JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 17 July 2007
Published in J Environ Qual 36:1281-1288 (2007)
DOI: 10.2134/jeq2006.0347
© 2007 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 PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McDowell, R. W.
Right arrow Articles by Robertson, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by McDowell, R. W.
Right arrow Articles by Robertson, F.
Agricola
Right arrow Articles by McDowell, R. W.
Right arrow Articles by Robertson, F.
Related Collections
Right arrow Phosphorus
Right arrow Solute Transport Models

TECHNICAL REPORTS

Landscape and Watershed Processes

Sources of Phosphorus Lost from a Grazed Pasture Receiving Simulated Rainfall

R. W. McDowella,*, D. M. Nashb and F. Robertsonc

a AgResearch Limited, Invermay Agricultural Centre, Private Bag 50034, Mosgiel, New Zealand
b Victorian Dep. of Primary Industries–Ellinbank, RMB 2460 Hazeldean Road, Ellinbank, Victoria 3821, Australia and e-Water CRC, Univ. of Canberra, ACT 2601, Australia
c Victorian Dep. of Primary Industries–Hamilton, Mount Napier Road, Hamilton, Victoria 3300, Australia

* Corresponding author (richard.mcdowell{at}agresearch.co.nz).

Received for publication August 30, 2006.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
Nutrients exported from grazing systems contribute to eutrophication of surface waters. In this study the contributions of soil, pasture-plants, and dung to P exports in overland flow were compared using simulated rainfall. The treatments were (i) grazed pasture-plants (isolated from soil by application of petrolatum to the soil surface), (ii) grazed pasture-plants and supporting soil, (iii) grazed pasture-plants and soil and treading, and (iv) grazed pasture-plants and soil and treading and dung. In general, dissolved reactive P (DRP) accounted for the majority of the P exported and P losses decreased in the order: treading and dung treatment > treading > pasture-plants and soil > pasture-plants. Very little dissolved organic P was lost in overland flow and the effects of treading diminished with time. Over a normal grazing cycle (30 d), the portion of P lost from pasture-plants was approximately half that lost from pasture-plants and soil, one-third that lost from treaded pasture-plants and soil, and one-quarter that lost from treaded pasture-plants, soil, and dung. The DRP in the pasture-plants treatment was approximately half that in the pasture-plants and soil treatment and suggests that a significant portion of the P exported from these systems is derived directly from pasture-plants. Due to higher proportions of particulate P (PP) in the treaded and dung treatments, DRP accounted for less of total P than in the pasture-plants and pasture-plants and soil treatments. Lower infiltration capacities probably caused by mechanical disaggregation at the soil surface are consistent with the higher proportions of PP in the treading treatments. These results were used to estimate P exports from a field trial site in Southland, New Zealand. The results suggested that P export attributable to fertilizer, dung, pasture-plants, and soil components were approximately 10, 30, 20, and 40%, respectively. These results suggest that since 90% of the P exports are derived from the soil–plant system and dung returns, managements to lessen P exports should continue to focus on maintaining soil P within the optimal range for pasture-plant production and maintaining soil surface properties that maximize infiltration and minimize overland flow.

Abbreviations: DM, dry matter • DRP, dissolved reactive phosphorus • DURP, dissolved unreactive phosphorus • PP, particulate phosphorus • SS, suspended sediment • TDP, total dissolved phosphorus • TP, total phosphorus • WSP, water-soluble phosphorus


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
PHOSPHORUS contributes to the eutrophication of surface waters and impaired water quality. Grazing of pasture is one of the many land uses contributing P to surface waters. However, in countries like Australia and New Zealand, pasture grazed every 20 to 60 d (on rotation) is often the major source of P entering surface waters in lowland catchments (Davies-Colley and Wilcock, 2004). Generally, the potential for P export from these systems is greatest during or soon after grazing and declines exponentially with time thereafter (McDowell et al., 2006). The sources of P exported during and after grazing may be a combination of fertilizer, excreta, and pasture-plant and soil pools. However, data on the relative contribution of each source is sparse.

Several studies have shown that P mobilized from dung is 10 times greater than from soil when dung or manure is wet, but declines quickly as dung dries (e.g., McDowell et al., 2006; Smith et al., 2001). The contribution from fertilizer declines quickly with time since application. For instance, McDowell et al. (2003a) showed that in soils with the same Olsen P concentration, potential P losses from superphosphate applied at 30 kg P ha–1 were greater than a control treatment with no P applied or a treatment with a less soluble form of P applied (e.g., reactive phosphate rock). The difference between treatments persisted for 60 d. However, the highest P mobilization potential (80%) for readily soluble P fertilizers occurs within the first few days after application (Barlow et al., 1998; Bush and Austin 2001; Nash and Halliwell 1999; Nash et al., 2004). The contribution from soil to P exports varies according to the size of the pool of P available for mobilization, which can be determined using gentle (no replacement power and pH neutral) extractants (McDowell and Condron 2004). This pool is influenced by a number of factors such as soil moisture, soil sorption capacity and strength, and treading.

Accounting for P exports from fertilizer and dung in a grazed pastoral system in Southland, New Zealand, McDowell et al. (2006) estimated that fertilizer made up 10% of exports, dung 30 to 40%, and the remainder was from soil P. These results are consistent with other calculations for fertilizer contributions to annual P exports of <10% (McDowell and Catto 2005), and dung contributions to flood irrigation losses from border dyke systems (Carey et al., 2004). However, a question remains around the proportion of P exported directly from soil and what is the influence of other factors such as the treading and plant damage via grazing.

One clue to answering this question has been provided by studies of nutrient losses from plant residues. Many workers have suggested that changes in P concentration in soils and waters of grassed catchments are partly due to the loss of P from pasture-plants in various stages of decomposition (Bromfield and Jones 1972; Schreiber and McDowell 1985). Commonly, these studies have drawn conclusions based on data from cut and/or dried pasture-plants (e.g., Wendt and Corey 1980; Timmons et al., 1970), that have shown significant losses of N and P during leaching: up to 80% of the total P held within the plant (Bromfield and Jones 1972). How such data on P leaching from plant residues relates to the mobilization of P directly from live plants and plants damaged by ripping during cattle grazing is unclear.

From the analysis of standing crops (sorghum [Sorghum sudanense Piper Stapf.], cotton [Gossypium hirsutum L.], and soybean [Glycine max (L.) Merr.]), Sharpley (1981) showed that P in canopy leachate contributed the major proportion (90%) of P transported in overland flow. The closest study to simulate the influence of grazing on pasture-plant–derived P exports is that of Mundy et al. (2003) who simulated grazing by cutting pasture-plants to 47 mm at a range of times before irrigation. In these studies the influence of the topsoil P pool and treading by cattle cannot be discounted. Hence, the question remains: What are the relative contributions of soil, pasture-plants, and animal-derived P to exports in overland flow?

Pastures in Australia and New Zealand are grazed on average 14 to 16 times a year. Often grazing coincides with high rainfall periods and the associated overland flow (Nash et al., 2000). Without an understanding of the relative contributions of soil, pasture-plants, and dung (assuming fertilizer losses can be minimized by good management) to the P exported from these systems, there are severe limitations on the development of effective and targeted remedial strategies. The objective of this article is to compare quantitatively the relative contributions of treading by cattle, dung, pasture-plants, and soil to P losses in overland flow from a grazed pasture system.


   Materials and Methods
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results and Discussion
 Conclusions
 REFERENCES
 
Soil Treatments
The soil (Waikiwi silt loam: USDA Taxonomy, Typic Dystrochrept; NZ Classification, Typic Firm Brown) was taken from the Woodlands research station near Invercargill, Southland, New Zealand. The site was a permanent ryegrass (Lolium perenne L.) and white clover (Trifolium repens L.). Soil and had a moisture content of 43% (v/v) ± 2%.

One hundred intact soil blocks were taken in 1 d, using a 1 m long, by 20 cm wide metal cutting blade to 10 cm deep. The pasture-plants covered 95% of the ground (ca. 1.8 mg dry matter [DM] ha–1). Soil blocks were placed in boxes 1 m long by 20 cm wide by 10 cm deep and with six 2-mm holes drilled in the bottom to allow some drainage. The 100 soil blocks were divided equally into one of four treatments:

  1. Grazed pasture-plants
  2. Grazed pasture-plants and soil
  3. Grazed pasture-plants and soil and treading
  4. Grazed pasture-plants and soil and treading and feces

These are referred to from here as Treatment 1, pasture-plants; Treatment 2, pasture-plants and soil; Treatment 3, treaded; and Treatment 4, dung. Grazing was simulated on all treatments by hand pulling pasture-plants until a cover of about 1300 kg DM ha–1 was reached (determined by rising plate meter). Pasture-plant samples were kept for analysis. Grazed pasture-plants in Treatment 1 were separated from the soil surface by pouring molten petrolatum along the side of the soil and box until level with the soil surface. This was left to set (approximately 1 h) and meant only grazed pasture-plants would interact with rainfall and overland flow. Treatment 2 was grazed and otherwise unaltered. Treatments 3 and 4 were each treaded on by an artificial cow hoof eight times (20 imprints m–2) to simulate treading during a 24- to 36-h grazing event (McDowell et al., 2003b). The artificial cow hoof was modeled on a 2-yr-old Friesian cow hoof and delivered 250 kPa of pressure over a 90 cm2 area (Di et al., 2001). Soil blocks in Treatment 4 each received a 0.5 kg dung pat (moisture content ca. 88%) placed within a 20-cm diameter metal ring at the upslope end of each box. Feces (40 kg) were collected on the first day of the experiment from fresh dung patches on a grazed paddock of a nearby dairy farm and thoroughly mixed before application. The application rate was equivalent to typical dung deposition over 0.2 m2 for a 24- to 36-h grazing period (Haynes and Williams 1993). The metal rings were removed, and soils left outside and inclined at 3% slope (similar to that found at the sampling site) and not watered before rain simulation. One set (four boxes) of each treatment was moved into an indoor artificial rainfall facility 0, 1, 3, 7, and 19 d after imposing treatments, subjected to simulated rainfall, and overland flow collected. After each simulation, soil blocks were subsampled (0–7.5 cm) and blocks discarded. Meteorological data was taken at hourly intervals from a site 150 m away, but no rain fell during the study period.

Overland Flow
Overland flow was generated by applying artificial rainfall (tap water, P less than detection limit of 0.002 mg P L–1) at 25 mm h–1 to each boxed soil inclined at 3% slope. The rainfall simulator uses one TeeJet 1/4HH-SS30WSQ nozzle (Spraying Systems Co., Wheaton, IL) approximately 250 cm above the soil surface to gain terminal velocity. The nozzle, plumbing, in line filter, and pressure gauge were fitted onto a 305 cm high by 305 cm wide by 305 cm deep aluminium frame with tarpaulins on each side to provide a wind screen. Simulated rainfall had drop-size, velocity, and impact energies approximating natural rainfall. The 25 mm h–1 rainfall-intensity has a return frequency of approximately two times a year for a 15 min event. The time taken to start overland flow, and to generate 1.5 L of flow, was noted. All flow was collected, mixed, and a subsample taken for analysis.

Water, Plant, Dung, and Soil Analyses
All analyses were measured in triplicate and P determinations made colorimetrically using the method of Watanabe and Olsen (1965). Water-soluble P (WSP) was measured in fresh dung using the method of Wolf et al. (2005), and pasture-plants collected during simulated grazing (0.2 g dry weight equivalent shaken with deionized water for 60 min, filtered through a 0.45-µm syringe filter and WSP determined on the filtrate). The method of McDowell and Condron (2004) was used to determine WSP in soil samples. Subsamples of dung, pasture-plants and soil were also dried and ground to pass a 1-mm sieve. Total P was determined via aqua regia (4:1 v v–1 concentrated HCl–HNO3; Crosland et al., 1995) digestion of 50 mg pasture-plants or dung and 100 mg soil. Soils were also analyzed for bicarbonate-extractable P (Olsen P; Olsen et al., 1954).

Macroporosity (percentage of pores >30 µm) measurements were made using the method outlined by Drewry and Paton (2000). Briefly, following simulated rainfall measurements, two undisturbed soil cores were collected from each soil box using stainless steel rings (10 cm diameter and 5 cm high). The inside of each ring was coated with petrolatum to avoid edge flow effects. Once inserted into the soil, cores were excavated and transported to the laboratory where plaster was applied to the underside and "peeled" away to give an unsmeared surface (Greenwood 1989). Earthworms were removed with formaldehyde, before saturating the cores and equilibrating them at 10 kPa on tension tables to determine macroporosity (pores likely to control water movement through soil). Dry bulk densities were calculated from oven dry weights.

Overland flow samples were immediately filtered (<0.45 µm) and analyzed for DRP within 24 h, and total dissolved P (TDP) after acidified persulphate digestion within 48 h (Eisenreich et al., 1975). An unfiltered sample was also digested and total P measured within 7 d. Fractions defined as dissolved unreactive P (largely organic P) DURP and PP were determined as TDP less DRP and TP less TDP, respectively. Suspended sediment (SS) was determined by weighing the residue left after filtration through a GF/F (0.7-µm pore size) glass fiber filter paper.

Statistical Analyses
Variation of hydrological parameters, P fractions and SS concentration with time were assessed with line fits using SPSS v10.0 (SPSS, 2001). Either a simple linear regression or a negative power function (y = {alpha}t-ß) were fitted to the change in concentration with time. Previous analysis indicated that these functions gave the best compromise between fit to the data well and the fewest number of parameters possible. For the power function, the parameters {alpha} and ß were fitted coefficients relating to the initial value of y (hydrological parameter, SS or P fraction) and the decrease in y as a function of time, t, respectively. Line-fits were generated by least-squares regression, with data weighted according to standard errors. Only fits with a coefficient of simple determination (r2) > 0.7 and P < 5% level or better are shown. An analysis of variance was also conducted for DRP, DOP, TDP, PP, TP, SS, time to overland flow, and time to produce 1.5 L of overland flow. This indicated that, overall, there was a significant (P < 0.05) effect of treatments, with time until rainfall simulation, and for the interaction between treatments and time.


   Results and Discussion
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
Conclusions
 REFERENCES
 
Soil Hydrology
Among the four treatments the time taken for overland flow to begin ranged from <1 min in the pasture-plants treatment to about 18 min in the pasture-plants and soil treatment (Fig. 1). For Treatment 1, the petrolatum barrier on the soil surface ensured that the infiltration rate was effectively zero, and that plants alone were exposed to the resulting infiltration-excess overland flow. All other treatments presumably had a combination of infiltration-excess and conditions where overland flow occurred due to rainfall on a saturated surface (i.e., saturation-excess overland flow).


Figure 1
View larger version (22K):
[in this window]
[in a new window]

  		Fig. 1. Variation in the mean time to overland flow, the mean time taken to generate 1.5 L of flow and mean suspended sediment concentration with time since the start of the experiment. Bars represent the standard error of the mean. The number of asterisks indicates significance of line fit: *, **, *** are P < 0.05, P < 0.01, and P < 0.001, respectively.

 
Typical infiltration rates for pastures in good structural condition are >50 mm h–1 unless compacted by machinery or animals (Singleton and Addison 1999; Koolen and Kuipers 1983). Measurements of saturated hydraulic conductivity in the soil studied here were well in excess of this rate (>75 mm h–1; Drewry et al., 2000). Consequently, in the pasture-plants and soil treatment, saturation-excess overland flow is likely to have generated most runoff. In contrast, for the two treaded treatments the time to overland flow increased as the time interval between treading and measurement increased. This infers that at the start of the experiment a mix of saturation- and infiltration-excess overland flow was occurring and that with time, infiltration rates recovered, increasing the proportion of saturation excess overland flow. This may be due in part to the well established recovery of macropores (pores >30 µm) in cattle-grazed pastures (Drewry et al., 2004), which in this experiment would increase both the soil infiltration rate and the soil water holding capacity.

Interestingly, the influence of porosity was evident in a graph showing the time taken for each treatment to generate 1.5 L of overland flow (Fig. 1). As expected, given a rainfall intensity of 25 mm h–1 and a surface area of 0.2 m2, Treatment 1 (pasture-plants) took 18 min to generate 1.5 L of runoff after the application of petrolatum prevented water infiltration with the soil. In contrast, 40 min was required to collect 1.5 L of runoff from Treatment 2. The difference in time between treatments was attributed to water infiltration before soil macropore space (Table 1) became saturated. These pores form the major part of the soil's water holding capacity with an estimated 2400 cm3 available to a 10-cm depth (assuming a macroporosity of 12%). At a rainfall intensity of 25 mm h–1, the estimated macropore volume would become saturated in about 29 min. Added to the time required to collect 1.5 L of runoff from the surface alone (infiltration-excess), a total of 47 min would elapse before 1.5 L was collected from pasture-plants plus soil. This was close to the results for the pasture-plants and pasture-plants + soil treatments (Fig. 1). However, as with time to overland flow, the two treaded treatments took less time and improved as the interval between treatment and experimentation increased. Such results are consistent with smearing of the soil surface restricting water infiltration and compaction decreasing macroporosity. Indeed, during a 1-h period, McDowell et al. (2003b) noted that macroporosity affected both time to overland flow and overland flow volume (i.e., greater macroporosity caused a decrease in the overland flow produced over their 1-h collection time).


View this table:
[in this window]
[in a new window]

  		Table 1. Mean general physical and chemical characteristics of the soil (10 cm depth unless specified), pasture and dung used in this study. Standard error of the mean is given in parentheses.

 
Phosphorus Losses
Data for the mean concentration of P fractions lost during each event are given in Fig. 2. Without exception, the mean concentration of P fractions declined with time in the two treaded treatments. However, of TP fractions, only DRP showed any significant decline with time in the pasture-plants and pasture-plants and soil treatments. The decline with time in the dung treatment has been noted several times for DRP and TP in overland flow or flood irrigation events following grazing (e.g., Carey et al., 2004; Smith and Monaghan, 2003; Mundy et al., 2003). The trend has been attributed to dung drying in between application and simulated rainfall, which decreases P desorption and access of rainfall to wet dung when a crust forms (McDowell, 2006) and may also affect the rate of P diffusion from internal to external surfaces. However, isolation of the contributions from dung and soil disturbance due to treading has not previously been possible.


Figure 2
View larger version (23K):
[in this window]
[in a new window]

  		Fig. 2. Variation in the mean concentration of P fractions with time since the start of the experiment. Bars represent the standard error of the mean. The number of asterisk indicates significance of line fit: *, **, *** are P < 0.05, P < 0.01, and P < 0.001, respectively.

 
From inspection of data in Fig. 2, the greatest loss of P fractions among treatments generally occurred from the treaded and dung treatment followed by the treaded > pasture-plants and soil > pasture-plants treatments. To determine a relative risk of P loss among treatments, a potential load (mg) for P loss was calculated assuming 1 L of overland flow per day for 30 d—a common period between grazing in these regions. The results in Table 2 compare the relative load among treatments, the contributing percentage of each P fraction to load (vertical), and the percentage of each P fraction lost from pasture-plants, soil, treading, and dung for each P fraction in the dung (Treatment 4; P + S + T + D) treatment (horizontal). Mean values given in Table 2 were obtained from the following equations: Pasture-plants (P) = Treatment 1; Soil (S) = Treatment 2 – Treatment 1; Treading (T) = Treatment 3 – Treatment 2 + Treatment 1; and Dung (D) = Treatment 4 – Treatment 1 + Treatment 2 + Treatment 3. In general, DRP accounted for most TP in all treatments. However, due to more PP in the treaded and dung treatments, DRP accounted for less TP than in the pasture-plants or pasture-plants and soil treatments. This implies that as a proportion of TP very little DURP was lost in overland flow. The concentration of PP in all treatments was related to the SS concentration (SS = 0.095PP + 0.0025; r2 – 0.972, P < 0.001). The greatest SS concentrations came from the dung treatment, probably due to the loss in overland flow of fine particles when dung was fresh (Fig. 1; McDowell et al., 2006).


View this table:
[in this window]
[in a new window]

  		Table 2. Matrix of the relative risk of P loss over 30 d (mg: given as the product of 1 L of flow per day and daily P concentrations via equations in Fig. 2, except TP, which is the sum of P fractions) among treatments showing the contributing percentage of each P fraction to risk (vertical), and the percentage of each P fraction lost from pasture-plants, soil, treading, and dung for each P fraction in the P + S + T + D treatment (horizontal). For example, in the pasture treatment 79% of TP lost was DRP, but pasture-plants accounted for 39% of DRP in the P + S + T + D treatment.

 
The DRP load in flow from the pasture-plants treatment was 53% that in the pasture-plants and soil treatment (Table 2). This suggests that a significant proportion of P mobilized following grazing could originate directly from the plants (excluding treading or dung deposition). This P may come directly from P stores in plants or from disrupted cells and xylem and phloem exposed to overland flow. The latter would be expected to decline gradually as plants repair themselves. The P in pasture-plants is usually very mobile (Lötscher and Hay 1996), and also largely in orthophosphate form (>60%, J. Scott, personal communication, 2007). Studies of decomposing residues and hayed-off pasture-plants (harding grass–subterraneum clover mix; Phalaris tuberose–Trifolium subterraneum L.) have shown from 68 to 90% of total P leached is water soluble molybdate-reactive (inorganic) P (Sharpley and Smith 1981, Bromfield and Jones 1972). Using these data, and a P concentration in grazed pasture-plants of 2.5 to 4.5 g kg–1, Nash and Halliwell (1999) estimated that the pasture-plants studied by Bromfield and Jones (1972) contained between 15 and 56 kg WSP ha–1. If the soil P concentration is maintained within the optimum range for maximum potential pasture-plant production, then a ryegrass and white clover herbage should typically contain 3 g P kg–1 dry matter (Cornforth 1984). This means that for a pasture producing 15 Mg dry matter ha–1 yr–1 and with a mean WSP concentration of 75%, 34 kg WSP ha–1 is potentially available for mobilization from pasture-plants. Clearly, not all of this will be available to overland flow as diffusion from the inside of plant tissues will limit supply (i.e., rate of P availability). However, this study demonstrates that in these grazing systems, direct mobilization from cut or grazed pasture-plants could be an important source of P loss should overland flow occur.

Field data on flood irrigation of border-dyke paddocks with mown pasture have confirmed the importance of pasture-plants to P loss. For instance, Nexhip et al. (1997) and Mundy et al. (2003) showed that in paddocks under flood irrigation, P concentrations in overland flow were greater 1 d after mowing pasture-plants (but not grazed) than from paddocks that had cattle grazing at 100 and 200 cows ha–1. In contrast, our data indicated that the proportion of P lost from pasture-plants was much less than from the dung treatment. This may be due to the resorption of P from overland flow either by SS (Sharpley et al., 1981) or the soil itself. It may also be due to the physical damage caused by the mower being greater than that caused by cattle treading, exposing additional stores of intracellular P, or the release of P from senescent pasture-plants left on the soil surface after mowing that would otherwise have been consumed by cattle. Previous studies of the Waikiwi silt loam (e.g., McDowell et al., 2003c) suggest that even at high soil Olsen P concentrations (>50 mg kg–1), sorption lowers P concentrations in overland flow. Further, it has also been postulated that treading damage exposes soil low in P and highly sorptive of P in overland flow (McDowell et al., 2003b). Unfortunately, this benefit may be nullified if treading damage causes too much soil disturbance and P loss is enhanced via PP loss. McDowell and Condron (2004) developed equations to predict DRP in overland flow from Olsen P and anion storage capacity (a measure of P retained by the soil). When using an Olsen P of 25 and anion storage capacity of 50%, the predicted concentration is 0.33 mg L–1, close to that measured (Fig. 2).

In this study the contributions of PP to TP in overland flow were greatest in the treading treatments. We used treading damage equivalent to a day's grazing. At higher rates of treading on a Pallic soil in Otago, McDowell et al. (2003b) noted soil disruption was more extensive and PP exports increased linearly with the number of treading imprints. The vulnerability of soil to physical damage and relative proportions of sediment and PP in overland flow will depend on a number of factors such as soil P concentration, erosivity and occurrence of overland flow, pasture-plant cover, slope, stocking rate, and soil type (Dunne et al., 1991). It follows that, while the physical impact of grazing cattle would be expected to increase P exports, the magnitude of the increase will be site specific. For example, volcanic-ash soils are commonly more resilient to treading than sedimentary soils (Climo and Richardson 1984; Nguyen et al., 1998; Singleton and Addison, 1999).

This study has also shown that dung deposited by grazing cattle increases the concentrations of DRP and PP in overland flow (Fig. 2), but increases PP proportionately more than DRP (Table 2). Kleinman et al. (2002a, 2002b) have shown that the concentration of TP in overland flow from soils with manure applied depends primarily on the WSP concentration of the manure. Assuming that this holds true for dung, P loss from dung may be underestimated as the WSP concentration in our dung was low (14%) compared with other studies (e.g., 30–60%; McDowell and Stewart, 2005; Kleinman et al., 2005). This probably reflects the low soil Olsen P concentration in which pasture was grown (25 vs. many dairy soils which commonly have Olsen P concentrations of >30 mg kg–1; Drewry et al., 2003). Pasture-plants grown on highly fertile soil, and the resulting dung, contain more WSP relative to TP than pasture-plants grown on, and dung derived from, poorer fertility soils (Nash and Halliwell 1999; J. Scott, personal communication, 2007).

Application to Source Identification
Conditions similar to that used in this study have been calibrated against P concentrations lost on an event basis in the field (McDowell and Condron, 2004). Assuming that the model conditions used in this study reflect conditions for P concentrations lost on an event basis with time since grazing, the data from this and related studies can be used to estimate relative sources and exports of P in the field. This calculation was done for a 2-yr trial in Southland, New Zealand. The soil on the trial site was a Pukemutu silt loam (NZ classification, Typic Fragic Pallic; USDA Taxonomy, Typic Hapludalf) and it was grazed with dairy cattle (Monaghan et al. (2002) and unpublished data). Phosphorus exported in overland flow from eight, 16 by 30 m hydrologically isolated plots was about 0.3 kg P ha–1 (12 events) in 2002, and 0.8 kg P ha–1 (20 events) in 2003. Runoff occurred between May and November and the mean soil Olsen P concentration averaged 30 mg kg–1 and ranged between 22 and 50 mg kg–1.

The equation of McDowell and Condron (2004) was used to determine the contribution of dissolved P in overland flow from treatments containing soil and pasture-plants; DRP (mg L–1) = 0.024(Olsen P/P retention) + 0.024 (Table 3, Fig. 3). The soil and pasture-plant P component included an additional 25% extra P to account for PP not estimated by the McDowell and Condron (2004) equation but found to be ≤25% during subsequent fractionation (McDowell, unpublished data, 2006). The equation derived in the present study for DRP export in overland flow from the pasture-plants treatment (Fig. 2) was used to estimate the pasture-plant only component. To estimate the contribution of P lost from dung, equations in Fig. 2 were combined with those generated by McDowell et al. (2006) and McDowell (2006). The application rate of dung used in these trials was similar to the 24 to 36 h grazing periods in the field trial. Finally, the mobilization and exports of P with time since superphosphate application was estimated with equations given in McDowell and Catto (2005). About 30 kg P ha–1 was applied to all 16 by 30 m plots in September each year (indicated by arrows in Fig. 3).


View this table:
[in this window]
[in a new window]

  		Table 3. Modeled individual P loads (percentage of sum in parenthesis) for soil, dung and fertilizer, the sum, and the actual load (all kg ha–1) of total P in overland flow from a dairy grazed trial in Southland, New Zealand for 2002 and 2003.

 

Figure 3
View larger version (24K):
[in this window]
[in a new window]

  		Fig. 3. Modeled load of P lost in overland flow from events measured during 2002 and 2003 for a dairy cattle grazed trial in Southland, New Zealand. Arrows indicate when superphosphate fertilizer was applied in September each year.

 
The sum of estimated total P exported was 138% of actual total P lost in 2002 (Table 3, Fig. 3), but almost equal to TP lost in 2003 (95%). Of the estimated P losses, fertilizer comprised 12 to 13%, while dung P losses were 28 to 38% of the estimated TP lost. Soil P was estimated to contribute 29 to 45% of total P, while the component from pasture-plants was estimated at 15 to 21%. In a previous analysis, McDowell et al. (2006) calculated the soil component of P losses to be 51 to 64% of total P. This example shows not only the application of data to indicate the potential for P losses at different times of year, but also the relative importance of each source and the importance of grazed pasture-plants to P losses. Since this may vary from region to region, these data are invaluable for considering how to better manage P loss. In general, the data implied that to minimize P losses, grazing should be timed to occur outside of periods of likely overland flow (including irrigation causing overland flow) due to higher plant P contributions in grazed pasture.


   Conclusions
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
REFERENCES
 
Separation of the sources of potential P loss in overland flow from soil blocks was done by difference of treatments containing pasture-plants only to pasture-plants and soil and to two treaded treatments (with and without dung). Treading decreased the time taken for initiation of overland flow and initially the time taken to generate the same volume of overland flow as untreaded soils. This delay decreased with time as soils recovered from treading damage and pores were again able to fill before saturation-excess overland flow occurred. Of the P fractions measured, DRP accounted for most TP lost, but less so in treaded treatments where PP became more important due to P associated with sediment and dung losses. When integrated over a period of 30 d, the treaded soil with dung applied lost the most P. The proportion of P lost from pasture-plants, the soil, treading, and dung varied according to treatment. In the grazed, treaded, and dung treatment, it was estimated that pasture plants, soil, treading, and dung accounted for roughly equal proportions of TP lost. Equations for the relative contributions of each source with time since grazing were then incorporated with fertilizer and grazing records for a field trial. This indicated that the proportion of P lost attributable to fertilizer, dung, pasture-plants, and soil components was on average about 10, 30, 20, and 40%, respectively. Since 90% of P losses in this system occurred from the soil–plant system and dung returns, then decreasing P losses should focus on maintaining soil P within the optimal range for pasture production and on soil physical condition to minimize overland flow.


   ACKNOWLEDGMENTS
 
This work was funded by the New Zealand Foundation for Research Science and Technology (contract no. C10X0320).


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results and Discussion
 Conclusions
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
R. W. McDowell, D. Nash, A. George, Q. J. Wang, and R. Duncan
Approaches for Quantifying and Managing Diffuse Phosphorus Exports at the Farm/Small Catchment Scale
J. Environ. Qual., August 24, 2009; 38(5): 1968 - 1980.
[Abstract] [Full Text] [PDF]

Home page
 Crop Sci.Home page
D. J. Soldat and A. M. Petrovic
The Fate and Transport of Phosphorus in Turfgrass Ecosystems
Crop Sci., November 24, 2008; 48(6): 2051 - 2065.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
W. J. Dougherty, P. J. Nicholls, P. J. Milham, E. J. Havilah, and R. A. Lawrie
Phosphorus Fertilizer and Grazing Management Effects on Phosphorus in Runoff from Dairy Pastures
J. Environ. Qual., March 1, 2008; 37(2): 417 - 428.
[Abstract] [Full Text] [PDF]

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 PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McDowell, R. W.
Right arrow Articles by Robertson, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by McDowell, R. W.
Right arrow Articles by Robertson, F.
Agricola
Right arrow Articles by McDowell, R. W.
Right arrow Articles by Robertson, F.
Related Collections
Right arrow Phosphorus
Right arrow Solute Transport Models

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