J Paleolimnol (2012) 48:559–569 DOI 10.1007/s10933-012-9631-4

Differential post-depositional mobility of phosphorus species in lake sediments M. L. Ostrofsky

Received: 27 January 2012 / Accepted: 3 July 2012 / Published online: 29 July 2012 Ó Springer Science+Business Media B.V. 2012

Abstract Historically, paleolimnologists have been cautious about interpreting sedimentary total phosphorus (P) profiles because of the well-documented post-depositional mobility of P. There is recent new attention given to the interpretation of component P fractions that are generally indicative of broad categories of chemical P species in sediments. Using homogenized sediments collected from 5 lakes with differing characteristics, the mobilities of total P, and of NH4Cl-, BD-, NaOH-, and HCl-extractible P were measured in short term incubations (15–24 weeks). Almost all of the observed mobility of total P could be explained by the mobility of reductant-soluble BD–P, with a smaller contribution from NaOH–P. In contrast, HCl–P (apatite) and organic-P showed no significant movement. These results reaffirm that sedimentary TP profiles should be interpreted with caution, and that component P species, particularly NH4Cl-, BD-, and NaOH–P are also prone to post-depositional mobility. In contrast, HCl–P and organic-P appear to be more reliable proxies for paleolimnological reconstructions. Keywords Sediments
Phosphorus
Phosphorus species
Mobility

M. L. Ostrofsky (&) Biology Department, Allegheny College, Meadville, PA 16335, USA e-mail: [email protected]

Introduction It is widely accepted that phosphorus (P) is an important limiting nutrient in freshwaters, and there is a rich and compelling literature linking various indicators of trophic state to water column phosphorus concentrations. The concentration of total phosphorus during spring turnover (TPspr) is often the sole independent variable in a number of models that predict biomass and productivity at numerous trophic levels (phytoplankton, zooplankton, benthos, fish) in lake communities (Peters 1986). Since many paleolimnological investigations are conducted to determine historical trophic state of lakes for both theoretical and practical reasons—to trace the ontogeny of lakes (Engstrom et al. 2000); to set attainable targets for lake remediation (Brezonik and Engstrom 1998); to document ecosystem responses to historical stressors—the sedimentary record of P accumulation should be of considerable interest to both paleolimnologists and neolimnologists. The sedimentation of P in lakes is assumed to be a first-order process (directly proportional to water column concentration) and that assumption has been integral to the historical development of phosphorus budget models (Vollenweider 1969; Snodgrass and O’Melia 1975; Dillon and Rigler 1975). Further, the magnitude of the proportionality constant is empirically related to areal water load (Kirchner and Dillon 1975; Ostrofsky 1978), so the accumulation or concentration of P in the sediments of any lake should

123

560

reflect water column phosphorus concentration. However, sediment P concentrations have not been shown to be very closely related to water column P in among-lake comparisons (McColl 1977; Ostrofsky 1987; Trolle et al. 2009). This lack of agreement is in part due to differences in P retention among lakes of different hydrologies. Even within lakes, Carignan and Flett (1981) demonstrated the potential for marked postdepositional P mobility in incubated sediments, and many lakes release considerable amounts of previously sedimented P to the overlying hypolimnetic water under both oxic and anoxic conditions (Nu¨rnberg et al. 1986; Jensen et al. 1992; Burger et al. 2007). Both observations suggest that any relationship between water column P and sediment P may be tenuous at best, and paleolimnologists have been understandably cautious about inferring historical trophic changes in lakes from sedimentary P profiles unless these inferences are corroborated by other lines of evidence (pigments, diatoms, C:N, biogenic silica, etc.; Schelske et al. 1986; Waters et al. 2005). Recently a number of paleolimnological investigations have used the concentrations of various chemical species of P in the sedimentary record of lakes to provide a temporal resolution to watershed soil development unavailable through soil chronosequences (Filippelli and Souch 1999; Filippelli et al. 2006; Norton et al. 2011) or to examine historical patterns of phosphorus loading (Brezonik and Engstrom 1998; Engstrom et al. 2009; Triplett et al. 2009). These investigations have assumed that the distribution of these P species in the lake sedimentary profile was stable with little post-depositional mobility or diagenesis. The chemical species of phosphorus in lake sediments and soils are operationally defined, that is they are defined by solubility in progressively more aggressive extractants. Commonly used procedures for partitioning P species in lake sediments are based on Hietljes and Lijklema (1980) or Psenner et al. (1988) or modifications that vary extractant strength or time (Hupfer et al. 1995). Typically solid-phase samples (soils, filtered seston, sediments, etc.) are subjected to sequential extractions with ammonium chloride, buffered dithionite (BD), sodium hydroxide, and hydrochloric acid. Ammonium chloride removes phosphorus compounds loosely attached to the surface of clay and calcite particles and includes any dissolved inorganic phosphorus compounds in the interstitial

123

J Paleolimnol (2012) 48:559–569

water. Buffered dithionite removes redox sensitive phosphorus bound largely to ferric and manganic hydroxides. Sodium hydroxide removes phosphorus compounds bound to oxides of Fe and Al. Hydrochloric acid removes phosphorus compounds bound to calcium—apatite. The difference between the sum of these inorganic fractions and total phosphorus, measured following a more rigorous digestion with perchloric acid or potassium persulfate or by combustion, is an estimate of organically bound and other refractory phosphorus species. Although migration of total P within the sediment column (Carignan and Flett 1981) and release to hypolimnetic waters (Nu¨rnberg et al. 1986) are commonly observed phenomena, it is unlikely that all component P fractions in the sediment behave equally. The assumed mechanism for P mobility depends on solid-phase P species entering the liquid phase in the interstitial water of sediments through desorption, redox shifts, or mineralization, and then moving along a diffusion gradient toward the oxidized surface layer where re-precipitation with iron, or biological uptake removes mobile P from the liquid phase maintaining the concentration gradient in the interstitial water. Reductant-soluble metal-bound-P has been correlated with sediment P release rates (Nu¨rnberg 1988; Ostrofsky et al. 1989; Søndergaard et al. 1993; Rydin 2000; Petticrew and Arocena 2001; Søndergaard et al. 2001) and is the most likely contributor to observed total P mobility, although the relatively high concentrations found within sedimentary profiles suggest incomplete reduction even in strongly reducing sediments (Filippelli et al. 2010). Organic-P is quite labile, and much of it is released soon after sedimentation (Hupfer et al. 1995) although the persistence of organic-P in sediment cores suggests that some is permanently stored in part due to the limited diffusion of oxygen into the sediments and the rapid exhaustion of alternate electron acceptors (NO3-, Mn4?,3?, Fe3?, SO42-) needed to mineralize organic material (Ga¨chter and Mu¨ller 2003) and in part due to the fact that when organic-P is calculated as the difference between total-P and the sum of the inorganic fractions it may include considerable amounts of recalcitrant inorganic P. Finally, apatiteP of terrigenous origin is particularly refractory, being neither very soluble (solubility product *10-58; Schlesinger 1991; Valsami-Jones et al. 1998) nor available for biological uptake (Williams et al. 1980).

J Paleolimnol (2012) 48:559–569

561

Consequently, we might anticipate large differences in the contribution of the various phosphorus species to the observed mobility of TP in sedimentary profiles. The purpose of this work is to test the hypothesis there are significant differences in the mobility of various P fractions in the sediment column so that while total sedimentary P is an unreliable proxy in the sedimentary profile, so too are at least some of the component P fractions. P species that are non-motile may prove to be more useful indicators of past lake or watershed conditions.

Methods Sediments were collected from natural lakes in northwestern Pennsylvania, in a region that was deglaciated 12–14,000 year BP. Initial settlement began in the late eighteenth century with maximum forest clearance (about 80–90 %) occurring between 1890 and 1900. Since then widespread agricultural abandonment, particularly after 1945, has allowed reforestation so that the area is now 60–70 % forested (Whitney and DeCant 2003). Surface sediments to a depth of 10–15 cm were collected using an Ekman dredge from or near the deepest region of 5 lakes (Table 1) to represent a range of sediment types. Crystal Lake has a very small watershed and its

Table 1 Characteristics of lakes/sediments used to assess post-depositional mobility of P species Lake/ sediment characteristic

Crystal

Edinboro

Ad/Aoa

4.6

39.3

15.2

8.1

181.9

TPbspr

23

30

36

14

41

Alkalinityc

62

67

96

67

37

Depthd

6

8.5

12.5

8.5

5.5

Dry Wt (%)

5.5

16.9

8.4

11.4

13.7

LOI575 (%)e LOI1000 (%)f

56 7.1

20 2.3

22 3.4

27 4.2

21 2.3

Pleasant

Sandy

Sugar

a

Drainage Basin Area/Lake Area

b

Lake water total phosphorus at spring overturn, lg L-1

c

mg L-1 as CaCO3

d

Water depth at collection site

e

Loss on Ignition (1 h. 575 °C)

f

Loss on Ignition (1 h. 1,000 °C)

unconsolidated sediments have very high organic content. Edinboro Lake had received secondary sewage treatment plant effluent for 24 years, ending in 2007. Lake Pleasant has the least current anthropogenic impact with 75 % of its shoreline protected by the Western Pennsylvania Conservancy, and recreational use restricted to motorless watercraft only. Sandy Lake sediments have anomalously high [Fe], a legacy of open-pit coal mining in its watershed (Ostrofsky and Schworm 2011). Sugar Lake has a very large drainage basin and is a warm monomictic lake due to its broad exposure and shallow depth, with a well-oxygenated sediment surface throughout the year. In the laboratory, freshly collected sediments were stirred for 60 s. to ensure homogeneity. An aliquot was placed in a pre-weighed aluminum weigh dish to determine water content and after drying to determine organic and carbonate content through loss-on-ignition (Dean 1974), and initial TP and P species. Fresh sediments were gently packed into replicate 2.54 cm i.d. 9 15 cm core tubes open at the top and sealed at the bottom with a rubber stopper, and submerged vertically in a 38-L aquarium filled (to 5–10 cm over the tops of the cores) with spring water that was continuously aerated with an aquarium air stone. Three cores were immediately extruded and P species analyzed to ensure that the sediments were initially homogeneous. Remaining cores were incubated undisturbed in the dark in a 10 °C cold room. After 15 weeks, three replicate Sugar Lake cores were removed from the aquarium, sediments were extruded in 1-cm sections, dried (60 °C), ground in a mortar and pestle, and P species fractionated. Based on this analysis, the remaining triplicate lake cores were allowed to incubate for an additional 8–9 weeks before being removed, extruded, dried, ground, and analyzed. P fractionation followed the sequence described by Hupfer et al. (1995). Approximately 75 mg of sediment from each 1-cm section from each of 3 cores from each lake was extracted with 10 mL of 1.0 M NH4Cl (pH 7) for 2 h, followed by freshly prepared 0.1 M bicarbonate/dithionite (1 h), 1 M NaOH (16 h) and 0.5 M HCl (24 h). A second sample from each section was analyzed for total P using the combustion technique described by Andersen (1976). Here, approximately 40 mg of dry sediment was combusted for 2 h at 575 °C, and the residue extracted in 1.0 M HCl. Phosphorus concentrations in all extracted and total P samples were determined using the heteropoly

123

562

J Paleolimnol (2012) 48:559–569

blue spectrophotometric method described by Strickland and Parsons (1968). The triplicate cores from each lake were considered replicates, and the mean P concentrations for total P and all fractions were analyzed using 1-way ANOVA, and followed by Fisher’s PLSD multiple comparisons test if appropriate. Significant differences in concentration among core depths were taken as evidence of P mobility.

Results The sediment collected from all 5 lakes was the typical dark brown gyttja common to small inland mesoeutrophic lakes with relatively high organic content (20–56 % as LOI, Table 1). Total P in the homogenized pre-incubation sediments ranged from 1 to 2.5 mg P g-1 sediment dry weight (Table 2). In all sediments the NH4Cl-P fraction contributed 0.5 % or less to sediment total P. BD–P was the most abundant inorganic fraction in all sediments except those from Sandy Lake where NaOH–P was the most abundant. HCl–P was a relatively constant fraction in all lakes, ranging from 3 to 5 % of TP. At the end of the incubation period (15 weeks for Sugar Lake cores, 23–24 weeks for Crystal, Edinboro, Pleasant and Sandy Lake cores) most cores exhibited a conspicuous lighter brown oxidized layer at the surface. This layer was typically only a few mm in thickness and was wholly contained in the 0–1 cm slice of the sectioned core. While no effort had been made to exclude organisms, there was no evidence of benthic activity. The increase in TP at the surface of all cores is evident (Fig. 1). In all lakes, there was significantly (p \ 0.0001, 1-way ANOVA) more total P in the top (0–1 cm depth) section of the core than in any of the other sections. In Sandy Lake only, there was also

significantly more TP at 1–2 cm and 2–3 cm than in any of the deeper sections. These results clearly indicate post-depositional mobility in the sedimentary column. Comparing the post-incubation total P concentration in the 0–1 cm layer to the initial preincubation concentration (Table 2) there was a 2.49 increase in the surface sediments from Crystal and Sandy lakes, a 2.19 increase in Lake Pleasant sediment, a 1.39 increase in Edinboro Lake sediment, and a 1.29 increase in Sugar Lake sediment. Sugar Lake likely would have shown a greater increase had the cores been incubated as long as those from the other four lakes. NH4Cl–P was the smallest P fraction in all sediments, generally contributing \0.5 % of the total P in any core. In the 4 lake sediments that were incubated the longest, the top section of the core (0–1 cm) had significantly more NH4Cl–P (Figs. 2, 3, 4, 5 and 6) than did any of the deeper sections (p \ 0.001, 1-way ANOVA), and in Crystal, Edinboro and Pleasant cores all of the deeper sections had the same amount (Fisher’s PLSD, p [ 0.05). In the Sandy Lake cores the 1–2 cm section had significantly more NH4Cl-P than did the deeper sections, but less than the surface section. In the Sugar Lake cores the surface section had significantly more NH4Cl–P than did most of the deeper sections, but the same amount as the 2–3, 3–4, 5–6, and 7–8 cm sections. Had the Sugar lake cores been incubated longer, they might have reflected the same pattern as cores from the other lakes. BD–P made a substantial contribution to the total sedimentary P in all lakes, ranging from 16 % in Sandy to 65 % in Sugar Lake. Again, in all lake cores there was significantly (p \ 0.001, 1-way ANOVA) more BD–P in the surface section of the core than in the lower sections (Figs. 2, 3, 4, 5 and 6). BD–P showed no differences in any of the deeper sections (1–2 to 11–12 cm) in the Crystal, Edinboro, Pleasant,

Table 2 Distribution of phosphorus species in the initial sediments collected from the 5 lakes NH4Cl–P

BD–P

NaOH–P

HCl–P

Lake

Total P

Crystal

1,437 (6.6)

Edinboro

2,392 (60.1)

9.1 (1.0)

Pleasant

2,380 (23.7)

12.2 (0.9)

Sandy

1,202 (24.3)

2.9 (0.1)

190.4 (3.4)

466.5 (7.5)

43.3 (1.9)

Sugar

2,461 (27.2)

7.7 (0.8)

1,589.4 (24.2)

465.1 (8.5)

85.9 (0.5)

1.1 (0.1)

586.6 (6.9)

210.6 (0.9)

44.5 (1.0)

1,074.0 (33.9)

714.4 (22.4)

107.6 (2.4)

1,358.2 (13.0)

642.6 (5.8)

97.8 (1.4)

All P species in ug P g-1 sediment dry weight. Means of three replicates (±SE)

123

J Paleolimnol (2012) 48:559–569

563

Fig. 1 Distribution of total P (lg P g-1 sediment dry weight) after 15 weeks (Sugar Lake) or 23–24 weeks (Crystal, Edinboro, Pleasant, and Sandy Lakes) incubation of homogenized

sediment. Means (filled circle) ± standard errors (whiskers) are plotted (n = 3). In most cases, standard error bars are masked by the mean symbol

and Sugar cores. In the Sandy Lake cores sections 3–4 to 11–12 were the same while the top three sections (0–1, 1–2, 2–3 cm) contained progressively less BD– P. The increase in the surface section over the uniform deeper sections ranged from 1.259 in Sugar Lake to 6.19 in Sandy Lake. NaOH–P was generally the second most abundant inorganic P fraction in the sediments except in the Ferich Sandy Lake sediments were it was the most abundant. Unlike NH4Cl–P and BD–P, there was not a consistent pattern to the distribution of NaOH–P with depth in the cores (Figs. 2, 3, 4, 5, 6). In Crystal, Sandy and Sugar Lake cores there was a significant (p \ 0.001) increase in the surface core section (1.8, 1.8, and 1.19, respectively). However, in Edinboro Lake sediment there were no differences with depth, and in Lake Pleasant sediment there was significantly less NaOH–P (0.79) in the 0–1 and 1–2 cm sections. HCl–P was the most stable of the inorganic P species, and while there were occasional differences in the concentrations at different depths within some

cores, there was no pattern to these differences except in the Sandy Lake cores where there was about 20 % less in the surface section than in the deeper sections. Organic-P concentrations showed greater variation among replicate cores than did any other P species, most likely because it was not measured directly but rather as the difference between the sum of the inorganic fractions and total P so that calculated organic-P values contained the accumulated variances of TP and each inorganic fraction. There were no significant differences with depth in the Edinboro, Pleasant, Sandy or Sugar Lake cores. Only in the Crystal Lake cores was there significantly more organic-P in the surface section than in the deeper sections (Figs. 2, 3, 4, 5, 6).

Discussion In all sediment cores there was a significant increase in the total-P concentration in the surface section that

123

J Paleolimnol (2012) 48:559–569

rg -P

l-P

H

H

O

C

aO N

N

BD

H

-P

4C

-P

l-P

564

0 1 2

Depth in Core (cm)

3 4 5 6 7 8 9 10 11 12 0

10

20

0

1000

2000

0

100

200

300

0

25

50

0

250

500

750

Crystal Lake ug P g-1 sed. dry wt.

Fig. 2 Distribution of NH4Cl-, BD-, NaOH-, HCl-, and Organic-P (lg P g-1 sediment dry weight) from incubated homogenized Crystal Lake sediment. Symbols as in Fig. 1

occurred over the relatively short incubation time (15, 23–24 weeks). These results are consistent with the observations of Carignan and Flett (1981), and highlight the danger of paleolimnological interpretation of sedimentary TP profiles. TP concentrations in the more recent surface sediments in any lake can be artificially elevated, augmented by P diffusion from deeper layers in the core. This higher concentration could be misinterpreted as increased P sedimentation from recent increases in water column P. Consequently, the caution with which sedimentary profiles of TP have been interpreted is well justified. The observed increases in TP were, however, almost exclusively due to the increase in BD–P. In Crystal Lake, for example, more than 80 % of the approximate 2,000 lg TP g-1 increase could be accounted for by the increase in BD–P. In Edinboro, Pleasant, Sandy, and Sugar Lakes, increases in BD–P in the surface contributed 100, 99, 70, and 77 % of the increase in TP, respectively. In those lake sediments where NaOH–P showed significant surficial increases in concentration, the fraction was very much a lesser

123

contributor to observed increases in TP. These results are consistent with the assumption that in the sedimentary environment previously bound phosphate in the interstitial pore-water can diffuse upward, and be re-precipitated as ferric and manganic hydroxides in the surface oxic zone. The relative increases in NH4Cl–P at the surface, as large as they were, contributed little to the observed changes in TP, yet are consistent with an increasing flux of dissolved interstitial P to the surface layers. The potential sources of this migrating interstitial P include desorption of P bound to clay and mineral particles, release as a result of dissolution of Fe–P complexes in a reducing environment, and mineralization of P-containing organic matter. However, in no case was there a significant decrease in TP in the subsurface that would account for the increase at the top of the core. In the Edinboro and Pleasant cores there was a decrease in the amount of NaOH–P in the 1–2 cm section, and in the Sandy core there was a decrease in HCl–P in the 0–3 cm sections. However these decreases alone were not enough to account for the increases in BD–P seen

J Paleolimnol (2012) 48:559–569

565

Fig. 3 As in Fig. 2, for Edinboro Lake sediment

in the 0–1 cm section of any core. Carignan and Flett (1981) observed a slight decrease in sedimentary TP immediately below the surface, and interpreted this as the source of the increase observed at the surface, although again more TP was added to the surface than could be accounted for by subsurface losses. In the absence of a statistically significant decrease in any P fraction with depth, we are left to conclude that small amounts from all depths were contributed more or less equally to the observed increase at the surface, but that the P fraction(s) contributing remain unknown. In contrast to the observed surficial increases in BD–P, NH4Cl–P, and often NaOH–P, the HCl–P fraction was consistently immobile. This fraction is assumed to be composed largely of detrital apatite, only sparingly soluble in the normal sedimentary environment, and derived from weathered soils. Similarly, organic-P showed no evidence for migration in four of the five lake sediments examined. Much of the organic-P in settling seston is remobilized soon after sedimentation (Hupfer et al. 1995), particularly poly- and pyro-phosphates (Reitzel et al. 2007), so that only the more refractory organic-P species accumulate

in sediments and long-term immobility is most likely due to poor diffusion of oxygen into deeper sediment layers, the rapid early exhaustion of alternate electron acceptors that would facilitate microbial mineralization of organic matter (Schlesinger 1991, Søndergaard et al. 2003) and the observation that an unknown fraction of organic-P may include inorganic P bound in mineral lattices (Ga¨chter and Mu¨ller 2003). The implications of these results are that sedimentary profiles of TP and of the NH4Cl-, BD-, and NaOHextractible fractions should continue to be interpreted with considerable caution. The HCl-extractible fraction, however, appears to be non-motile in the sediment column and would more accurately reflect conditions in the lake’s watershed at the time of sedimentation. The origin of all soil P was contained in igneous rock, notably the primary mineral apatite (Ca5(PO4)3(F, Cl, OH)) (Schlesinger 1991). Carbonation weathering releases PO4 that is subsequently used biologically and converted to organic-P, or reacts with other soil minerals (e.g., Al, Fe, Mn) and accumulates in occluded forms. Soil chronosequences have demonstrated decreases in apatite-P and

123

566

Fig. 4 As in Fig. 2, for Lake Pleasant sediment

Fig. 5 As in Fig. 2, for Sandy Lake sediment

123

J Paleolimnol (2012) 48:559–569

J Paleolimnol (2012) 48:559–569

567

Fig. 6 As in Fig. 2, for Sugar Lake sediment

increases in occluded- and organic-P as soil develops (Walker and Syers 1976), and these changes have been reflected in the sedimentary profile of a number of small, oligotrophic, headwater lakes where diagenetic overprinting was assumed to be minimal across millennial time scales (Filippelli and Souch 1999, Norton et al. 2011). However, these changes in the distribution of soil P species may be quite dynamic. The acidification of young soils is in large part due to the development of vegetation (Boyle 2007), and forested soils typically have lower pH than nonforested soils (Grossmann and Mladenoff 2008). Land use changes that affect soil pH have a direct effect on the dissolution of apatite. Garcia-Montiel et al. (2000) found that deforestation led to increases in soil pH and an abrupt increase in HCl–P (apatite-P), and reforestation of previously open sites decreases soil pH and which presumably leads to decreases in apatite-P (Chen et al. 2008). Deforested soils not only contain more apatite-P, but are also more prone to erosion with transport of soil particles and contained apatite to lakes.

Because apatite-P is responsive to land-use changes, and is immobile in the sedimentary profile of lakes, it makes a convenient proxy in the sedimentary record. Filippelli et al. (2010), for example, were able to reconstruct historical cycles of agriculture and reforestation driven by changes in climate and by depopulation following the Spanish Conquest and subsequent repopulation, using changes in the relative proportions of organic-, mineral- (apatite-), and occluded- (metal-bound-) P in the sedimentary record of Lake Zoncho. This analysis assumed that forested soils contain less mineral-P (apatite) than agricultural soils due to higher acidity, and that agricultural soils contain less organic-P due to regular removal of biomass via crop harvests and that these changes in soil P species are detectable over relatively short time spans. If HCl–P is assumed to reflect detrital apatite, any watershed disturbance that increases soil erosion would increase apatite-P loading to the lake, and subsequent sedimentation. Consequently, an increase in HCl–P flux to sediments might be induced by fire,

123

568

severe storm events, road construction, initial land clearance for agriculture or repeated tilling of row crops. The apparent immobility of HCl–P in the sediments suggests that it would be a reliable proxy for paleolimnological reconstructions of lake and watershed histories. The large surficial increases observed in NH4Cl-, BD-, and NaOH-extractible P suggest that these P fractions should continue to be interpreted with caution. Acknowledgments I thank P. Guilizzoni and two anonymous reviewers for their thoughtful comments on the manuscript.

References Andersen JM (1976) An ignition method for determination of total phosphorus in lake sediments. Water Res 10:329–331 Boyle JF (2007) Loss of apatite caused irreversible earlyHolocene lake acidification. Holocene 17:543–547 Brezonik PL, Engstrom DR (1998) Modern and historic accumulation rates of phosphorus in Lake Okeechobee, Florida. J Paleolimnol 20:31–46 Burger DF, Hamilton DP, Pilditch CA, Gibbs MM (2007) Benthic nutrient fluxes in a eutrophic, polymictic lake. Hydrobiology 584:13–25 Carignan R, Flett RJ (1981) Postdepositional mobility of phosphorus in lake sediments. Limnol Oceanogr 26:361–366 Chen CR, Condron LM, Xu ZH (2008) Impacts of grassland afforestation with coniferous trees on soil phosphorus dynamics and associated microbial processes: a review. For Ecol Manag 255:396–409 Dean WE (1974) Determination of carbonate and organic matter in calcareous sediments and sedimentary rocks by loss on ignition: comparison with other methods. J Sed Petrol 44:242–248 Dillon PJ, Rigler FH (1975) A simple method for predicting the capacity of a lake for development based on lake trophic status. J Fish Res Bd Canada 32:1519–1531 Engstrom DR, Fritz SC, Almendinger JE, Juggins S (2000) Chemical and biological trends during lake evolution in recently deglaciated terrain. Nature 408:161–166 Engstrom DR, Almendinger JE, Wolin JA (2009) Historical changes in sediment and phosphorus loading to the upper Mississippi River: mass-balance reconstructions from the sediment of Lake Pepin. J Paleolimnol 41:563–588 Filippelli GM, Souch C (1999) Effects of climate and landscape development on the terrestrial phosphorus cycle. Geology 27:171–174 Filippelli GM, Souch C, Menounos B, Slater-Atwater S, Jull AJT, Slaymaker O (2006) Alpine lake sediment records of the impact of glaciation and climate change on the biogeochemical cycling of soil nutrients. Quat Res 66:158–166 Filippelli GM, Souch C, Horn SP, Newkirk D (2010) The preColumbian footprint on terrestrial nutrient cycling in Costa Rica: insights from phosphorus in a lake sediment record. J Paleolimnol 43:843–856

123

J Paleolimnol (2012) 48:559–569 Ga¨chter R, Mu¨ller B (2003) Why the phosphorus retention of lakes does not necessarily depend on the oxygen supply to their sediment surface. Limnol Oceanogr 48:929–933 Garcia-Montiel DC, Neill C, Melillo J, Thomas S, Steudler PA, Cerri CC (2000) Soil phosphorus transformations following forest clearing for pasture in the Brazilian Amazon. Soil Sci Soc Am J 64:1792–1804 Grossmann EB, Mladenoff DJ (2008) Farms, fires, and forestry: disturbance legacies in the soils of the Northwest Wisconsin (USA) Sand Plain. For Ecol Manag 256:827–836 Hietljes AHM, Lijklema L (1980) Fractionation of inorganic phosphates in calcareous sediments. J Environ Qual 9:405–407 Hupfer M, Ga¨chter R, Giovanoli R (1995) Transformation of phosphorus species in settling seston and during early diagenesis. Aquat Sci 57:305–324 Jensen HS, Kristensen P, Jeppesen E, Skytthe A (1992) Iron:phosphorus ratio in surface sediment as an indicator of phosphate release from aerobic sediments in shallow lakes. Hydrobiol 235(236):731–743 Kirchner WB, Dillon PJ (1975) An empirical method of estimating the retention of phosphorus in lakes. Water Resour Res 11:181–182 McColl RHS (1977) Chemistry of sediments in relation to trophic conditions in eight Rotorua Lakes, New Zealand. J Mar Freshwat Res 11:509–523 Norton SA, Perry RH, Saros JE, Jacobson GL, Fernandez IJ, Kopa´cˇek J, Wilson TA, SanClements MD (2011) The controls on phosphorus availability in a Boreal lake ecosystem since deglaciation. J Paleolimnol 46:107–122 Nu¨rnberg GK (1988) Prediction of phosphorus release rates from total and reductant-soluble phosphorus in anoxic lake sediments. Can J Fish Aquat Sci 45:453–462 Nu¨rnberg GK, Shaw M, Dillon PJ, McQueen DJ (1986) Internal phosphorus load in an oligotrophic Precambrian Shield lake with an anoxic hypolimnion. Can J Fish Aquat Sci 43:574–580 Ostrofsky ML (1978) Modification of phosphorus retention models for use with lakes with low areal water loading. J Fish Res Bd Canada 35:1532–1536 Ostrofsky ML (1987) Phosphorus species in the surficial sediments of lakes in eastern North America. Can J Fish Aquat Sci 44:960–966 Ostrofsky ML, Schworm AE (2011) A history of acid mine contamination, recovery, and eutrophication in Sandy Lake, Pennsylvania. J Paleolimnol 46:229–242 Ostrofsky ML, Osborne DA, Zebulske TJ (1989) Relationship between anaerobic sediment phosphorus release rates and sedimentary phosphorus species. Can J Fish Aquat Sci 46:416–419 Peters RH (1986) The role of prediction in limnology. Limnol Oceanogr 31:1143–1159 Petticrew EL, Arocena JM (2001) Evaluation of iron-phosphate as a source of internal lake phosphorus loadings. Sci Tot Environ 266:87–93 Psenner R, Bostro¨m B, Dinka M, Pettersson K, Pucsko R, Sager M (1988) Fractionation of phosphorus in suspended matter and sediment. Ergeb Limnol 30:98–109 Reitzel K, Ahlgren J, DeBrabandere H, Waldeba¨ck M, Gogloo A, Tranvik L, Rydin E (2007) Degradation rates of organic phosphorus in lake sediment. Biogeochem 82:15–28

J Paleolimnol (2012) 48:559–569 Rydin E (2000) Potentially mobile phosphorus in Lake Erken sediment. Wat Res 34:2037–2042 Schelske CL, Conley DJ, Stoermer EF, Newberry TL, Campbell CD (1986) Biogenic silica and phosphorus accumulation in sediment as indices of eutrophication in the Laurentian Great Lakes. Hydrobiol 143:79–86 Schlesinger WH (1991) Biogeochemistry: an analysis of global change. Academic Press, New York Snodgrass WJ, O’Melia CR (1975) Predictive model for phosphorus in lakes. Environ Sci Technol 9:937–944 Søndergaard M, Kristensen P, Jeppesen E (1993) Eight years of internal phosphorus loading and changes in the sediment phosphorus profile of Lake Søbygaard, Denmark. Hydrobiology 253:345–356 Søndergaard M, Jensen JP, Jeppesen E (2001) Retention and internal loading of phosphorus in shallow, eutrophic lakes. Sci World 1:427–442 Søndergaard M, Jensen JP, Jeppesen E (2003) Role of sediment and internal loading of phosphorus in shallow lakes. Hydrobiology 506–509:135–145 Strickland JDH, Parsons TR (1968) A practical handbook of seawater analysis. Fisheries Research Board of Canada, Ottawa (Bull. 167) Triplett LD, Engstrom DR, Elund MB (2009) A whole-basin stratigraphic record of sediment and phosphorus loading to the St. Croix River, USA. J Paleolimnol 41:659–677

569 Trolle D, Zhu G, Hamilton D, Luo L, McBride C, Zhang L (2009) The influence of water quality and sediment geochemistry on the horizontal and vertical distribution of phosphorus and nitrogen in sediments of a large, shallow lake. Hydrobiology 627:31–44 Valsami-Jones E, Ragnarsdottir KV, Puntis A, Bosbach D, Kemp AJ, Cressey G (1998) The dissolution of apatite in the presence of aqueous metal cations at pH 2–7. Chem Geol 151:215–233 Vollenweider RA (1969) Mo¨glichkeiten und Grenzen elementarer Modelle der Stoffbilanz von Seen. Arch Hydrobiol 66:1–36 Walker TW, Syers JK (1976) The fate of phosphorus during pedogenesis. Geoderma 15:1–19 Waters MN, Schelske CL, Kenney WF, Chapman AD (2005) The use of sedimentary algal pigments to infer historic algal communities in Lake Apopka, Florida. J Paleolimnol 33:53–71 Whitney GG, DeCant JP (2003) Physical and historical determinants of the pre- and post-settlement forests of northwestern Pennsylvania. Can J For Res 33:1683–1697 Williams JDH, Shear H, Thomas RL (1980) Availability to Scenedesmus quadricauda of different forms of phosphorus in sedimentary materials from the Great Lakes. Limnol Oceanogr 25:1–11

123