| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Agriculture and Agri-Food Canada, P.O. Box 3000, Lethbridge, AB, T1J 4B1 Canada
b Lethbridge Community College, 3000 College Drive South, Lethbridge, AB, T1K 1L6 Canada
c Department of Renewable Resources, General Services Building, University of Alberta, Edmonton, AB, T6G 2H1 Canada
d Alberta Agriculture, Food and Rural Development, Lethbridge, AB, T1J 4V6 Canada
* Corresponding author (millerjj{at}agr.gc.ca)
Received for publication October 12, 2005.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
0.05) higher TP, DRP, and NH4–N concentrations, and higher DRP and TN loads under FM than CM after 2 yr of manure application. The TP loads were also higher under FM than CM at the 83 Mg ha–1 rate in 2000, and DRP loads were higher for FM than CM at this high rate when averaged over both years. Application rate had a significant effect on TP and DRP concentrations in runoff. In addition, the slope values of the regressions between TP and DRP in runoff versus application rate were considerably higher for FM in 2000 than for FM in 1999, and CM in both 1999 and 2000. Significant positive relationships were found for TP and DRP in runoff versus soil Kelowna-extractable P and soil water-extractable P for FM and CM in 2000, indicating that interaction of runoff with the soil controlled the release of P. Total P and DRP were the variables most affected by the treatments. Overall, our study found that application of CM rather than FM to cropland may lower certain forms of P and N in surface runoff, but this is dependent on the interaction with year, application rate, or both.
Abbreviations: CM, composted manure • DRP, dissolved reactive phosphorus • FM, fresh manure • KEP, Kelowna-extractable phosphorus • PP, particulate phosphorus • TN, total nitrogen • TP, total phosphorus • WEP, water-extractable phosphorus
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Since composting produces a stable humus-like material due to biochemical changes from microbial action (Larney et al., 2001), we would expect greater short-term release of organic-derived nutrients (e.g., N, P) from FM into surface runoff, and a slower, long-term release of these nutrients from CM into surface runoff. Composting of feedlot cattle manure generally results in a decrease in the carbon to nitrogen ratio, TN, NH4–N, and total C; and an increase in NO3–N, electrical conductivity, available P, and TP (Eghball et al., 1997; Larney et al., 2001). In addition, fine homogeneous compost particles versus heterogeneous fine to large fresh manure particles may result in differential release of nutrients into runoff. The proportion of various P forms in FM and CM, and interaction of manure with the soil, can also affect P release during rainfall (Sharpley and Moyer, 2000). Therefore, manure type might be expected to have an effect on P and N in surface runoff.
Few studies have compared P or N in surface runoff under fresh and composted cattle manure application (Eghball and Gilley, 1999; Qu et al., 1999). The former authors applied beef cattle manure (12–126 Mg ha–1 dry wt.), based on N and P crop requirements, to a silty clay loam soil in Nebraska cropped to sorghum [Sorghum bicolor (L.) Moench] and winter wheat (Triticum aestivum L.) under tilled (disked) and no-till conditions. They measured the P and N concentrations and mass loads in surface runoff generated using a rainfall simulator for dry and wet initial soil moisture conditions. For concentrations of P and N in runoff, there was no or little amendment type effect on dissolved P, bioavailable P, or TN. Particulate P, TP, and NO3–N concentrations tended to be higher under CM than FM, whereas NH4–N tended to be higher under FM than CM. There were no or few amendment type effects on mass loads of P and N in runoff. The exception was for NO3–N loads, where values were generally higher under CM than FM. Overall, Eghball and Gilley (1999) found that concentrations and loads of P and N were dependent on crop residue type, antecedent soil moisture conditions, and the year of the study.
Qu et al. (1999) measured the quality of surface runoff from fresh and composted dairy cattle manure applied to snow in a laboratory experiment. They found that biological oxygen demand (BOD5), soluble C, and TN were significantly higher for FM than CM. They concluded that CM had a lower pollution potential than FM because composting had converted the most degradable and soluble organic compounds in the manure to more stable compounds, which were not as easily released into runoff. We are unaware of any field studies in the Great Plains region of North America that have compared the N and P content of surface runoff under increasing rates of fresh and composted beef cattle manure.
Research on the effect of application rate of cattle manure on surface runoff quality has focused more on comparisons between one rate (manured treatment) of manure and the unamended control (Mueller et al., 1984; Ginting et al., 1998; Bundy et al., 2001; Eghball et al., 2002; Kleinman et al., 2002; Andraski and Bundy, 2003; Kleinman et al., 2004; Little et al., 2005) rather than on the effect of increasing rates of manure (Mathers et al., 1977; Eghball and Gilley, 1999; Kleinman and Sharpley, 2003). Mathers et al. (1977) applied beef cattle manure at 0, 22, and 67 Mg ha–1 (wet wt.) to a clay loam soil in Texas that was cropped to sorghum and furrow-irrigated. They found that runoff quality (NO3, P, Cl) was not affected by manure rate. Eghball and Gilley (1999) reported that P-based manure or compost application (12–18 Mg ha–1) and disked N-based (49–126 Mg ha–1) manure or compost application resulted in dissolved P runoff concentrations of <1 mg L–1 (critical limit), indicating that runoff quality was unaffected by these manure rates. In contrast, dairy cattle manure was applied at six rates from 0 to 150 kg TP ha–1 to two soils (fine- and coarse-loamy) in packed runoff boxes in the laboratory (Kleinman and Sharpley, 2003). They found that runoff P was significantly related to application rate (r2= 0.50–0.98) due to increased concentrations of dissolved reactive phosphorus (DRP) in runoff; and as application rate increased, so did the contribution of DRP to TP.
The objective of this study was to compare the effect of increasing rates of fresh versus composted manure on the P and N content of surface runoff.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The field was divided into 28 plots, each measuring 9 x 20 m, laid out in four blocks of seven plots each in a randomized complete-block design across the slope. Within each block, each plot was treated with one of three rates of FM or CM or remained untreated as a control. Selected chemical properties of the FM and CM applied to our plots are shown in Table 1, and additional details on the feedlot manure have been reported by Miller et al. (2004). Seven treatments were applied, including a control (CON), as well as FM and CM at three rates each (14, 42, 83 Mg ha–1 dry wt.). Beef manure application rates in this area range from approximately 25 to 113 Mg ha–1, with a mean value of 75 Mg ha–1 (Porcupine Corral Cleaning Ltd., personal communication, 2005). Therefore, the highest application rate of 83 Mg ha–1 applied in our study was just slightly higher than the average rate applied by local commercial manure applicators.
|
Rainfall Simulations, Runoff, and Soil Sampling
Rainfall simulations were conducted at one location in each plot before the site was seeded to a cereal crop. A Guelph Rainfall Simulator (Tossell et al., 1987) was used to simulate a rainfall intensity of 103 mm h–1 from a height of 0.80 m on a 1- x 1-m area. Lower rainfall intensities were initially tried on these plots, but runoff could not be generated, and testing found that runoff could only be generated within a reasonable time period at 103 mm h–1. This rainfall intensity was within the range of values (64–145 mm h–1) used by other researchers studying P transport in runoff (Mueller et al., 1984; Eghball and Gilley, 1999). Rainfall simulations were conducted between 8 June to 9 July 1999 and between 8 May to 12 June 2000. The duration of rainfall varied between 15 and 28 min, representing return periods between 18 and >50 yr. Demineralized water was used for the rainfall simulations. The site was seeded to barley after the rainfall simulations in 1999.
For all rainfall simulations, 19 consecutive runoff samples of approximately 700 mL each were collected after runoff commenced. Every odd-numbered sample of the 19 samples was used for quantifying NO3–N and NH4–N in 1999 and 2000, and statistical analyses conducted on the mean value of these subsamples. For TP, particulate phosphorus (PP), DRP, and total suspended solids (TSS), a subsample was taken from each of the 19 samples taken during runoff, the composite sample mixed, and a subsample taken for chemical analyses. Water samples were also taken from the rainfall simulator tank during each simulation and analyzed for chemical properties. Water samples were stored in plastic bottles, preserved with acid if required (American Public Health Association, 1995), and then stored at –20°C until analyzed.
Soil samples were taken from the 0- to 15-cm depth within each plot before manure application in the fall (October) of 1998 to determine background levels of electrical conductivity (EC), pH, C to N ratio, total C, total phosphorus (TP), Kelowna-extractable P, total nitrogen (TN), NH4–N, and NO3–N before amendment application. Three surface (0–5 cm) soil samples were collected adjacent to the rainfall simulator area before rainfall and one composite sample formed to determine water- and Kelowna-extractable P. The inorganic C content (mainly calcite, dolomite, and magnesian calcites) of surface soils in this area is less than 0.20 g 100 g–1 (Janzen, 1987). Mean plant-extractable Fe is approximately 98 mg kg–1 in soils of the Dark Brown soil zone in Alberta (Alberta Agriculture, Food and Rural Development, 2006). Extractable Al (1:10 extracts, 0.1 M BaCl2) in soils of southern Alberta ranges from 0 to 35.1 mg kg–1, with a mean value of 5.4 mg kg–1 (G. Dinwoodie, personal communication, 2006).
Chemical Analyses
Ammonium N in runoff water was analyzed using the automated salicylate method (Rhine et al., 1998), and NO3–N using the automated hydrazine reduction method (Kempers and Luft, 1988). Dissolved reactive P was determined using the automated ascorbic acid method on the autoanalyzer (Technicon Industrial Systems, 1974). Total N was determined using the persulfate digestion method (Method 4500-NorgD; American Public Health Association, 1995), followed by analysis of NO3–N using the method outlined above. Total P was determined using the persulfate digestion method (Method 4500-P B; American Public Health Association, 1995), followed by analysis of ortho-P using the method described above for DRP. Particulate P was determined as the difference between TP and DRP. Total suspended solids were determined by filtering a 100-mL water sample, oven drying the filter, and then weighing the filter (American Public Health Association, 1995).
Soil samples were air-dried and ground to pass through a 2-mm sieve. Electrical conductivity and pH were determined on 1:2 soil to water extracts. Total C and N were determined using the Dumas automated combustion technique (McGill and Figueiredo, 1993) using a CNS analyzer (Carla Erba, Milan, Italy). Total P was determined by a wet-oxidation procedure (Parkinson and Allen, 1975), and ortho-P was analyzed on the autoanalyzer as described above. Water-extractable phosphorus (WEP) in the soil was determined by mixing 1 g of soil with 25 mL of deionized water, shaking for 1 h, centrifuging for 5 min at 266 m s–1, and filtering (Pote et al., 1996). Kelowna-available phosphorus (KEP) in the soil (van Lierop, 1988) was determined on a 1:10 (2.5 g soil to 25 mL extract solution) extract after shaking at low speed for 15 min and filtering (20–25 µm). The composition of the Kelowna extract was ammonium fluoride (0.015 M NH4F) and glacial acetic acid (0.25 M CH3COOH). All soil extracts were analyzed for ortho-P as described above for DRP. Quality control for water and soil analyses included blanks, spikes, and duplicates.
Data and Statistical Analyses
Mean values of the chemical variables were used in the statistical analysis. Average runoff mass loads (g ha–1 h–1) were calculated based on the mean concentration, total runoff volume, time, and area of runoff (1 m2). The chemical data were analyzed using SAS (SAS Institute, 1989) and a mixed model analysis (Littell et al., 1998). Different covariance model types were tested to obtain the best covariance structure for the mixed model. The seven individual treatments were compared for statistical differences in least-squares mean (LSM) values. Mean values of N and P in the control versus amended plots were tested for significant differences using an ESTIMATE statement. The seven individual treatments per replicate were the following: control, FM-14 Mg ha–1, FM-42, FM-83, CM-14, CM-42, and CM-83. The main treatment effects of manure type (FM, CM), rate (14, 42, 83 Mg ha–1), and year (1999, 2000) were assessed for N and P variables. Treatment effects were considered significant at the P 0.05 level. Linear and quadratic regression was used to determine relationships between chemical variables in runoff versus N and P in the soil.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The chemical properties of the soil (0- to 15-cm depth) in 1998 before any manure applications are shown in Table 1. The CV values of the soil properties ranged from 1 to 30%, indicating relatively uniform chemical properties in the plots before the experiment. Olson (2004) reported no significant differences in gravimetric soil water content among the seven treatments in 1999 and 2000. Mean soil (0–15 cm) test P (KEP) before amendment application was 44.7 mg kg–1 (Table 1), which is considered high for Alberta soils (Alberta Agriculture, Food and Rural Development, 1997). Available N (NO3–N + NH4–N) in the soil (0–15 cm) before amendment application was 20.6 mg kg–1 (Table 1), and is considered to be a medium level of available N in these soils (R. McKenzie, personal communication, 2006).
Precipitation, Irrigation, and Runoff
Approximately 2.4 times more rainfall and irrigation water fell on these plots during the month before rainfall simulations in 1999 than in 2000, but similar amounts of water fell on these plots during the rainfall simulations in 1999 and 2000 (Table 2). Similar precipitation and irrigation in both years during the time periods of rainfall simulations was consistent with no significant difference in antecedent, gravimetric soil water content of the surface (0–3 cm) soil among the seven treatments (Olson, 2004). Total cumulative runoff depths for all seven treatments ranged from 11.4 to 12.7 mm, with no significant differences among treatments (data not shown). Runoff depths were also similar for FM (11.7 mm) and CM (11.9 mm) when averaged over both years. The average runoff coefficient (rate of runoff to rainfall rate applied) at the end of runoff was 0.47 in both 1999 and 2000, with no significant differences among treatments (Olson et al., 2005).
|
|
|
|
|
Manure application rate had a significant effect on TP concentration in runoff, where mean values were higher at the 83 Mg ha–1 than 14 Mg ha–1 rate (Table 3). When the control treatment was included as the 0 Mg ha–1 rate, linear relationships with R2 values of >0.84 were observed between manure application rate and TP runoff concentrations (Fig. 3a). The slope of the line was greatest for FM in 2000, where TP increased 0.064 mg L–1 for every 1 Mg ha–1 increase in manure rate. Eghball and Gilley (1999) found that TP concentrations in runoff under FM and CM increased from 156 to 192% when application rate increased from 0 to 126 Mg ha–1 dry weight (34–99 kg TP ha–1), and that TP loads increased from 227 to 370%. Kleinman and Sharpley (2003) reported significant (R2 = 0.50–0.95) linear relationships between TP concentrations in runoff and TP application rate (0–150 kg TP ha–1) for two dairy manured soils. They reported regression slopes ranging from 0.001 to 0.004 mg L–1 for every 1 kg of TP applied per hectare. The range in annual total P application of FM or CM to our plots (56–490 kg TP ha–1) was considerably higher than the TP application rate used by other researchers (Eghball and Gilley, 1999; Kleinman and Sharpley, 2003).
|
Particulate Phosphorus
There were no significant differences among the seven treatments for particulate phosphorus (PP) concentrations and loads when averaged over the two years (data not shown). There was a significant type x year interaction on PP concentration (Table 3). Values were higher for CM than FM in 1999, but were similar in 2000 (Fig. 1b). There were no main or interaction treatment effects on PP loads (Table 4). Significant positive linear relationships were observed between PP and total suspended solids (TSS) for FM and CM in 2000 (Fig. 4a), indicating that suspended soil particles controlled runoff PP. The relationship was much stronger for CM (r2 = 0.90) than FM (r2 = 0.45), and the increase in PP per unit increase in TSS was higher for CM (slope = 0.68) than for FM (slope = 0.44) (Fig. 4a). We speculate that this may have been due to the finer and more uniform particle size of CM, and this may have resulted in greater release of particles from CM. Eghball and Gilley (1999) reported that PP concentrations were higher under CM than FM for dry and wet soil moisture conditions under sorghum residue in 1996, but values were similar for dry and wet conditions under wheat residue in 1997. Similar to our study, they also found no manure type effect on PP loads in runoff. Manure application rate had no significant effect on PP concentrations and loads in runoff (Tables 3 and 4).
|
There were significant relationships between DRP runoff concentrations versus soil Kelowna-extractable P and soil water-extractable P for FM and CM in 2000 (Fig. 4b and 4c); and between DRP runoff loads versus KEP and WEP in soil for FM in 2000 (data not shown). This suggests the total amount of P applied and subsequent soil test P are important for the DRP concentrations and loads. The slope values for the regressions of DRP concentration in runoff versus soil KEP were higher for FM (0.0096) than CM (0.00068) in 2000 (Fig. 4b); and the slope values of soil WEP versus DRP concentration were also higher for FM (0.045) than CM (0.0067) in 2000 (Fig. 4c). Greater release of P to runoff from soil under FM than CM may be related to the greater proportion of water-extractable inorganic P in FM than CM (Sharpley and Moyer, 2000).
Extraction coefficients or slope values for DRP concentrations in runoff versus soil test P are dependent on the soil extractant used (Fang et al., 2002; Wright et al., 2003). In our study, the Kelowna solution extracted considerably more P than water (Fig. 4b and 4c), which was consistent with the findings of others (Wright et al., 2003). Wright et al. (2003) reported that regressions of soil KEP versus flow-weighted mean concentrations of dissolved inorganic P in runoff for 25 soil types in Alberta (laboratory rainfall simulations) had equilibrium slope values of 0.0023, and that regressions of soil WEP (1:5 extracts) versus runoff dissolved inorganic P had equilibrium slope values of 0.0711. Pote et al. (1996) reported that regressions of soil WEP (1:25 extracts) versus DRP had slope values of 0.012. Extraction coefficients of 0.04218 have been reported between runoff DRP and water-extractable P in calcareous soils (0–15 cm) in Texas (Torbert et al., 2002). Overall, our extraction coefficients for soil KEP or WEP versus runoff DRP for FM were generally similar to extraction coefficients reported by these other studies, but our slope values for CM were generally lower than the values cited by these other studies.
Manure rate had a significant effect on DRP concentration in runoff, where values were higher at the 83 Mg ha–1 than the 14 Mg ha–1 rate (Table 3). When the control treatment was included as the 0 Mg ha–1 rate, linear relationships with R2 values of >0.85 were found between DRP in runoff and manure rate for CM in 1999, as well as FM and CM in 2000, whereas a quadratic relationship was found for FM in 1999 (Fig. 3b). Similar to TP, the slope was highest for FM in 2000, where it increased 0.058 mg L–1 for every unit increase in manure rate. Eghball and Gilley (1999) reported that DRP concentrations and loads in runoff increased between 8 and 15 times when application rate was increased from 0 to 126 Mg ha–1. Kleinman and Sharpley (2003) found a positive linear relationship between runoff DRP versus TP application rate for dairy manure, with slope values ranging from 0.002 to 0.005 mg L–1 for every 1 kg P ha–1 applied.
Ratio of Dissolved Reactive Phosphorus to Total Phosphorus
The DRP to TP ratio for all CON and treatment plots in both years ranged from 0.042 to 0.88, with a mean ratio of 0.25 (data not shown). There was no significant treatment effect on DRP to TP ratio in 1999 (data not shown). In 2000, the ratio was significantly higher for the FM-83 treatment (0.73) than all other treatments (0.12–0.45) except for CM-83 (0.52). The DRP to TP ratio was significantly higher for FM (0.33) than CM (0.23) when averaged over both years. Manure application rate significantly increased the DRP to TP ratio, and the ratio was higher in 2000 (0.38) than in 1999 (0.17). Kleinman and Sharpley (2003) also reported that the contribution of DRP to TP increased with manure rate.
Total Nitrogen
There were no significant differences among the seven treatments for TN concentration when averaged over both years (data not shown). There were no main or interaction treatment effects on TN concentration (Table 3). The TN loads were similar among the seven treatments in 1999, and were highest for FM-83 in 2000 (data not shown). There was a significant type x year effect on TN loads (Table 4), where values were higher for FM than CM in 2000 (Fig. 2e). Our finding of no or few significant manure type effects on total runoff N agreed with some studies but not with others. Eghball and Gilley (1999) reported similar TN concentrations and loads in runoff under FM and CM application (12–126 Mg ha–1 dry wt.) for dry and wet soil moisture conditions under wheat residue. In contrast, Qu et al. (1999) reported significantly higher TN concentrations in runoff under FM (67.5 Mg ha–1 wet wt.) than CM (24.4 Mg ha–1 wet wt.), but their amendments were applied to snow and there was no soil for the amendments to interact with. The latter authors proposed that greater TN in runoff under FM than CM was due to the composting process converting easily degradable organic N compounds into more stable N compounds. DeLuca and DeLuca (1997) reported that although the N content of CM is generally lower than that of FM, more of the N in CM exists in stable organic forms, resulting in lower N losses after field application.
The TN concentrations (1.2 mg L–1) in runoff from the unamended control plots slightly exceeded the water quality guideline of 1.0 mg L–1 TN for protection of aquatic life (Alberta Environment, 1999). Mean runoff concentrations of TN for FM (1.9 mg L–1) and CM (1.1 mg L–1) also slightly exceeded this water quality guideline (Table 3).
Nitrate Nitrogen and Ammonium Nitrogen
There were no significant differences among the seven treatments for NO3–N concentrations or loads when averaged over both years (data not shown). There were no significant main or interaction effects on either concentrations (Table 3) or loads (Table 4) of NO3–N. In comparison, Eghball and Gilley (1999) found that NO3–N concentrations and loads in runoff were generally higher under CM than FM.
Ammonium N concentrations were highest for the CM-42 treatment in 1999, and values were relatively similar in 2000 (data not shown). Average mass NH4–N loads were highest for CM-83 in 1999, and were highest for FM-83 in 2000 (data not shown). There was a significant type x year effect on NH4–N concentration (Table 3), where values were significantly higher for FM than CM in 2000 (Fig. 1d). There was also a significant type x year effect on average mass loads of NH4–N (Table 4), where values were higher for CM than FM in 1999 (Fig. 2f). There were significant relationships between runoff NH4–N concentrations (r2 = 0.71) and loads (r2 = 0.70) versus soil NH4–N for FM in 2000 (data not shown). Eghball and Gilley (1999) reported higher NH4–N concentrations for FM than CM under sorghum residues, but found no difference under wheat residue. They also found no manure type effect on NH4–N loads in runoff. Similar to TN, we observed no application rate effects on NO3–N or NH4–N concentrations or loads in our study. Eghball and Gilley (1999) also reported dramatically higher NO3–N, NH4–N, and TN in runoff under FM and CM in Nebraska than found in our study and by Little et al. (2005) in southern Alberta.
The NO3–N concentration of the runoff from the unamended control plots (0.12 mg L–1) was below the recommended federal guideline of 10 mg L–1 for protection of community water supplies for drinking water for humans (Canadian Council of Ministers of the Environment, 2002). Nitrate N concentrations in runoff from FM and CM plots (Table 3) were also well below this water quality guideline. However, mean NH4–N concentrations in runoff from the unamended control plots in 1999 (0.068 mg L–1) and 2000 (0.060 mg L–1) and mean concentrations from the FM and CM plots (Table 3) exceeded the recommended maximum guideline for total ammonia N in water (0.0192 mg L–1 NH3–N for mean temperature of 20°C and mean pH of 9.38 in runoff water) (Canadian Council of Ministers of the Environment, 2002).
Treatment Effects on Phosphorus and Nitrogen
When comparing concentrations of the six P and N variables for the six manured treatment plots, our results indicated trends for highest concentrations of TP, DRP, and NH4–N for the FM-83 treatment in 2000, and highest NH4–N concentrations for the CM-42 treatment in 1999. Average mass loads for TP, DRP, TN, and NH4–N were also higher for the FM-83 treatment in 2000, and NH4–N load was highest for the CM-83 treatment in 1999.
For main treatment effects on concentrations and average mass loads of N and P variables in runoff, year affected the most P and N variables, followed by type and rate. For interaction effects on concentrations and loads of N and P in runoff, type x year affected the most P and N variables. Total P and DRP concentrations in runoff were the chemical concentration variables most affected by the treatments.
Although composting of cattle manure generally results in an increase in TP and available P (Eghball et al., 1997; Larney et al., 2001; Sharpley and Moyer, 2000), our study generally found higher TP and DRP concentrations and loads in runoff under FM than CM after the second year of application or at the highest application rate. Significant positive relationships between DRP in runoff versus soil KEP or WEP (Fig. 4b and 4c) indicated that the interaction of the organic amendments with the soil, and subsequent interaction of soil–amendment mixture with runoff, was the controlling factor for DRP in runoff. The average effective depth of soil that interacts with rainfall and overland flow in releasing soil chemicals such as P to runoff ranges between 20 and 30 mm (Ahuja et al., 1981); but this depth may vary depending on rainfall intensity, slope, and management (Sharpley, 1985).
Sharpley and Moyer (2000) noted that although the water-extractable P concentration of FM or CM may be used to estimate P release to water, the presence of soil may mask differences in P in the original amendments and the soil may dominate release of P into water such as runoff. Since the FM used in our study had higher TP, KEP, TN, NH4–N, and NO3–N than the CM (Table 1), we would expect the same trend in surface runoff. However, this trend did not occur, and was attributed to the masking effect of the soil. High levels of soil test P and a medium level of soil available N before amendment application, in conjunction with no significant differences between average values of N and P in the control versus amended treatments (Tables 3 and 4), indicated that high to medium levels of soil test P and N in the initial soil may have enhanced this masking effect, and contributed to a lack of amendment treatment response on this soil. However, treatment responses still occurred for the treatment factors, particularly in the second year, and at the highest application rate.
Higher DRP and TP in our runoff under FM than CM may also be due to the greater percentage of TP as water-extractable inorganic and organic P in FM than CM (Sharpley and Moyer, 2000). They reported that 51% of TP in fresh dairy cattle manure was in water-extractable inorganic P form compared to only 15% for composted dairy cattle manure. Similarly, 12% of TP in fresh dairy manure was in water-extractable organic P form compared to only 1% for composted dairy manure. They also reported that under five consecutive simulated rainfalls, greater amounts of TP were leached from fresh dairy manure (58%) than from composted dairy manure (15%). Their findings are consistent with our trend of higher TP in runoff under FM than CM in 2000 (Fig. 3a).
Phosphorus in runoff is related to P in the soil solution, and the concentration of solution P in soil is controlled by the processes of precipitation–dissolution, adsorption–desorption, and mineralization–immobilization (Frossard et al., 2000). In high pH soils such as the calcareous soils of our study, exchangeable Ca and Ca compounds rather than Fe and Al compounds control soil solution P (Havlin et al., 1999). Sharpley et al. (1984) reported that increasing application of cattle manure to a clay loam soil decreased the P sorption index of surface soil, indicating that adsorption of P would be decreased and more solution P available for runoff. They also found that cattle manure increased the proportion of inorganic soil P in the loosely bound (labile P) and non-occluded fractions (P associated with Fe and Al) compared with the unamended control, and that the increases were greater for the non-occluded fractions. However, the issue of whether desorption or dissolution is the main processes controlling solution P cannot usually be resolved; as the saturation of sorbing surfaces (desorption), and the solubility product of the least soluble P compound (dissolution), determines the P supply at equilibrium (Thiessen and Moir, 1993). Sharpley et al. (1981) examined the kinetics of P release to runoff by desorption, and found that soluble P in runoff was dependent on the initial P content of the soil, mass of soil in the interacting zone, water to soil ratio, and total rainfall. Havlin et al. (1999) reported that adsorption likely dominates P retention at low solution P concentrations, whereas precipitation likely dominates when solution P concentration exceeds the solubility product of the mineral.
The finding of higher NH4–N concentrations in runoff from FM than CM plots after 2 yr of manure application, and a significant positive relationship between runoff NH4–N concentrations and loads and soil NH4–N, suggested that runoff NH4–N was also controlled by interaction with the soil rather than by differences in the original amendment materials.
There were no significant relationships between any of the P and N dependent variables and the depth of runoff (data not shown). Runoff depths were similar for FM (11.7 mm) and CM (11.9 mm) when averaged over both years, as was the total cumulative depth of runoff. Olson (2004) compared the hydrologic response of the same treatments in our study, and found few or no significant differences in initial abstraction, average runoff rates, and runoff coefficients. Our finding of a potential for greater release of P and N into surface runoff from cropland amended with fresh than composted beef manure in the second year or at the highest application rate (or both), indicated that the water quality differences were more likely related to chemical interactions with the soil rather than to differences in hydrology. Our evidence for this was significant relationships between runoff P and N versus soil P and N, and no significant relationships between runoff P and N versus hydrological variables.
Two consecutive years of FM application at 83 Mg ha–1 increased TP concentration in runoff to as high as 6.2 mg L–1, whereas 2 yr of CM application at this rate increased total runoff P to only 1.9 mg L–1. Similarly, two consecutive years of FM application at 83 Mg ha–1 increased DRP concentrations in runoff to as high as 4.7 mg L–1, whereas 2 yr of CM application at this rate increased DRP concentrations to only 1.0 mg L–1. In addition, FM application at 83 Mg ha–1 increased TN concentration in runoff to as high as 2.98 mg L–1 (average over 2 yr), whereas CM application at this rate increased TN to only 1.21 mg L–1. These results suggest CM has the potential to lower some forms of P and N in runoff compared to FM, but this is dependent on year, rate, or both of these factors.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The TP, TN, and NH4–N concentrations in runoff from control, FM, and CM plots, and DRP concentrations in runoff from FM plots, exceeded the maximum recommended water quality guidelines. In contrast, NO3–N concentrations in runoff from all plots, and DRP concentrations in control and CM plots, did not exceed the guidelines. Application of CM rather than FM to cropland may be a potential management tool to control P and N in surface runoff, but may be dependent on the interaction with year, application rate, or both of these factors. Further research is required to examine the longer-term effects of FM and CM on P and N in runoff.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. A. Volf, G. R. Ontkean, D. R. Bennett, D. S. Chanasyk, and J. J. Miller Phosphorus Losses in Simulated Rainfall Runoff from Manured Soils of Alberta J. Environ. Qual., April 5, 2007; 36(3): 730 - 741. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
