S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 4 ( 2 0 0 8 ) 2 6 9 – 2 7 5

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r. com/ loca te / sc i to tenv

Sediment geochemistry of Al, Fe, and P for two historicallyacidic, oligotrophic Maine lakes

Tiffany A. Wilsona,⁎, Stephen A. Nortonb, Bjorn A. Lakea, Aria Amirbahmana

aDepartment of Civil and Environmental Engineering, Boardman Hall, University of Maine, Orono, Maine 04469 USAbDepartment of Earth Sciences, Bryand Global Sciences Center, University of Maine, Orono, Maine 04469 USA

A R T I C L E I N F O

⁎ Corresponding author. Tel.: +1 207 581 3401E-mail address: Tiffany.Wilson@umit.main

0048-9697/$ – see front matter © 2008 Elsevidoi:10.1016/j.scitotenv.2008.06.061

A B S T R A C T

Available online 28 August 2008

Phosphorus (P) may be liberated from lake sediments by reductive dissolution of Fe(OH)3(S)during periods of hypolimnetic anoxia. P, however, remains adsorbed to Al(OH)3(S)regardless of redox conditions. During chronic or episodic acidification of a catchment,ionic Al is mobilized from soils to receiving waters. A fraction of the mobilized Al mayprecipitate as a consequence of higher pH of the receiving waters. We hypothesized thatphosphorus retention in lake sediments is directly related to the magnitude of Al loading inresponse to low pH in the watershed. We studied cores representing over 200 years ofsediment accumulation in historically acidic Mud Pond and Little Long Pond in easternMaine, USA. Sequential chemical extractions of sediment were used to assess the history ofAl, Fe, and P interactions. Mud Pond is a first-order pond with a pH of ~4.7, having acidifiedslightly in response to anthropogenic acidification from ~1930. The inlet stream to MudPond has dissolved Al concentrations often exceeding 500 μg/L, of which more than half isorganically-bound. Mud Pond drains into Little Long Pond, a second-order pond with ahistorical pH of b6, and which has shown little pH or alkalinity response to increases ordecreases in atmospheric SO4

2− input.Sequential extractions show that Al and P are predominantly in the 0.1 M NaOH-extractablefraction in the sediments from both ponds throughout the cores. The concentration of thelikely biogenic and non-reactive P within the NaOH fraction increases up core from b30% to~60%. Extractable Fe (b20% of extractable Al) is mainly in the 0.1 M NaOH-extractablefraction, except for the top few cm, which are predominantly in the bicarbonate-dithionitereducible fraction. Accumulation rates of sediment, Al, Fe, and P in both ponds haveincreased in the last 50–60 yr, but fractions remain in the same proportion. Throughout bothsediment cores the molar ratio of specific Al:P fractions greatly exceeds 25, and molar ratioof specific Al:Fe fractions greatly exceeds 3, the thresholds proposed by Kopáček et al.[Kopáček J, Borovec J, Hejzlar J, Ulrich K-U, Norton SA, Amirbahman A. Aluminum control ofphosphorus sorption by lake sediments. Environ Sci Technol 2005; 39: 8784–89.] for P releaseduring anoxia. The data illustrate a continuous association of P with Al in both ponds duringthe last two centuries, likely due to the persistent natural acidity of the catchments.

© 2008 Elsevier B.V. All rights reserved.

Keywords:AluminumIronLake sedimentOligotrophyPhosphorusSediment fractionation

; fax: +1 207 581 3888.e.edu (T.A. Wilson).

er B.V. All rights reserved.

270 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 4 ( 2 0 0 8 ) 2 6 9 – 2 7 5

1. Introduction

Chronic and episodic acidification greatly impacts the chem-istry of freshwater catchments and may have an effect onnutrient cycling in lakes (Steinberg and Wright, 1994).Phosphorus (P) concentrations and bioavailability are well-studied for surface waters because P is typically the limitingnutrient for primary productivity, and therefore, controls thetrophic status of a lake (Schindler, 1977). P is transported tolakes by catchment drainage, generally associated with solidphases of iron (Fe), aluminum (Al), calcium (Ca), and humicsubstances (Ulrich and Pöthig, 2000a). Lakes that develophypolimnetic anoxia during summer stratification mayexperience P release from bottom sediments due to reductivedissolution of Fe(III) and associated P (Amirbahman et al.,2003), and/or possible release of stored polyphosphate frombacteria under anaerobic conditions (Hupfer and Gächter,1995). Lakes with elevated inputs of Al due to acidification arelikely to precipitate Al as Al(OH)3, because alkalinity produc-tion in the lake and/or mixing of waters results in an increasein pH and a decrease in the solubility of Al(OH)3, which mayresult in the removal of P from the water column (Kopáčeket al., 2000). Precipitated Al may also originate from photo-oxidation of dissolved organically-bound Al (Kopáček et al.,2006). Precipitation of Al and associated P is pH-dependent butnot redox-dependent. Therefore, Al immobilizes P even duringanoxia, whereas Fe does not (Ulrich and Pöthig, 2000b).

There are limited data that describe the effect of aciddeposition on P concentrations in surface waters. Ogburn andBrezonik (1986) explained the oligotrophication hypothesis(low P in the water column of acidic lakes) as being due to anoptimum P sorption by sediment at pH 4.0–5.5. Huser andRydin (2005) describe an example of acido-oligotrophicationresulting from increased P immobilization by complexationand precipitation of Al–P in the sediment of Lake Gårdsjön,Sweden, from sediment dated 1950 to 2001, where theincreased P immobilization corresponded to a decrease in

Fig. 1 –Sequential extraction method for lake sediment based on

lake pH and increase in Al input. Ulrich and Pöthig (2000a)similarly found that sediment P in reservoirs with acidifiedcatchments was mainly associated with Al. Kopáček et al.(2005) developed an empirical relationship from 43 lakesediment studies, which sets a threshold for release ofsediment P based on ratios of operationally-defined molarfractions of Al, Fe, and P. The fractions are characterized bytheir solubility during a sequential chemical extractionprocedure based on that of Psenner and Pucsko (1988),shown in Fig. 1. The ratios, (AlNaOH~25 : FeBD) N3 or (AlNaOH~25 :P(NH4Cl+BD)) N25, define the thresholds abovewhich P is unlikelyto be released from sediment during hypolimnetic anoxia(Kopáček et al., 2005). Recently, Lake et al. (2007) demonstratedthe validity of the proposed model by Kopáček et al. (2005) inthe field.

2. Materials and methods

2.1. Site description

Sediment cores were collected in March 2005 from Mud Pond(MP) and downstream Little Long Pond (LLP), in eastern Maine,USA (Fig. 2). MP (N44.634°, W68.088°, elevation 102 m) is a1.6 ha, first-order pondwith a total catchment drainage area of0.65 km2. LLP (N44.638°, W68.080°, elevation 72 m) is a 24.3 ha,second-order pond with a total catchment drainage area of2.36 km2, including drainage from MP. Both catchments arecurrently forested (spruce, fir, and red oak) and undeveloped,though there are records of logging and forest fire prior to 1905(Davis et al., 1994). MP is 16.2 m deep and has a residence timeof ca. 1 month. LLP is 25.3 m. deep and has a residence time ofca. 1 year (Maine DEP, 2006). Water residence time is pondvolume divided by the average annual rainfall on thecatchment, including the lake surface. Both ponds developanoxic hypolimnia during summer and winter density strati-fication, and are typically ice-covered from December to April.Both catchments contain steep slopes with coarse, shallow,

Psenner and Pucsko (1988) and Hieltjes and Lijklema (1980).

Fig. 2 –Little Long Pond and Mud Pond site locations.

271S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 4 ( 2 0 0 8 ) 2 6 9 – 2 7 5

organic-rich soils overlying thin granitic till and some exposedgranitic bedrock. Thus surface waters are poorly buffered andacid-sensitive.

Chemistry for MP and LLP and the analytical methods arein Table 1. Only intermittent data exist on the chemistry andhydrology of the inlets and outlets. MP is an oligo-meso-trophic, naturally acidic, and fishless lake, with a pH b5. Thelow pH is caused primarily by organic acids (up to 8 mg/Ldissolved organic carbon, DOC) from catchment drainage, andsecondarily by excess SO4

2− from atmospheric deposition(Davis et al., 1994). Al is mobilized by low pH catchmentinputs, is largely complexed with organic acids, and reachesconcentrations up to 500 μg/L in the water column of MP. TotalP (TP) is typically b10 μg/L in the summer epilimnion and up to18 μg/L in the hypolimnion. During late summer anoxia,substantial Fe and SO4

2− reduction generates alkalinity in thehypolimnion, though the effect on pH is minimal perhaps dueto buffering by the dissolved organic matter. LLP is oligo-trophic and naturally acidic with a pH ~6, and has lower DOCthan MP. LLP has shown little to no pH response to anthro-pogenic atmospheric input. Al concentrations in the water

Table 1 – Ranges of annual mean values (1983–2005) for selecte(MP) ponds

LLP(epilimnion)

LLP(hypolimni

pH 5.47–6.04 5.19–5.61ANC, μeq/L 1.1–16.9 7.9–20.3SO4

2−, μeq/L 59–80 62–79DOC, mg/L 0.8–2.4 0.8–3.1Al, μg/L 10–71 46–84P, μg/L 1–4 5–12Fe, μg/L 4 29

Methods included gran titration for ANC (acid neutralizing capacity); ion cFe (total dissolved) by ICP-AES or graphite furnace atomic absorption (pre-values were only measured in 2005 (Maine DEP, 2006).

column are historically b100 μg/L in both the epilimnion andhypolimnion, though the major inlet concentration is in thehundreds of μg/L; TP is b5 μg/L in the epilimnion and inlets,and up to 12 μg/L in the hypolimnion of LLP. Hypolimneticanoxia in LLP results in little Fe or SO4

2− reduction, and nosignificant alkalinity or pH change.

2.2. Sediment collection and processing

Sediment was obtained at points 13 m deep in MP and 24 mdeep in LLP during ice-cover (March 2005), using a 10 cmdiameter, stationary piston corer with an acrylic core tube(Davis and Doyle, 1967). Cores were taken near but not at thedeepest points of the ponds in order to avoid samplingpreviously disturbed sites, e.g., those of Davis et al. (1994).Sediment cores were sectioned at the lakes using stainlesssteel and plastic ware. Sediment was sectioned in the field in0.5 cm intervals from 0 to 10 cm, in 1 cm intervals from 10 to30 cm, and in 2 cm intervals from 30+ cm. Samples wereplaced directly intoWhirlpack™ bags and stored at 4 °C in thedark.

d water chemistry parameters of Little Long (LLP) and Mud

on)MP

(epilimnion)MP

(hypolimnion)

4.61–4.87 4.61–5.08(−27.2)–(−8.2) (−29.8)–20.372–105 56–1102.9–5.8 3.4–6.3235–433 310–4394–8 7–1733 514

hromatography for SO42−; DOC (dissolved organic carbon) by IR; Al and

1999); P (total) by persulfate digestion/molybdate-blue colorimetry. Fe

Fig. 4 –Sedimentation rate (g/cm2/year) in Little Long (LLP)and Mud (MP) ponds, Maine, calculated using 210Pb dating.Sedimentation rates are considered to be constant ~pre-1900.The dashed line indicates the extent of 210Pb dating; datesprior to ~1850 are extrapolated. Symbols represent datapoints.

272 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 4 ( 2 0 0 8 ) 2 6 9 – 2 7 5

At the laboratory, 0.5 g wet sediment aliquots were takenfrom several intervals for diatom analysis, and 10 to 20 galiquots were dried at 100 °C in ceramic crucibles for percentwater determination. The remaining sediment was frozenat −20 °C until processed for further analysis. Dried sedimentaliquots were homogenized with ceramic mortar and pestleand percent loss on ignition (%LOI) was determined by heating~0.5 g of dried sediment at 550 °C for 4 h. Dry sediment wasdated by analyzing the gamma activity of 210Pb. The constantrate of supply (CRS) model of Appleby and Oldfield (1978) wasused to calculate the age of the midpoints (Fig. 3) of thesediment increments based on determination of supportedand unsupported 210Pb activities.

2.3. Sediment chemistry

Sequential extraction (Fig. 1) was performed on 1 to 2 g of wetsediment to determine operationally-defined fractions of Al,Fe, and P: A) 1 M NH4Cl, pH 7, at 25 °C for 1 h to obtain theexchangeable fraction, B) 0.11 M bicarbonate-dithionite (BD)at 40 °C for 30 min to release the reducible portion that isprimarily Fe and Mn (oxyhydr)oxides and associated P, C)0.1 M NaOH at 25 °C for 16 h to dissolve Al and Fe(III)hydroxides and oxyhydroxides and associated P, as well asorganic-P, D) 0.5MHCl at 25 °C for 16 h to dissolve acid-solubleminerals, and E) 1 M NaOH at 85 °C for ≥24 h to solubilizeresidual material. The procedure by Psenner and Pucsko(1988) has been utilized in several similar studies (e.g., Huserand Rydin, 2005; Coolidge, 2004; Kopáček et al., 2000) using 1MNaOH. We found that 0.1 M NaOH, based on comparableprocedures of Hieltjes and Lijklema (1980), Ahlgren et al.(2005), and others, has a sufficiently high pH to accomplishthe third fractionation step. Total extractable concentrationsare the sum of concentrations from the five sequentialextractions.

Extractions were performed on 21 sediment core incre-ments from LLP and 22 from MP, with higher frequency in thetop 10 cm of the core. Batches of 16 samples were extracted

Fig. 3 – 210Pb dating profiles for Little Long (LLP) and Mud (MP)ponds, Maine. Vertical bars are 210Pb counting errors.Horizontal bars represent the thickness of the sedimentincrement.

together and each batch contained a duplicate sample and areagent blank. For each step, the extraction solution wasadded to the wet sediment in a 50 mL plastic centrifuge tube,capped and shaken in a water bath at the desired temperatureand for the appropriate duration, and centrifuged at~3000 rpm for 15 min. The supernatant was collected into apolypropylene bottle. The same sediment was then rinsedwith extraction solution, centrifuged, decanted and mixedwith the supernatant collected previously.

Concentrations of Al and Fe in the extracts weredetermined using inductively coupled plasma atomic emis-sion spectrometry (ICP-AES; Perkin-Elmer model 3300XL).Total P (TP) was analyzed by ICP-AES (for the BD, NaOH25,HCl, and NaOH85 solutions) and by the molybdate-bluespectrophotometric method (Murphy and Riley, 1962) (forthe NH4Cl extraction). The NaOH25 fraction was also ana-lyzed using the molybdate-blue method for reactive P(NaOH25-rP), which is the P in solution as orthophosphateand is assumed to be mainly associated with sediment Al.The difference between NaOH25-TP and NaOH25-rP isNaOH25-nrP (non-reactive P) (Psenner and Pucsko, 1988).NaOH25-nrP is likely the organic-P and bacteria-incorporatedP fraction (Ahlgren et al., 2005).

3. Results

Using 210Pb dating (Fig. 3) we calculated sediment massaccumulation rates in LLP and MP during the last ~100 yr.Lack of precision of 210Pb dates in older sediment requires us toassume a constant (baseline) accumulation rate prior to thistime period, based on the lower part of the dated record.Accumulation rate is higher in LLP than in MP as describedpreviously by Norton et al. (1992). Integrated unsupported210Pb and stable anthropogenic Pb in MP are both about 0.1 ofthat of LLP, suggesting substantial bypassing of sediment in

Table 2 – Percent of total extractable Al, Fe, TP and rP inthe individual extraction steps, averaged for all sedimentincrements

NH4Cl BD NaOH25 HCl NaOH85

Little Long PondAl b1 1 80 (87–83) 12 (7–9) 7 (6–7)Fe b1 37 (32–57) 48 (50–35) 9 (13–5) 6 (5–2)TP b1 0 97 1 2rP ND ND 41 (58–25) ND ND

Mud PondAl b1 2 (0.5–6) 79 (76–80) 7 (2–9) 12 (14–12)Fe b1 32 (37–33) 31 (26–34) 21 (21–19) 15 (16–13)TP b1 3 95 (96–91) 1 1rP ND ND 35 (61–12) ND ND

Numbers in parentheses are ranges of %TE from bottom interval totop interval. ND=no data.

273S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 4 ( 2 0 0 8 ) 2 6 9 – 2 7 5

MP. Both ponds exhibit increasing sedimentation rate sincethe 1950s (Fig. 4). Total extractable Al, Fe and P also showincreases in accumulation rate over this time period, butwhennormalized for mass accumulation rate, the resulting flux ofthese elements shows a much less pronounced increase, ifany (Fig. 5). LOI is higher in MP than in LLP, with an average of52% and 41% in the two lakes, respectively. LOI quantifies theloss of organicmatter andwater fromdehydration of Fe and Alhydroxides in the sediment. LOI is relatively constantthroughout both cores, excluding a few deeper intervals (24–26 cm) of MP, where the LOI is distinctly lower. The 210Pb datesof these intervals were extrapolated to be earlier than 1600.Those same intervals also have lower water content thansurrounding intervals.

In both cores, the highest percentage of extractable Al andP are in the NaOH25 fraction. Table 2 summarizes theproportions of total extractable Al, Fe, and P (fractionatedinto rP and nrP) in each sequential extraction fraction. LLPand MP exhibit similar fractionation patterns with approxi-mately 80% of Al and ≥95% of P in the NaOH25 fraction. Theremaining ~20% of Al is nearly evenly divided between themineral (HCl soluble) and the recalcitrant (NaOH85) fractions.A negligible proportion of P is found in the other fractions,notably, P associated with the BD fraction that is releasedunder anoxic conditions (Coolidge, 2004; Kopáček et al., 2005).Both the rP and nrP fractions are significant in the NaOH25

extract. In both LLP and MP cores, the proportion of nrPdecreases and rP increases with increasing sediment age

Fig. 5 –Flux, normalized to sedimentation rate, of totalextractable Al, Fe and P in Little Long Pond (LLP) and MudPond (MP) based on sequential extraction chemistry fromsediment cores collected in March of 2005.

(Fig. 6). Reducible Fe (BD-soluble) is highest in the top few cmof sediment. In deeper sediment, most Fe occurs in theNaOH25 fraction. Reducible Fe is likely a transient that occursat the sediment–water interface. Compared to MP, the flux oftotal extractable elements to LLP sediment is higher for Al (2–5 times), Fe (10 times), and P (3–4 times). The molar ratio of Alas Al (oxy)hydroxides (AlNaOH25) to labile and reducible P

Fig. 6 –Sediment phosphorus from Little Long (LLP) and Mud(MP) Ponds extracted by 0.1 M NaOH at 25 °C as the third stepof a sequential extraction based on Psenner and Pucsko(1988) and Hieltjes and Lijklema (1980). The reactive portion(NaOH(25 °)-rP) is ortho-P that is potentially complexed withAl as Al–P. The non-reactive portion (NaOH(25 °)-nrP) isbiogenic-P.

Fig. 7 –Molar ratio (dry weight basis) of AlNaOH25 to PNH4Cl+BD

(upper graph) and AlNaOH25 to FeBD (lower graph) extractedfrom dated sediment intervals from Little Long (LLP) andMud(MP) ponds. The estimated minimum threshold ratios for Pretention by sediment during anoxia are 25 and 3,respectively (Kopáček et al., 2005) and are indicated by theheavy horizontal lines. These oligotrophic systems exceedthe minimum ratios through time. The dashed line indicatesthe extent of 210Pb dating, and dates prior to ~1850 areextrapolated.

274 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 4 ( 2 0 0 8 ) 2 6 9 – 2 7 5

(PNH4Cl+BD) in the sediment exceeds 25 throughout the coresfromboth ponds (Fig. 7). There is a 30–50%decline in that ratioin sediment younger than 1970.

4. Discussion

The chronic historical acidity of the Mud and Little Long Pondcatchments proposed by Davis et al. (1994) has maintainedconditions for high Al mobility throughout the period of timeexamined in our cores. Despite some recovery from atmo-spheric SO4

2− input since the Davis study, both ponds remain ata low pH. Historically small loads of P to these ponds make itdifficult to assess any changes in P flux to the ponds related toacidification. We focus on the chemical associations of Al, Feand P deposited to the sediment over the last two centuries.

Both ponds show a clear historical relationship betweensediment Al and P as indicated by concentrations of the two

elements in the operationally-defined extractionmethod. Thedata support the oligotrophication hypothesis mentionedearlier. The flux of Al to the profundal sediments of bothponds is several times higher than those of Fe and P,suggesting that removal of P from the water column andstorage in sediment may be dominated by Al. The pH of thetwo ponds differs, but the mechanisms of P removal, reten-tion, and burial by Al(OH)3 seem to be similar. Mud Pond has apH b5, but has a greater input of inorganic Al which mayprecipitate as Al(OH)3 and remove orthophosphate from thewater column (Kopáček et al., 2006). Although LLP has lowerinput concentrations of inorganic Al, the hypolimnetic pH of 5to 6 is likely to precipitate Al as Al(OH)3 and remove orthopho-sphate. Allochthonous and autochthonous sources of P to bothponds are historically low, suggesting that the small pool ofbiologically available P is readily controllable by Al.

The increase in sedimentation rates (Fig. 4) was notexpected because the catchments have remained relativelyundisturbed by humans for a century. A non-anthropogenicdisturbance such as the beaver dam at the major outlet of MPwould increase the water residence time and affect sedimen-tation but this explanation does not apply to LLP. A systematicbias of 210Pb activity could skew the results but the bias wouldhave to apply to both cores. We do not have a satisfactoryexplanation for the increasing mass accumulation rates.

Reduction and oxidation of Fe occurs in the seasonalhypolimnia of these ponds and is likely involved with a higherproportion of P release than of P capture. LLP sediment almostreaches the threshold for P release because of significantamounts of reducible Fe in the uppermost layers. However,any P that is mobilized with Fe(III) reduction is likely adsorbedby abundant Al. The top few cms of sediment become Fe-enriched with oxidative precipitation of previously mobilizedFe(II), after the lake turnover.

Declines in the AlNaOH25:PNH4Cl+BD ratios (Fig. 7) suggest adecrease in inorganic Al, and/or an increase in labile P fromthe catchment during the past ~20 yr. Early 20th century landdisturbances (fire and logging) and subsequent years ordecades of catchment response may be factors that haveaffected the mobilization of Al by altering the proportionof inorganic Al to Al-organic complexes. Changes in Al andP fluxes are masked by the increase in sedimentation rates.Al(OH)3 concentrations in the sediment are high enoughrelative to P that a small decrease of Al input would notchange the mechanisms controlling P release.

The increase in the proportion of biogenic P (NaOH25-nrPused as a proxy) compared to orthophosphate (NaOH25-rP) inyounger sediments (Fig. 6) suggests a higher input of organic-Pin recent sediments or diageneticmineralizationofP, increasingwith age (depth). The half life of these biogenic-P compounds isin the range of 10 to 20+ years (Ahlgren et al., 2005).

5. Conclusions

Sediments from MP and LLP contain higher concentrations ofextractable Al than Fe and P throughout the period of timerepresented by the cores. After normalization with respect tothe sedimentation rate, Al flux has increased in LLP but showsno decrease in the SO4

2− recovery period (post-1970s). In MP,

275S C I E N C E O F T H E T O T A L E N V I R O N M E N T 4 0 4 ( 2 0 0 8 ) 2 6 9 – 2 7 5

the normalized Al flux to the sediment shows no obvioustrend with respect to historic pH perturbations, but is highestin ~1940. Normalized P and Fe fluxes to the sediment remainfairly constant throughout the cores with the exception ofsurface enrichment of Fe in the top few cm. We cannotspeculate about the effects of recovery from acidification on Ploading, because these sites are inherently acidic and remainso to the present day. However, the data from theseoligotrophic systems suggest that Al may be the predominantabiotic control of sediment P retention.

Acknowledgements

We thank J. Cangelosi (Sawyer Environmental ChemistryResearch Lab, University of Maine), H. Goss and M. Laird(Department of Earth Sciences, University of Maine) forinvaluable field assistance in collecting the sediment cores.Dr. C. T. Hess and V. Guisseppe (Department of Physics,University of Maine) provided the 210Pb data for sedimentchronology. The U.S. National Science Foundation (DEB-0415348) provided financial support for this study.

R E F E R E N C E S

Ahlgren J, Tranvik L, Gogoll A, Waldebäck M, Markides K, Rydin E.Sediment depth attenuation of biogenic phosphoruscompounds measured by 31P NMR. Environ Sci Technol2005;39:867–72.

Amirbahman A, Pearce AR, Bouchard R, Norton SA, Kahl SK.Relationship between hypolimnetic phosphorus and ironrelease from eleven lakes in Maine, USA. Biogeochem2003;65:369–86.

Appleby PG, Oldfield F. The Calculation of 210Pb dates assuming aconstant rate of supply of unsupported 210Pb to the sediment.Catena 1978;5:1–8.

Coolidge, K.M. Phosphorus cycling in Maine lakes: a sedimentaryanalysis. Unpublished Thesis, University of Maine, Orono, ME,2004, 109 pp.

Davis RB, Doyle RW. A piston corer for upper sediment in lakes.Limnol Oceanogr 1967;14:643–8.

Davis RB, Anderson DS, Norton SA, Whiting MC. Acidity of twelvenorthern New England (U.S.A.) lakes in recent centuries.J Paleolimnol 1994;12:103–54.

Hieltjes AHM, Lijklema L. Fractionation of inorganic phosphates incalcareous sediments. J Environ Qual 1980;9:405–7.

Hupfer M, Gächter R. Polyphosphate in lake sediments: 31P NMRspectroscopy as a tool for its identification. Limnol Oceanogr1995;40:610–7.

Huser BJ, Rydin E. Phosphorus inactivation by aluminum in LakesGårdsjön and Härsvatten sediment during the industrialacidification period in Sweden. Can J Fish Aquat Sci2005;62:1702–9.

Kopáček J, Hejzlar J, Borovec J, Porcal P, Kotorová I. Phosphorusinactivation by aluminum in the water column andsediments: lowering of in-lake phosphorus availability inan acidified watershed-lake ecosystem. Limnol Oceanogr2000;45:212–25.

Kopáček J, Borovec J, Hejzlar J, Ulrich KU, Norton SA, AmirbahmanA. Aluminum control of phosphorus sorption by lakesediments. Environ Sci Technol 2005;39:8784–9.

Kopáček J, Marešová M, Norton SA, Porcal P, Veselý J.Photochemical source of metals for sediments. Environ SciTechnol 2006;40:4455–9.

Lake BA, Coolidge KM, Norton SA, Amirbahman A. Factorscontributing to the internal loading of phosphorus from anoxicsediments in six Maine, USA, lakes. Sci Total Environ2007;373:534–41.

Maine Department of Environmental Protection. PEARL(Public Educational Access to Resources on Lakes); 2006.http://www.pearl.maine.edu.

Murphy J, Riley JP. A modified single-solution method for thedetermination of phosphate in natural waters. Anal Chim Acta1962;27:31–6.

Norton SN, Bienert RW, Binford MW, Kahl JS. Stratigraphy of totalmetals in PIRLA sediment cores. J Paleolimnol 1992;7:191–214.

Ogburn RW, Brezonik PL. Examination of the oligotrophicationhypothesis: phosphorus cycling in an acidic Florida lake. WaterAir Soil Pollut 1986;30:1001–6.

Psenner R, Pucsko R. Phosphorus fractionation: advantages andlimits of the method for the study of sediment P origins andinteractions. Arch Hydrobiol Beih Ergebn Limnol1988;30:43–59.

Schindler DW. Evolution of phosphorus limitation in lakes.Science 1977;195:260–2.

Steinberg CEW, Wright RF, editors. Acidification of FreshwaterEcosystems: Implications for the Future. Chichester: Wiley;1994. 404pp.

Ulrich KU, Pöthig R. Evidence for aluminium precipitation andphosphorus inactivation in acidified watershed-reservoirecosystems. Silva Gabreta 2000a;4:185–98.

Ulrich KU, Pöthig R. Precipitation of aluminium and phosphateaffected by acidification. Acta Hydrochim Hydrobiol2000b;28:313–22.