Phosphorus Transfer in Runoff Following Application of Fertilizer ...

Published January, 2001

Phosphorus Transfer in Runoff Following Application of Fertilizer, Manure, and Sewage Sludge Paul J. A. Withers,* Stephen D. Clay, and Victor G. Breeze ABSTRACT

face runoff in rural watersheds, causing eutrophication (Haygarth and Jarvis, 1999). Although of no agronomic or financial consequence to the farmer, nonpoint P loads in runoff can become a readily available P source to aquatic biota. Manures and sludges are of particular concern because they are increasingly applied to limited land areas and result in an excess loading of P to the soil when application rates are based on crop N requirements, particularly for sewage sludges, which tend to have a lower N to P ratio than manures (Sharpley et al., 1998). Land application of sewage sludge may therefore increase nonpoint P transfers compared with other P amendments, especially at high application rates and/ or under conditions of high runoff potential (Kirkham, 1982; Melanen et al., 1985; Mostaghimi et al., 1988). Control over such incidental P losses requires an understanding of the conditions under, and time periods over, which losses occur and the processes involved. Differences in P transfer to water following application of P amendments might be expected depending on the type of material applied (McLeod and Hegg, 1984; Bushee et al., 1998), the rate and timing of application (Edwards and Daniel, 1993; Smith et al., 1998), the frequency and/or timing of rainfall events after application (Sharpley, 1997), the method employed (Mueller et al., 1984), or the prescence of a buffer zone (Heathman et al., 1995). Measurements of P transfer rates in land runoff from experimental plots receiving manures have therefore been shown to be very variable (Sharpley et al., 1998; Khaleel et al., 1980). Differences in P transfer have been mostly related to site conditions and effects on runoff volumes, rather than to inherent differences in the relative P availabilities of the materials applied. For example, data presented by Frossard et al. (1996) suggest differences in the P availability of different types of sewage sludge, depending on sludge treatment, which might affect their potential for P release in storm runoff. If such differences exist, which is supported by limited experimental data (McLeod and Hegg, 1984; Melanen et al., 1985), this may influence the types of P amendments, or application rates, used in different areas. This paper reports on a field experiment that compared the transfers of P in surface runoff from field plots amended with inorganic fertilizer, cattle manure, and sewage sludge, and the extent to which differences in P transfer could be accounted for by differences in the P availability of the materials applied as estimated in the laboratory.

Phosphorus (P) transfer in surface runoff from field plots receiving either no P, triplesuperphoshate (TSP), liquid cattle manure (LCS), liquid anaerobically digested sludge (LDS), or dewatered sludge cake (DSC) was compared over a 2-yr period. Dissolved inorganic P concentrations in runoff increased from 0.1 to 0.2 mg L⫺1 on control and sludge-treated plots to 3.8 and 6.5 mg L⫺1 following application of LCS and TSP, respectively, to a cereal crop in spring. When incorporated into the soil in autumn, runoff dissolved P concentrations were typically ⬍0.5 mg L⫺1 across all plots, and particulate P remained the dominant P form. When surface-applied in autumn to a consolidated seedbed, direct loss of LCS and LDS increased both runoff volume and P transfers, but release of dissolved P occurred only from LCS. The largest P concentrations (⬎70 mg L⫺1) were recorded following TSP application without any increase in runoff volume, while application of bulky DSC significantly reduced total P transfers by 70% compared with the control due to a reduced runoff volume. Treatment effects in each monitoring period were most pronounced in the first runoff event. Differences in the release of P from the different P sources were related to the amounts of P extracted by either water or sodium bicarbonate in the order TSP ⬎ LCS ⬎ LDS ⬎ DSC. The results suggest there is a lower risk of P transfer in land runoff following application of sludge compared with other agricultural P amendments at similar P rates.

I

norganic P fertilizers and livestock manures are routinely applied to agricultural land to meet crop P requirements, boost soil P fertility, or to avoid manure storage. Alternative sources of P, such as sewage sludge, have also become more available and their application to land is forecast to expand in some regions as other routes of disposal become increasingly restricted (Davies, 1996). The resulting build-up of soil P fertility has given farmers greater flexibility in the rate and timing of these different P amendments. For example, a significant number of UK farmers now topdress P fertilizer to the growing crop in order to save time at drilling. These topdressings often supply the P requirements of all or part of the crop rotation rather than the specific requirement of the crop being grown. Farmers are also encouraged to apply manures to growing crops in spring to help maximize manure N utilization and reduce N leaching losses associated with incorporation prior to sowing in autumn (Smith and Chambers, 1993). These trends might be considered to have increased the risk of incidental P transfer in surface and/or subsur-

Paul J.A. Withers, ADAS Bridgets, Martyr Worthy, Winchester SO21 1AP, UK; Stephen D. Clay, Severn Trent Water Limited, Process Engineering, Alpha House, Warwick Technology Park, Heathcote Road, Warwick CV34 6DA, UK; and Victor G. Breeze, ADAS Rosemaund, Preston Wynne, Hereford HR1 3PG, UK. Received 30 Nov. 1999. *Corresponding author ([email protected]).

Abbreviations: DSC, dewatered sludge cake; DUP, dissolved unreactive phosphorus; LCS, liquid cattle manure; LDS, liquid digested sludge; MRP, molybdate-reactive phosphorus; PP, particulate phosphorus; TDP, total dissolved P; TP, total phosphorus; TSP, triplesuperphosphate.

Published in J. Environ. Qual. 30:180–188 (2001).

180

181

WITHERS ET AL.: PHOSPHORUS TRANSFER IN SURFACE RUNOFF

Table 1. Mean and range in nutrient composition of the manure treatments, and the proportions of manure total P extracted by water, sodium bicarbonate (NaHCO3), sodium hydroxide (NaOH), and hydrochloric acid (HCl). Values are means of three replicates except for dry solids (DS) and total P, which are means of five replicates. Liquid cattle manure (LCS)

kg⫺1)

Dry solids (g N (g kg⫺1 DS) P (g kg⫺1 DS) N to P ratio Water P NaHCO3 P NaOH P HCl P

Liquid digested sludge (LDS)

Mean

Range

Mean

46 58 12.8 4.7

(35–72) (55–61) (11.3–13.8) (4.3–5.4)

71 43 16.5 2.7

60 91 46 100

(41–82) (93–100) (31–63) (97–100)

Range

Mean

Range

(24–126) (29–58) (13.1–21.4) (2.1–3.3) Percent of total P

227 32 10.6 2.8

(183–272) (32–33) (10.2–13.1) (2.5–3.0)

(4–13) (23–53) (24–36) (89–100)

0.8 12 37 66

(0.3–1.5) (11–14) (32–41) (63–69)

8 35 30 94

MATERIALS AND METHODS Site Establishment and Sample Collection Fifteen adjacent experimental plots were established during February 1995 on a south-facing, 5⬚ slope of a field at ADAS Rosemaund, Herefordshire, UK (Grid Reference SO554486). The silty clay loam soil (Argillic Albaqualf) is developed over Old Red Sandstone and disperses readily during heavy rain. The site is representative of well-managed, fertile arable land within the surrounding area and receives an average annual rainfall of 787 mm yr⫺1. The initial concentration of Olsenextractable P in the soil ranged from 16 to 25 mg L⫺1. The plots measured 2 m wide and 16 m long and were hydrologically isolated by a gravel trench at the upslope perimeter and by 110-mm smooth-bore, nonperforated, plastic pipe laid along the length of each plot. The interface of the pipe with the soil was sealed with bentonite to prevent surface runoff entering adjacent plots. Surface runoff was collected in 110mm gutter pipe located at the end of each plot and fed by connecting pipes into a plastic reception tank. After each significant storm event, the runoff volume was measured and, after thorough stirring, a 250-mL subsample was taken for determination of P content. Runoff was collected mostly from storm events giving 10 to 15 mm or more of rain over a 24to 48-h period (Fig. 1). The collection tanks were emptied after each event.

Treatments and Experimental Design The experiment compared inorganic P fertilizer applied as TSP, LCS, LDS, and DSC, with a control receiving no P. Both sludges had undergone primary filtering and secondary biological treatment prior to either anaerobic digestion (LDS) or dewatering (DSC), and were representative of sludges recycled to land in the UK. All materials were obtained from the same source each year, except that a second source of DSC was applied in autumn 1996. The range in chemical composition of the organic amendments is given in Table 1. The treatments were arranged in a randomized block design

Dewatered sludge cake (DSC)

with three replicates and applied by hand at rates that were intended to supply comparable amounts of P as determined by preliminary analysis of test samples taken before the treatments were applied. The P rates were chosen to represent up to the maximum recommended in the UK under the Code of Good Agricultural Practice for the Protection of Water (Ministry of Agriculture, Fisheries and Food, 1998), which is equivalent to 250 kg ha⫺1 of N. Actual rates applied were calculated from analysis results of samples taken at the time of application, and sometimes differed from the target rates due to variation in the chemical composition of the materials applied and difficulties in handling them (Table 2). Plots receiving solid P amendments (TSP and DSC) tended to receive more P than plots receiving liquid materials (LCS and LDS). Overall, the TSP, LCS, LDS, and DSC treatments supplied 330, 186, 150, and 329 kg P ha⫺1, respectively. All plots were cropped to cereals each year and received similar amounts of agrochemicals (other than P), and were managed in the same manner as the rest of the field.

Runoff Monitoring Program Two runoff events were monitored before treatments were applied to ensure uniformity of runoff volumes between plots. Thereafter, storm runoff was collected during four monitoring test periods (Periods 1–4) following treatment application (Table 2). In April 1995, the treatments were surface-applied to a growing crop of winter wheat (Triticum aestivum L.) at the tillering stage (Period 1). For the second monitoring period, the treatments were incorporated into the surface 20 cm of soil with a rotavator before sowing with winter wheat in December 1995. Each plot was plowed before treatment application. The treatments were applied again in late May 1996 (except LCS and LDS, which were applied on 3 June 1996) when the crop was at the stem extension stage (Period 3). For the fourth monitoring period, the treatments were applied in November 1996 to the surface of a seedbed that had been prepared after conventional plowing and tine cultivation, sown with winter barley (Hordeum vulgare L.), and subsequently rolled. Appli-

Table 2. Treatment application details, monitoring dates, and number of storm events sampled during each test period. Test periods Period Period Period Period

1 2 3 4

Treatment application dates

Method of application

Total P applied TSP†

LCS†

DLS†

28 Apr. 1995 30 Nov. 1995 22 May 1996§ 1 Nov. 1996

surface-applied incorporated surface-applied surface-applied

60 90 90 90

60 40 37 49

54 35 33 28

DSC†

Runoff monitoring periods‡

Number of storm events monitored

35 144 58 92

18 May to 7 Aug. 1995 20 Dec. 1995 to 23 Apr. 1996 23 May to 12 Aug. 1996 4 Nov. 1996 to 25 Feb. 1997

4 8 2 7

† TSP, triplesuperphosphate; LCS, liquid cattle manure; LDS, liquid digested sludge; DSC, dewatered sludge cake. ‡ Dates of first and last storm runoff events measured. § Application of the LCS and LDS treatments was delayed until 3 June 1996.

Fig. 1. Monitored runoff events (indicated by down arrows) in relation to daily rainfall during (a) Period 1, (b) Period 2, (c) Period 3, and (d) Period 4 following treatment application. Runoff volumes (L plot⫺1) collected from control plots for each event are also shown.

182 J. ENVIRON. QUAL., VOL. 30, JANUARY–FEBRUARY 2001


cation rates of the liquid treatments in this last test period had to be reduced to avoid direct runoff losses from the consolidated surface left by rolling (Table 2). The winter barley crop did not extablish well following treatment application and the site was plowed-under in March 1997, and resown to spring barley. Three storm events were monitored after plowing to test for any residual soil effects from the cumulative build-up of soil P, as measured in soil samples taken from each plot after the crop was harvested in 1997.

Analytical Methods Runoff samples collected after each storm event were analyzed for dissolved molybdate-reactive phosphorus (MRP), total dissolved phosphorus (TDP), and total phosphorus (TP). Total P was determined by inductively coupled plasma–optical emission spectroscopy (ICP–OES) after nitric and hydrochloric acid (aqua regia) digestion. Dissolved fractions (MRP and TDP) were determined after filtering through a 0.45-␮m millipore filter, MRP was determined colorimetrically according to the method of Murphy and Riley (1962), and TDP was determined directly by ICP–OES. Total dissolved P concentrations were generally well above the 100 ␮g L⫺1 detection limit of ICP–OES suggested by Rowland and Haygarth (1997). The difference between TP and TDP was assumed to be particulate phosphorus (PP), and that between MRP and TDP was termed dissolved unreactive phosphorus (DUP) and assumed to be largely of organic origin. The amounts of inorganic P extracted from the organic amendments by distilled water, 0.5 M NaHCO3, 0.1 M NaOH, and 1 M HCl were determined colorimetrically following a nonsequential adaptation of the method of Hedley et al. (1982). Water-soluble P in the soil was determined colorimetrically at a 1:10 soil to solution ratio. Soil test P was determined by the method of Olsen et al. (1954) and total P in soil by ICP–OES following aqua regia digestion.

Data Handling Treatment effects on mean values of measured parameters were determined by analysis of variance techniques (Genstat 5 Committee, 1993). Data values for Period 4 were logarithmically transformed prior to Genstat analysis. Treatment effects on cumulative P loads assumed that the monitoring periods were independent of each other; this assumption was based on the fact that all plots were plowed to 20 to 25 cm between periods and the lack of any effect of residual soil P build-up on P concentrations in runoff at the end of the experiment. To allow better comparison between treatments, the amounts of P collected over each monitoring period were adjusted according to the total runoff volume and reported as volumeadjusted (flow-weighted) concentrations. Valid treatment comparisons were restricted to the first, second, and fourth monitoring periods, since there was only one event comparing all treatments during the third monitoring period and this event occurred very late in the growing season.

RESULTS AND DISCUSSION Manure Analysis Laboratory analysis revealed large differences in the amounts of P extracted from the organic amendments by water, NaHCO3, and HCl, while differences in NaOHextractable P were relatively small (Table 1). Water and NaHCO3 extracted considerably more P from the cattle manure than from the sludges, while HCl-soluble P was greater in the liquid amendments (LCS and LDS) than

183

in the two solid DSC samples. The high P solubility of LCS in both NaHCO3 and HCl indicates a dominance of inorganic forms, while the large amounts of P extracted by water (60% of total P) suggest this material has potential for rapid P release in storm runoff. Although the P in the liquid sludge was also largely soluble in HCl, the proportions extracted by NaHCO3 (35% of total P) and water (⬍10% of total P) were noticeably lower than for the cattle manure, especially for watersoluble P. In contrast, the dewatered sludges were only partially soluble in HCl, contained little NaHCO3– extractable P (⬍15% of total P), and virtually no watersoluble P (ⱕ1% of total P). The latter is probably due to the removal of soluble P forms in the sludge during the dewatering process. Differences in P solubility between the sludges and the cattle manure may be due to differences in the proportions of sparingly soluble calcium phosphate compounds present (Sommers et al., 1976; Hinedi et al., 1989; Frossard et al., 1996). Since TSP is ⬎95% soluble in water, the relative differences in the potential P release to runoff can be expected to be in the order TSP ⬎ LCS ⬎ LDS ⬎ DSC.

First Monitoring Period The first storm event producing runoff was recorded on 18 May, 3 wk after treatment application. During this event, the amounts of total P collected from plots receiving TSP and LCS were significantly (P ⬍ 0.05) greater than those from plots receiving no P, LDS, or DSC. This contrasts strongly with the amounts of P collected in plot runoff prior to treatment application, which were relatively uniform across all plots (Fig. 2). These large treatment effects were due to an increase in the concentrations of dissolved (⬍0.45 ␮m) P in the runoff rather than to an increase in particulate P concentrations (Fig. 3). For example, average MRP concentrations of 0.14, 6.49, 3.77, 0.21, and 0.20 mg L⫺1 were recorded from plots receiving no P, TSP, LCS, LDS, or DSC, respectively. There were no treatment effects on DUP concentrations that were ⬍0.1 mg L⫺1. Consequently, TDP represented 81 and 65% of TP loads (mg plot⫺1) collected on TSP- and LCS-treated plots, respectively, but remained close to 30% on other plots, including those receiving sewage sludge. Concentrations of dissolved P in the two subsequent runoff events in May remained elevated on TSP- and LCS-treated plots, averaging close to 1 mg L⫺1 compared with values of ca. 0.3 mg L⫺1 on other plots (Fig. 3). However, the differences in MRP concentrations were much smaller than observed in the first event, and treatment effects on runoff TP loads were significant at only the 10% level. This is demonstrated by a return to a shallower gradient of cumulative P loss after the first storm event (Fig. 2). Following a prolonged period with very little rain, a further heavy storm on 7 August generated a small amount of P-rich runoff from all plots, including the control (average 3 L plot⫺1). For example, TDP concentrations averaged 8.7 mg L⫺1, and accounted for 77% of TP concentrations. This event therefore had a strong influence on cumulative P transfer

184

J. ENVIRON. QUAL., VOL. 30, JANUARY–FEBRUARY 2001

Fig. 2. Cumulative loss of total phosphorus (TP) in relation to cumulative runoff before and after treatment application in Period 1. The down arrow represents the date of treatment application. TSP, triplesuperphosphate; LCS, liquid cow manure; LDS, liquid digested sludge; DSC, dewatered sludge cake.

over the monitoring period (Fig. 2) and suggests that the P amendments were no longer contributing P in runoff. The cumulative amounts of total P collected over the first three storm events were threefold higher than the control following TSP and LCS application, while amounts of TDP collected from these two treatments were 12-fold higher than the control (Table 3). Total amounts of P collected from sludge-treated plots were not significantly different from the control. Cumulative differences between treatments were not related to differences in runoff volumes, which were generally very uniform across all plots. Runoff volumes collected from TSP-treated plots were slightly lower than for other treatments during the first two events (Fig. 2), but not

significantly so (P ⬎ 0.05). Comparison of volumeadjusted P concentrations confirmed the lack of any significant release of P from the sludge materials (Table 3), which was consistent with the differences in potential P release measured in the laboratory. Similar relative differences in runoff P between sludge, manure, and inorganic fertilizer were observed by McLeod and Hegg (1984). The application rate of TSP used was typical of what might be applied on commercial farms in the UK as a topdressing to match crop P requirements over part of a crop rotation, and the LCS loading rate was well within recommended limits (Ministry of Agriculture, Fisheries and Food, 1998). Such practices clearly represent a potential eutrophication hazard in riparian areas,

Fig. 3. Concentrations of total dissolved phosphorus (TDP) and particulate phosphorus (PP) in runoff collected in the first three storm events following treatment application in Period 1. Error bars indicate least significant difference, P ⬍ 0.05. C, control; TSP, triplesuperphosphate; LCS, liquid cow manure; LDS, liquid digested sludge; DSC, dewatered sludge cake.

185


Table 3. Treatment effects on cumulative total dissolved phosphorus (TDP) and particulate phosphorus (PP) loads, flow-weighted total phosphorus (TP) concentrations, and the proportion occurring in dissolved form for Periods 1, 2, and 4. Values in Period 1 cover the first three events only, values in Period 4 are log-10 transformed. Period 1 (surface-applied in spring) Cumulative load Treatment†

TDP

PP

Control TSP LCS LDS DSC LSD Significance

mg plot⫺1 5 19 63 19 62 26 8 16 8 25 40.5 12.8 * NS‡

TP conc.

TDP conc.

mg L⫺1 0.89 4.79 3.99 0.87 1.19 2.548 *

% of TP 24 74 62 34 28 22.5 **

Period 2 (soil-incorporated in autumn) Cumulative load TDP

PP

mg plot⫺1 23 79 48 65 57 78 37 78 27 88 19.5 22.3 * NS

TP conc.

TDP conc.

mg L⫺1 0.60 0.70 0.84 0.69 0.77 0.221 NS

% of TP 22 40 42 33 23 7.4 **

Period 4 (surface-applied in autumn) Cumulative load TDP

PP

mg plot⫺1 1.70 2.45 3.54 2.74 3.55 3.61 2.53 3.18 1.39 1.93 0.390 0.345 *** ***

TP conc.

TDP conc.

mg L⫺1 3.41 4.37 4.29 3.96 3.16 0.145 ***

% of TP 17 86 47 19 23 17.4 ***

* Significant at the 0.05 probability level. ** Significant at the 0.01 probability level. *** Significant at the 0.001 probability level. † TSP, triplesuperphosphate; LCS, liquid cattle manure; LDS, liquid digested sludge; DSC, dewatered sludge cake. ‡ Not significant.

especially in view of the high dissolved P component of the runoff.

Second Monitoring Period Compared with the very high concentrations recorded in Period 1, average runoff TP concentrations recorded during the second monitoring period were uniformly low (0.6–0.8 mg L⫺1). Total amounts of P collected also fell within the relatively narrow range of 102 to 135 mg plot⫺1 (Table 3, Fig. 4) due to incorporation of the treatments before sowing. However, in the first runoff event, which again occurred 3 wk after treatment application, six- and eightfold greater dissolved P concentrations were recorded on plots receiving TSP and LCS, respectively, while runoff P concentrations on LDS plots were also increased threefold, compared with both the untreated control and the DSC-treated plots (Table 4). Dissolved P concentrations on treated plots remained slightly elevated in subsequent events, especially on plots receiving TSP and LCS, and during the larger

storms, but these tended not to be statistically significant (P ⬎ 0.05). Concentrations of DUP remained low during this monitoring period (⬍0.1 mg L⫺1). Cumulative amounts of TDP collected in runoff over the eight events monitored, and the volume-adjusted concentrations of TDP, were significantly larger than the control following TSP and LCS application (Table 3). Smaller increases in TDP transfers from plots receiving liquid sludge were not significant, except when expressed as a proportion of total P. Application of the dewatered sludge did not increase dissolved P transfer. Particulate P remained the dominant form transported from all plots during this period and the increases in TDP concentrations were not sufficient to cause a significant (P ⬍ 0.05) treatment effect on TP concentrations or loads (Table 3). As in the first monitoring period, there was no statistically significant effect of the P treatments on either runoff volumes or on PP concentrations in the runoff. The main treatment effect was, therefore, to slightly increase the proportion of the loss occurring in dissolved form, particularly in the first run-

Fig. 4. Cumulative loss of total phosphorus (TP) in relation to cumulative runoff after treatment application in Period 2. TSP, triplesuperphosphate; LCS, liquid cow manure; LDS, liquid digested sludge; DSC, dewatered sludge cake.

186


Table 4. Treatment effects on runoff volume and total dissolved phosphorus (TDP), particulate phosphorus (PP), and total phosphorus (TP) concentrations in the first runoff event following incorporation (Period 2) or surface application (Period 4). Values in Period 4 are log-10 transformed. Period 2 (soil-incorporated)

Period 4 (surface-applied)

Concentration Treatment†

Runoff L

Control TSP LCS LDS DSC LSD Significance

TDP

plot⫺1 19 14 16 15 14 6.9 NS‡

PP

Concentration TP

L⫺1

0.09 0.58 0.74 0.34 0.17 0.375 *

mg 0.98 0.99 1.08 0.71 1.39 0.417 NS

Runoff

TDP

plot⫺1

1.07 1.57 1.82 1.06 1.56 0.583 NS

L 1.44 1.28 2.12 1.61 0.75 0.221 ***

PP

TP

L⫺1

2.96 4.99 4.07 3.51 2.71 0.370 ***

mg 3.59 3.86 4.28 4.35 3.33 0.411 **

3.68 5.02 4.53 4.41 3.42 0.276 ***

* Significant at the 0.05 probability level. ** Significant at the 0.01 probability level. *** Significant at the 0.001 probability level. † TSP, triplesuperphosphate; LCS, liquid cattle manure; LDS, liquid digested sludge; DSC, dewatered sludge cake. ‡ Not significant.

off event (Table 4). The notable lack of any increase in dissolved P concentrations following application of DSC, despite the very large rate of P applied compared with other treatments (Table 2), confirms the low water solubility of P in this sludge found by laboratory analysis. As reported by Dunnigan and Dick (1980), incorporation of the P amendments effectively reduced P transfers compared with surface application, and minimized any differences in potential P release to runoff between the different treatments applied. However, other work has shown that, in the absence of any crop cover, or under more extreme rainfall intensities, the soil disturbance caused by incorporation may also increase P transfers due to accelerated transport of soil particles (Mostaghimi et al., 1992).

Third Monitoring Period Although 11 mm of rain fell the day after surface application of the TSP and DSC treatments on 22 May 1996, concentrations of MRP, TDP, and TP were similar across all plots, averaging 1.1, 1.3, and 2.6 mg L⫺1, respectively. Only on one TSP-treated plot was an exces-

sively large TDP concentration measured (25 mg L⫺1). Little rain fell after the application of the liquid treatments (3 June) until 12 August. Runoff volumes and P concentrations during this latter event were similar across all plots, with TDP accounting for 62% of the total P transfer. Runoff TDP and TP concentrations averaged 4.8 and 6.9 mg L⫺1, respectively, and were as large as those found in August runoff during the first monitoring period, despite the greater volume of runoff collected (17 L plot⫺1). These large concentrations may be due to a P mineralization effect of rain on dry soils (Magid and Nielsen, 1992) and/or the prescence of senescing crop tissue (Schreiber and McDowell, 1985).

Fourth Monitoring Period Lower rainfall intensities were needed to generate plot runoff during the fourth monitoring period than during earlier periods (Fig. 1) due to the consolidated nature of the soil surface before treatment application. Also, in contrast to other monitoring periods, there was a significant (P ⬍ 0.01) treatment effect on the amount of runoff collected both during individual storm events,

Fig. 5. Cumulative loss of total phosphorus (TP) in relation to cumulative runoff after treatment application in Period 4. TSP, triplesuperphosphate; LCS, liquid cow manure; LDS, liquid digested sludge; DSC, dewatered sludge cake. Note the logarithmic scale.


Table 5. Treatment effects on the amounts of water-soluble, Olsen-extractable, and total P in the topsoil at the end of the experiment. Treatment† Control TSP LCS LDS CSC LSD Significance

Total P applied kg ha⫺1 0 330 186 150 329

Total P 362 485 442 418 522 53.8 **

Olsenextractable P mg kg⫺1 18 52 30 28 32 7.4 ***

Watersoluble P 1.9 14.3 6.1 4.4 5.9 2.64 ***

** Significant at the 0.01 probability level. *** Significant at the 0.001 probability level. † TSP, triplesuperphosphate; LCS, liquid cattle manure; LDS, liquid digested sludge; DSC, dewatered sludge cake.

and cumulatively over the whole monitoring period (Fig. 5). In the first runoff event, which occurred within a few days of application, plots receiving liquid amendments, especially LCS, generated significantly more runoff, and plots receiving DSC generated less runoff, than plots receiving either no P or TSP (Table 4). The increased runoff volumes were due to a direct loss of manure and sludge from the consolidated surface left by rolling, and resulted in large increases in concentrations of both dissolved and particulate P. For example, MRP concentrations increased from 0.7 mg L⫺1 on control plots to 74, 9, and 3 mg L⫺1 on TSP-, LCS-, and LDS-treated plots, respectively (Table 4). Dissolved unreactive P concentrations were also elevated after TSP and LCS application. The very high amounts of dissolved P collected from TSP-treated plots were obtained without any increase in surface runoff. Treatment differences in P transfer were maintained for approximately 1 mo after application (data not shown). Cumulative TDP and PP loads in runoff over the whole monitoring period were all enhanced following TSP, LCS, and LDS application, and PP loads were reduced following DSC application (Table 3). However, comparison of volume-adjusted P concentrations indicated that while there was a large release of dissolved P from the TSP and LCS, the proportion of total P transfer occurring in dissolved form following LDS or DSC application was not increased. Once again, the results indicate a poor or negligible release of dissolved P from the sludges, and the differences in P transfer between the two sludges and the control were largely due to differences in runoff volume and PP transfer, and not due to any significant release of dissolved P in the runoff (Table 3). Beneficial reductions in runoff and PP transfer following application of high dry matter sludge have been reported previously (Kladivko and Nelson, 1979; Deizman et al., 1989).

Residual Effects In the period directly after plowing and reseeding in spring 1997, there was no significant (P ⬍ 0.05) treatment effect on either runoff volume or the P content of the runoff. Total dissolved P concentrations were slightly elevated on TSP- and DSC-treated plots (1.9 and 1.8 mg L⫺1, respectively) compared with other plots

187

(0.9–1.3 mg L⫺1) in the first runoff event after plowing, but this was significant only at the 10% level (data not shown). These plots had received the most P and showed higher concentrations of total P in the soil at the end of the experiment (Table 5). In the two subsequent events, runoff P concentrations were similar across all plots. All treatments increased both water-soluble and Olsen-extractable P concentrations in the soil compared with the control, but the increase was much larger on TSP-treated plots (Table 5). There was no difference in water- or Olsen-extractable P between the manure and sludge treatments despite the differences in total P input and in their laboratory-determined potential P availabilities (Table 1). Expressed as a proportion of the increase in soil total P, the amounts of P recovered in Olsen-extractable form were 28, 15, 18, and 9% for TSP, LCS, LDS, and DSC, respectively, and the amounts of P recovered in water-extractable form were 10, 5, 4, and 3% for TSP, LCS, LDS, and DSC, respectively. The results suggest that a large proportion of the water- and sodium bicarbonate-extractable P in LCS had accumulated in a more slowly exchangeable form in the soil. This may be related to the amounts of Ca applied in the different P sources (Sharpley et al., 1998). The lower recovery of P in DSC compared with TSP is consistent with its low solubility in water and sodium bicarbonate.

CONCLUSIONS A comparison of the release of P in surface runoff following application of different P amendments to a dispersive silty clay loam soil has indicated that the risk of P transfer to watercourses from agricultural land amended with liquid and dewatered sewage sludge is less than when amended with either inorganic P fertilizer or liquid cattle manure, due to their lower P solubility in water and/or sodium bicarbonate. The amounts of inorganic P fertilizer used in this experiment were chosen to represent large single-dose applications to meet the P requirements of part (Period 1) or all (other periods) of a crop rotation, and to be comparable with the P rates applied in other amendments. The results show that large fertilizer topdressings to the soil surface represent a significant eutrophication risk. However, fertilizer application rates to meet the P requirements of an individual crop would be lower than N-driven manure or sludge P loading rates, and P release in runoff would be expected to be correspondingly lower than with organic amendments. The differences in P solubility between the organic amendments were not apparent in soil test P analysis. The main effect of the P amendments was to increase dissolved inorganic P concentrations in runoff and this was most pronounced in the first runoff event following application. The extent of the increase was dependent not only on the type of P amendment applied, but also the time and method of application. Substantial release of P in runoff occurred when P amendments were applied to the soil surface, especially where the latter was

188


consolidated by rolling after seedbed preparation and sowing. Under these conditions, application of bulky sludges reduced runoff and particulate P loss compared with the control. The lowest P release in runoff occurred when the P amendments were incorporated into the soil, although this may reduce flexibility in the timing of P applications and increase the risk of N losses by leaching. This practice may also lead to increased erosional P transfers at other sites depending on the degree of surface protection provided and rainfall intensity. ACKNOWLEDGMENTS This experiment formed part of a research project on the environmental impact of sewage sludge application to agricultural land funded by the UK Water Industry Research Limited.

REFERENCES Bushee, E.L., D.R. Edwards, and P.A. Moore, Jr. 1998. Quality of runoff from plots treated with municipal sludge and horse bedding. Trans. ASAE 41:1035–1041. Davies, R.D. 1996. The impact of EU and UK environmental pressures on the future of sludge treatment and disposal. J. Chartered Inst. Water Environ. Manage. 10:65–69. Deizman, M.M., S. Mostaghimi, T.A. Dillaha, and C.D. Heatwole. 1989. Tillage effects on phosphorus losses from sludge-amended soils. J. Soil Water Conserv. 44(3):247–251. Dunnigan, E.P., and R.P. Dick. 1980. Nutrient and coliform losses in runoff from fertilized and sewage sludge-treated soil. J. Environ. Qual. 9:243–250. Edwards, D.R., and T.C. Daniel. 1993. Runoff quality impacts of swine manure applied to fescue plots. Trans. ASAE 36:81–90. Frossard, E., S. Sinaj, and P. Dufour. 1996. Phosphorus in urban sewage sludges as assessed by isotopic exchange. Soil Sci. Soc. Am. J. 60:179–182. Genstat 5 Committee. 1993. Genstat 5 Release 3 reference manual. Clarendon Press, Oxford, UK. Haygarth, P.M., and S.C. Jarvis. 1999. Transfer of phosphorus from agricultural soils. Adv. Agron. 66:195–249. Heathman, G.C., A.N. Sharpley, S.J. Smith, and J.S. Robinson. 1995. Poultry litter application and water quality in Oklahoma. Fert. Res. 40:165–173. Hedley, M.J., J.W.B. Stewart, and B.S. Chauhan. 1982. Changes in inorganic soil phosphorus fractions induced by cultivation practices and by laboratory incubations. Soil Sci. Soc. Am. Proc. 46:970–976. Hinedi, Z.R., A.C. Chang, and R.W.K. Lee. 1989. Characterization of phosphorus in sludge extracts using phosphorus-31 nuclear magnetic resonance spectroscopy. J. Environ. Qual. 18:323–329. Khaleel, R., K.R. Reddy, and M.R. Overcash. 1980. Transport of potential pollutants in runoff water from land areas receiving animal wastes: A review. Water Res. 14:421–436.

Kirkham, M.B. 1982. Agricultural use of phosphorus in sewage sludge. Adv. Agron. 35:129–163. Kladivko, E.J., and D.W. Nelson. 1979. Surface runoff from sludgeamended soils. J. Water Pollut. Control Fed. 51:100–110. Magid, J., and N.E. Nielsen. 1992. Seasonal variation in organic and inorganic phosphorus fractions of temperate-climate sandy soils. Plant Soil 144:155–165. McLeod, R.V., and R.O. Hegg. 1984. Pasture runoff quality from application of inorganic and organic nitrogen sources. J. Environ. Qual. 13:122–126. Melanen, M., A. Jaakkola, M. Melkas, M. Ahtiainen, and J. Matinvesi. 1985. Leaching resulting from land application of sewage sludge and manure. Publ. Water Res. Inst. no. 61. Finland National Board of Waters, Helsinki. Ministry of Agriculture, Fisheries and Food. 1998. Code of good agricultural practice for the protection of water. 2nd ed. MAFF Publ., London. Mostaghimi, S., M.M. Deizman, T.A. Dillaha, C.D. Heatwole, and J.V. Perumprai. 1988. Tillage effects on runoff water quality from sludge-amended soils. Bull. 162. Virginia Water Resour. Res. Center, Blacksburg. Mostaghimi, S., T.M. Younos, and U.S. Tim. 1992. Effects of sludge and chemical fertilizer application on runoff water quality. Water Resour. Bull. 28:545–552. Mueller, D.H., R.C. Wendt, and T.C. Daniel. 1984. Phosphorus losses as affected by tillage and manure application. Soil Sci. Soc. Am. J. 48:901–905. Murphy, J., and J.D. Riley. 1962. A modified single solution for determination of phosphate in natural waters. Anal. Chim. Acta 27:31–36. Olsen, S.R., C.V. Cole, F.S. Watanabe, and L.A. Dean. 1954. Estimation of available phosphorus in soils by extraction with sodium bicarbonate. USDA Circ. 939. USDA, Washington, DC. Rowland, A.P., and P.M. Haygarth. 1997. Determination of total dissolved phosphorus in soil solutions. J. Environ. Qual. 26:410–415. Schreiber, J.D., and L.L. McDowell. 1985. Leaching of nitrogen, phosphorus and organic carbon from wheat straw residues: I. Rainfall intensity. J. Environ. Qual. 14:251–256. Sharpley, A.N. 1997. Rainfall frequency and nitrogen and phosphorus runoff from soil amended with poultry litter. J. Environ. Qual. 26:1127–1132. Sharpley, A.N., J.J. Meisinger, A. Breeuwsma, J.T. Sims, T.C. Daniel, and J.S. Schepers. 1998. Impacts of animal manure management on ground and surface water quality. p. 173–242. In J.L. Hatfield and B.A. Stewart (ed.) Animal waste utilization: Effective use of manure as a soil resource. Ann Arbor Press, Chelsea, MI. Smith, K.A., A.G. Chalmers, B.J. Chambers, and P. Christie. 1998. Organic manure phosphorus accumulation, mobility and management. Soil Use Manage. 14:154–159. Smith, K.A., and B.J. Chambers. 1993. Utilizing the nitrogen content of organic manures on farms—Problems and practical solutions. Soil Use Manage. 9:105–112. Sommers, L.E., D.W. Nelson, and K.J. Yost. 1976. Variable nature of chemical composition of sewage sludge. J. Environ. Qual. 5:303–306.