Soil Phosphorus Forms in Hydrologically Isolated Wetlands and Surrounding Pasture Uplands

Technical reportsTECHNICAL: Wetlands REPORTS and Aquatic Processes

Alexander W. Cheesman* University of Florida/IFAS Ed J. Dunne St. Johns River Water Management District Benjamin L. Turner Smithsonian Tropical Research Institute K. Ramesh Reddy University of Florida/IFAS

iffuse phosphorus (P) loads from agricultural ecosystems Newly created and restored wetlands play an important role impact the ecological functioning of many inland water- in sequestering excess nutrients at the landscape scale. In D evaluating the long-term efficacy of nutrient management ways and wetland systems (Khan and Ansari, 2005; Verhoeven et strategies to increase wetland capacity for sequestering P, al., 2006). Yet wetlands within agricultural landscapes also offer information is needed on the forms of P found across the a potential solution, by acting as a water and nutrient storage upland–wetland transition. To assess this, we studied soils system at the landscape scale (Paludan et al., 2002; Perkins et al., (0–10 cm) from four wetlands within cow–calf pastures north 2005; Mitsch and Day, 2006; Moreno et al., 2007). The func- of Lake Okeechobee, FL. Wetlands contained significantly (P < 0.05) greater concentrations of organic matter (219 g tional role that wetlands play within the landscape has been rec- C kg−1), total P (371 mg P kg−1), and metals (Al, Fe) relative ognized in the United States at the federal level by programs such to surrounding pasture. When calculated on an aerial basis, as the U.S. Army Corps of Engineers and USEPA-administered wetland surface soils contained significantly greater amounts of −1 −1 Compensatory Mitigation for Losses of Aquatic Resources (www. total P (236 kg ha ) compared with upland soils (114 kg ha ), epa.gov/wetlandmitigation/), and the NRCS-administered which was linked to the concomitant increase in organic matter with increasing hydroperiod. The concentration of P forms, Wetland Reserve Program (www.nrcs.usda.gov/programs/wrp/). determined by extraction with anion exchange membranes, 1 At the state level, this has translated to the adoption of state- mol L−1 HCl, and an alkaline extract (0.25 mol L−1 NaOH and specific programs such as Florida’s Lake Okeechobee Isolated −1 50 mmol L ethylenediaminetetraacetic acid [EDTA]) showed Wetland Restoration Program. A cost-share program under the significant differences between uplands and wetlands but did mandate of the Lake Okeechobee Protection Program (Fla. Stat. not alter as a proportion of total P. Speciation of NaOH– EDTA extracts by solution 31P nuclear magnetic resonance §373.4595, 2009), this initiative seeks to enhance and restore spectroscopy revealed that organic P was dominated by wetlands to retain increased amounts of water and P within the phosphomonoesters in both wetland and pasture soils but that four priority basins (S-65D, S-65E, S-154, and S-191) north of myo-inositol hexakisphosphate was not detected in any sample. Lake Okeechobee (Zhang et al., 2009). The tight coupling of total C and P in the sandy soils of the Within the south-central Florida region, isolated wetlands region suggests that the successful management of historically isolated wetlands for P sequestration depends on the long-term represent shallow (∼1 m) depressions within a landscape of accumulation and stabilization of soil organic matter. low (0–2%) topographic relief (Reddy et al., 1996; Capece et al., 2007). Although a significant component of the landscape, ∼13,000 ha (13.8% of the land area) within the four priority basins (McKee, 2005), these wetlands historically had little con- nectivity to surrounding surface waters. Anthropogenic altera- tion of drainage patterns during the expansion of cattle ranching (Steinman and Rosen, 2000) means that a large proportion of these wetlands (including the study sites) are now classed as ‘head of ditch’ wetlands, with some degree of channelized outflow (Flaig and Reddy, 1995). It is estimated that the surface soils (0–10 cm) of these systems store ∼290 kg P ha−1 (McKee, 2005), with recent Copyright © 2010 by the American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America. All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including pho- A.W. Cheesman and K.R. Reddy, Univ. of Florida/IFAS, Wetland Biogeochemistry Lab., tocopying, recording, or any information storage and retrieval system, Soil and Water Science Dep., 106 Newell Hall, Gainesville, FL 32611; E.J. Dunne, St. without permission in writing from the publisher. Johns River Water Management District, Division of Environmental Sciences, Dep. of Water Resources, 4049 Reid St., Palatka, FL 32177; B.L. Turner, Smithsonian Tropical J. Environ. Qual. Research Institute, Apartado 0843-03092, Balboa, Ancón, Republic of Panama. doi:10.2134/jeq2009.0398 Assigned to Associate Editor Robert Malone. Published online 25 May 2010. Received 7 Oct. 2009. Abbreviations: AEM, anion exchange membrane; DDI, distilled deionized water; EDTA, *Corresponding author ([email protected]). ethylenediaminetetraacetic acid; MDP, methylenediphosphonic acid; NMR, nuclear

5585 Guilford Rd., Madison, WI 53711 USA by AEM method; PM, soil microbial P determined by hexanol fumigation. studies showing significant increases in both soil and total eco- vary systematically across the upland–wetland transition in system P storage with increased hydroperiod (Dunne et al., response to differences in biogeochemical properties. 2007). It is believed that the restoration of former hydrological Here we compare the composition and availability of P conditions will lead to increased water and nutrient storage in between wetland and upland soils of an agricultural cow–calf the landscape and, therefore, reduce the P loading to down- grazing system in the subtropics. Our specific objectives were stream waters that ultimately contribute to Lake Okeechobee. (i) to determine basic soil physicochemical and P characteris- It is important to determine the impact of restoration efforts tics in wetland and uplands soils of grazed subtropical pasture; on current P storage in both wetland and upland soils that (ii) to quantify organic P composition across the upland–wet- experience an increased hydroperiod following restoration. land interface using solution 31P nuclear magnetic resonance Wetland ecosystems have surface soils that contain greater (NMR) spectroscopy; and (iii) to determine the effect of amounts of organic matter relative to both underlying soils hydrology on P composition and storage, with the aim of pre- and adjacent terrestrial ecosystems (Axt and Walbridge, 1999; dicting impacts of increased hydroperiod. Pant and Reddy, 2001; Gathumbi et al., 2005). This is due to increased organic matter accumulation within wetlands, a result Materials and Methods of high plant productivity, their receiving landscape position, Site Description and a relatively slow decomposition rate mediated by anaerobic conditions (Craft and Richardson, 1993; DeBusk and Four study sites were located on two cow–calf ranches, Larson- Reddy, 2003). Such conditions can promote the accumulation Dixie and Beaty, north of Lake Okeechobee in south-central of organic P relative to inorganic P, leading to the generaliza- Florida (Fig. 1). Surrounding pasture uplands are unimproved tion that P in wetland soils is dominated by organic fractions cow–calf operations with typically low stocking densities of ∼1 −1 (Newman and Robinson, 1999; Reddy et al., 2005). In a meta- animal unit ha (Gornak and Zhang, 1999; M. Flinchum, per- analysis of grassland soils, organic P was found to represent 26 sonal communication, 2004). The unfenced wetlands, two on to 57% of total P (Harrison, 1987). Similarly, studies of specific each ranch, were similar in size (1 to 2 ha) and supported simi- landscape types have shown considerable variation; for example, lar vegetation communities, with deep marsh zones (see soil between 31 and 75% of total P occurred as organic P in a series sampling) dominated by open water species [e.g., Pontedaria of 29 temperate pasture soils (Turner et al., 2003b). Such varia- cordata var. lancifolia (Muhl.)], shallow marsh zones dominated tion can be attributed to differences in land-use history, local cli- by flood-tolerant species (e.g.,Juncus effusus L.), and surround- mate, and pedogenic development (Harrison, 1987; Sumann et ing pasture uplands being predominantly Paspalum notatum al., 1998; McDowell and Stewart, 2006; Turner et al., 2007) and Flugge. Soils at Larson Dixie are mapped as siliceous, hyper- highlights the need to determine changes in P forms between thermic Spodic Psammaquents (Basinger series). Soils in the wetlands and uplands in the context of a given climatic zone and agricultural management history. The biogeochemical turnover of organic P is dependent on its stability in the environment, a result of both abiotic and biotic processes affected by landscape position (Wetzel, 1999; Celi and Barberis, 2005). Studies in a range of upland soils have shown organic P to be dominated by phosphomonoesters (Chapuis-Lardy et al., 2001; Turner et al., 2003a; McDowell and Stewart, 2006; Murphy et al., 2009), whereas recent work has indicated a greater prevalence of phosphodiesters in organic matter dominated wetlands (Turner and Newman, 2005; Turner et al., 2006a). Studies of upland soils have attributed a positive correlation between the proportion of organic P found as phosphodiesters (e.g., nucleic acids, phospholipids) and annual precipitation to the increased recalcitrance of phosphodiesters under wetter conditions (Tate and Newman, 1982; Condron et al., 1990). It is presumed that similar mechanisms, alongside the reduced stability of redox sensitive iron-bound inositol hexakisphosphate (Suzumura and Kamatani, 1995; Heighton et al., 2008) may account for the differences observed Fig. 1. Location of study sites showing Florida outline with area of interest north between the organic P forms of uplands and wetlands of Lake Okeechobee and detail of ranch sites containing two study wetlands each, observed to date (Turner and Newman, 2005). In within priority basins north of Lake Okeechobee. Beaty Ranch is within subba- addition to a greater prevalence of organic P within sin S-65D of the Kissimmee River watershed, with hydraulic connection through Cypress slough to Canal C38 and then to the Lake. Larson Dixie ranch is located in wetlands, it seems likely that organic P forms will subbasin S-154 within the Taylor Creek watershed, connected through Mosquito Creek to Taylor Creek itself.

Journal of Environmental Quality • Volume 39 • July–August 2010 Beaty uplands are sandy siliceous hyperthermic Aeric Alaquods extraction for 31P NMR analysis (Cade-Menun and Preston, (Myakka sands), while the study wetlands are delineated as a 1996). “Basinger and Placid soils depressional association” with Placid In the AEM–NMR method, duplicate fresh samples (3.5 g soils being sandy, siliceous, hyperthermic Typic Humaquepts. dry weight equivalent) were measured into preweighed 250 mL All soils are sandy textured with high infiltration rates yet low high density polyethylene centrifuge bottles and distilled deion- internal drainage given the typically high water table (Lewis et ized water (DDI) was added to bring the water content to 74 al., 2003). mL. One set of samples received a further 1 mL of DDI (nonfu- migated), and a parallel set received 1 mL hexanol (fumigated). Soil Sampling This allows for the determination of both exchangeable P (PAEM) In March 2007, during a period of draw down, each wetland and, by difference, a measure of fumigation-released or “micro- area was stratified into three zones based on soil, vegetation, bial” P (PM) (Kouno et al., 1995; Myers et al., 1999). All samples and hydrologic indicators: pasture upland, shallow marsh, and received a single 6.25- by 1.5-cm AEM strip (BDH Prolabo, deep marsh. Within each zone, three locations were selected product number 551642S, VWR International, Lutterworth, randomly for quadrat sampling. Within each quadrat, three UK), which were prepared by shaking 25 strips for 24 h in three −1 intact soil cores (diameter 7.5 cm by 10 cm) were collected sequential changes of 200 mL of 0.5 mol L NaHCO3. The and amalgamated. Samples were transported to the laboratory bottles were sealed and samples shaken on a reciprocating table on ice and homogenized before sorting. Roots and recogniz- for 24 h. The AEM strips were then removed, rinsed of adhering able organic fragments >2 mm were removed by hand and the material with DDI, shaken dry, and the phosphate eluted by −1 soil stored at 4°C. Subsamples were oven dried (70°C, 72 h), shaking for 3 h in 50 mL of 0.25 mol L H2SO4 (Turner and sieved, and ground for total elemental analysis (see below). Romero, 2009; Cheesman et al., 2010). The concentration of phosphate was determined using a discrete autoanalyzer (AQ2+, Hydroperiod Determination SEAL Analytical, Fareham, UK) and standard molybdate colo- Soil sampling elevations were compared to the elevation of a rimetry (USEPA, 1993). The AEM strips recovered ∼100% of groundwater well located in the center of each wetland. Using an orthophosphate standard solution (500 mg P membrane–1) wetland water level elevations, we determined how many days and 97% of an orthophosphate spike added to replicate soil per year a specific soil sampling location was saturated and we samples (100 mg P to 3.5 g soil membrane–1). It should be noted termed this estimate “hydroperiod.” Water levels in Larson that in addition to inorganic orthophosphate, the membrane Dixie wells were recorded from July 2003 to April 2006 and in strips may have recovered small concentrations of labile organic Beaty wells from December 2003 to March 2006. P (Cheesman et al., 2010). After removing the AEM strips, DDI was added to the Soil Physical and Chemical Properties unfumigated samples to bring the water content to 100 mL. − Soil water content was determined as weight loss following To this, 5 mL of a solution containing 5.25 mol L 1 NaOH − drying at 70°C for 72 h. Soil pH was determined on a 1:2 soil- and 1.05 mol L 1 EDTA was added to give a final concentra- − − to-water suspension using a glass electrode, and soil bulk den- tion of 0.25 mol L 1 NaOH and 50 mmol L 1 EDTA in a 1:30 sity was calculated using the known sample volume (3 cores, soil-to-solution ratio. The samples were shaken for 4 h and diameter 7.5 cm by 10 cm) and determined water content. then centrifuged at 7000 rpm (maximum relative centrifugal Total P was determined by combustion of soil at 550°C in a force ∼7500 g) (Sorvall RC6, SLA 1500 Rotor; Thermo Fisher muffle furnace for 4 h, dissolution of the ash in 6 mol L−1 HCl Scientific, Waltham, MA) for 10 min. The supernatant was

(Andersen, 1976), and then detection of molybdate reactive decanted and each sample analyzed for total P (NaOHTP) using P using a segmented flow analyzer (AAII Technicon, SEAL a modified double-acid digest (Rowland and Haygarth, 1997). Analytical, UK) and standard molybdate colorimetry (USEPA, In brief, 3 mL of extract was pipetted into a boiling tube to

1993). A subsample of the acid solution was analyzed for Al, which 1 mL of concentrated H2SO4 and 1 mL of concentrated

Ca, Fe, and Mg on an inductively coupled plasma–optical HNO3 were added. After placing on a heat block to reduce emission spectrometer (ICP–OES) (Thermo Jarrell Ash ICAP the volume to ∼0.5 mL, samples were refluxed at 550°C for at 61E, Franklin, MA). Total soil C and N were measured by least 3 h before being diluted by a factor of 100 in DDI and combustion and gas chromatography using a Flash EA1112 analyzed for orthophosphate by molybdate colorimetry (see (Thermo Scientific, Waltham, MA). above). Alkaline extracts were not analyzed for molybdate reactive P due to error associated with phosphate detection by this Phosphorus Characterization procedure in high organic matter soils (Turner et al., 2006b). Phosphorus forms were determined by two parallel methods: Solution 31P NMR Spectroscopy (i) a single-step process to determine acid extractable inorganic P by extraction in 1 mol L−1 HCl for 3 h (Reddy et al., 1998) Replicate alkaline extracts were combined on an equal volume and (ii) a combination of anion exchange membranes (AEM) basis within each of the three zones in a given wetland, result- and solution 31P NMR spectroscopy. This method utilizes ing in 12 amalgamated samples for solution 31P NMR analysis. HCO − loaded AEM strips as an initial extraction procedure Each extract (15 mL) was spiked with 1 mL methylenediphos- 3 − for exchangeable P (Myers et al., 1999), followed by the addi- phonic acid (MDP) (50 mg P mL 1) as an internal standard, tion of a solution containing NaOH and ethylenediamine- frozen at −80°C, and lyophilized. tetraacetic acid (EDTA) to proceed with a standard alkaline Approximately 100 mg of lyophilized material was resus-

pended in 0.1 mL D2O and 0.9 mL of a solution containing 1 mol L−1 NaOH and 100 mmol L−1 EDTA, and loaded into Results and Discussion a 5-mm NMR tube. Solution 31P spectra were acquired using a Bruker Avance 500 MHz spectrometer, with a broadband Soil Physicochemical Characteristics probe operating at 202.45 MHz. Spectra were accumulated As expected, hydroperiod varied between landscape positions, using waltz decoupling (zgig program) with a 4.0-ms pulse with the deep marsh zone being flooded on average 63% of (∼30°), 0.4-s acquisition time and 1.0-s delay. Between 18,000 the year, the shallow marsh being flooded 25% of the year, and 60,000 scans were needed for good signal-to-noise ratio and the uplands being rarely if ever flooded (Table 2). Deep dependent on the P concentration in the sample marsh zones had significantly lower pH, lower bulk density, Spectra were analyzed with NMR Utility Software (NUTS) and higher organic matter concentration (as indicated by loss initially using 15 Hz line broadening. Spectra were phased, on ignition) than other landscape positions (P < 0.05). Of the corrected for base line shift and calibrated to the MDP inter- total metals analyzed, Mg was at or below the practical quan- nal standard (chemical shift 17.47 ppm determined when tification limit for the majority of samples (data not shown). a spiked sample was calibrated against externally held 85% Calcium showed no significant trends across landscape posi-

H3PO4 set at 0 ppm). Spectra were integrated over specific tion, whereas Al and Fe concentrations showed a significant intervals to determine functional P groups (Table 1) based on increase within wetlands (P < 0.05). Post hoc analysis sug- literature values (Turner et al., 2003c). The region between gested that deep marsh areas were the most distinct, contain- 3 and 8 ppm was further analyzed using the deconvolution ing concentrations of up to twice as much Fe and 10 times as utility of NUTS software. A best-fit deconvolution of the much Al compared with the surrounding uplands. This mirrors spectra was acquired using 3 Hz line broadening and peak positive correlations between organic matter and Fe/Al within picking parameters of 5% of maximum peak height and 0.5 other wetland systems (Darke and Walbridge, 2000), and we for the root mean squared noise parameter. The region was hypothesize that this represents stable organo-metal complexes split into orthophosphate (average assignment 6.241 ppm, in the more organic deep marsh soils. standard deviation 0.001) and phosphomonoesters (all other The expected gradient in accumulating organic matter was peaks determined by the algorithm in the region 3 to 8 ppm). supported by analysis of total C and total N (Table 2), both Peak proportions from deconvolution were then applied to of which increased significantly (P < 0.05) with landscape the interval integration determined in the 15 Hz spectra. position toward the wetlands deep marsh. There was also sig- Repeated integration of the same spectrum provided confi- nificant (P < 0.05) interaction observed between landscape dence in the detection of signals within the pyrophosphate position and wetland site, attributed to longer hydroperiods − region equivalent to 1 mg P kg 1 soil. within the Beaty wetlands deep marsh (average 239 d) compared with Larsons (194 d) (data not shown). Data Analysis For most soil properties, significant differences among All statistical tests were performed in SPSS for Windows version landscape positions were the result of distinct deep marsh char- 17.0.0 statistical software (SPSS, 2008). Both visual inspec- acteristics. With the exception of total N, post hoc analysis tion and the Shapiro–Wilk test were used to test normality. If demonstrated only limited, nonsignificant differences between required, a natural log transformation was applied before sta- the pasture upland and shallow marsh zones (see “Impact of tistical analysis. Basic physiochemical characteristics and forms Hydroperiod” below). of P were analyzed with a simple univariate ANOVA with landscape position as the main factor and wetland site (Larson Soil Phosphorus Pools East/West, Beaty North/South) as a random factor. Post hoc There was a significant difference in total P concentration analysis (Tukey test) was applied to explore significant relation- across landscape position, with an approximate threefold ships determined between landscape positions. Relationships increase from 117 mg P kg−1 in the pasture uplands to 371 between calculated hydroperiod and biogeochemical character- mg P kg−1 in the deep marsh (P < 0.05) (Table 3, Fig. 2). This istics were explored by the fitting of linear, exponential, and significant gradient was also observed in the HCl-extractable linear segmented curves regressions to the data. inorganic P, which ranged from 23 mg P kg−1 (20% total P) in the uplands to 61 mg P kg−1 (16% total P) in the deep marsh.

Table 1. Specific integral ranges used in the classification of solution31 P nuclear magnetic resonance spectra for lyophilized material resuspended in 0.9 −1 −1 mL (1 mol L NaOH 100 mmol L ethylenediaminetetraacetic acid) + 0.1 mL D2O. Internal standard methylenediphosphonic acid (MDP) calibrated to 17.46 ppm. Region 8 to 3 ppm is determined as orthophosphate or phosphomonoesters after deconvolution of spectra at 3 Hz line broadening. Chemical shift Attributed phosphorus group 21.0 to 20.0 Phosphonates 18.0 to 16.8 MDP 8.0 to 3.0 Orthophosphate + phosphomonoesters Line broadening = 3Hz Orthophosphate 6.24 ± 0.01 ppm Phosphomonoesters = [8 to 3] minus orthophosphate 3.0 to −1.5 Phosphodiesters 0.5 to –1.5 DNA [3.0 to −1.5] minus DNA Other phosphodiesters −3.0 to −5.0 Pyrophosphate and end-chain polyphosphate functional groups −18.0 to −21.0 Polyphosphate mid chain functional groups

Journal of Environmental Quality • Volume 39 • July–August 2010 The AEM–NMR method extracted on average Table 2. Soil characteristics and nutrients determined in samples across landscape posi- 53% ± 11% of total P in all soils sampled, with no tion. Values represent means (n = 12) ± 1 SD. significant bias in the recovery rate among land- Pasture Shallow Deep scape positions. Phosphorus not extracted (residual upland marsh marsh a b c P) was considered a distinct unidentified pool, Hydroperiod (days)*** 2 ± 6† 92 ± 67 231 ± 57 a a b hypothesized to consist of recalcitrant organic and pH* 4.8 ± 0.6 4.7 ± 0.4 4.3 ± 0.3 −3 a a b alkaline-stable mineral forms (Cade-Menun, 2005; Bulk density (g cm ) * 0.98 ± 0.10 0.92 ± 0.21 0.65 ± 0.19 −1 a a b Turner et al., 2005). Organic matter‡ (g kg ) * 86 ± 23 123 ± 71 219 ± 66 −1 Exchangeable phosphate showed a significant Total elements (g kg ) a a b increase from 4.4 mg P kg−1 (3.8% total P) in Carbon* 44 ± 12 58 ± 34 112 ± 37 a b c the uplands to 23.9 mg P kg−1 (6.4% total P) in Nitrogen* 2.8 ± 0.5 3.6 ± 1.2 4.8 ± 0.9 the deep marsh (P < 0.05). Microbial P showed Calcium 1.1 ± 0.5 1.2 ± 1.0 1.9 ± 0.6 a a b a significant increase in concentration but a Iron* 0.7 ± 0.3 0.6 ± 0.4 1.1 ± 0.3 a a b reduction in its proportion of total P (P < 0.05), Aluminum*** 0.3 ± 0.2 0.7 ± 0.8 3.6 ± 0.8 from 22 mg P kg−1 (19% total P) in uplands to * Significant at the 0.05 probability level. 37 mg P kg−1 (10% total P) in the deep marsh. *** Significant at the 0.001 probability level. † Different letters in a row signify significantly different averages based on Tukey HSD test. Although the concentration of NaOHTP changed, its proportional contribution to the total P did ‡ Organic matter estimated by loss on ignition at 550°C for 4 h. not differ significantly with landscape position, ranging from 56 mg P kg−1 (49% total P) in the Table 3. Phosphorus forms across landscape position determined by acid extraction and anion exchange membrane–nuclear magnetic resonance (AEM–NMR) extraction −1 uplands to 184 mg P kg (51% total P) in the method. Values represent means (n = 12) ± 1 SD. deep marsh. This pattern was repeated in the Pasture Shallow Deep residual P fraction, with a significant (P < 0.05) upland marsh marsh difference in concentration, yet similar propor- Total P*(mg kg−1) 117a ± 25† 171a ± 107 371b ± 61 −1 tions of total P, from 56 mg P kg (47% total HCl-Pi‡ (% total P) 20 ± 5 19 ± 4 16 ± 2 −1 P) in uplands to 163 mg P kg (43% total P) in AEM-NMR Extraction (% total P) the deep marsh. Although concentrations of all PAEM§ 4 ± 2 6 ± 3 6 ± 2 P pools determined by the AEM–NMR method a a b PM§* 19 ± 5 16 ± 10 10 ± 4 (P , P , NaOH , and residual) showed signifi- AEM M TP NaOHTP¶ 49 ± 10 42 ± 8 51 ± 13 cant differences across landscape positions, the Residual# 47 ± 11 52 ± 8 43 ± 13 difference between pasture uplands and shallow * Significant at the 0.05 probability level. marsh were not significant. Given the trends in † Different letters in a row signify significantly different averages based on Tukey HSD test. data across landscape position, we attributed this ‡ Inorganic P extracted in 1 mol L−1 HCl. lack of significance to the high variability within § Exchangeable (P ) and microbial (P ) determined by AEM extraction. soils collected from the shallow marsh zone (see AEM M “Impact of Hydroperiod” below and Fig. 3). ¶ NaOH–ethylenediaminetetraacetic acid extracted total P. # Total P – (PAEM + NaOHTP). Solution 31P NMR Spectroscopy 31 was detected at low concentrations in certain deep marsh soils Solution P NMR spectra showed the presence of orthophos- − (average 1.5 mg P kg 1 soil). phate (PO ), phosphomonoesters, phosphodiesters, pyro- 4 The concentrations of phosphomonoesters, DNA, and phosphate, and trace concentrations of phosphonates and orthophosphate increased significantly from uplands to deep polyphosphates (Fig. 4, Table 4). Approximately 73% of the marsh (P < 0.05; Table 4). However, there was a striking lack extracted P was organic, with phosphomonoesters consti- of distinction between the relative proportions of P forms tuting the major fraction, ranging from 50 mg P kg−1 in the identified across landscape position. Instead of the expected uplands to 127 mg P kg−1 in the deep marsh. The lack of a changes in forms and magnitudes of P between wetlands and characteristic 1:2:2:1 signature within the phosphomonoes- uplands, all sites contained similar forms of P, leading us ter region indicated the absence of detectable concentrations to reject our initial hypothesis. The proportion of P identi- of myo-inositol hexakisphosphate (Turner et al., 2003d). The fied as phosphodiesters did not increase in the deep marsh remaining organic P included phosphodiesters (11–13% total as expected from previous studies of wetland soils (Turner soil P), dominated by DNA, with the remainder representing and Newman, 2005; Turner et al., 2006a). In addition, the various alkali-stable phospholipids. Due to the rapid hydroly- marked lack of the distinctive peak signature associated with sis of certain compounds in alkaline extracts (i.e., RNA and the various isomers of inositol hexakisphosphate, even in some phospholipids), the proportion of organic P determined upland pasture samples, was in contrast to other studies of as phosphodiesters is likely to be underestimated (Turner et al., upland grasslands (Tate and Newman, 1982; Turner et al., 2003c). Total inorganic P constituted between 16 and 19% of 2003b; McDowell and Stewart, 2006; Murphy et al., 2009). total P. This was dominated by orthophosphate, whereas pyro- The abiotic sorption and stabilization of inositol phosphates phosphate was detected in all spectra at concentrations ranging in such environments is attributed to their interaction with from 3.4 mg P kg−1 (2.0% total P) in the shallow marsh to 6.7 clay minerals and amorphous Fe and Al oxides (Celi and mg P kg−1 (1.8% total P) in the deep marsh. Polyphosphate The SPSS curve estimating procedure was applied to test the significance of linear regressions between the calculated hydroperiod and various soil P characteristics. There was a positive linear relationship between soil total P and hydroperiod (R2 = 0.734; P < 0.001; Fig. 4a). The potential for a more complex segmented relationship with a critical threshold hydroperiod at approximately 100 d was explored by the fitting of a segmented linear regression; however, there was not a substantially better fit to the data (R2 = 0.758). Phosphorus forms determined by solution 31P NMR spectroscopy were plotted against the average hydroperiod of the amalgamated soils used to generate the spectra. Total organic P determined by 31P NMR analysis showed a significant positive linear relationship with hydroperiod (R2 = 0.837; P < 0.001; Fig. 4b). Similar relationships were found for phosphomonoesters and phosphodiesters (Fig. 4c, d). Although total organic P Fig. 2. Phosphorus pools determined by anion exchange membrane– nuclear magnetic resonance (AEM–NMR) method across the three and its major forms showed similar positive trends in concen- landscape positions. Values are mean (n = 12) ± 1 SD. No significant tration with calculated hydroperiod, we initially hypothesized differences were found between Pasture Upland and Shallow Marsh. that increasing hydroperiod would result in increased stabili- (*) denotes significant differences between pools of the different landscape positions (P < 0.05). EDTA, ethylenediaminetetraacetic acid. zation of organic P and, as a result, lead to an increase in its proportion of total P, as well as the preferential stabilization of forms characteristic of wetter soils (i.e., phosphodiesters). However, we observed no evidence of a significant relationship between hydroperiod and organic P (as a proportion of total P), or the relative proportions of phosphodiesters to phosphomonoesters. Across landscape position, there was a general increase in total soil P toward the wetland deep marsh, but the concomitant increase in all partitioned forms of P suggests a mecha- nism of general accumulation, as opposed to a preferential stabilization of specific forms. The highly significant (P < 0.001) linear relationship between total C and total P (Fig. 5) suggests that the accumulation of P is concurrent with the accumulation of organic matter. Previous analysis of manure- impacted surface soils in the region has shown the upper soil horizons to consist of uncoated quartz sand grains and the low clay concentrations to be dominated by noncrystal- Fig. 3. Example solution 31P nuclear magnetic resonance spectra from line silica (Harris et al., 1994; Harris et al., 1996). Although amalgamated samples from the Beaty North wetland. Spectra acquired deeper horizons may interact with P strongly (Graetz and using a Bruker Avance 500 MHz with a 4.00 µs pulse (~30°), 0.4 s acquisition time and 1 s delay. Spectra referenced and integrated against Nair, 1995), mineral components within the surface soils internal methylenediphosphonic acid (MDP) standard (17.46 ppm), have a very low P binding capacity, resulting in P dynamics displayed here using 15 Hz line broadening and scaling to MDP. driven by an association with organic matter. The accumulation of organic matter in response to Barberis, 2007; Heighton et al., 2008). In addition to data increased hydroperiod may be a result of anaerobosis lead- presented, which report the low Fe and Al concentrations ing to the accumulation of endogenous organic matter or as throughout the system (Table 2), existing soil surveys (Lewis the status of wetlands as receiving bodies in the landscape. et al., 2003) and detailed particle-size distribution analysis The transportation of particulate P in runoff has long been (Bhadha and Jawitz, 2010) show only minimal clay content recognized as a significant component of P transfer (Daniel (<2.5%) in the uplands. Therefore, it is likely that inositol et al., 1998) and, therefore, a focus for management prac- phosphates are not physically stabilized within these soils and tices to reduce P loss from agricultural landscapes (Sharpley instead are hydrolyzed rapidly. et al., 2001; McDowell et al., 2007). The forms of P trans- Impact of Hydroperiod ported in surface runoff are dependent on soil biophysical characteristics and site hydrological properties (Heathwaite The use of a priori categorization using vegetation, soils, and and Dils, 2000; Ballantine et al., 2009). The direct deter- hydrology of the wetland–upland landscape continuum is a mination of surface runoff at dispersed wetland fringes qualitative approach to classifying the different landscape posi- with low overall topographic relief is problematic, although tions (deep marsh, shallow marsh, and upland). The use of a indirect modeling suggests that overland flow may play a calculated hydroperiod is a quantitative approach, giving an considerable role in the water budget of these study wet- absolute value for a hydroperiod of a given location. lands (J.H. Min, personal communication, 2010). Direct Journal of Environmental Quality • Volume 39 • July–August 2010 overland flow accounts for up to 25% of water inputs, often associated with isolated high-intensity precipitation events. Such high energy rainfall events contribute dis- proportionately to increased movement of particulate P in highly sloped tilled silt or loam soils (Jin et al., 2009). Yet we are unaware of studies that monitor particulate movement under conditions that emulate the site conditions, i.e., that include con- tiguous groundcover, low topographic relief, and soils dominated by siliceous sand grains and variously sized organic matter particles. In addition, all study wetlands are unfenced within grazed pastures. Given known cattle use of wetlands throughout the year (Pandey et al., 2009) and associated increases in potential sedimentation, vegetation turnover (McDowell et al., 2007), and direct fecal additions, the influence of cattle on organic matter accumulation may be significant. The importance of organic matter stabilization of P, even in the low organic matter upland soils, as well as the potential transloca- tion of this organic matter, could account for the absence of any distinct shifts in P forms Fig. 4. Phosphorus characteristics plotted against hydroperiod: (a) total phosphorus for seen across the landscape transition. Therefore, all samples, (b) total organic P determined by 31P nuclear magnetic resonance analysis, (c) further work is required to investigate how phosphomonoesters, and (d) phosphodiesters. endogenous and exogenous inputs influence Table 4. Phosphorus forms determined by solution 31P nuclear magnetic resonance (NMR) the C and P dynamics of these depressional spectroscopy of amalgamated alkaline extracts and quantified by integration against inter- wetland systems. nal standard methylenediphosphonic acid. Values from each landscape position represent means (n = 4) ± 1 SD. Phosphorus Storage Pasture Shallow Deep Phosphorus form If restoration of hydrologic conditions in iso- upland marsh marsh lated wetlands is to be an effective P manage- ——————— mg kg−1 ——————— ment practice, wetlands must show evidence Phosphonate 0.6 ± 1.2 trace 0.9 ± 1.4 a a b of increased nutrient storage on an aerial Orthophosphate* 18 ± 6† 24 ± 11 57 ± 9 a a b basis (kg ha−1) relative to other landscape Phosphomonoester* 50 ± 5 58 ± 16 127 ± 21 a a b positions. Although moderated by a reduced DNA* 6.4 ± 1.9 14.2 ± 2.2 29.7 ± 11.0 bulk density in wetlands relative to uplands Other phosphodiesters 8.8 ± 7.2 10.1 ± 0.7 12.5 ± 7.6 (Table 2), storage of both total and organic Pyrophosphate 5.3 ± 2.6 3.4 ± 2.4 6.7 ± 3.8 P is significantly (P < 0.05) elevated within Polyphosphate nd‡ nd 1.5 ± 2.9 the surface soils of the wetlands deep marsh Organic P (% of total NMR) 74 75 72 (Table 5), with a twofold increase in total P * Significant at the 0.05 probability level. from 114 kg ha−1 in the unimproved pasture † Different letters in a row signify significantly different averages based on Tukey HSD test. to 236 kg ha−1 in the deep marsh areas. The ‡ nd, not detected by 31P NMR spectroscopy. marginal, but not significant, increase in P content between pasture uplands and shallow Conclusions marsh suggests that meaningful nutrient storage requires a Isolated wetlands exhibited an accumulation of organic substantial increase in hydroperiod. In addition, comparing matter compared with surrounding pasture uplands, which current landscape storage does not take into account the was closely associated with an accumulation of total P. time scale required for its development. It can be assumed However, this was not related to differences in soil P forms that an increase in hydroperiod would require a period of predicted from differential stability across the landscape. adjustment during which P and C would be accreted. The Instead, total P accumulation was concomitant with an rate and controlling mechanisms of this accretion would increase in all P forms, including exchangeable P, microbial have significant impact on the efficacy of restored wetlands P, and forms identified by alkaline extraction and 31P NMR to reduce downstream P loading. spectroscopy. Successful management of isolated wetlands for the onsite sequestering of P from agricultural runoff will Capece, J.C., K.L. Campbell, P.J. Bohlen, D.A. Graetz, and K.M. Portier. 2007. Soil phosphorus, cattle stocking rates, and water quality in subtropical pastures in Florida, USA. Rangeland Ecol. Manag. 60:19–30. Celi, L., and E. Barberis. 2005. Abiotic stabilization of organic phosphorus in the environment. p. 113–132. In B.L. Turner et al. (ed.) Organic phosphorus in the environment, CABI Publ., Wallingford, UK. Celi, L., and E. Barberis. 2007. Abiotic reactions of inositol phosphates in soil. p. 207–220. In B.L. Turner et al. (ed.) Inositol phosphates: Linking agriculture and the environment. CABI Publ., Wallingford, UK. Chapuis-Lardy, L., M. Brossard, and H. Quiquampoix. 2001. Assessing organic phosphorus status of Cerrado oxisols (Brazil) using 31P NMR spectroscopy and phosphomonoesterase activity measurement. Can. J. Soil Sci. 81:591–601. Cheesman, A.W., B.L. Turner, and K.R. Reddy. 2010. Interaction of organic and condensed inorganic phosphorus compounds with anion-exchange membranes: Implications for soil phosphorus analysis. Soil Sci. Soc. Am. J. (in press). Condron, L.M., E. Frossard, H. Tiessen, R.H. Newman, and J.W.B. Stew- art. 1990. Chemical nature of organic phosphorus in cultivated and uncultivated soils under different environmental conditions. J. Soil Sci. 41:41–50. Fig. 5. Total soil carbon plotted against phosphorus across Craft, C.B., and C.J. Richardson. 1993. Peat accretion and N, P, and organic C landscape positions (n = 36) from four wetland sites showing a accumulation in nutrient-enriched and unenriched Everglades peatlands. significant (P < 0.001) positive relationship (R2 = 0.819). Ecol. Appl. 3:446–458. Daniel, T.C., A.N. Sharpley, and J.L. Lemunyon. 1998. Agricultural phosphorus and eutrophication: A symposium overview. J. Environ. Qual. 27:251–257. Table 5. Storage of total (n = 12) and organic P (n = 4) as determined by Darke, A.K., and M.R. Walbridge. 2000. Al and Fe biogeochemistry in a flood- solution 31P nuclear magnetic resonance analysis within the top 10 cm plain forest: Implications for P retention. Biogeochemistry 51:1–32. of soil, across landscape positions. Values from each landscape posi- DeBusk, W.F., and K.R. Reddy. 2003. Nutrient and hydrology effects on tion represent means ± 1 SD. soil respiration in a northern Everglades marsh. J. Environ. Qual. 32:702–710. Phosphorus form Pasture Shallow Deep Dunne, E.J., J. Smith, D.B. Perkins, M.W. Clark, J.W. Jawitz, and K.R. Reddy. (kg ha−1) upland marsh marsh 2007. Phosphorus storages in historically isolated wetland ecosystems Total P* 114a ± 27† 141a ± 50 236b ± 62 and surrounding pasture uplands. Ecol. Eng. 31:16–28. Organic P* 65a ± 10 75a ± 19 108b ± 33 Flaig, E.G., and K.R. Reddy. 1995. Fate of phosphorus in the Lake Okeechobee watershed, Florida, USA: Overview and recommendations. Ecol. Eng. * Significant at the 0.05 probability level. 5:127–142. † Different letters signify significantly different averages based on Tukey Fla. Stat. §373.4595. 2009. Northern Everglades and estuaries protection pro- HSD test. gram. Available at http://www.leg.state.fl.us (verified 13 May 2010). Gathumbi, S.M., P.J. Bohlen, and D.A. Graetz. 2005. Nutrient enrichment of depend on the accumulation and stabilization of organic wetland vegetation and sediments in subtropical pastures. Soil Sci. Soc. Am. J. 69:539–548. matter. This study suggests that a substantial hydroperiod Gornak, S.I., and J. Zhang. 1999. A summary of landowner surveys and water increase is required to sequester significantly increased quality data from the northern Lake Okeechobee watershed. Appl. Eng. amounts of P. Agric. 15:121–127. Graetz, D.A., and V.D. Nair. 1995. Fate of phosphorus in Florida Spodosols Acknowledgments contaminated with cattle manure. Ecol. Eng. 5:163–181. Harris, W.G., H.D. Wang, and K.R. Reddy. 1994. Dairy manure influence on We thank Yu Wang, University of Florida, and Alex Blumenfeld, soil and sediment composition: Implications for phosphorus retention. University of Idaho, for analytical support. The project was supported J. Environ. Qual. 23:1071–1081. by a grant from the USDA–CREES National Research Initiative (No. Harris, W.G., R.D. Rhue, G. Kidder, R.B. Brown, and R. Littell. 1996. Phos- 2004-35107-14918) and the Florida Department of Agriculture and phorus retention as related to morphology of sandy coastal plain soil materials. Soil Sci. Soc. Am. J. 60:1513–1521. Consumer Services. Four anonymous reviewers provided constructive Harrison, A.F. 1987. Soil organic phosphorus: A review of world literature. comments that helped to improve the final manuscript. CAB Int., Wallingford, UK. Heathwaite, A.L., and R.M. Dils. 2000. Characterising phosphorus loss in References surface and subsurface hydrological pathways. Sci. Total Environ. 251:523–538. Andersen, J.M. 1976. An ignition method for determination of total phospho- Heighton, L., W.F. Schmidt, and R.L. Siefert. 2008. Kinetic and equilib- rus in lake sediments. Water Res. 10:329–331. rium constants of phytic acid and ferric and ferrous phytate derived Axt, J.R., and M.R. Walbridge. 1999. Phosphate removal capacity of palus- from nuclear magnetic resonance spectroscopy. J. Agric. Food Chem. trine forested wetlands and adjacent uplands in Virginia. Soil Sci. Soc. 56:9543–9547. Am. J. 63:1019–1031. Jin, K., W.M. Cornelis, D. Gabriels, M. Baert, H.J. Wu, W. Schiettecatte, Ballantine, D., D.E. Walling, and G.J.L. Leeks. 2009. Mobilisation and trans- D.X. Cai, S. De Neve, J.Y. Jin, R. Hartmann, and G. Hofman. 2009. port of sediment-associated phosphorus by surface runoff. Water Air Soil Residue cover and rainfall intensity effects on runoff soil organic carbon Pollut. 196:311–320. losses. Catena 78:81–86. Bhadha, J.H., and J.W. Jawitz. 2010. Characterizing deep soils from an im- Khan, F.A., and A.A. Ansari. 2005. Eutrophication: An ecological vision. Bot. pacted subtropical isolated wetland: Implications for phosphorus stor- Rev. 71:449–482. age. J. Soils Sediments 10:514–525. Kouno, K., Y. Tuchiya, and T. Ando. 1995. Measurement of soil microbial bio- Cade-Menun, B.J. 2005. Using phosphorus-31 nuclear magnetic resonance mass phosphorus by an anion-exchange membrane method. Soil Biol. spectroscopy to characterize organic phosphorus in environmental sam- Biochem. 27:1353–1357. ples. p. 21–44. In B.L. Turner et al. (ed.) Organic phosphorus in the Lewis, D.L., K.J. Liudahl, C.V. Noble, and L.J. Carter. 2003. Soil survey of environment. CABI Publ., Wallingford, UK. Okeechobee County, Florida. Available at http://websoilsurvey.nrcs. Cade-Menun, B.J., and C.M. Preston. 1996. A comparison of soil extraction usda.gov (verified 13 May 2010). USDA–NRCS, Washington, DC. procedures for 31P NMR spectroscopy. Soil Sci. 161:770–785.

Journal of Environmental Quality • Volume 39 • July–August 2010 McDowell, R.W., D.M. Nash, and F. Robertson. 2007. Sources of phosphorus Steinman, A.D., and B.H. Rosen. 2000. Lotic-lentic linkages associated with lost from a grazed pasture receiving simulated rainfall. J. Environ. Qual. Lake Okeechobee, Florida. J. North Am. Benthol. Soc. 19:733–741. 36:1281–1288. Sumann, M., W. Amelung, L. Haumaier, and W. Zech. 1998. Climatic effects McDowell, R.W., and I. Stewart. 2006. The phosphorus composition of con- on soil organic phosphorus in the North American Great Plains identi- trasting soils in pastoral, native and forest management in Otago, New fied by phosphorus-31 nuclear magnetic resonance. Soil Sci. Soc. Am. J. Zealand: Sequential extraction and 31P NMR. Geoderma 130:176–189. 62:1580–1586. McKee, K.L. 2005. Predicting soil phosphorus storage in historically isolated Suzumura, M., and A. Kamatani. 1995. Mineralization of inositol hexaphos- wetlands within the Lake Okeechobee priority basins. Master’s thesis, phate in aerobic and anaerobic marine-sediments- implications for the Univ. of Florida, Gainesville. phosphorus cycle. Geochim. Cosmochim. Acta 59:1021–1026. Mitsch, W.J., and J.W. Day. 2006. Restoration of wetlands in the Mississippi- Tate, K.R., and R.H. Newman. 1982. Phosphorus fractions of a climose- Ohio-Missouri (MOM) river basin: Experience and needed research. quence of soils in New-Zealand tussock grassland. Soil Biol. Biochem. Ecol. Eng. 26:55–69. 14:191–196. Moreno, D., C. Pedrocchi, F.A. Comin, M. Garcia, and A. Cabezas. 2007. Turner, B.L., B.J. Cade-Menun, L.M. Condron, and S. Newman. 2005. Ex- Creating wetlands for the improvement of water quality and landscape traction of soil organic phosphorus. Talanta 66:294–306. restoration in semi-arid zones degraded by intensive agricultural use. Turner, B.L., B.J. Cade-Menun, and D.T. Westermann. 2003a. Organic phos- Ecol. Eng. 30:103–111. phorus composition and potential bioavailability in semi-arid arable soils Murphy, P.N.C., A. Bell, and B.L. Turner. 2009. Phosphorus speciation in of the western United States. Soil Sci. Soc. Am. J. 67:1168–1179. temperate basaltic grassland soils by solution 31P NMR spectroscopy. Turner, B.L., L.M. Condron, S.J. Richardson, D.A. Peltzer, and V.J. Allison. Eur. J. Soil Sci. 60:638–651. 2007. Soil organic phosphorus transformations during pedogenesis. Eco- Myers, R.G., S.J. Thien, and G.M. Pierzynski. 1999. Using an ion sink to ex- systems 10:1166–1181. tract microbial phosphorus from soil. Soil Sci. Soc. Am. J. 63:1229–1237. Turner, B.L., N. Mahieu, and L.M. Condron. 2003b. The phosphorus compo- Newman, S., and J.S. Robinson. 1999. Forms of organic phosphorus in water, sition of temperate pasture soils determined by NaOH–EDTA extraction soils and sediments. p. 207–224. In K.R. Reddy et al. (ed.) Phosphorus and solution 31P NMR spectroscopy. Org. Geochem. 34:1199–1210. biogeochemistry in subtropical ecosystems. CRC Press, Boca Raton, FL. Turner, B.L., N. Mahieu, and L.M. Condron. 2003c. Phosphorus-31 nuclear Paludan, C., F.E. Alexeyev, H. Drews, S. Fleischer, A. Fuglsang, T. Kindt, P. magnetic resonance spectral assignments of phosphorus compounds in Kowalski, M. Moos, A. Radlowki, G. Stromfors, V. Westberg, and K. soil NaOH–EDTA extracts. Soil Sci. Soc. Am. J. 67:497–510. Wolter. 2002. Wetland management to reduce Baltic Sea eutrophication. Turner, B.L., N. Mahieu, and L.M. Condron. 2003d. Quantification of myo- Water Sci. Technol. 45:87–94. inositol hexakisphosphate in alkaline soil extracts by solution 31P NMR Pandey, V., G.A. Kiker, K.L. Campbell, M.J. Williams, and S.W. Coleman. spectroscopy and spectral deconvolution. Soil Sci. 168:469–478. 2009. GPS monitoring of cattle location near water features in south Turner, B.L., and S. Newman. 2005. Phosphorus cycling in wetland soils: The Florida. Appl. Eng. Agric. 25:551–562. importance of phosphate diesters. J. Environ. Qual. 34:1921–1929. Pant, H.K., and K.R. Reddy. 2001. Hydrologic influence on stability of or- Turner, B.L., S. Newman, and J.M. Newman. 2006a. Organic phosphorus ganic phosphorus in wetland detritus. J. Environ. Qual. 30:668–674. sequestration in subtropical treatment wetlands. Environ. Sci. Technol. Perkins, D.B., A.E. Olsen, and J.W. Jawitz. 2005. Spatially distributed iso- 40:727–733. lated wetlands for watershed-scale treatment of agricultural runoff. p. Turner, B.L., S. Newman, and K.R. Reddy. 2006b. Overestimation of organic 80–91. In E.J. Dunne et al. (ed.) Nutrient management in agricultural phosphorus in wetland soils by alkaline extraction and molybdate colo- watersheds: A wetlands solution. Wageningen Academic Publishers, Wa- rimetry. Environ. Sci. Technol. 40:3349–3354. geningen, the Netherlands. Turner, B.L., and T.E. Romero. 2009. Short-term changes in extractable inor- Reddy, K.R., E.G. Flaig, and D.A. Graetz. 1996. Phosphorus storage capacity ganic nutrients during storage of tropical rain forest soils. Soil Sci. Soc. of uplands, wetlands and streams of the Lake Okeechobee watershed, Am. J. 73:1972–1979. Florida. Agric. Ecosyst. Environ. 59:203–216. USEPA. 1993. Methods for chemical analysis of water and wastes. Environ- Reddy, K.R., Y. Wang, W.F. Debusk, M.M. Fisher, and S. Newman. 1998. mental Monitoring Support Laboratory, Cincinnati, OH. Forms of soil phosphorus in selected hydrologic units of the Florida Ev- Verhoeven, J.T.A., B. Arheimer, C. Yin, and M.M. Hefting. 2006. Regional erglades. Soil Sci. Soc. Am. J. 62:1134–1147. and global concerns over wetlands and water quality. Trends Ecol. Evol. Reddy, K.R., R.G. Wetzel, and R.H. Kadlec. 2005. Biogeochemistry of phos- 21:96–103. phorus in wetlands. p. 263–316. In J.T. Sims and A.N. Sharpley (ed.) Wetzel, R.G. 1999. Organic phosphorus mineralization in soils and sediments. Phosphorus: Agriculture and the environment. ASA, Madison, WI. p. 225–245. In K.R. Reddy et al. (ed.) Phosphorus biogeochemistry in Rowland, A.P., and P.M. Haygarth. 1997. Determination of total dissolved subtropical ecosystems. CRC Press, Boca Raton, FL. phosphorus in soil solutions. J. Environ. Qual. 26:410–415. Zhang, J., T. James, and P. McCormick. 2009. Lake Okeechobee Protection Sharpley, A.N., P. Kleinman, and R. McDowell. 2001. Innovative manage- Program- State of the Lake and Watershed. In G. Redfield (ed.) South ment of agricultural phosphorus to protect soil and water resources. Florida environmental report. South Florida Water Management Dis- Commun. Soil Sci. Plant Anal. 32:1071–1100. trict, West Palm Beach, FL. SPSS. 2008. SPSS for windows, Rel. 17.0.0. SPSS Inc., Chicago, IL.