uVudtSm

2

2MtFwCpl

e c o l o g i c a l e n g i n e e r

hort-term P uptake rates. These experiments were conductedn conjunction with the GPP and R measurements describedbove. Bottles were first partially filled with ambient waternd amended with a solution of NaH2PO4. Target SRP con-entration treatments ranged from 50 to 120 �g P L−1 abovembient SRP levels. Approximately 1-g WW samples of peri-hyton mat were added to individual BOD bottles and theottles then filled completely with ambient water. ControlOD bottles without periphyton were spiked and incubated toetermine loss of P from the water attributable to sinks otherhan periphyton uptake (e.g., uptake by water column bacteriand phytoplankton or adsorption to the bottle walls). Bottlesere incubated as described above for productivity measure-ents. A single BOD bottle from each SRP concentration/mat

ype treatment (three treatments run on March 1994 and tworeatments run on January 1995) was collected after 0.5, 1 andh of incubation. Water samples from these bottles were fil-

ered through a 0.45 �m membrane filter in the field, storedn the dark at 4 ◦C and analyzed within 48 h for SRP usingtandard methods (EPA 365.2; USEPA, 1983). Samples of ambi-nt water were processed in the same manner to determinehe marsh background SRP concentration on each date. Peri-hyton material from each BOD bottle was returned to the

aboratory and analyzed for AFDM as described above.Periphyton P uptake rates were normalized for differences

n biomass:

= ([(Ci − Ct) − �Cc]/V)/(AFDM)t

(1)

here U is the periphyton P uptake rate (�g P g−1 AFDM h−1);is the duration of the incubation (h); V is the volume of theOD bottle (L); Ci is the initial SRP concentration in the BODottle adjusted for the ambient water SRP concentration andhe SRP spike addition (�g P L−1); Ct is the SRP concentrationn the BOD bottle at the end of incubation (�g P L−1); �Cc is thehange in SRP concentration observed in the control BOD bot-le (�g P L−1) and AFDM is periphyton biomass (g). Note thatur measurements of P uptake included both direct assimila-ion of P by the periphyton and sorption of P onto the surfacef the mat (see Scinto and Reddy, 2003).

Statistical differences in P uptake rates were analyzedsing a 2-way analysis of variance (ANOVA; PROC GLM, SASersion 6, SAS Institute, Cary, NC). For a given mat type, Pptake rates were similar among BOD bottles incubated forifferent times and containing different initial SRP concentra-ions. Therefore, data were pooled over incubation times andRP concentration treatments for each mat type and experi-ent date before performing the ANOVA.

.4. Longer–term studies

.4.1. Phosphorus uptakeesocosms (1.2 m tall × 1.5 m diameter fiberglass cylinders)

hat allowed transmission of >70% of ambient PAR (SUN-LITE®

iberglass Solar Components Corporation, Manchester, NH)

ere used to isolate nine 1.8 m2 plots in an open-water area ofell 4. The plots were devoid of emergent and floating macro-hytes, contained no established periphyton mats and had

ow to moderate water-column net productivity (see above).

2 7 ( 2 0 0 6 ) 279–289 281

Therefore, major potential sinks for P were the soil and anyperiphyton that developed during the experiment. Each meso-cosm was pushed 10 cm into the sediment and secured to fourPVC poles driven deep into the peat. Mesocosms were perfo-rated with numerous 3-cm diameter holes to facilitate waterexchange between the plots and the surrounding marsh. Wefound in preliminary trials that the periphyton community inmesocosms without holes quickly changed from the commu-nity composition in the surrounding marsh but observed noobvious species shift in mesocosms with holes. Mesocosms ofa similar design have been used successfully by the District innutrient enrichment experiments conducted elsewhere in theEverglades (e.g., McCormick and O’Dell, 1996). Acrylic dowels(7.5 cm long × 0.6 cm diameter; surface area = 14.42 cm2) weresuspended from monofilament line secured between polesinside each mesocosm to provide an inert, standardized sur-face for periphyton growth.

Triplicate mesocosms were dosed weekly for five con-secutive weeks with one of the following nutrient amend-ments: (1) controls with no added nutrients; (2) P added asNaH2PO4 at 0.25 g P m−2 week−1 and (3) P added as NaH2PO4

at 0.25 g P m−2 week−1 and nitrogen (N) added as NH4Clat 0.83 g N m−2 week−1 and NaNO3 at 0.42 g N m−2 week−1.Because inflows to the ENRP were enriched with both Pand N, the combined enrichment treatment was includedto provide the most realistic assessment of periphytonresponses to stormwater pulses. Nutrient doses were appliedas a single concentrated pulse and dispersed by gentlymixing the water in each mesocosm with a paddle toachieve near instantaneous concentrations of approximately300 �g PO4 L−1, 1000 �g NH4 L−1 and 500 �g NO3 L−1. Thesenutrient levels were selected based on the expected maxi-mum nutrient inflow to the STAs. Mesocosms were wrappedin heavy-gauge plastic sheeting before dosing to isolate waterinside the enclosure and allow for nutrient uptake by theperiphyton. The sheeting was removed approximately 24 hafter dosing to reinstate free exchange of water between themesocosm and the surrounding marsh. A similar strategyof pulsed nutrient addition to mesocosms was employed byMcCormick and O’Dell (1996). After 5 weeks of dosing, addi-tions of both P and N were increased to twice a week inall treatments for two additional weeks to assess periphytonresponses to higher nutrient loading rates. Ambient SRP andTP concentrations during this experiment ranged between4–15 and 38–90 �g L−1, respectively. Total Kjeldahl nitrogenranged between 2.02 and 2.62 mg L−1 and ammonium andnitrate ranged between 11–156 and <4–32 �g L−1, respectively.

Periphyton P uptake was measured in all treatmentsweekly, except during the second week when fieldwork wascurtailed due to a passing tropical storm. A single acrylicdowel from each mesocosm was placed in a BOD bottlecontaining ambient water and spiked with P (50 �g P L−1 asNaH2PO4). Control BOD bottles without periphyton wereprepared in the same fashion. All bottles were incubatedin situ just below the water surface for 1.5–2 h after whichtime water samples were collected, filtered and analyzed

for SRP (EPA 365.2; USEPA, 1983). Acrylic rods were returnedto the laboratory where the periphyton was removed andprocessed for AFDM (SM 10300C; APHA, 1989); AFDM onthe rods was expressed on a unit-area (m2) basis. Changes

r i n g

282 e c o l o g i c a l e n g i n e e

in P concentrations were normalized for periphyton AFDMand incubation time and corrected for P adsorption to thebottle walls and plankton uptake in the control bottles whencalculating periphyton P uptake rates using Eq. (1).

Statistical differences in P uptake among nutrient treat-ments on each date were analyzed using one-way ANOVAs.Post hoc multiple mean comparisons (REGWF option, PROCGLM, SAS) were used to detect differences among individualtreatments on individual dates where significance in the over-all ANOVA F value (p < 0.05) was found.

2.4.2. Periphyton decomposition and phosphorus releaseTwo concurrent experiments were conducted to assess ratesof periphyton decomposition and P release under anaerobicand aerobic conditions. Anaerobic decomposition was mea-sured by incubating pre-weighed samples (WW) of live Chloro-phycean and cyanobacterial mat in situ in BOD bottles filledwith ambient water. The BOD bottles were covered with alu-minum foil and placed on the marsh bottom at water depthsof 10–30 cm. Triplicate BOD bottles of each mat type werecollected after 1, 2, 3, 6, 8, 10, 22, 27, and 35 days of incu-bation. Because the Chlorophycean mats decomposed moreslowly than the cyanobacterial mats, additional BOD bottlescontaining Chlorophycean mats were collected after 42 daysof incubation. Upon retrieval, the DO concentration and watertemperature in each BOD bottle and the water temperaturein the surrounding marsh were measured using a field oxy-gen meter and probe (YSI model 57; YSI Corp., Yellow Springs,OH). A water sample was collected from each BOD bottle andanalyzed for SRP and the remaining periphyton mat in the bot-tle was filtered onto a pre-ashed/pre-weighed glass fiber filter(Whatman GF/C) and analyzed for AFDM. Additional samplesof each mat type collected at the same time as samples forincubation were returned to the laboratory and processed forAFDM to develop WW:AFDM regression relationships. Theseregressions were used to estimate the initial (day 0) AFDM con-tent of periphyton samples incubated in the BOD bottles.

Aerobic decomposition was measured by incubating bothperiphyton mat types in uncapped BOD bottles in a darkenedlaboratory incubator at 30 ◦C, which approximated the averagedaily temperature in the marsh during this study. Bottles wereinitially filled with ambient water and aerated continuously byinserting aquarium tubing into each bottle and adjusting air-flow to maintain slow bubbling. Aluminum foil was securedover the mouth of each BOD bottle to minimize evaporativewater loss, which was replenished periodically with distilled,deionized water. Triplicate BOD bottles of each periphytontype were processed after 2, 3, 6, 9, 13, 16, and 28 days ofincubation. Additional BOD bottles containing chlorophyceanmats were processed after day 34. Soluble reactive P and DOlevels were measured in all BOD bottles and the AFDM con-tent of the remaining periphyton was determined as describedabove for the anaerobic decomposition experiment.

Decomposition rate coefficients were estimated for bothperiphyton mat types under aerobic and anaerobic conditions

by fitting a first-order exponential function to each data setusing non-linear regression (Stumm and Morgan, 1996):

Mt = M1 exp(−kt) (2)

2 7 ( 2 0 0 6 ) 279–289

where M1 is periphyton AFDM on the first collection daterelative to AFDM at the start of the incubation (%), Mt peri-phyton AFDM remaining at time = t relative to AFDM at thestart of the incubation (%), k the decomposition coefficient(day−1) and t is time (days). No asymptote term was includedin this equation as it was assumed that virtually all peri-phyton AFDM was decomposable under both aerobic andanaerobic conditions given sufficient time. By contrast, vas-cular plants produce recalcitrant compounds such as ligninsthat undergo little or no decay under anaerobic conditions.Statistical differences in rate coefficients between periphy-ton types and oxygen regimes were detected using a t-testfor parallelism of slopes (Kleinbaum and Kupper, 1978; Zar,1999).

3. Results

3.1. Short-term experiments

3.1.1. Characterization of periphyton matsThe taxonomic assemblage of the periphyton mat types usedin short-term experiments was dominated by only a few taxa.Chlorophycean mats were largely composed of Rhizocloniumspp. (class Chlorophyceae). Diatom (class Bacillariophyceae)epiphytes also were present, although Rhizoclonium clearlycomprised the bulk of the biomass. Microscopic examina-tion revealed that many of the Rhizoclonium filaments wereencrusted with mineral deposits suggestive of carbonate pre-cipitation. Cyanobacterial mats were dominated by Oscillatoriaprinceps Vaucher and Shizothrix calcicola (Agardh) Gomont (classCyanophyceae). This mat type lacked both epiphytic microal-gae and visible mineral deposits on the algal filaments.

The TC content of Chlorophycean mats was significantlygreater than that for cyanobacterial mats (p < 0.001, Fig. 1a).Because both mat types had similar TN and TP contents(p > 0.05, Fig. 1b and c), the Chlorophycean mat had higherC:N and C:P mass ratios compared to cyanobacterial mats.Cyanobacterial and Chlorophycean mats had N:P molar ratiosof 25.9:1 and 26.9:1, respectively, which suggested a balancebetween N and P (see Table 1 in Pietro et al., 2006). Differencesbetween mat types for both TCa content (Fig. 1d) and percentash (Fig. 1e) were not significant (p > 0.05). These measures ofmineral content were considered more conclusive than ourmicroscopic examination, which suggested greater mineralprecipitation on Chlorophycean mats.

The GPP of cyanobacterial mats was approximately 3-foldhigher and significantly different from that of Chlorophyceanmats (p < 0.001, Fig. 1f). Respiration rates were not statisticallydifferent (p > 0.05, Fig. 1f) and considerably lower than the cor-responding GPP for both mat types. While cyanobacterial matswere more productive, both mat assemblages appeared to bein a healthy physiological state.

3.1.2. Phosphorus uptakeCyanobacterial mats had significantly higher P uptake rates

than Chlorophycean mats (Fig. 2; p = 0.035). Within sam-pling dates, mean P uptake by cyanobacterial mats (80 and164 �g P g−1 AFDM h−1 in March 1994 and January 1995, respec-tively) was more than twice that measured for Chlorophycean

e c o l o g i c a l e n g i n e e r i n g 2 7 ( 2 0 0 6 ) 279–289 283

Fig. 1 – (a–f) Characteristics of cyanobacterial (open bars) and Chlorophycean (solid bars) periphyton mats used inshort-term BOD bottle experiments. Bars represent means of fiveand gross primary productivity, differences between mat types w

Fig. 2 – Phosphorus uptake rates for cyanobacterial (openbars) and Chlorophycean (solid bars) mats measured inshort-term BOD bottle experiments. Bars represent meansof five replicate samples ±1 S.E. Differences in P uptakerates were statistically significant for both mat type andsampling date.

replicate samples ±1 S.E. Except for total carbon contentere not statistically significant (see text for details).

mats (33 and 61 �g P g−1 AFDM h−1). The overall mean P uptakerate for both mat types in January 1995 was 2-fold higher andsignificantly different from the overall mean rate in March1994 (p = 0.034). However, our sampling regime was unableto determine whether this variation was simply stochasticin nature or reflected seasonal and/or other deterministicchanges in environmental conditions.

3.2. Longer–term mesocosm study

Periphyton biomass on acrylic dowels increased in all treat-ments during the first month of the dosing experiment(Fig. 3A). Dowels in all treatments were initially colonized bya thin film of diatoms, but became dominated by filamen-tous green algae (Oedogonium spp. and Stigeoclonium spp. [classChlorophyceae]) by day 25. Periphyton biomass decreased dur-ing the final week of the study in all treatments. This decreasecoincided with the growth of dense filamentous algae on themonofilament line used to suspend the dowels, which reduced

light reaching the dowels to 5% of surface PAR. In comparison,routine monitoring of light penetration in open-water areas ofthe ENRP found 72% (median value) of surface irradiance at adepth of 10 cm (Chimney et al., 2006).

284 e c o l o g i c a l e n g i n e e r i n g

Fig. 3 – (A–C) Biomass, total phosphorus content and Puptake rates for periphyton growing on acrylic dowels inunenriched (open bars) and enriched mesocosms thatreceived weekly pulses of P (gray bars) or P + N (black bars).Bars represent means of three replicate samples ±1 S.E.except for P content as noted in the text. Statisticallysignificant differences among treatments within a givendate are indicated by different letters (a and b) (see text fordetails). No P + N enriched samples were analyzed on day

ing cyanobacterial mats had substantially lower daily meanDO concentrations (0.5–2.2 mg L−1) compared to BOD bottleswith Chlorophycean mats (4.0–6.0 mg L−1) by day 6 despitesimilar aeration regimes (Table 1). Similarly, BOD bottles con-

Fig. 4 – Relationship between periphyton phosphorusuptake and areal biomass on acrylic rods in unenriched

47.

Mean periphyton P content ranged from 0.9 to 2.4 mg P g−1

DM during the dosing experiment (Fig. 3B). The mean peri-phyton TN:TP molar ratio was 26.7:1, suggesting that P wasin balance with N (Pietro et al., 2006). The lack of signifi-cant differences in periphyton biomass between enriched andunenriched enclosures (Fig. 3A) further suggests that neithernutrient was the primary factor limiting algal growth. Periphy-ton P content was generally higher in the nutrient enrichedtreatments than in the unenriched controls, suggesting thatat least a portion of the added P accumulated in the periphytonmat. However, because total periphyton biomass in the enclo-sures was not measured, mass balance calculations could notbe performed to determine the proportion of added P that wasincorporated into the periphyton mat. Compared with unen-riched mats, nutrient enriched periphyton had a significantly

higher P content (p < 0.05) on days 25 (P and P + N treatments),40 (P + N treatment) and 47 (P treatment; no P + N enrichedsamples were analyzed on this date). Only a single P and P + Nsample was analyzed on day 32, therefore, differences in TP

2 7 ( 2 0 0 6 ) 279–289

content could not be determined statistically, although thepattern of differences among treatments was similar to thatobserved on other dates.

Phosphorus uptake rates by periphyton on artificial sub-strates in the mesocosms (Fig. 3C) were within the range of Puptake rates measured in the short-term experiments (Fig. 2).Phosphorus uptake in the mesocosms decreased by more than5-fold in all treatments during the first 40 days of sampling andthen increased slightly on the final sampling date (Fig. 3C).This decrease in P uptake was weakly correlated with anincrease in periphyton biomass on the acrylic dowels acrossall treatments (Fig. 4). Differences in P uptake among treat-ments seemed unrelated to the nutrient additions, althoughthe control mesocosms had the highest mean P uptake rate onthree of the six sampling dates. Differences in P uptake ratesbetween the control and the P and P + N treatments were notsignificantly different throughout the mesocosm study excepton day 40 (p < 0.001) (Fig. 3C).

3.3. Decomposition rates and phosphorus release

We assumed that water temperature in the in situ BOD bottlesdid not differ appreciably from the surrounding marsh. Marshwater temperatures during BOD bottle incubation ranged from24 to 30 ◦C (mean temperature = 26.6 ◦C). The in situ bottlesquickly became anaerobic (DO <0.5 mg L−1), while BOD bot-tles aerated in the laboratory generally maintained higherDO concentrations throughout the incubation period (Fig. 5).Decomposing cyanobacterial mats exerted a greater oxygendemand in both experiments. Laboratory BOD bottles contain-

(open circles) and enriched mesocosms that receivedweekly pulses of P (gray squares) or P + N (black triangles).Solid line is the regression line and dotted line indicates 95% confidence intervals.

e c o l o g i c a l e n g i n e e r i n g 2 7 ( 2 0 0 6 ) 279–289 285

Table 1 – Comparison of periphyton decomposition coefficients, remaining periphyton biomass at the end of incubation,range of low dissolved oxygen concentrations and maximum soluble reactive phosphorus concentrations in longer termBOD bottle experiments run under anaerobic and aerobic conditions

Experiment type k (day−1) Remaining biomass (%) Low DO (mg L−1) Max SRP (�g P L−1)

Cyanobacterial mat Anaerobic 0.1408 25 <0.5 363519 0.5–2.2 2003

36 <0.5 57486 4.0–6.0 37

tsm

CdcdbTrctfitobAi

cbco

FbC(V

Fig. 6 – Decomposition of cyanobacterial (circles) andChlorophycean (squares) mats in darkened BOD bottlesincubated under anaerobic (closed symbols) and aerobicconditions (open symbols). Decomposition measured as

Aerobic 0.1095

Chlorophycean mat Anaerobic 0.0250Aerobic 0.0066

aining cyanobacteria incubated in situ became anaerobicooner (<24 h) than BOD bottles containing Chlorophyceanats (24–48 h).Cyanobacterial mats decomposed more rapidly than

hlorophycean mats under both anaerobic and aerobic con-itions (Fig. 6). Under anaerobic conditions, no recognizableyanobacterial mat remained in the BOD bottles after 10ays of incubation. Decline of biomass in these bottles sta-ilized at approximately 25% of original mat AFDM (Table 1).his remaining material was assumed mostly bacteria andefractory remains of cyanobacterial cells. Biomass loss foryanobacteria was somewhat slower under aerobic condi-ions, although recognizable mat was still lost by day 16 andnal AFDM after 28 days of incubation was only 19% of the ini-ial amount. In contrast, Chlorophycean mats retained muchf their integrity throughout the experiments and substantialiomass remained at the end of both the aerobic (86% of initialFDM at day 34) and anaerobic (36% of initial AFDM at day 41)

ncubations.Non-linear regression analysis (Eq. (2)) was used to cal-

ulate rate coefficients for periphyton decomposition underoth aerobic and anaerobic conditions (Fig. 6 inset). Data fromyanobacterial mats were limited to the period when rec-gnizable mat was found in the BOD bottles; recognizable

ig. 5 – Dissolved oxygen concentrations in darkened BODottles containing cyanobacterial (circles) andhlorophycean (squares) mats incubated under anaerobic

closed symbols) and aerobic conditions (open symbols).alues represent means of three replicate samples ±1 S.E.

percent loss of initial periphyton ash free dry mass placedin the BOD bottles. Values represent means of threereplicate samples ±1 S.E. Insert shows decomposition rate

coefficients for cyanobacterial (hatched bars) andChlorophycean (gray bars) mats.

material was present throughout the incubation period in thebottles containing Chlorophycean mats. Both anaerobic andaerobic decomposition rates for cyanobacterial mats (0.1408and 0.1095 day−1, respectively; Table 1) were at least fourtimes greater than corresponding rates for Chlorophyceanmats (0.0250 and 0.0066 day−1, respectively). The slower rateof biomass loss under aerobic conditions for each mat typeand the initial increase in biomass in the anaerobic Chloro-phycean treatment prior to the onset of anaerobiosis wasattributed to bacterial growth within the BOD bottles that com-pensated for the loss of periphyton tissue. We assumed thatbacterial growth would be faster under aerobic than anaer-obic conditions. Significant differences in decomposition ratecoefficients were detected between both periphyton mat types(p < 0.001) and oxygen regimes (p < 0.001).

Decomposing cyanobacterial mats released substantiallymore P than Chlorophycean mats (Fig. 7). Mean SRP concentra-tions in BOD bottles with cyanobacterial mats incubated underanaerobic conditions reached ∼3600 �g P L−1 by day 10, and

remained ≥ 2500 �g P L−1 for the remainder of the experiment,while SRP concentrations in aerobic bottles with cyanobacte-rial mats reached ∼2000 �g P L−1 by day 16 (Table 1). In contrast,SRP levels in anaerobic bottles with Chlorophycean mats never

286 e c o l o g i c a l e n g i n e e r i n

Fig. 7 – Changes in soluble reactive phosphorusconcentration in darkened BOD bottles containingcyanobacterial (circles) and Chlorophycean (squares) matsincubated under anaerobic (closed symbols) and aerobic

conditions in the laboratory (open symbols). Valuesrepresent means of three replicate samples ±1S.E.

exceeded 600 �g P L−1 and reached only 37 �g P L−1 in aero-bic bottles. The amount of SRP released from each mat typeduring decomposition was considerably lower under aerobicconditions.

4. Discussion

Evaluations of the efficacy of created wetlands to removenutrients have focused largely on the role of macrophytesand sediments as storage compartments. While the poten-tial importance of periphyton as a short-term sink for P hasbeen acknowledged (Richardson and Craft, 1993), the P stor-age capacity of this compartment is considered small relativeto vascular plants and sediments. Furthermore, it is typicallyassumed that periphyton tissue decomposes rapidly, and thus,that algal P is a labile pool that contributes little to the long-term storage capacity of wetlands (Reddy et al., 1999). Resultsfrom this study indicate that periphyton can play an importantrole in both short-term nutrient removal and storage. Specif-ically, our data show that: (1) periphyton can rapidly removepulses of SRP from the water column; (2) the P storage capac-ity of this compartment is not always quickly saturated, evenin response to repeated, large pulses of P over several weeks;and (3) algal decomposition and P release rates can vary greatlyamong different taxonomic assemblages and are not alwaysfaster than those of wetland macrophytes. These findingssupport the need to examine further the role of periphyton-nutrient fluxes in treatment wetlands and to evaluate current

paradigms for their operation.

Both periphyton assemblages investigated in this studyexhibited rapid P uptake when exposed to the range of Pconcentrations expected during STA operation. The P con-

g 2 7 ( 2 0 0 6 ) 279–289

tent of periphyton in the ENRP generally ranged from 1 to2 mg P g−1 DM, which represented a 30,000 to 60,000-fold nutri-ent bioconcentration relative to the average P content of theambient water (30 �g P L−1 or 0.03 �g P g−1). The capacity ofperiphyton to rapidly remove and P from the water columnand concentrate it in their tissues is supported by other workin the Everglades. Ten days after enclosing plots of Evergladeshabitat and exposing them to a spike of radiolabelled P (32P),Davis (1982) found floating periphyton mats contained 32Pin concentrations at least 30-fold higher on a weight-relatedbasis compared with other ecosystem compartments, includ-ing sediments and various macrophyte tissues. Water columnconcentrations of 32P were considerably lower in enclosedplots that contained a periphyton mat (2–4% of recovered 32Pin the water column) compared to plots without mats (16% ofrecovered 32P in the water column). Noe et al. (2003) and Scintoand Reddy (2003) reported similarly high uptake rates andcapacities to assimilate P in the Everglades. Pietro et al. (2006)demonstrated in a mesocosm study conducted in the ENRPthat the Ceratophyllum/periphyton complex had an exception-ally high affinity for P. All these studies implicate periphytonuptake as an important mechanism for rapid removal of P inwetlands.

Emergent macrophytes and sediments represent the twolargest potential storage compartments for P in most wet-lands, including the STAs. Although periphyton biomass rarelyexceeds macrophyte standing crop, periphyton can still playan important role in P removal in wetland treatment sys-tems because of their high turnover rates. Davis (1982) foundthat floating periphyton mats concentrated 32P at almost 30times the rate of adventitious roots of Typha at P-enrichedlocations in the Everglades. In littoral macrophyte beds, epi-phytic growths of Cladophora (a taxon related to Rhizoclonium)removed P from the water at twice the rate of the host plant,Potamogeton, although the latter accounted for over 90% ofthe biomass (Howard-Williams and Allanson, 1981). Sedimentsorption represents a major sink for P in wetlands with sed-iments high in mineral content (Richardson and Marshall,1986). However, the STAs have been constructed in areas withhighly oxidized peat soils that had been subjected to decadesof fertilization to support agricultural production. The lowP sorption capacity of these enriched sediments (Reddy andGraetz, 1991) suggests that this compartment will not con-tribute significantly to P removal at the low inflow P concen-trations expected during STA operation. Therefore, periphytonlikely represent an important short-term sink for P pulsesmoving through these wetlands during storm events based onour results and those of Davis’ radiotracer studies if conditionsin the STAs permit sufficient periphyton biomass accrual.

In contrast to predictions (e.g., Richardson and Craft, 1993),P loading rates that equaled or exceeded anticipated peakP inflow concentrations to the STAs had little effect on Puptake rates by periphyton mats in our experiments. Periphy-ton biomass increased in enclosures throughout much of thedosing experiment, thereby providing a progressively largerstorage compartment for added P. Decreases in P removal

rates during the course of this experiment were attributed toincreases in areal periphyton biomass on the acrylic rods andconsequent reductions in the rate of nutrient diffusion into themat as demonstrated in other aquatic systems (e.g., Stevenson

g

plants (e.g., Webster and Benfield, 1986; Reddy and Debusk,1991; Enriquez et al., 1993; Chimney and Pietro, 2006) and theseratios are known to vary as a function of taxonomic group andphysiological condition (Tezuka, 1989). The portion of algal

2 7 ( 2 0 0 6 ) 279–289 287

biomass that is refractory differs widely among different taxa(0–86%; Fallon and Brock, 1979; Jewell and McCarty, 1971; Ottenet al., 1992) and influences the rate of decomposition.

Algal and cyanobacterial cells are composed of variousstructural (e.g., cell wall) and non-structural (e.g., storageproducts) components that degrade at varying rates (Moranet al., 1989). Otten et al. (1992) studied the decay kineticsof the cyanobacterium Oscillatoria limnetica and identified alabile fraction, which decomposed completely within 20 days,and a refractory fraction, which comprised 7–24% of the totalbiomass and had a decay rate of only 0.005 day−1. However,because a disproportionate amount of algal cellular P may bepresent in labile forms, first-order kinetics (exponential decay)may at best only approximate P release rates. Although theP fractions of ENRP periphyton assemblages were not deter-mined, previous studies of nutrient losses from plant tissuesfound that rates of P release during decomposition are consid-erably higher than for C or N (Toth, 1987; Reddy et al., 1994).Therefore, short-term studies such as our experiments shouldprovide reasonably accurate estimates of P release.

Algal decomposition tends to be faster than rates formacrophytes, although some overlap is apparent. Emergentmacrophytes decayed more slowly than either phytoplankton(Chlorella) or macroalgae (Ulva) in a temperate estuary (Twilleyet al., 1986). The slower rate of plant decay was associatedwith a higher proportion of structural tissue (e.g., celluloseand lignin) and a higher C:N ratio compared with the algae.Davis (1991) found rates of decay of Cladium jamaicense andTypha domingensis leaves to be extremely slow in the Ever-glades, with approximately half of the initial litter remainingafter 2 years of decomposition. Leaf tissue C:N ratios for Cla-dium and Typha in the habitats used by Davis (1982) averaged70:1 and 58:1, respectively (Miao, SFWMD, personal communi-cation). Decomposition rates for Sagittaria lancifolia (C:N = 13:1)and Typha spp. (C:N = 26:1) in a eutrophic constructed marshin central Florida were much faster, with 50% of the biomassbeing lost in 5–11 and 25–48 days, respectively (Reddy et al.,1994). The mean half-life of Najas guadalupensis/Ceratophyllumdemersum (C:N = 14:1), Pistia stratiotes (C:N = 18:1), Eichhorinacrassipes (C:N = 32:1) and Typha spp. (C:N = 69:1) litter incu-bated in decomposition bags within the ENRP was 12, 14, 36and 117 days, respectively (Chimney and Pietro, 2006). In thepresent study, 50% of the biomass of anaerobically decompos-ing cyanobacteria (C:N = 9:1) and Chlorophycean (C:N = 20:1)mats was lost in 4 and 35 days, respectively. Thus, the rateat which algae decompose is not uniformly faster than formacrophytes; rather, the relative rates of plant and algaldecomposition appear to be influenced by tissue elementalcontent.

The quantitative importance of periphyton in wetlandsnutrient cycling ultimately is constrained by the ability to pro-duce biomass, which is in turn affected by the availabilityof nutrients, light, and suitable substrata for growth. Algaluptake may represent an important biological mechanismfor rapid P removal in treatment wetlands that receive largeamounts of P over short time periods as stormwater pulses.

e c o l o g i c a l e n g i n e e r i n

and Glover, 1993). However, rates of P removal were similar inenriched and unenriched enclosures throughout most of the7-week experiment, including the final sampling date despitea doubling of the P content of the enriched periphyton. Mostinflow to the STAs is from intermittent stormwater discharges,and our findings, while preliminary, indicate that the capac-ity of the periphyton to remove incoming P does not diminisheven after successive nutrient additions.

Both the periphyton N:P molar ratios and the lack ofdifference in periphyton biomass and productivity betweenenriched and unenriched enclosures indicate that P was notlimiting to periphyton growth in the ENRP during our exper-iments. Thus, increased periphyton P content may resultfrom the accumulation of intracellular storage compounds(e.g., polyphosphate bodies) and/or increased precipitation ofP with carbonate on the surface of the mat. Luxury storesof P can accumulate and be released much faster than Puptake related to growth (Shapiro, 1967). In contrast, P removalthrough coprecipitation with carbonate onto periphyton sur-faces is largely irreversible. The ash content of the periphytonassemblages studied here was well within the range of val-ues reported for Everglades calcareous periphyton (49–81%,Browder et al., 1982) and suggests that some P precipitationoccurred on ENRP periphyton. Clearly, the form of P accumu-lated by periphyton will determine its fate within a wetlandand may influence P retention efficiency of the STAs. Furtherstudy is required to determine the significance of different Premoval processes mediated by periphyton and the residencetimes of these various forms of P in these treatment wetlands.

Net P retention by periphyton is an equilibrium betweenP uptake and incorporation into tissue growth versus cellulardecomposition and P release back to the water column. Therate of periphyton decomposition in our experiments variedwidely (k = 0.0066–0.1408 day−1) depending on the taxonomicassemblage and environmental conditions (Table 1). Theserates were within the range of values reported for other algae.Decomposition rates of the unicellular green alga Chlorella var-ied between 0.02 and 0.09 day−1 in laboratory cultures inoc-ulated with a natural bacterial assemblage (DePinto, 1974).Decomposition rates of three green algal taxa ranged between0.08 and 0.2 day−1 (Sudo et al., 1978). First-order decompositionrates of Anabaena in Lake Mendota, Wisconsin (USA) rangedbetween 0.005 and 0.11 day−1 (Fallon and Brock, 1979). Fiftypercent of the dry mass of the estuarine alga Ulva was lostfrom decomposition bags after approximately 35 days, yield-ing a short-term loss rate of 0.014 day−1 (Twilley et al., 1986).Algal decomposition rates can vary substantially dependingupon taxon, physiological condition and water quality condi-tions (Rodgers, 1979).

The rapid decomposition of the cyanobacterial mat, andsubsequent release of P back to the water in this study, wasconsistent with its lower C:N and C:P ratios, compared withslower decomposition rates and higher C:N and C:P ratios forChlorophycean mats.

Numerous studies have found a negative correlationbetween decomposition rate and the C:N ratio of aquatic

However, rates of periphyton P cycling and retention can varysubstantially depending on taxonomic composition. Furtherwork is warranted to determine P storage potential of periphy-ton, including a more thorough understanding of P loss rates

r i

r

288 e c o l o g i c a l e n g i n e e

and the amount of P stored in different organic and inorganic(e.g., carbonate) fractions within algal mats.

Acknowledgments

Chad Kennedy helped design and install the ENRP meso-cosms. John Backus, Chad Kennedy, Jim Laing, and Sue New-man assisted with field collections. Tom Fontaine, Sue New-man, Ramesh Reddy, Dave Rudnick, Al Steinman, and threeanonymous reviewers provided substantive comments thatimproved drafts of this manuscript.

e f e r e n c e s

APHA, 1989. Standard Methods for the Examination of Water andWastewater, 17th ed. American Public Health Association,American Water Works Association and Water PollutionControl Federation, Washington, DC.

Browder, J., Cottrell, D., Brown, M., Newman, M., Edwards, R.,Yuska, J., Browder, M., Krakoski, J., 1982. Biomass and primaryproductivity of microphytes and macrophytes in periphytonhabitats of the southern Everglades. Report T-662. SouthFlorida Research Center, Homestead, FL.

Chimney, M.J., Goforth, G., 2006. History and description of theEverglades Nutrient Removal Project, a subtropicalconstructed wetland in south Florida (USA). Ecol. Eng. 27,268–278.

Chimney, M.J., Pietro, K.C., 2006. Decomposition of macrophytelitter in a subtropical constructed wetland in south Florida(USA). Ecol. Eng. 27, 301–321.

Chimney, M.J., Wenkert, L., Pietro, K.C., 2006. Patterns of verticalstratification in a subtropical constructed wetland in southFlorida (USA). Ecol. Eng. 27, 322–330.

Davis, S.M., 1982. Patterns of Radiophosphorus Accumulation inthe Everglades After its Introduction into Surface Water. Tech.Pub. 82-2, South Florida Water Management District, WestPalm Beach, FL.

Davis, S.M., 1991. Growth, decomposition, and nutrient retentionof Cladium jamaicense Crantz and Typha domingensis Pers. in theFlorida Everglades. Aquat. Bot. 40, 203–224.

DePinto, J.V., 1974. Studies on phosphorus and nitrogenregeneration: the effect of aerobic bacteria on phytoplanktondecomposition and succession in freshwater lakes. Ph.D.Dissertation. University of Notre Dame, Notre Dame, IN.

Enriquez, S., Duarte, C.M., Sand-Jensen, K., 1993. Patterns indecomposition rates among photosynthetic organisms: theimportance of detritus C:N:P content. Oecologia 94, 457–471.

Fallon, R.D., Brock, T.D., 1979. Decomposition of blue-green algal(cyanobacterial) blooms in Lake Mendota, Wisconsin. Appl.Environ. Microbiol. 37, 820–830.

Guardo, M., Fink, L., Fontaine, T.D., Newman, S., Chimney, M.,Bearzotti, R., Goforth, G., 1995. Large-scale constructedwetlands for nutrient removal from stormwater runoff: anEverglades restoration project. Environ. Manage. 19, 879–889.

Howard-Williams, C., Allanson, B.R., 1981. Phosphorus cycling ina dense Potamogeton pectinatus L. bed. Oecologia 49, 56–66.

Jewell, W.J., McCarty, P.L., 1971. Aerobic decomposition of algae.Environ. Sci. Technol. 5, 1023–1031.

Kadlec, R.H., Knight, R.L., 1996. Treatment Wetlands. Lewis

Publishers, Boca Raton, FL.

Kleinbaum, D.G., Kupper, L.L., 1978. Applied Regression Analysisand Other Multivariable Methods. Duxbury Press, Boston, MA.

Knight, R.L., 2003. Efficacy of periphyton-dominated wetlands forphosphorus removal in the Florida Everglades. In: Vymazal, J.

n g 2 7 ( 2 0 0 6 ) 279–289

(Ed.), Wetlands: Nutrients, Metals and Mass Cycling. BackhuysPublishers, Leiden, The Netherlands, pp. 87–106.

McCormick, P.V., O’Dell, M., 1996. Quantifying periphytonresponses to phosphorus in the Florida Everglades: asynoptic-experimental approach. J. N. Am. Benthol. Soc. 15,450–468.

Moran, M.A., Benner, R., Hodson, R.E., 1989. Kinetics of microbialdegradation of vascular plant material in two wetlandecosystems. Oecologia 79, 158–167.

Newman, S., Pietro, K., 2001. Phosphorus storage and release inresponse to flooding: implications for Everglades stormwatertreatment areas. Ecol. Eng. 18, 23–38.

Nichols, D.S., 1983. Capacity of natural wetlands to removenutrients from wastewater. J. Water Pollut. Control Fed. 55,495–505.

Noe, G.B., Scinto, L.J., Taylor, J., Childers, D.L., Jones, R.D., 2003.Phosphorus cycling and partitioning in an oligotrophicEverglades wetland ecosystem: a radioisotope tracing study.Freshwater Biol. 48, 1993–2008.

Otsuki, A., Wetzel, R.G., 1972. Coprecipitation of phosphate withcarbonates in a marl lake. Limnol. Oceanogr. 17, 763–767.

Otten, J.H., Gons, H.J., Rijkeboer, M., 1992. Dynamics ofphytoplankton detritus in a shallow eutrophic lake(Lake Loosdrecht, The Netherlands). Hydrobiologia 233,61–67.

Pietro, K.C., Chimney, M.J., Steinman, A.D., 2006. Phosphorusremoval by the Ceratophyllum/periphyton complex ina south Florida (USA) freshwater marsh. Ecol. Eng. 27,290–300.

Reddy, K.R., D’Angelo, E., Fisher, M.M., Ann, Y., Olila, O.G.,Schipper, L.A., 1994. Nutrient storage and movement in theLake Apopka marsh. Final Report to the St. John’s WaterManagement District, Palatka, FL.

Reddy, K.R., Debusk, W.F., 1991. Decomposition of water hyacinthdetritus in eutrophic lake water. Hydrobiologia 211,101–109.

Reddy, K.R., Graetz, D.A., 1991. Phosphorus dynamics in theeverglades nutrient removal system. Final Report Submittedto the South Florida Water Management District, West PalmBeach, FL.

Reddy, K.R., Kadlec, R.H., Flag, E., Gale, P.M., 1999. Phosphorusretention in streams and wetlands: a review. Crit. Rev.Environ. Sci. Tech. 29, 83–146.

Reed, S.C., Crites, R.W., Middlebrooks, E.J., 1995. Natural Systemsfor Waste Management and Treatment, second ed.McGraw-Hill, Inc., New York, NY.

Richardson, C.J., Craft, C.B., 1993. Effective phosphorus retentionin wetlands: fact or fiction? In: Moshiri, G.A. (Ed.), ConstructedWetlands for Water Quality Improvement. Lewis Publishers,Boca Raton, FL, pp. 271–282.

Richardson, C.J., Marshall, P.E., 1986. Processes controllingmovement, storage, and export of phosphorus in a fenpeatland. Ecol. Monogr. 56, 279–302.

Richardson, C.J., Qian, S., Craft, C.B., Qualls, R.G., 1997. Predictivemodels for phosphorus retention in wetlands. Wet. Ecol.Manage. 4, 159–175.

Rodgers, P.W., 1979. Kinetic Studies of PhytoplanktonDecomposition. Ph.D. Dissertation. Clarkson College ofTechnology, Omaha, NB.

Scinto, L.J., Reddy, K.R., 2003. Biotic and abiotic uptake andrelease of phosphorus by periphyton in a subtropicalfreshwater wetland. Aquat. Bot. 77, 203–222.

Shapiro, J., 1967. Induced rapid release and uptake of phosphateby microorganisms. Science 155, 1269–1271.

Stevenson, R.J., Glover, R., 1993. Effects of algal density and

current on ion transport through periphyton communities.Limnol. Oceanogr. 38, 276–1281.

Stumm, W., Morgan, J.J., 1996. Aquatic Chemistry, third ed. JohnWiley and Sons, Inc., New York, NY.

i n g

S

T

T

T

e c o l o g i c a l e n g i n e e r

udo, R., Ohtake, H., Aiba, S., Mori, T., 1978. Some ecologicalobservations on the decomposition of periphytic algae andaquatic plants. Water Res. 12, 179.

ezuka, Y., 1989. The C:N:P ratio of phytoplankton determines therelative amounts of dissolved inorganic nitrogen andphosphorus released during aerobic decomposition.Hydrobiologia 173, 55–62.

oth, L.A., 1987. Effects of Hydrologic Regimes on LifetimeProduction and Nutrient Dynamics of Sawgrass. Tech. Pub.87-6, South Florida Water Management District, West Palm

Beach, FL.

willey, R.R., Ejdung, G., Romare, P., Kemp, W.M., 1986. Acomparative study of decomposition, oxygen consumptionand nutrient release for selected aquatic plants occurring inan estuarine environment. Oikos 47, 190–198.

2 7 ( 2 0 0 6 ) 279–289 289

USEPA, 1983. Methods for Chemical Analysis of Water andWastes. U.S. Environmental Protection Agency. EnvironmentalMonitoring and Support Laboratory, Cincinnati, OH.

USEPA, 1986. Test Methods for Evaluating Solid Waste. Physicaland Chemical Methods, SW-846, third ed. U.S. EnvironmentalProtection Agency, Washington, DC.

Verhoeven, J.T.A., Meuleman, A.F.M., 1999. Wetlands forwastewater treatment: opportunities and limitations. Ecol.Eng. 12, 5–12.

Webster, J.R., Benfield, E.F., 1986. Vascular plant breakdown in

freshwater ecosystems. Annu. Rev. Ecol. Syst. 17, 567–594.

Wetzel, R.G., Likens, G.E., 2000. Limnological Analyses, third ed.Springer-Verlag, New York, NY.

Zar, J.H., 1999. Biostatistical Analysis, fourth ed. Prentice Hall,Upper Saddle River, NJ.