| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Institute of Grassland and Environmental Research, North Wyke, Devon, EX20 2SB, UK
b Dep. of Geography, Univ. of Sheffield, Winter Street, Sheffield S10 2TN, UK
* Corresponding author (neil.preedy{at}bbsrc.ac.uk)
Received for publication October 19, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
3 mm h-1) resulted in total phosphorus (TP) exports between 1.8 and 2.3 kg ha-1. A single 4-h period (with overland flow) accounted for 33 to 46% of overall loads from the P-amended treatments. Concentrations in discharge from TSP + Slurry and TSP peaked at 11000 µg TP L-1 (67–68% as reactive P < 0.45 µm [RP<0.45]). Slurry peaked at 7000 µg TP L-1, 66% as particulate TP (>0.45 µm) and 20% as RP<0.45. Even in subsurface flow, concentrations exceeded 3000 µg TP L-1 for all P-amended treatments. Incidental TP concentrations in plot discharge were up to 110-fold higher than those considered eutrophic in inland waters. We suggest that targeting short-term management decisions for P applications is the most immediately viable method to mitigate P loss and benefit the environment.
Abbreviations: Q, grand mean discharge for each sampling period • RP, molybdate reactive phosphorus • TPunf, total phosphorus in unfiltered sample • TP<0.45, total phosphorus filtered through a 0.45-µm filter • TP>0.45, total phosphorus greater than 0.45 µm • UP, unreactive phosphorus, calculated as the difference of total phosphorus and molybdate reactive phosphorus
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Grasslands in the UK, which account for 70% of the agricultural land area, receive high P inputs in the form of fertilizer, animal feeds, and manure. These P inputs are often in excess of P outputs in crops and livestock (Haygarth et al., 1998a), which can lead to an increase in soil P concentration (Sims et al., 2000). However, areas with high soil P levels do not necessarily indicate areas that are susceptible to P transfer. The interaction between soil and hydrology is more important (Gburek and Sharpley, 1998; Haygarth et al., 2000; Gburek et al., 2000). In the UK, much grassland is associated with impermeable soils where overland flow, shallow subsurface flow, and artificial land drains (if installed) are the dominant hydrological pathways (Armstrong et al., 1984). The relationship between overland flow and P transfer is well documented (e.g., Sharpley, 1985), but only recently has the transfer of P through natural macropores and artificial land drains received more attention (Haygarth et al., 1998b; Heathwaite and Dils, 2000; Simard et al., 2000). Annual P exports from conventionally managed drained and undrained grassland have been reported to be in the range 0.5 to 2 kg P ha-1 (Haygarth and Jarvis, 1995; Haygarth et al., 1998b). Furthermore, Simard et al. (2000) found P concentrations in a composite of overland plus lateral subsurface (0–30 cm) and tile drain (85 cm) flow to be consistently higher than the 100 µg L-1 required to promote growth of blue-green algae in freshwater (Vollenweider, 1968).
Haygarth and Jarvis (1999) have defined three modes by which mobile P species start their journey from soils to waterways: solubilization, physical detachment, and direct transfer of recent P amendments (incidental transfer). Solubilization is the dissolution and subsequent displacement of soil solution P, commonly referred to as leaching. Physical transfer occurs when water flow carries nondissolved P, for example in association with soil particles during soil erosion (Dils and Heathwaite, 1996; Ryden et al., 1973). Incidental losses of P can occur when fertilizer or manure applications are coincident with the onset of rainfall. In reality, incidental losses will include solubilization and physical detachment; however, Haygarth and Jarvis (1999) have argued that, conceptually, incidental transfer should be kept separate because of the unique circumstances leading up to its occurrence and control.
Circumstances for incidental transfers of P are not uncommon in the UK. Frequent low-intensity rainfall can offer few opportunities to apply fertilizer or manure when soil and weather conditions are good. Furthermore, farmers are often under pressure to apply slurry (even to wet soils) to reduce pressure on limited storage facilities (Edwards and Withers, 1998; Withers et al., 2000). There is much anecdotal evidence of slurry applications that occur in defiance of good agricultural practice. As well as spreading onto wet or frozen land, farmers will use other methods to empty storage tanks, such as ejecting slurry from roadways and farm tracks onto hillslopes of adjacent fields. Additionally, incidental P transfer can occur from other P source areas within a catchment such as manure heaps, areas around animal housing, and zones of concentrated animal activity (e.g., near stream channels). Some experimental evidence of incidental P transfers has been observed following application of fertilizer (Scholefield and Stone, 1995; Haygarth and Jarvis, 1997), different forms of manure (Misselbrook et al., 1995; Wang et al., 1996), and direct returns from grazing animals (Sharpley and Syers, 1979). However, none of these studies have attempted to define the circumstances of incidental transfer in an objective and wholly orthogonal experimental situation. Our aim was to experimentally examine circumstances of incidental transfer from different P amendments, focusing particularly on mobilization mechanisms, P forms, and load potentials.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Lysimeter Design and Treatments
The site contains 24 plots (each 3 x 10 m) positioned in three blocks of eight down a 5° hillslope. Each plot is hydrologically isolated by a gravel-filled trench (85 cm) at the top of the plot and plastic sheeting to a depth of 50 cm along each side (Fig. 1)
. Any vertical drainage below 27 cm is impeded by the clay horizon (hydraulic conductivity 0.005 m d-1). A composite of overland flow and lateral subsurface flow (0–27 cm) is collected in a gravel trench at the base of each plot. Discharge over time was recorded using individually calibrated tipping buckets connected to a reed switch and counter (Scholefield and Stone, 1995).
|
The applications were made to the plots by hand. The soil was wet and further rainfall was forecast. These conditions are typical of our field site and many other grassland areas within the UK for several months of the year.
Rainfall, Plot Discharge, and Phosphorus Load
Rainfall was monitored hourly from a meteorological station 100 m from the site (Fig. 2a)
. Discharge from each plot was determined using individual tipping bucket data recorded for each sampling period. The exception to this was at the end of the 169 h, when discharge was determined by timing flow into a measuring vessel. Based on the randomized block design (three blocks, four treatments), discharge data from the 24 plots were subject to multivariate analysis for each sampling period. On all occasions the discharges from treatments were not significantly different (P > 0.05) and the grand mean (Q) was accepted. We have used Q as a record of plot discharge for all sampling periods. The values for Q represent average discharges over time and are plotted at the midpoint of each sampling period (i.e., 2, 6, 10 h, etc.) (Fig. 2a).
|
|
The nomenclature for P forms and fractions used in this study follow the operationally defined classifications described by Haygarth and Sharpley (2000). The terms for P forms are TP, RP, and unreactive P (UP) (calculated as the difference of TP and RP) and the fractions are unfiltered (unf), >0.45 µm, and <0.45 µm.
Analytical Methods
Analysis of TP in slurry was determined by oxidation in boiling aqua regia (15.5 M HNO3 and 10.3 M HCl mixture in a ratio of 1:3) under reflux for 3 h followed by inductively coupled plasma–optical emission spectroscopy (ICP–OES) determination of P (MEWAM, 1981). Total P in the discharge was determined using a persulfate digest combined with a modification of the Murphy and Riley (1962) method described by Rowland and Haygarth (1997). Following persulfate digestion in an autoclave (120°C, 103 kPa), 20 mL of sample was added to 1 mL of a colorimetric reagent and the formation of a molybdenum blue complex was measured at 880 nm. The high concentrations in many samples from the P-amended treatments required dilution (up to a factor of 100). This was performed after filtration but prior to TP determination. Dilution had the additional effect of reducing the concentration of particulate matter (especially slurry particles) that O'Connor and Syers (1975) have shown can result in underestimation of TP.
Additionally, aliquots from samples taken at 24 and 72 h were vacuum-filtered (80 kPa) through Millipore (Beford, MA) cellulose-nitrate membrane filters (0.45 µm). The filter membranes were prerinsed with ultrapure water and sample. Total P and molybdate reactive P were determined on unfiltered (TPunf, RPunf) and filtered (TP<0.45, RP<0.45) samples. Reactive P was determined using a similar procedure to those described for TP, but excluded the persulfate digest step and used 4 mL of colorimetric reagent to 20 mL of sample (Denison et al., 1998). This method was also used to determine the reactive water-soluble P content of five subsamples of the initial slurry source, diluted by a factor of 2000.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Total Phosphorus Concentrations
Immediately before the application of P, concentrations of TPunf in discharge were <100 µg L-1 for all treatments (Fig. 2b). However, following P application the increases in TPunf were rapid (Fig. 2b). Concentrations from the Slurry treatment exceeded 4000 µg TP L-1 after only 4 h and by 20 h concentrations from all the amended treatments were in excess of 3000 µg L-1. Peak TPunf for all treatments were coincident with high rainfall and Q at 24, 28, and 32 h. Concentrations for the TSP and TSP + Slurry treatments reached up to 11000 µg TP L-1, although standard errors for both treatments were considerable (Fig. 2b). In comparison, standard errors for the Slurry treatment were small and TPunf remained at around 7000 µg L-1. After 36 h TPunf declined for all treatments. From this point on TPunf from Slurry and TSP + Slurry were higher than the TSP treatment and the TSP + Slurry yielded consistently higher TPunf for the remainder of the experiment. Total P concentrations from Zero P showed some variation with Q, but within a relatively small range (54–191 µg L-1).
Loads
Total P transfers during the 7-d experimental period were equivalent to 1863, 2297, and 1805 g ha-1 from the TSP, TSP + Slurry, and Slurry treatments respectively, while the Zero P treatment yielded only 60 g ha-1 (Table 1). Maximum P loads were transferred with peak Q between 28 and 32 h (Table 1, Fig. 2a). This period accounted for 18% of the total discharge and rainfall during the experiment. However, 46% of the overall P load from the TSP treatment, 38% from TSP + Slurry, 33% from Slurry, and 27% from Zero P were transferred during this time. After the first 20 h the highest load was transferred from the TSP + Slurry treatment. Overall, P transfer accounted for 6 to 8% of the total P applied and loads from the P-amended treatments were 30- to 40-fold higher than P exported from the Zero P treatment.
Phosphorus Fractions and Forms at Low and High Discharge
The size fractions and forms of P in discharge varied between treatments and also within treatments at 24 h (high Q, some overland flow) and 72 h (low Q, subsurface flow) (Table 2). Reactive P<0.45 dominated P transfer from the TSP and TSP + Slurry treatments at both high Q (61–64%) and low Q (64–70%). In contrast, there was a more even contribution of all P forms and fractions from the Slurry treatment, although UP>0.45 contributed most at both discharge rates (49% at high Q and 39% at low Q). However, RP was more important than UP in the dissolved fraction (<0.45 µm), accounting for 20% of TPunf at high Q and 26% at low Q. For the Zero P treatment, particulate TP>0.45 had the greatest contribution at high and low Q, although the proportion of RP<0.45 was greater at low Q. For all treatments, particulate P (>0.45 µm) had a greater contribution to P transfer at high Q, whereas the contribution of dissolved P (<0.45 µm) increased at low Q.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The close agreement between total rainfall (48.8 mm) and total discharge (48 mm) (Fig. 2a, Table 1) during the experiment confirmed the effectiveness of plot isolation. This also agrees with the observations of Armstrong and Garwood (1991) who reported that, when the soil is saturated, the hydrological response to rainfall is dominated by rapid overland and subsurface flow (0–27 cm). In fact, the low hydraulic conductivity of the soil at depth >15 cm suggests that rapid flow response is probably from shallow subsurface pathways (0–15 cm), with overland flow generated whenever the capacity of subsurface pathways (0–15 cm) is exceeded. The actual contribution of overland flow, both to total discharge and to P transfer, could not be accurately quantified during this experiment. However, overland flow was observed at 24, 28, 32, 40, 54, and 125 h, and was probably a major component of plot discharge at 28 h, when rainfall intensity was 3 mm h-1, and at 32 h, which was preceded by 9 mm of rainfall in 4 h (Table 1). In this experiment the peaks in TPunf for all treatments were coincident with the observations of overland flow, the most important period being between 28 and 32 h. The TSP and TSP + Slurry treatments peaked at 11000 µg TP L-1, Slurry at 7000 µg TP L-1, and Zero P at 191 µg TP L-1. This 4-h period, which accounted for only 18% of rainfall and Q, resulted in 46% of the overall TPunf load from the TSP treatment compared with 38, 33, and 27% from the TSP + Slurry, Slurry, and Zero P treatments, respectively (Table 1). The relationship between overland flow and P transfer is well documented. Overland flow erodes and transports P associated with particulate material (Ryden et al., 1973) and soluble P is readily desorbed into overland flow from the P-rich soil surface (Ahuja et al., 1981; Sharpley, 1985).
Apart from the aforementioned observations of overland flow, the majority of discharge over the duration of the experiment occurred as lateral subsurface flow (0–27 cm). In the subsurface pathways, macropore flow is important during rapid response to rainfall and slower responses are the result of water percolating through the soil matrix. However, contrary to the traditional perspective that P is relatively immobile within the soil profile, TPunf in subsurface flows was also high. There was little rainfall at the very start of the experiment and Q remained low (Fig. 2a). However, after only 4 h, TPunf from all amended treatments had increased from <100 µg L-1 to >2000 µg L-1, with the Slurry treatment >4000 µg L-1 (probably due to its rapid infiltration) (Fig. 2b). By 20 h all amended treatments were yielding concentrations >3000 µg TP L-1. The peak TPunf in subsurface flow occurred at 46 h from the TSP + Slurry treatment (5000 µg L-1) (Fig. 2b). After this time TPunf in subsurface flow declined for all treatments (Fig. 2b). Even at 125 h, when rainfall and Q were similar to those at 46 h, TPunf from the TSP + Slurry treatment had decreased to 2000 µg L-1.
Overall, the P loads transferred from all amended treatments were extremely high. However, there were differences in P loads transferred within and between treatments depending on the hydrological circumstances. An explanation for the high TPunf (up to 11000 µg L-1) recorded between 24 and 32 h from the treatments receiving TSP (Fig. 2b), could be that TSP granules at the soil surface were directly entrained or rapidly solubilized by overland flow (Heathwaite et al., 1998). This idea of solubilzation is further supported by the fractionation of samples at high Q (Table 2). The contribution of RP<0.45 at this time accounted for 61 and 64% of TPunf from the TSP and TSP + Slurry treatments, respectively. However, after 32 h and even during other periods of high Q (40 and 54 h, Fig. 2a), TPunf from the TSP had reduced considerably, probably because solubilized P was becoming adsorbed to the soil. At low Q, TPunf from the TSP treatment (966 µg L-1) had reduced 11-fold from the peak concentration at 28 h. However, this value was still high relative to the Zero P treatment (Fig. 2b). Furthermore, the fact that RP<0.45 accounted for 64% of TPunf getting through the soil system at this time (Table 2) is perhaps surprising considering the adsorption potential of the P-deficient bulk soil. Yet this is consistent with work by Jensen et al. (1998), who found that dissolved inorganic P can be efficiently transported in large macropores and avoid adsorption to the bulk soil.
All measured P forms and fractions contributed to P transfer from the Slurry treatment (Table 2). However, it appears that the physical detachment of P associated with slurry material was the most important mechanism for P transfer from the Slurry treatment. Again, the highest concentrations of P (7000 µg L-1) were at 24, 28, and 32 h, when overland flow was observed. Particulate P (TP>0.45) contributed most to P transfer at both high Q (66%) and low Q (55%), most of which was UP>0.45 (49 and 39%, respectively). Reactive P was associated with both particulate (RP>0.45) and dissolved (RP<0.45) fractions. However, it is interesting to note that total water-soluble P (RPunf) accounted for 68% of TPunf in the original slurry matrix but it only contributed 37% (high Q) and 42% (low Q) of TPunf in discharge (Table 2). This may indicate the preferential transfer of UP over RP, perhaps because RP has greater potential to be adsorbed to the soil. Additionally, during subsurface flow at 72 h, TPunf (1810 µg L-1) was only 3.5-fold less than peak concentration at 28 h (Fig. 2b). The higher TPunf from the Slurry treatment compared with the TSP treatment at this time is consistent with work by Chardon et al. (1997) and Edwards et al. (1996), who have demonstrated the high mobility of organic P forms in subsurface pathways.
Phosphorus transfer from the TSP + Slurry treatment followed a similar pattern to the TSP and Slurry treatments. Up to 32 h, P fractions and concentrations from the Slurry treatment were almost identical to the TSP treatment (Fig. 2b, Table 2). Then, after 32 h (mostly subsurface flow), TPunf from the TSP + Slurry treatment behaved similarly to the Slurry treatment (although concentrations were higher from the TSP + Slurry treatment). Fractionation of samples at low Q showed 70% of TP to be RP<0.45 compared with only 26% from the Slurry treatment (Table 2). This could be explained by an interaction between solubilized TSP and slurry particles. By plotting the observed TSP + Slurry TPunf values (Fig. 2b) against expected values (calculated as the mean of paired TSP and Slurry values), we see that the observed values are consistently higher (Fig. 3) . This may be a function of TSP-derived P becoming adsorbed to the surface of the more mobile slurry particles. Indeed, this concept is consistent with recent work (e.g., Donald et al., 1993; Kretzschmar et al., 1999; Stamm et al., 1998) that suggests that P is easily adsorbed to colloids and particles, which can then act as vehicles for P transfer in overland and subsurface hydrological pathways.
|
Further investigations at a range of integrated scales are required to assess the effect of incidental P transfer. At the plot scale we need to make the distinction between P forms and fractions in overland and subsurface hydrological pathways for a range of antecedent conditions. While moving up the scale we need to understand the implications of applying P amendments to different areas in relation to the hydrological connectivity within a catchment. For this experiment, P amendments were applied to the soil when an immediate response was expected. However, we also need to understand how long fertilizer and manure remain susceptible to incidental transfer when applied in good conditions. The rates at which P is incorporated into the soil system, leaving it less susceptible to direct transfer, also require investigation for different P amendments at a range of antecedent conditions.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The highest P loads were coincident with high rainfall and discharge. At these times, physical detachment was important for mobilizing P (especially P > 0.45 µm from the Slurry treatment), whereas the solubilization of TSP granules at the soil surface meant that RP<0.45 was the most important P fraction for the TSP and TSP + Slurry treatments. The high UP concentrations from the treatments receiving slurry augment previous work showing organic P forms to be highly mobile in subsurface pathways. Furthermore, the potential of slurry particles to act as vehicles for P transfer has been illustrated by the TSP + Slurry treatment. The high RP concentrations in subsurface pathways also support previous findings, which show rapid macropore flow to be an efficient pathway for RP transfer through P-deficient soil horizons.
Highlighting the risk associated with incidental P transfers and targeting short-term decision-making is perhaps the most immediately viable method for mitigating P loss. At the very least, farmers who are under pressure to recycle manures should be informed of the benefits of using local knowledge of hydrology to apply P to areas of low transfer potential. Further research is needed to assess the time that new P inputs remain vulnerable to incidental transfer. For example, we need to investigate how quickly P is incorporated into the soil system and the risk associated with applications made in different soil moisture conditions.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
|
|
P. Kleinman, D. Sullivan, A. Wolf, R. Brandt, Z. Dou, H. Elliott, J. Kovar, A. Leytem, R. Maguire, P. Moore, et al. Selection of a Water-Extractable Phosphorus Test for Manures and Biosolids as an Indicator of Runoff Loss Potential J. Environ. Qual., July 17, 2007; 36(5): 1357 - 1367. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. van der Perk, P. N. Owens, L. K. Deeks, B. G. Rawlins, P. M. Haygarth, and K. J. Beven Controls on Catchment-Scale Patterns of Phosphorus in Soil, Streambed Sediment, and Stream Water J. Environ. Qual., April 5, 2007; 36(3): 694 - 708. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X.-L. Huang, Y. Chen, and M. Shenker Solid Phosphorus Phase in Aluminum- and Iron-Treated Biosolids J. Environ. Qual., March 1, 2007; 36(2): 549 - 556. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. A. Elliott, R. C. Brandt, P. J. A. Kleinman, A. N. Sharpley, and D. B. Beegle Estimating Source Coefficients for Phosphorus Site Indices J. Environ. Qual., October 27, 2006; 35(6): 2195 - 2201. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Lazarovitch, A. Ben-Gal, and U. Shani An Automated Rotating Lysimeter System for Greenhouse Evapotranspiration Studies Vadose Zone J., May 26, 2006; 5(2): 801 - 804. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Schelde, L. W. de Jonge, C. Kjaergaard, M. Laegdsmand, and G. H. Rubaek Effects of Manure Application and Plowing on Transport of Colloids and Phosphorus to Tile Drains Vadose Zone J., March 8, 2006; 5(1): 445 - 458. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Heathwaite, P. Haygarth, R. Matthews, N. Preedy, and P. Butler Evaluating Colloidal Phosphorus Delivery to Surface Waters from Diffuse Agricultural Sources J. Environ. Qual., January 1, 2005; 34(1): 287 - 298. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. R. Hart, B. F. Quin, and M. L. Nguyen Phosphorus Runoff from Agricultural Land and Direct Fertilizer Effects: A Review J. Environ. Qual., November 1, 2004; 33(6): 1954 - 1972. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. J. Dougherty, N. K. Fleming, J. W. Cox, and D. J. Chittleborough Phosphorus Transfer in Surface Runoff from Intensive Pasture Systems at Various Scales: A Review J. Environ. Qual., November 1, 2004; 33(6): 1973 - 1988. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Dou, G. Y. Zhang, W. L. Stout, J. D. Toth, and J. D. Ferguson Efficacy of Alum and Coal Combustion By-Products in Stabilizing Manure Phosphorus J. Environ. Qual., July 1, 2003; 32(4): 1490 - 1497. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Tabbara Phosphorus Loss to Runoff Water Twenty-Four Hours after Application of Liquid Swine Manure or Fertilizer J. Environ. Qual., May 1, 2003; 32(3): 1044 - 1052. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME |
