(2000) Relating Soil Phosphorus Indices to Potential ...

1 Relating soil Phosphorus Indices to Potential Phosphorus release to WaterP. S. Hooda,* A. R. Rendell, A. C. Edwards, P. J. A. Withers, M. N. Aitken, and V. W. TruesdaleABSTRACTR elationships between soil test Phosphorus (STP) and release ofP in surface and subsurface runoff are needed to help identify sourceareas for implementing management strategies to limit P loss to determine whether soil P release could be predicted either bySTP values, sorption-desorption Indices , or the degree of soil satura-tion with Phosphorus (DSSP), 11 sites with contrasting chemical prop-erties and management histories were sampled from long-term fieldtrials in the UK. Each site offered up to three treatments, resultingin a total of 29 soil samples. The results showed that the amount ofP desorbed using a successive dilution procedure had no relationshipwith either total soil P content or P sorption capacity.

2 The mostsignificant property was the extent of P saturation. There was littledesorption for DSSP values below 10%; above this point, the amountof P desorbed increased linearly with the DSSP. Five STP methods(Olsen, Mehlich-3, acidified ammonium oxalate-oxalic acid, Fc2O,-coated paper strip, and distilled water ) were compared to predict theireffectiveness in predicting Potential P release to water . While STPvalues obtained using acidified ammonium oxalate proved to be leasteffective, those extracted with water correlated best with the amountof P desorbed, accounting for 96% of the variability in differential Prelease from the transport in agricultural runoff can ac-celerate eutrophication in surface waters ( Hooda, Rendell, and Truesdale, School of Biologicaland Molecular Sciences, Oxford Brookes Univ., Oxford OX3 OBP,UK; Edwards, Macaulay Land Use Research Institute, Craigie-buckler, Aberdeen AB15 8QH; Withers, ADAS Bridgets, Win-chester SO21 1AP; and Aitken, Scottish Agricultural College,Auchincruive, Ayr KA6 5HW, UK.)

3 Received 11 Aug. 1999. '''Corre-sponding author in J. Environ. Qual. 29:1166-1171 (2000).and Menzel, 1987). The loss of P from agricultural soilsin surface runoff or subsurface drainage therefore canhave potentially serious implications for water magnitude of P loss from soils is influenced byvarious factors, including land use (Nearing et al., 1993), soil type (Miller, 1979; Hooda et al., 1997), soil P (Heck-rath et al., 1995; Pote et al., 1996; Hooda et al., 1999),amount and intensity of rainfall (Edwards and Daniel,1993), and time and rate of fertilizer and manure appli-cation (Edwards and Daniel, 1993; Hooda et al., 1999).However, soil P is the underlying source of P in agricul-tural runoff, and increases in soil P status due to surplusP inputs therefore represent a long-term eutrophica-tion Potential loss of P from soils has been assessedusing STP methods ( , Olsen, Mehlich-3, and aceticacid) that were originally designed to estimate plantavailable P (Miller et al.

4 , 1993). A certain degree ofrelationship has been demonstrated to exist betweenSTP and the loss of P in runoff. These relationshipsgenerally have been tested for situations having similarsoil type (Heckrath et al., 1995; Austin et al., 1996;Pote et al., 1996; Hooda et al., 1999). However, recentresearch suggested that an approach based upon STPmight lead to ambiguous conclusions, particularly whendifferent soils are compared, due to differences in thenature of P-adsorbing surfaces. For example, Sharpley(1995), assessing P loss from a group of 10 soils amendedwith poultry litter, showed that two soils of 200 mg kg"1 Abbreviations: STP, soil test Phosphorus ; DSSP, degree of soil satura-tion with Phosphorus ; DRP, dissolved reactive Phosphorus ; PSI, phos-phorus sorption index; PSM, Phosphorus sorption ET AL.: Phosphorus release TO WATERT able 1. Land use, range of P inputs, and classification of the soils from 11 long-term field seriesLand use and management detailsRange of P Classificationkg P ha-1 yr-1 StiflingGrassland; fertilized and/or cattle slurry applied45-75 Silty clay loamTypic HaplaqueptDessGrass-barley rotation; no fertilizer or fertilized0--50 Sandy clay loamAlflc FragiortodInschGrass-barley rotation; no fertilizer or fertilized0-40 Sandy loamTypic CryochreptFoudlandGrass-barley rotation; no fertilizer or fertilized0-40 Sandy loamTypic FragiortodStonehavenGrass-arable rotation; no fertilizer or fertilized0-40 Sandy loamTypic CryochreptBromyardCereals; no fertilizer, fertilized, or cattle slurry applied0-82 Silty clay loamTypic HapludalfCuckneyCereals; fertilized, no fertilizer, or poultry litter applied0-300 Loamy sandTypic UdipsammentBecclesCereals; fertilized or no fertilizer0-44 Clay loamTypic AlbaqualfPeebles-1 Grassland.

5 No fertilizer or sewage sludge applied0-130 Sandy loamTypic DystrochreptCapringtonGrassland; no fertilizer or sewage sludge applied0-130 Clay loamTypic HaplaqueptPeebles-2 Grassland; sewage sludge applied115-355 Sandy clay loamTypic HumaqueptMehlich-3 P supported a dissolved reactive Phosphorus (DRP) concentration of mg -1 and mg et al. (1997), monitoring P outputs in six agricul-tural catchments in Scotland, provided a more tangibleevidence of soil extractable P being an ambiguous indexof assessing P loss in agricultural runoff. The P lossesin runoff from well-drained podzols were more thanfour times smaller than those from the poorly drainedgleysols, despite the fact that the former group of soilshad four times larger M acetic acid soil extractable-P tests might not be expectedto provide a good estimation of the P loss in runoff,because these (agronomic) tests extract a large pool P accessible to growing plants rather than the P fractionthat might be potentially released to draining of the ease with which P is desorbed fromdifferent soils and the factors that affect the processtherefore might help better assess the soils that arepotentially vulnerable to P loss to water .

6 For example,the amount of P that can be easily desorbed is morelikely to be dependent on the DSSP than the extract-able-P content, as the soil -specific soil -phosphate bind-ing strength will decrease with increasing P loading (Ry-den and Syers, 1977; Hooda et al., 1999). A measure ofthe DSSP therefore might provide a better estimate ofthe Potential for P loss to water than soil total- andextractable-P or P sorption capacity (Sharpley, 1995);however, this approach has not been fully paper reports a study designed specifically totest the relationships between a range of soil test meth-ods with potentially desorbable P using soils collectedfrom long-term field trials receiving fertilizer, cattleslurry, or sewage sludge as the source of P inputs. Thespecific objectives of the work were to (i) examine therelationship between soil total P, sorption-desorptionindices, DSSP, and the amount of P that can be poten-tially released to water and (ii) determine which of thecommonly used soil tests best predict Potential loss ofP to water , particularly when compared across a rangeof AND METHODSLand Use, soil Characteristics, and AnalysesA range of soils with contrasting chemical properties andmanagement histories under both grassland and arable farm-ing systems were collected from 11 long-term field trials fromvarious locations in the UK.

7 The land use, ranges of P inputs,management histories, and classifications of the soils are listedin Table 1. Some of the grassland sites had received cattleslurry or superphosphate or a combination of the two; othershad received P inputs through sewage sludge applications. Thearable sites had either superphosphate, cattle slurry, or poultrylitter as the source of P. Each soil offered up to three treat-ments, resulting in a total of 29 soil samples. Representativesurface samples (0-20 cm) of the 11 sites (Table 1) were lected, air-dried, gently ground, and passed through a 2-ramsieve prior to use in these size distribution, pH, organic matter, and calciumcarbonate equivalent were determined using standard proce-dures given in Rowell (1994). For total P, soils were firstdigested in 60% HC104 followed by a mixture of concentratedHNO3 and HC104 before being diluted with distilled waterand filtered through acid-resistant filter paper. Inorganic P inthe soils was extracted with M H2SO4, and organic P wasassumed to be the difference between total and inorganic P(Olsen and Sommers, 1982).

8 Extractable P in the soil samples was determined using fivedifferent methods and extractants: Mehlich-3 (Mehlich, 1984),Olsen (Olsen et al., 1954), iron oxide-impregnated paper strips(Sharpley, 1993), acidified ammonium oxalate-oxalic acid(Evans and Wilson, 1985), and distilled water (Sissingh, 1971).The Olsen and Mehlich-3 methods are commonly used for soilfertility advisory purposes. The ammonium oxalate method isdesigned to remove the P associated with amorphous oxidesof Fe and AI (the two principal components of P sorption innon-calcareous soils) while iron oxide-impregnated strips actas a sink and have been shown to correlate with algal biomass(Sharpley, 1993). This method therefore may provide a mea-sure of the potentially bioavailable P in runoff from soils. Thedistilled water method is designed to extract easily desorbableP (Sissingh, 1971). The extraction procedure involved adding2 mL distilled water to each soil sample weighing g in 50-mL centrifuge tubes.

9 Following a contact time of 20 h, a further25 mL distilled water was added and the mixtures were shakenfor 16 h before centrifugation at 27 000 g and filtration throughWhatman no. 42 filter , iron, and aluminum concentrations in the am-monium oxalate extracts were determined using inductivelycoupled plasma atomic emission spectroscopy. Phosphorusconcentrations in all other digests/extracts were determinedusing the colorimetric method of Murphy and Riley (1962).Where required, samples were neutralized prior to results are the means of two individual digestions/extrac-tions and are expressed on an oven dry (105 C) mass average values of selected soil properties are listed inTable 2. The DSSP was computed using the P, Fe, and AIcontents (mmol kg-1) extracted with the acidified ammoniumoxalate-oxalic acid:DSSP (%) [P] 100[Fe + All[1]Breeuwsma and Silva (1992) calculated DSSP as a functionof the ratio of [P] [Fe + All. They reported mmol P as1168J.]]

10 ENVIRON. QUAL., VOL. 29, JULY-AUGUST 2000 Table 2. Selected physical and chemical characteristics of the studied seriesClayOM ~CCE$Feox + Aloe pHTotal P rangeg kg-tmg P OM, organic matter.~CCE, CaCO~ equivalent. Feo~ and AIo~ = ammonium oxalate-extractable Fe and maximum amount of P that can be sorbed per mmol of[Fe + All. However, this approach of calculating a sorptionmaximum as a function of [Fe +A1] is arbitrary, as P sorp-tion maxima will increase with the time of equilibration andthe initial concentration ( Chardon, 1999, personal com-munication). Any such factor therefore should not be appliedarbitrarily unless it has been pre-determined for individualsituations ( , soil type, Potential environmental impact).Adsorption-Desorption StudiesPhosphorus sorption and desorption characteristics of thesoils were studied using the equilibrium-based Freundlich andLangmuir models. Although these models are empirical innature and cannot be used for delineating sorption or desorp-tion mechanisms at soil surfaces (Sparks, 1995), they provideuseful Indices for comparing different soils.]