JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 7 November 2005
Published in J Environ Qual 34:2118-2128 (2005)
DOI: 10.2134/jeq2005.0013
© 2005 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cox, J. W.
Right arrow Articles by van Leeuwen, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cox, J. W.
Right arrow Articles by van Leeuwen, J.
Agricola
Right arrow Articles by Cox, J. W.
Right arrow Articles by van Leeuwen, J.
Related Collections
Right arrow Water Quality
Right arrow Preferential Flow
Right arrow Phosphorus
Right arrow Runoff
Right arrow Nutrient Management

Phosphorus Workshop

Using Gypsum to Reduce Phosphorus in Runoff from Subcatchments in South Australia

J. W. Coxa,*, J. Varcoeb, D. J. Chittleboroughb and J. van Leeuwena

a CSIRO Land and Water, PMB 2, Glen Osmond, SA, 5064, Australia
b School of Earth and Environmental Sciences, University of Adelaide, SA 5005, Australia

* Corresponding author (jim.cox{at}csiro.au)

Received for publication January 15, 2005.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Concentrations of phosphorus (P) in runoff from agricultural catchments in southern Australia are high and well above national and international limits. Phosphorus was found to exit two subcatchments of 3.6 and 4.2 ha in the Adelaide hills via both overland flow and interflow. The subcatchments had texture-contrast soils with high inputs of superphosphate and were openly grazed by cattle all year. Interflow at the boundary of the B and C soil horizons accounted for as much as half the total water flow that was measured (overland flow, A–B interflow, and B–C interflow). The average flow-weighted concentration of total P within overland flow was as high as 0.25 mg L–1, and 0.05 mg L–1 in B–C interflow. In most years P loss was in the dissolved (<0.45 µm) form. In some years, interflow was the major pathway for P loss off these catchments. The B–C interflow cannot be discounted when searching for management options to reduce P loss from texture-contrast soils to waterways. Preliminary laboratory experiments showed promise that gypsum could modify agricultural soils and reduce the concentrations of P (and dissolved organic C) in runoff before it enters public water supply reservoirs. In this study, gypsum, applied at a rate of 15 Mg ha–1 to the 4.2-ha subcatchment, substantially modified the soil chemistry, and thereby soil structure. The size and stability of structural aggregates increased markedly and this change affected not only the A but also the upper B horizons, to a profile depth of approximately 50 cm. However, the impact of these physicochemical changes on P concentrations in runoff was not marked. Average profile P concentrations were only slightly lower in the runoff from the subcatchment following treatment. The high subsoil macroporosity of the gypsum-treated subcatchment caused an increase in the proportion of runoff by interflow.

Abbreviations: CEC, cation exchange capacity • DP, dissolved phosphorus • EP, extractable phosphorus • ESP, exchangeable sodium percentage • G0 and G15, soils treated with 0 and approximately 15 Mg ha–1 of gypsum, respectively • ICP–OES, inductively coupled plasma–optical emission spectrometry • MRP, molybdate-reactive phosphorus • PP, particulate phosphorus • TP, total phosphorus • XRF, X-ray fluorescence


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE FULL IMPACT of diffuse sources of P from fertilized agricultural lands in South Australia, particularly the movement of particulate (>0.45 µm) P in overland flow of eroded soil material, is now well understood, and point sources of P to streams have substantially been reduced as a result of legislation (Davis et al., 1998). Management aimed at minimizing erosion and thereby P loss was introduced (such as no-till cropping; Malinda, 1995), and buffer strips (vegetation planted at strategic points between where P is applied and a waterway; e.g., Daniels and Gilliam, 1996) were installed. In some Australian regions this management was effective in reducing particulate P in runoff to streams whereas in others, stream P levels continued to increase. The reasons for this were not initially clear but thought to be related to the unusual nature of southern Australian soils (strong texture-contrast) and peculiarity of climate (xeric; i.e., strong seasonal wetting and drying cycles). The reason for this process in those South Australian landscapes with texture-contrast soils was partly explained by Fleming and Cox (2001), who measured the relative proportions of particulate phosphorus (PP) versus dissolved phosphorus (DP) in runoff from agricultural catchments. They showed that the concentration of DP in runoff was similar each month. In contrast, PP dominated runoff at the start of the growing season when the sandy loam over clay soils had minimal vegetation cover and were vulnerable to erosion. Increased rainfall intensity later in the growing season, when pastures were established, did not significantly increase PP (Fleming and Cox, 1998). Whereas conservation tillage practices such as surface application of fertilizers (McIsaac et al., 1995) can lead to increased soluble P losses, PP fractions can represent 3 to 40 times more P than the soluble fractions from some cropland (Bundy et al., 2001). In South Australian pasture systems, however, PP dominates runoff from parts of the landscape where easily erodible (dispersive) sodic soils have formed (Cox et al., 2002). These are usually waterlogged (seepage) areas within the catchment and it has been found that an order of magnitude more P and sediment is lost under seepage conditions than under drainage conditions (Zheng et al., 2004). Cox et al. (2000) found that PP could move through clay subsoils within macropores. The amount of P mobilized was controlled by the "residence time," the time taken for water to move through a soil with macropores.

Kirkby et al. (1997) found that P loss was influenced by the wetness of the soil. Phosphorus in drainage from wet soil cores was significantly less than P lost through dry soils. However, soils that are saturated for long periods of time tend to have more P-loss potential than well-drained soils (Baxter et al., 2002), because ferric phosphate minerals can be reduced to ferrous phosphates that are more soluble and susceptible to losses via overland flow.

It was thought that to reduce P loss to surface water storages across the agricultural regions of southern Australia, management must do more than just reduce erosion at the start of the growing season. Management must reduce sodicity or at least stabilize sodic soils, keeping them from becoming saturated, and increase the residence time of P, thereby allowing mineral and organic fractions sufficient time to sorb P. Soil chemical amendments, such as red mud (a by-product of bauxite mining, e.g., Ho et al., 1989; Vlahos et al., 1989; Kayaalp et al., 1998), which Summers et al. (1993) found to retain up to 70% more P than untreated soil (at an application rate of 80 Mg ha–1) in a large scale field trial, and polyacrylamides (e.g., Lentz et al., 1998), which can bind P to soil, have also been tested. Recent work (Churchman, 2002; Kleinig et al., 2003) has shown that clays can be modified using poly-DADMAC (poly-diallyldimethyl ammonium chloride), a water-soluble polymer that is used extensively in drinking water treatment, so that the clays acquire a positive charge and are able to sorb substantial amounts of P from water. Substantial reductions in P loss in overland flow have been recently demonstrated in small-scale field trials where poly-DADMAC was applied by spraying. However, the cost of spreading polyacrylamides and poly-DADMAC on a large scale is currently prohibitive. Similarly, the major drawback of using red mud is the cost in transport, as no local sources are available in South Australia. Also, some of the industrial by-products high in Al and Fe are quite efficient in sorbing P but have the disadvantage of containing heavy metals in toxic amounts (McDowell, 2004).

Calcium amendments in the form of calcium carbonate or gypsum have not been extensively studied as soil amendments to attenuate P loss (e.g., Sova, 1996). In a column leaching study, Nelson et al. (1991) reduced dissolved organic C by 44% and DP by 88% in extracts from Alfisols with a sandy loam surface texture treated with gypsum. They attributed this to a combination of three possible mechanisms: (i) Ca acting as a "bridge" which links clay and organic matter, (ii) formation of insoluble Ca–organic matter complexes, and (iii) the flocculating effect on clay–organic complexes of the high electrolyte concentration of the gypsum.

Gypsum was selected as the preferred amendment to reduce P loss from agricultural catchments in South Australia based on: (i) a proven history of yield improvement when applied to some soils (e.g., Sumner, 1995), (ii) being a good source of Ca because of the relatively high solubility of gypsum (compared with liming agents), and (iii) its ready availability and relative cheapness in South Australia which gives it a potentially cost-effective advantage. Previous work by Moore and Miller (1994), Coale et al. (1994), and Stout et al. (1998) indicated that gypsum may be successful in reducing P solubility (through enhancing Ca-P precipitation).

The aims of this research were to determine the pathways that water and P move from grazed agricultural subcatchments with texture-contrast soils, to measure the timing, quantity, and forms of the P in the flow paths, and to determine the effect of a large quantity of surface-applied gypsum on the timing, quantity, forms, and pathways of flow of P.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Field Sites
Two subcatchments were selected in the Adelaide hills within the watershed of the Mt. Bold reservoir, a major source of the domestic water for the city of Adelaide, South Australia. The subcatchments were adjacent to one another and hydrologically isolated by a central valley (Fig. 1) . They were surveyed using a laser theodolite, with a 1-m resolution. The eastern subcatchment (Mt. Bold East, G15) was 4.2 ha in size and the western subcatchment (Mt. Bold West, G0; Fig. 2) was 3.6 ha in size. Both subcatchments were similar in slope (20%) with improved pastures, and a cover of eucalypts on the upper slopes.



View larger version (81K):
[in this window]
[in a new window]
  		Fig. 1. Location of the eastern and the western subcatchments used in this study relative to the Mt. Bold public water supply reservoir. The eastern subcatchment (G15) was treated with 15 Mg ha–1 gypsum and the western subcatchment (G0) was untreated.

 


View larger version (73K):
[in this window]
[in a new window]
  		Fig. 2. Instrumentation in the subcatchments (full details of instrumentation are in Stevens et al., 1999).

 
Fertilizer had been applied to both subcatchments infrequently since clearing of native vegetation (Mark Quigley, lease holder, personal communication, 1996). The current leaseholder applied single superphosphate at a rate of approximately 8 kg P ha–1 yr–1 from 1991 to 1996. There are no fences separating the two subcatchments so cattle were free to openly graze both subcatchments all year. The instrumentation and area within 3 m of the water collection and sampling points were fenced to prevent disturbance. Average long-term annual rainfall in the subcatchments was approximately 750 mm per year.

Each site was treated with single superphosphate at a rate of approximately 15 kg P ha–1 yr–1 for the first 3 yr after instrumentation was installed (i.e., 1996–1998). In early June 1998 the G15 subcatchment was treated with approximately 15 Mg ha–1 of gypsum and G0 was left untreated. Between 1999 and 2000 no P fertilizer was applied, then in 2001 a further 5 kg P ha–1 was applied to both subcatchments at the start of winter (early June) based on standard farmer practice. Initial (1996) available P (bicarbonate-extractable P) was reported by Stevens et al. (1999) as being in the 20 to 40 mg P kg–1 range in the A1 horizon and negligible in the lower horizons.

In 1996, sixteen 15-cm-diameter intact soil cores were removed from the subcatchments for description, chemical analyses, and soil classification [methods in Kirkby et al. (1997) and Cox et al. (2000)]. Soil cores were also collected for specific chemical analyses after gypsum application to the G15 site. In addition, some of the soils were re-sampled and re-analyzed 4 yr after gypsum treatment. Lack of large-scale replication of soil samples for some chemical analyses may be an issue due to variation in the 0- to 10-cm depth (see Results); however, it is less likely to be a problem below 10 cm. Field sampling methods used were those of McDonald et al. (1984) and soils were classified using the Australian Soil Classification (Isbell, 1996). Laboratory methods for standard soil chemistry (e.g., pH, electrical conductivity, cation exchange capacity [CEC], exchangeable bases, and exchangeable sodium percentage [ESP]) are in Rayment and Higginson (1992). Mineralogy was studied by X-ray diffraction (XRD) and element abundance by X-ray fluorescence (XRF) spectroscopy. Soil P was extracted using a sequential extraction procedure based on the method proposed by Hedley and Stewart (1982), as described by Rayment and Higginson (1992). Only selected soil chemical data are presented in this paper.

Soil Aggregates
Soil aggregates used for stability measurements were sampled at depths of 0 to 10, 20 to 30, and 40 to 50 cm. Initial sampling was of large, fist-size aggregates. Subsequently these were gently broken into smaller aggregates of approximately 1 to 2 cm in diameter. Approximately 25 g of these aggregates (7–10 in total) were weighed and placed on a nest of sieves in order (top to bottom) of mesh diameters 2 mm, 1 mm, 500 µm, and 250 µm. A separate sample of approximately 25 g was weighed and oven-dried at 105°C for 24 h to determine the approximate moisture content of the aggregates used in the stability assessment. The nest of sieves was placed on a mechanical oscillator within a cylinder. Water was added to the cylinder to a level that was seated at the base of the uppermost sieve containing the aggregates. The mechanical oscillator moved in a vertical plane within the water column with a stroke length of 2 cm and a frequency of 30 strokes per minute. Aggregates in the top sieve were completely immersed at the lowest point. The soil aggregates from the upper 10 cm of both soils were found to be extremely stable, and so the total agitation time was 1.5 h. The less stable soil aggregates at 20 to 30 and 40 to 50 cm were agitated for 10 min. Following agitation the sieves were allowed to drain for 10 min. The material collected in each of the sieves was carefully washed from each sieve into preweighed containers, oven-dried at 105°C for 3 d, and reweighed.

General Water Sampling Procedure
Overland flow was measured with a calibrated 150 mm RBC flume (Fig. 2; Clemmens et al., 1984). Lateral flow along the interface of the A and B horizons (A–B interflow) and lateral flow along the interface of the B and C horizons (B–C interflow) was intercepted with tube drainage and measured by a tipping bucket attached to a datalogger. Water samples were obtained by flow proportional automatic samplers and selected samples were analyzed within 24 h of a runoff event by inductively coupled plasma–optical emission spectrometry (ICP–OES). Both filtered (sample passed through a 0.45-µm Whatman [Maidstone, UK] cellulose nitrate membrane filter) and nonfiltered samples were analyzed. Full details of the water sampling instrumentation and analytical procedures are presented in Stevens et al. (1999).

Reactive phosphate (or molybdate-reactive phosphorus [MRP]) was analyzed by an automated modification of the colorimetric molybdenum blue method (Murphy and Riley, 1962) on the dissolved (<0.45-µm) fraction. Total P in 1998 was measured by MRP of samples neutralized following nitric perchloric acid digestion of unfiltered water samples. This was done as the minimum detection limit of the ICP–OES was higher than many MRP concentrations from the runoff samples in 1998.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil Morphological Changes and Aggregate Stability
Soils of the Mt. Bold subcatchments have a moderate textural contrast between the A and B horizons. A horizons have a silty clay loam texture and are generally <0.2 m thick. They overlie a slightly acid, well-structured B horizon with many medium to coarse macropores. The soils have formed in a micaceous siltstone parent rock. Soil profile depth in both subcatchments is about 1 to 1.5 m over the highly weathered bedrock but the soils are shallower on the upper slope. The soils are Dermosols with Chromosols on the mid- and upper slopes.

Differences in soil chemistry between the two subcatchments and changes over time are discussed below. However, the most striking difference between the soils of G15 and G0, 4 yr after gypsum application on G15, was a clearly darker G15 profile (10YR 3/1–4/1 moist) in comparison to the G0 profile (10Y/R 5/3–6/3 moist). The color change occurred to a depth of at least 0.5 m, or well into the B horizon (Fig. 3) . Photographs were taken in comparable topographic positions (mid-slope).



View larger version (109K):
[in this window]
[in a new window]
  		Fig. 3. Photograph of the G15 gypsum-treated profile (left) and the G0 untreated profile (4 yr after treatment).

 
Results of soil aggregate stability after treatment are presented in Table 1. At the surface the aggregate stability is very similar between G15 and G0, with 98.3 and 97.1% of the water stable aggregates respectively being >2 mm, and <1.5% in both soils being smaller than 0.25 mm in diameter. With increasing depth (20–30 and 40–50 cm) there was a relative increase in the stability of soil aggregates of >2 mm from the G15 subcatchment compared to G0. At the 40- to 50-cm depth, 76.7% of the water stable aggregates in G15 were >2 mm compared with only 32.1% in G0.


View this table:
[in this window]
[in a new window]
  		Table 1. Estimation of water stability of soil aggregates after treatment.

 
Table 2 presents exchangeable Na, K, Mg, and Ca, CEC, and ESP before and following gypsum addition. The exchangeable cations increased in the soils of both subcatchments to 50 cm due to the addition of increased amounts of P fertilizer (single superphosphate). Highest ESP was measured below 112 cm in G0. The CEC of the gypsum-treated soil (G15) had increased substantially and was greater, throughout the profile, than the untreated soil (G0). Exchange of Ca for both Mg and Na led to an increase in the exchangeable Ca to Mg ratio and a substantial decrease in the ESP in the gypsum-treated soil. It was thought this outcome should have important implications for PP in runoff (discussed below).


View this table:
[in this window]
[in a new window]
  		Table 2. Typical exchangeable cations, cation exchange capacity (CEC), and exchangeable sodium percentage (ESP) measured on soil cores taken from the subcatchments (before and after treatment).

 
The soil digests and XRF analysis showed that in the surface 5 cm both P and S were higher in the G0 than the G15 soils; however, significant variation was observed in both the two upper horizons (to 10 cm) for both elements (Fig. 4) . Below 10 cm, P and S were much higher in the G15 soils. Variability in concentration notably decreased with depth in both the G15 and G0 soils.



View larger version (16K):
[in this window]
[in a new window]
  		Fig. 4. Distribution of P and S after nitric perchloric acid digestion and using X-ray fluorescence (XRF) spectroscopy.

 
The distributions of NaHCO3, NaOH, NaOH (sonicated), and HCl-extractable P fractions are shown in Fig. 5 . Figure 6 shows the sum of extractable phosphorus (EP) fractions, total phosphorus (TP), and residual phosphorus (RP; TP – EP). With the exception of the upper 5 cm of G0, EP was greater in all horizons of the treated soils. All extractable fractions were greater in the top 5 cm of the G0 soils, with the NaOH-extractable P being greater in G0 to a depth of 10 cm. The NaHCO3–extractable P was also greater in the G0 profile in the upper 5 cm and below 30 cm. The HCl-extractable P was also marginally greater in the surface of G0, but below 10 cm was below detection limits. In the G15 soils concentrations fell below detectable levels below 20 cm. At depths greater than 20 cm, EP was greater in the G15 soils. Generally, RP (Fig. 6) was greater throughout the G0 profile.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 5. Distribution with depth from soil cores collected throughout the subcatchments for HCO3–extractable P (1), NaOH-extractable P (2), sonicated NaOH-extractable P (3), and HCl-extractable P (4).

 


View larger version (23K):
[in this window]
[in a new window]
  		Fig. 6. Typical changes in extractable phosphorus (EP), total phosphorus (TP), and residual phosphorus (RP) from soil cores collected in both subcatchments.

 
Subcatchment Hydrology, 1996–1999
Figure 7 shows the rainfall, overland flow, and A–B and B–C interflow for G0 and G15 from 1996 to 1999. The runoff (and rainfall) data are presented as the daily sum. Runoff in both subcatchments was initially dominated by overland flow, with B–C flow generally exceeding A–B flow (particularly in G15). Runoff commenced in G0, through all monitored soil horizons. These early (generally autumn–early winter) flows were dominated in volume by G0 (i.e., G0 runoff > G15 runoff). However, by mid-winter (generally 4–6 wk following initial runoff) G15 volumes exceeded G0 in total flow, overland flow, and B–C interflow, with G0 A–B interflow generally greater than G15 A–B interflow. Figure 8 shows the total flow each year in each flow path; overland flow, A–B interflow, and B–C interflow, for both G0 and G15. The only possible treatment effect following gypsum application in 1998 was that the flows became more evenly distributed between flow paths in G15; that is, the relative proportion of overland flow contributing to total runoff was strongly reduced. In general the amount of interflow particularly along the B–C boundary was extraordinarily high compared to expectations.



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 7. Rainfall, overland flow, and A–B and B–C interflow from 1996 to 1999 from G0 and G15 (soils treated with 0 and approximately 15 Mg ha–1 of gypsum, respectively). Gypsum was applied to G15 on 1 June 1998.

 


View larger version (26K):
[in this window]
[in a new window]
  		Fig. 8. Flow paths for G0 and G15 (soils treated with 0 and approximately 15 Mg ha–1 of gypsum, respectively) from 1996 to 1999. Rainfall was 145, 542, 284, and 345 mm for the period of monitoring each year.

 
Runoff Chemical Characteristics before Gypsum Treatment
A summary of the average monthly runoff chemistry (<0.45 µm) (i.e., pH, DP, Ca, Fe, and Al) in the first year, 1996, is presented in Fig. 9 . This provides a comparison between the chemical compositions of subcatchments before gypsum treatment. Runoff compositions for 1997 are not presented but were comparable.



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 9. Average monthly runoff chemistry of overland flow, and B–C interflow from G0 and G15 (soils treated with 0 and approximately 15 Mg ha–1 of gypsum, respectively) in 1996.

 
In 1996 the average monthly pH levels of the overland flow and B–C interflow were similar in G15 and G0, ranging between 5.9 and 7. Maximum average monthly DP concentrations were observed in the June overland flow of G15 (approximately 1 mg L–1). Apart from June, when the DP concentrations in the B–C interflow were greater in G0 (approximately 0.7 compared with approximately 0.2 mg L–1), DP volumes in all other flow events were very similar. Calcium concentrations in runoff were also very similar throughout 1996, averaging approximately 10 mg L–1. Concentrations of Al and Fe were significantly higher in B–C interflow from G15 in autumn (April) than in runoff from G0. The most likely source of this would be fluxes of clay minerals (mainly kaolinite). At other times all runoff from G0 and G15 averaged much lower than 5 mg L–1, generally <1.5 mg L–1, Al and Fe.

From the water chemistry it was concluded that there were negligible differences between G0 and G15 before gypsum treatment. From this it was assumed that differences observed in the water chemistry between G0 and G15 following gypsum application were a consequence of gypsum application, and not a result of inherent differences between the subcatchments themselves.

Major Element Concentrations in Runoff after Gypsum Treatment
Consistent differences in runoff composition of Fe, P, Al, Mn, Ca, Mg, Na, S, and K (<0.45-µm fraction) were evident between G15 and G0 (data not shown). These differences were consistent from the first flow following gypsum treatment (June 1998). The flows from G15 were consistently higher in Ca, S, Mn, Mg, and Na, and were lower in Fe, Al, and P with no notable difference in K compared to G0. The electrical conductivity of the G15 samples was also consistently higher albeit with a slight decrease over time. The pH of the G15 runoff was also consistently 0.25 to 1.5 units lower than G0, and over 1998–1999 the average pH was 0.5 units lower in G15. In contrast to this trend, the pH of the G15 sample 4 yr after gypsum treatment was approximately 0.25 pH units higher than that of G0. This may indicate a degree of natural variation in pH, and if so indicates the impact of gypsum on runoff pH is not clear.

Phosphorus Concentrations in Runoff after Gypsum Treatment
The mean concentrations of P, as TP and MRP during each storm (flow) event in G15 and G0 from 1998 to 2000, are presented in Table 3. The average TP and MRP over that period are also shown for the same period. In G15, average TP was 35.3% lower in overland flow, and 11.4% lower in the B–C interflow than in G0 (A–B horizon interflow was not analyzed for MRP). The ratio of MRP to TP in both overland flow and B–C interflow was lower in G15 (0.71 and 0.41 for overland flow and B–C interflow respectively) compared with G0 (0.80 and 0.59, respectively). Additionally the ratio was always lower for both subcatchments in the B–C interflow than the overland flow. Low MRP to TP ratios are indicative of relatively high PP.


View this table:
[in this window]
[in a new window]
  		Table 3. Concentrations of molybdate-reactive phosphorus (MRP) and total phosphorus (TP) in overland flow and B–C interflow from G0 and G15 (soils treated with 0 and approximately 15 Mg ha–1 of gypsum, respectively) from 1998 to 2000.{dagger}

 
Following P fertilizer application in mid-June 1998, P concentrations were initially highest in the G15 runoff in the period immediately following application (25 June–8 July). This would appear to be a consequence of hydrologic factors rather than gypsum treatment, as the volumes and flow rates were notably higher, particularly from 25 to 30 June 1998 in G15 B–C interflow. Succeeding these events, most concentrations of both MRP and TP were higher in the G0 B–C interflow, and overland flows (mid-July 1998–October 2000). Differences were most evident in the early flow events of 1999 (May–July) when P concentrations in G0 were between 5- and 10-fold higher.

In 2000 only very low concentrations of P were observed in both subcatchments, indicating the pool of water-soluble P had been largely depleted in both subcatchments. Although only trace amounts were detected, concentrations of MRP in 2000 were still marginally lower in G15 than G0 (approximately 0.03 compared with approximately 0.05 mg L–1).

Differences in Flow-Weighted Phosphorus in Runoff from Subcatchments, 1996–1999
Figure 10 shows the flow-weighted DP and TP concentrations in overland flow and B–C horizon interflow for the two subcatchments. In overland flow the proportion of DP to TP did not differ between subcatchments over time. Most of the P in runoff was as DP with the exception of 1997 when there was a large increase in the proportion of PP in both subcatchments.



View larger version (29K):
[in this window]
[in a new window]
  		Fig. 10. Average dissolved P and total P in overland flow and B–C interflow from G0 and G15 (soils treated with 0 and approximately 15 Mg ha–1 of gypsum, respectively) from 1996 to 1999.

 
Concentrations of TP in overland flow from G15 were much lower than from G0 in 1998 and 1999. In 1996, TP in overland flow from G15 was much higher than in 1998 and 1999. However, this was also the case in 1997, before the treatment. The affect of the gypsum can be seen in the B–C interflow concentrations that increased markedly in 1998 in G15.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Soil Chemistry
The addition of gypsum increased the CEC throughout the soil profile. This was achieved through an overall increase in exchangeable Ca, which occurred in excess of the simple exchange of Ca for Na, K, and Mg as confirmed by the results of Ca exchangeable cations. The overall effect of gypsum on cation properties is a progressive exchange of cations by Ca, as gypsum was leached down the profile. The Ca in the gypsum displaced Mn, K, Mg, and Al from the upper 5 to 10 cm to lower in the profile (20 cm for Mn and Al, 30 cm for K, and 40 cm for Mg). Exchangeable Na was lower and less variable throughout the profile in treated than untreated soils, suggesting that Ca for Na exchange had taken place to at least 50 cm.

The capacity of the soil to retain P was enhanced by the application of gypsum. The distribution of P in the soil profile reflected the distribution of total and exchangeable Ca and Al, with the total and extractable P greater in the surface 5 cm of the untreated soils. The lower P content of the upper 5 cm of the treated soil may be explained by a number of factors including enhanced microbial activity; cation leaching in G15; differences in mineralogy (a much higher coarse sand content in the treated soils); exchange of Al3+ by Ca2+, reducing net surface charge, or to a lesser extent (because of the pH of these soils) competition between SO42– and PO43– (e.g., Barrow et al., 1980). Below 5 cm, the treated soils had greater TP and EP concentrations, with RP being the only fraction greater in the G0 soils. A common perception of increased P retention in the presence of high concentrations of Ca is via the precipitation of Ca phosphates. This would seem unlikely at the low pH of these soils, a prediction confirmed by relatively indifferent concentrations of HCl-extractable P in both soils. The differences between P content of treated and untreated soils is characterized by greater NaOH-extractable (or chemisorbed) P in treated soils. This suggests that the mechanism of P retention would be through chemisorption of P on Fe and Al either through adsorption or surface enhanced heterogeneous precipitation. The increase in exchangeable Al below 10 cm (for example) is evidence that adsorption may be an important mechanism by which P is retained and this adsorption is promoted by increased surface charge. Additionally the data support the mass exchange of Ca for Fe and Al, modeled and described for example by Stout et al. (1998).

Runoff Chemistry
Increased pH of runoff following gypsum application has been commonly observed (Sumner, 1993), and was explained by the exchange of SO42– for OH– (Shainberg et al., 1989). The lower pH of the runoff from the G15 observed in this study indicates that exchange of SO42– for OH– was exceeded by the exchange of Ca2+ for H+ (Shainberg et al., 1989).

Increased leaching of Ca and S was observed in G15, as expected, given the dissolution of applied gypsum. There was a steady decline in concentrations in the <0.45-µm fraction over the five seasons, from Ca = approximately 400 mg L–1 and S = approximately 400 mg L–1 in 1998 to Ca = approximately 100 mg L–1 and S = approximately 120 to 180 mg L–1 in 2001–2002. This was indicative of a reduced pool of available gypsum for leaching. The ratio of Ca to S in gypsum by weight is approximately 1.25. The ratios observed in solution were considerably less than this (average 0.9 for 1998–1999), with the highest Ca to S ratio of 1.1 in overland flow in July and August 1998. The lowest value of 0.5 was measured in October 1999 and September 2002. There appears to have been a seasonal effect of time, as this ratio was observed to decrease over the period of the autumn break to the end of seasonal flow in both 1998 and 1999.

The high proportional loss of S indicates it was preferentially leached through the profile compared with Ca. This indicates significant adsorption or exchange of Ca, which would account for the increased leaching of Na, Mg, and Mn and would confirm the observed pH decrease. Other studies have reported increased leaching of K (Shainberg et al., 1989; O'Brien and Sumner, 1988; Pavan et al., 1984). Increased exchangeable Ca accompanied decreased exchangeable K throughout the profile. Additionally in the G15 profile Ca exchangeable (using 0.01 M CaCl2) Fe and Al was decreased in the upper profile compared to G0 (Fe, 0–5 cm and Al, 0–10 cm), with Ca exchangeable Al increasing down profile in G15. The decrease in exchangeable Al, Fe, and K accompanied by decreased runoff concentration of these cations suggested they might be removed in precipitates. O'Brien and Sumner (1988) gave similar explanations for lower Al concentration through the precipitation of alunite [KAl3(SO4)2(OH)6]. Chemical speciation calculations confirmed that this is possible, along with the precipitation of basaluminite [Al4SO4(OH)10]. This may further account for the reduction in K concentration (i.e, alunite precipitation). The overall decrease in exchangeable Al and Fe, particularly in the upper 10 to 20 cm, indicates that the concentration of these constituents in the soil solution at that depth would have been higher than that indicated by the concentration in the runoff collected. So it is possible that the precipitation of Al and Fe hydroxides and oxides (and phosphates; i.e, strengite FePO4·2H2O) may have been greater in G15 than in G0.

Gypsum appears to have modified the proportion of P in each transport pathway. There does not appear to be an increase in PP as found by Stout et al. (2000). They found enhancement by the net increase in CEC of the soil (from increased Ca throughout the profile, and displaced cations Al and Mg lower in the profile). This may be facilitated by the direct precipitation of phosphate minerals, namely Fe phosphates such as strengite and Mn phosphates (e.g., MnHPO4). The precipitation of other mineral phases, most significantly the Al-sulfates, alunite, and basaluminite may provide additional reactive surfaces for phosphate adsorption.

Aluminum was released into solution following a mass exchange with Ca in the upper A profile, and accumulated in the B horizon. The accumulation may involve Al-sulfate precipitation, and increases in exchangeable Al suggest that the soluble Al species had been adsorbed to mineral surfaces, as occurs with coagulation and flocculation with alum.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
The impacts of gypsum application on the physicochemistry of texture-contrast soil have been demonstrated in this study. However, the impacts of gypsum on P in runoff were not as marked. There was a limited reduction in anion adsorption capacity at the soil surface (0–10 cm), combined with vastly increased capacity at depth (10+ cm). The addition of gypsum can increase the ability of the soils to retain P by the exchange of Ca for Al, Fe, and Mn. The release of these cations into solution in the upper 10 cm of soil and their subsequent accumulation in lower horizons have increased the retention capacity of the soil at depth. The decrease in the amount of P in the surface soil would normally be important from an environmental point of view, in as much as high P levels in surface soils, regardless of the form of P (dissolved or particulate), will be more susceptible to loss through erosion in overland flow.

By improving soil structure, the gypsum appears to alter the relative proportions of P movement through overland flow and interflow. Gypsum may enhance P movement through the subsoil probably via macropores as evidenced by high volumes of interflow measured above the C horizon. Much of the P in interflow was in the dissolved form.


   ACKNOWLEDGMENTS
 
The research was partly funded by Land & Water Australia. Jon Varcoe received a Ph.D. studentship from the Cooperative Research Centre for Water Quality and Treatment. Jim Cox received a research fellowship from the Organisation for Economic Co-operation and Development.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Cox, J. W.
Right arrow Articles by van Leeuwen, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Cox, J. W.
Right arrow Articles by van Leeuwen, J.
Agricola
Right arrow Articles by Cox, J. W.
Right arrow Articles by van Leeuwen, J.
Related Collections
Right arrow Water Quality
Right arrow Preferential Flow
Right arrow Phosphorus
Right arrow Runoff
Right arrow Nutrient Management

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome