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


     

Published online 26 April 2006
Published in J Environ Qual 35:928-937 (2006)
DOI: 10.2134/jeq2005.0409
© 2006 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zvomuya, F.
Right arrow Articles by Olson, B. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zvomuya, F.
Right arrow Articles by Olson, B. M.
Agricola
Right arrow Articles by Zvomuya, F.
Right arrow Articles by Olson, B. M.
Related Collections
Right arrow Soil Fertility and Productivity
Right arrow Animal Waste
Right arrow Nutrient Management
Right arrow Phosphorus
Right arrow Soil Chemistry

TECHNICAL REPORTS

Waste Management

Predicting Phosphorus Availability from Soil-Applied Composted and Non-Composted Cattle Feedlot Manure

Francis Zvomuyaa, Bobbi L. Helgasonb, Francis J. Larneya,*, H. Henry Janzena, Olalekan O. Akinremic and Barry M. Olsond

a Agriculture and Agri-Food Canada, Lethbridge Research Centre, 5403 1st Avenue South, Lethbridge, AB, Canada T1J 4B1
b Land Resource Unit, Agriculture and Agri-Food Canada, 51 Campus Drive, Saskatoon, SK, Canada S7N 5A8
c Department of Soil Science, University of Manitoba, Winnipeg, MB, Canada R3T 2N2
d Alberta Agriculture, Food and Rural Development, 5401 1st Avenue South, Lethbridge, AB, Canada T1J 4V6

* Corresponding author (larney{at}agr.gc.ca)

Received for publication October 26, 2005.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Prediction of phosphorus (P) availability from soil-applied composts and manure is important for agronomic and environmental reasons. This study utilized chemical properties of eight composted and two non-composted beef cattle (Bos taurus) manures to predict cumulative phosphorus uptake (CPU) during a 363-d controlled environment chamber bioassay. Ten growth cycles of canola (Brassica napus L.) were raised in pots containing 2 kg of a Dark Brown Chernozemic clay loam soil (fine-loamy, mixed, Typic Haploboroll) mixed with 0.04 kg of the amendments. Inorganic P fertilizer (KH2PO4) and an unamended control were included for comparison. All treatments received a nutrient solution containing an adequate supply of all essential nutrients, except P, which was supplied by the amendments. Cumulative P uptake was similar for composted (74 mg kg–1 soil) and non-composted manures (60 mg kg–1 soil) and for the latter and the fertilizer (40 mg kg–1 soil). However, the CPU was significantly higher for organic amendments than the control (24 mg kg–1 soil) and for composted manure than the fertilizer. Apparent phosphorus recovery (APR) from composted manure (24%) was significantly lower than that from non-composted manure (33%), but there was no significant difference in APR between the organic amendments and the fertilizer (27%). Partial least squares (PLS) regression indicated that only two parameters [total water-extractable phosphorus (TPH2O) and total phosphorus (TP) concentration of amendments] were adequate to model amendment-derived cumulative phosphorus uptake (ACPU), explaining 81% of the variation in ACPU. These results suggest that P availability from soil-applied composted and non-composted manures can be adequately predicted from a few simple amendment chemical measurements. Accurate prediction of P availability and plant P recovery may help tailor manure and compost applications to plant needs and minimize the buildup of bioavailable P, which can contribute to eutrophication of sensitive aquatic systems.

Abbreviations: ACPU, amendment-derived cumulative phosphorus uptake • APR, apparent phosphorus recovery • CPU, cumulative phosphorus uptake • MKP, modified Kelowna-extractable phosphorus • OLS, ordinary least squares • OM, organic matter • Pi, inorganic phosphorus • PiH2O, total water-extractable inorganic phosphorus • PiHCl, total HCl-extractable inorganic phosphorus • PiHCO3, total NaHCO3–extractable inorganic phosphorus • PiNaOH, total NaOH-extractable inorganic phosphorus • PLS, partial least squares • Po, organic phosphorus • PoH2O, total water-extractable organic phosphorus • PoHCl, total HCl-extractable organic phosphorus • PoHCO3, total NaHCO3–extractable organic phosphorus • PoNaOH, total NaOH-extractable organic phosphorus • Pres, residual phosphorus • PRESS, predicted residual sum of squares • SBM, straw-bedded manure • TC, total carbon • TP, total phosphorus • TPH2O, total water-extractable phosphorus • TPHCl, total HCl-extractable phosphorus • TPHCO3, total NaHCO3–extractable phosphorus • TPi, total inorganic phosphorus • TPNaOH, total NaOH-extractable phosphorus • TPo, total organic phosphorus • VIP, variable importance on prediction • WBM, wood chip–bedded manure


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
LAND APPLICATION of animal manures is coming under increasing scrutiny because of growing public concern over the long-term effects on soil P accumulation and potential contribution to surface water eutrophication (Sims et al., 2000). The spatial separation of crop and livestock production systems and the high cost of transporting manure have resulted in excessive land applications of manure in agricultural fields near confined feeding operations (e.g., beef cattle feedlots) (Kellogg et al., 2000). Composting, which is gaining increased acceptance in western Canada's cattle feedlot industry, is one strategy to reduce total manure volume and mass. Composting allows transportation of manure over greater distances from areas of nutrient surplus (in the vicinity of feedlots) to areas of deficiency (Larney et al., 2000). Composting also improves the market value of feedlot manure as a soil amendment and fertilizer (DeLuca and DeLuca, 1997) and eliminates pathogens (Larney et al., 2003b) and weed seeds (Larney and Blackshaw, 2003) typically found in manure.

There is growing interest in the use of organic amendments for reclamation of degraded soils. In Alberta, Canada alone, an estimated 40000 ha of land are disturbed by industrial activity (including oil and natural gas exploration, pipelines, gravel pits, and coal mines) annually (Janz, 1999). Recent research in southern Alberta has demonstrated the beneficial effects of manure and compost in restoring crop productivity to degraded soils (Larney et al., 2003a).

Phosphorus content of non-composted and composted cattle manure can vary widely due to the influence of type of feeding operation, water content, type and amount of bedding, animal age, feed and supplements used, and duration of storage (Barnett, 1994; Dou et al., 2000). Composting results in higher P concentrations because P is conserved while CO2 and NH3 losses result in 30 to 50% reduction in mass of C and N (DeLuca and DeLuca, 1997).

Although P-based manure loading rates and long-term applications are based primarily on total phosphorus (TP) content, it is the inorganic phosphorus (Pi) forms that directly determine plant uptake and runoff losses. Dissolved Pi is readily available for uptake by algae and aquatic plants and therefore presents an immediate risk to receiving water bodies (Dou et al., 2000). The organic phosphorus (Po) fraction becomes available for plant uptake after it is mineralized (Richardson et al., 2000). As much as 50% of the TP in manure is in the organic form (Barnett, 1994). However, the degree and extent of P mineralization and availability following soil application of composts and manures is poorly understood. Quantification of P availability once the amendments are soil-applied may lead to better manure P management and minimize the adverse environmental effects associated with manure application on cropland.

While the reduction in C (Eghball, 2002; Helgason et al., 2005) and N (Eghball, 2000) mineralization rates following soil application of composted versus fresh manure is well documented, only recently has there been increasing focus on the role of composting on P mineralization rate and availability to plants (Dao et al., 2001; Sikora and Enkiri, 2005). Similar water-extractable phosphorus (WEP) and Mehlich 3–extractable P concentrations have been measured in composted and non-composted poultry manure (Dao et al., 2001; Sikora and Enkiri, 2005). Eghball et al. (2002) reported slightly higher P availability for non-composted (85%) compared with composted beef feedlot manure (73%), based on the soil test P changes and plant P uptake one year after application.

Organic P mineralization results in the accumulation of plant-available P in soils, and this increases the potential for P loss and degradation of aquatic systems. Therefore, regardless of the condition (composted or non-composted) of manure when soil-applied, there is a need to develop practical methods of predicting plant P uptake from the amendments to better manage P addition for reduced pollution of sensitive ecosystems. Phosphorus availability from soil-applied manures and composts can be measured directly using crop uptake in pot and field experiments (Eghball et al., 2002) or indirectly using laboratory incubation of the amendments in soil (Eghball et al., 2005). However, these field and laboratory studies have limited practicability because they are time consuming and laborious.

There is evidence from recent studies that N and P indices obtained by rapid fractionation methods can be successfully used as indicators of plant-available nutrients in organic amendments. For example, Velthof et al. (1998) demonstrated in a pot experiment that plant availability of P from composts, manures, and plant materials could be predicted from iron oxide–extractable P.

While dependence of P availability on such factors as compost maturity (Henry and Harrison, 1996), competition between orthophosphate and organic anions for sorption sites (Sharpley and Sisak, 1997), sorption kinetics, and environmental conditions is well documented, it is likely that some of the sequentially extracted P fractions may adequately predict P availability of soil-applied compost. Sequential P extraction has long been widely used to partition extractable soil P into various functional inorganic and organic fractions (Chang and Jackson, 1957; Hedley et al., 1982). Recently, the fractionation procedure developed by Hedley et al. (1982) has been adapted and modified for sequential extraction of organic amendment P (Ajiboye et al., 2004; Dou et al., 2000; Sharpley and Moyer, 2000).

The overall objective of this study was to develop a quantitative model using P fractions and other chemical parameters to predict P availability for plant uptake from soil-applied composted and non-composted beef cattle feedlot manure.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Composted and Non-Composted Beef Cattle Manures
Sampling
Eight composted feedlot manures from southern Alberta were collected during the curing phase for use in this study. The composts were all generated in open-air windrows that were turned five to seven times during the thermophilic phase using commercially available composting equipment. Active thermophilic composting lasted 3 to 4 mo and was followed by a further 3- to 4-mo curing phase. In Alberta, the timing and frequency of turning during composting is not standardized. All composted materials contained straw bedding, with the exception of one compost (Compost 4), which contained wood-chip bedding. The composts were typical of commercial feedlot manure composts available for land application in southern Alberta. Two non-composted feedlot manure samples (from different sources than the manures from which the composts were made)—one straw-bedded manure (SBM) and one wood chip–bedded manure (WBM)—were included for comparison.

Chemical Analysis
Available N concentration (NO3–N + NO2–N + NH4–N) was determined on wet amendment samples (10 g) using an AutoAnalyzer II (Technicon, Tarrytown, NY) following extraction with 2 M KCl (200 mL). Amendment subsamples were oven-dried (60°C for 5 d) and fine-ground (<0.15 mm) for total C and N analysis by dry combustion using an automated CNS analyzer (Carlo Erba, Milan, Italy). It should be noted that oven-drying results in NH3 volatilization and, therefore, some underestimation of total N content (Wood and Hall, 1991; Sistani et al., 2001). Nevertheless, our total N results were not adjusted since the main focus of the study was P availability. Amendment ash content was determined by combustion at 650°C for 24 h. Organic matter content (%) was estimated by subtracting ash content (%) from 100.

Total P concentration was determined after digestion of finely ground amendment samples with concentrated H2SO4, H2O2, Li2SO4, and Se powder (Parkinson and Allen, 1975). Phosphorus concentration in the digests was measured colorimetrically by the ammonium molybdate-ascorbic acid method (Murphy and Riley, 1962) at a wavelength of 882 nm using an AutoAnalyzer II SC Colorimeter (Technicon).

Phosphorus Fractionation
Fractionation of composted and non-composted manure P was performed using the sequential extraction procedure described by Ajiboye et al. (2004). The method is based on the use of progressively more aggressive extractants in the order: deionized water, 0.5 M NaHCO3 (pH 8.5), 0.1 M NaOH, and 1 M HCl. Water- (PH2O) and 0.5 M NaHCO3–extractable phosphorus (PHCO3) fractions are considered labile or bioavailable, while 0.1 M NaOH–extractable phosphorus (PNaOH) is viewed as the moderately labile fraction consisting of Pi chemisorbed to Fe and Al compounds and Po associated with humic compounds (Cross and Schlesinger, 1995). The 1 M HCl–extractable phosphorus (PHCl) fraction may be considered non-labile or moderately recalcitrant, and consists of Ca-bound (apatite-associated) Pi. Residual total phosphorus (Pres) was assumed to be highly recalcitrant Po.

Triplicate 0.3-g oven-dry composted or non-composted manure samples were weighed into 50-mL centrifuge tubes and sequentially extracted with 30 mL of deionized water, 0.5 M NaHCO3 (pH 8.5), 0.1 M NaOH, and 1 M HCl, as described by Ajiboye et al. (2004). After 16 h shaking on an end-to-end shaker at 150 excursions per min, the suspensions were centrifuged at 10000 rpm for 15 min, followed by vacuum-filtration using 0.45-µm cellulose membranes. The extracts were analyzed colorimetrically (Murphy and Riley, 1962) on an Ultrospec 3100 pro UV/visible spectrophotometer (Biochrom, Cambridge, UK) at 882 nm without digestion for Pi determination or following digestion with H2SO4–H2O2 for TP determination. Total P determination was also performed on subsamples of the initial material and on the residue (Pres) remaining after sequential extraction. Organic P in each extract was estimated as the difference between TP and Pi. All extract Pi values were summed to give total Pi, while total Po was estimated as the sum of all extract Po values plus the residual P (Ajiboye et al., 2004). Evidence from recent research suggests that much of the Pres in the Hedley fractionation is organic (Levy and Schlesinger, 1999). Despite the widespread use of the terms inorganic P and organic P, a few limitations associated with Pi and Po determination as described above are noteworthy. Although the inorganic P fraction is primarily orthophosphate, it can include acid-labile organic and condensed P compounds (Dick and Tabatabai, 1977). Similarly, the organic P fraction may also include inorganic polyphosphates (Shand et al., 2000).

Bioassay
Soil Properties
A Dark Brown Chernozemic clay loam soil (fine-loamy, mixed, Typic Haploboroll) was air dried and sieved to <2 mm. Triplicate 10-g samples of the soil were extracted with 200 mL of 2 M KCl and analyzed for available N (NO3–N + NO2–N + NH4–N) concentration using an AutoAnalyzer II (Technicon). Available P concentration was determined by the modified Kelowna method in which 5 g air-dry soil were extracted with 50 mL of a solution containing 0.015 M NH4F, 1.0 M NH4OAc, and 0.5 M HOAc (Ashworth and Mrazek, 1995). Phosphorus (i.e., modified Kelowna-extractable P, MKP) concentration was measured using a 305D Digital Detector (Astoria, Clackamas, OR) as described above. Soil particle size distribution was determined using the hydrometer method (Gee and Bauder, 1986). Soil pH was measured with an Accumet pH meter 50 (Fisher Scientific, Hampton, NH) in a 1:2 soil to water suspension. Selected baseline soil properties are presented in Table 1.


View this table:
[in this window]
[in a new window]
  		Table 1. Properties of loam soil used in the bioassay.

 
Experimental Setup
Forty grams (dry wt.) of the eight composted and two non-composted manures were mixed thoroughly with 2 kg dry soil to give a rate of 20 g amendment kg–1 soil (approximately 24 Mg ha–1 dry wt., assuming soil bulk density of 1.2 Mg m–3 and incorporation depth of 0.1 m). Inorganic P fertilizer (60 mg P kg–1 as KH2PO4) and non-fertilized control treatments were included for comparison, giving a total of 12 treatments. The treated soils were put into 2-L pots and placed in a controlled environment chamber. The amendment rate reflects compost and manure application rates normally used on-farm (Gagnon and Simard, 1999) and in the reclamation of abandoned oil and natural gas wellsites in agricultural areas in Alberta (Larney et al., 2003a). All treatments were replicated three times in a randomized complete block design.

Three days after amendment addition, 15 canola seeds were planted in the amended and unamended soils in each pot. Ultra pure water (233 g kg–1 soil) was added to all treatments to bring the soil moisture content to approximately field capacity. All treatments received initial full-strength nutrient solutions from which P was omitted (Table 2). Solutions were prepared separately for each of the macronutrients (N, K, and S) and added in 20-mL aliquots to each treatment. Micronutrients (Zn, Mn, Mo, B, Cu, and Fe), on the other hand, were mixed and added in one solution (4-mL aliquot). Throughout the study, the pots were weighed every 2 d and watered as needed to bring the soil moisture back to field moisture capacity. Following germination, seedlings were thinned to three plants per pot. The controlled environment chamber was maintained at 20°C under a 16 h light–8 h dark photoperiod with a light intensity of 264 µmol m–2 s–1 during the entire study.


View this table:
[in this window]
[in a new window]
  		Table 2. Composition of nutrient solution added to all treatments at full strength before planting and at 50% dilution after each of nine subsequent harvests during the bioassay.

 
Canola plants were harvested 30 d after planting by cutting the shoots at the soil surface using a scalpel. Each pot was emptied into an aluminum pan and the roots were carefully separated from soil. Finer roots were left in the soil due to difficulty in their complete removal. The shoots and roots were dried at 60°C for 48 h. Soil from each pot was thoroughly mixed, and a 20-g (dry wt.) subsample was taken and air-dried. The remaining soils were stored in the controlled environment chamber for 24 h, after which macro- and micronutrients were applied at half the rate used during the first planting (Table 2). The soils were then re-potted and re-seeded to canola as described previously. This cycle was repeated nine times, resulting in 10 crops over a period of 363 d.

Plant Tissue Analysis
Plant samples were oven dried at 60°C before chemical analysis. Fine-ground samples (<0.15 mm) were analyzed for total C and N by dry combustion using an automated CNS analyzer (Carlo Erba). Total P concentration was measured colorimetrically (Murphy and Riley, 1962) using an AutoAnalyzer II SC Colorimeter (Technicon) following digestion with concentrated H2SO4 and H2O2 (Thomas et al., 1967).

Plant Phosphorus Uptake
Total P uptake for each treatment was estimated by summing the products of shoot dry weight by shoot P concentration and root dry weight by root P concentration:

Formula 1[1]
where DW and P are tissue dry weight (kg) and P concentration (g kg–1), respectively, and subscripts s and r denote shoot and root material. The P concentrations and root dry weights were adjusted for adhering soil using the method described by Janzen et al. (2002). Briefly, roots from Harvest 7 were sampled and cleaned thoroughly to remove adhering soil. The root material was then fine ground and analyzed for total C and P content. The amount of adhering soil was then corrected for mathematically by determining the "dilution" of C concentration of the root tissue by soil through a back-calculation using the total C concentration of the soil. The same average dilution factor determined from the three root samples was then applied to all root dry matter and P concentration determinations. Given the concentration of C and P in the soil (Cs and Ps, respectively) and the C concentration in the thoroughly cleaned root tissue (Cr), the corrected P concentration in "soil-less" roots (Pr) was calculated using the following equation:

Formula 2[2]
where the subscript t denotes the total sample (root + soil), and all concentrations are in units of g g–1.

For each treatment, total P uptake values from all harvests were summed to give the cumulative phosphorus uptake (CPU) (mg kg–1 soil) over the 363-d period:

Formula 3[3]
where i is the harvest number and 10 is the total number of harvests.

Since amendment-derived P uptake is difficult to accurately determine in the absence of isotopic techniques, apparent phosphorus recovery (APR) rather than P uptake efficiency was used as an efficiency index for plant uptake of amendment-derived P. Apparent P recovery was calculated as the amount of P per unit of amendment phosphorus (TPamend) taken up by amendment-treated plants in excess of that taken up by unamended plants:

Formula 4[4]
where ACPU, the amendment-derived cumulative phosphorus uptake, was estimated by subtracting the CPU of the control treatment from that of the amended soil. In this calculation, it was assumed that the total P uptake from the control treatment represented the efficiency of non-amendment P uptake for all treatments; therefore, the difference in uptake between the control and the amendment-treated plants would be related to efficiency of amendment P uptake. It is possible that biochemical processes from amendment application may enhance the availability of P already present in the soil. Nevertheless, on a comparative basis, Eq. [4] provides an insight into the relative ranking of the amendments with respect to P recovery by plants growing on the amended soils.

Statistical Analysis
Analysis of Variance
Data were analyzed using the General Linear Models (GLM) procedure of SAS (SAS Institute, 2005) to test for amendment effects on biomass, P uptake, and APR. The UNIVARIATE procedure (SAS Institute, 2005) indicated that APR data did not conform to a normal distribution, according to the Shapiro–Wilk statistic (W, P ≤ 0.05). The APR data were therefore normalized via log-transformation before parametric analysis. Means were separated using the LSMEANS statement of SAS and adjusted according to the Tukey's studentized range test.

Partial Least Squares Analysis
Quantitative relationships between the response variable (ACPU) and amendment chemical parameters were explored using the method of partial least squares projection to latent structures (PLS). The PLS methodology was chosen because it is able to handle smaller numbers of observations with many independent variables, even when these display a high degree of multicollinearity, and it can also tolerate missing data. Ordinary least squares (OLS) multiple linear regression can present problems under such scenarios (Neter et al., 1996).

The PLS analysis was performed for ACPU using the PLS procedure of SAS (SAS Institute, 2005; Tobias, 1995). Initially, all amendment chemical parameters (Table 6) were included as predictor variables in the PLS model. Predictors with the greatest influence in explaining variability in the independent variables, or variable importance on prediction (VIP), were then selected based on the criterion of VIP > 0.8 (Wold, 1995). This was followed by sequential exclusion of predictor variables with the least impact on the model, based on loading weights and scores, until the highest R2 was obtained.


View this table:
[in this window]
[in a new window]
  		Table 6. Partial least squares (PLS) regression coefficient and variable influence on projection (VIP) values for 22 predictors in the preliminary model for estimation of cumulative amendment P uptake in the bioassay.

 
The number of PLS factors were selected using a cross validation method (CV = TESTSET option of PROC PLS) in which the original data set was divided into two groups: a training or calibration set and a test or validation set. The number of extracted factors with the minimum predicted residual sum of squares (PRESS) statistic was chosen as the optimum. Using the CVTEST option of PROC PLS, the optimum or minimizing number of factors was compared to the PRESS for fewer factors to test whether there was a significant difference. In the absence of a significant difference, the model with fewer factors was chosen. The predictive strength of the model was assessed by linear regression of measured values in the response variables versus the predicted values obtained in the cross-validation procedure. To interpret the results in terms of the original variables rather than the components (latent variables), the PLS solution was transformed into an OLS form using the SOLUTION option of PROC PLS. The result was a set of "pseudo"-regression coefficients in the original dependent variables.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Amendment Chemical Properties
Chemical properties differed considerably among the composted and non-composted beef cattle manures used in this study (Table 3). Mean total C concentration of the composted manures was 198 ± 47 g kg–1, which was lower than that for the non-composted manures (367 ± 71 g kg–1). Conversely, mean ash content of the composted manures (616 g kg–1) was nearly double that for the non-composted manures (311 ± 52 g kg–1). The higher ash content of the composts was expected and was due to the decomposition of OM and concomitant emission of CO2 and gaseous forms of N during composting (DeLuca and DeLuca, 1997). While total N concentrations were similar for composted (mean 17 ± 5 g kg–1) and non-composted manures (17 ± 4 g kg–1), total available N (NH4–N + NO3–N) was higher for non-composted (mean 1746 ± 127 mg kg–1) than composted manures (mean 1216 ± 534 mg kg–1). Similarly, the percentage of total N that was in the available form was higher for the non-composted (10.3%) than composted manures (7.0%). Mean NH4–N and NO3–N were 526 and 690 mg kg–1, respectively, for the composted manures, and 1690 and 56 mg kg–1, respectively, for the non-composted manures, reflecting expected changes in the NH4–N to NO3–N ratio following composting of beef cattle manure. The high coefficients of variation for compost NH4–N and NO3–N are typical of commercially available composts in southern Alberta. Variability in total available N results from differences in cattle N excretion rates and dilution ratios with bedding or soil while the NH4–N to NO3–N ratio is largely a function of the degree of compost maturity. Geometric mean NH4–N to NO3–N ratios were 0.58 for composted manures and 31.4 for non-composted manures. The NH4–N to NO3–N ratios for five of the eight composted manures (0.5–3) were indicative of mature products (Bernal et al., 1998; Woods End Research Laboratory, 2000).


View this table:
[in this window]
[in a new window]
  		Table 3. Chemical properties of composts and manures used in the bioassay.{dagger}

 
Carbon to nitrogen ratios were lower for the composted (mean 12) than non-composted manures (mean 22). Wood chip–bedded manure had higher C to N ratio (28.9) than SBM (16.1). Similar differences in C to N ratios between non-composted WBM and non-composted SBM (Miller et al., 2003) and between composted WBM and composted SBM (Zvomuya et al., 2005) have been reported in southern Alberta. Miller et al. (2003) attributed the differences to the higher C to N ratio for wood chips (mean 320) than straw (111) used as bedding material. Carbon to nitrogen ratios of >20 are considered optimum for successful composting (Rynk, 1992).

Amendment Phosphorus Fractionation
Total P concentration (dry wt. basis) in the composted manures averaged 8.5 ± 3.2 g kg–1, compared with 5.5 ± 1.91 g kg–1 in the non-composted manures (Table 3). Mean C to P ratio, on the other hand, was three times higher for non-composted than composted manures. These differences were expected since, during composting, P is conserved while CO2 and NH3 losses result in dry matter loss (DeLuca and DeLuca, 1997). The wide variation among composted manures likely reflects differences in the type of feeding operation, water content, type and amount of bedding, animal age, feed and supplements used, and duration of storage of the raw manure (Barnett, 1994; Dou et al., 2000). On average, non-composted manure contained a greater percentage of PH2O (29% of TP, 1.56 ± 0.42 g kg–1) and PHCO3 (34%, 1.82 ± 0.88 g kg–1) than composted manure, with 23% (1.99 ± 1.35 g kg–1) and 27% (2.27 ± 0.6 g kg–1), respectively (Table 4). The moderately labile PNaOH fraction was similar for non-composted and composted manures (both 9% of TP). HCl-extractable P, which is considered moderately recalcitrant, was higher for the composted (32%) than non-composted manures (16% of TP). Studies on the effect of composting on PH2O have given conflicting results. Sharpley and Moyer (2000) reported a reduction in PH2O following composting of poultry manure, ostensibly due to the diluting effect of the added C sources, whereas Dao et al. (2001) and Sikora and Enkiri (2005) found no significant effect of composting on PH2O.


View this table:
[in this window]
[in a new window]
  		Table 4. Phosphorus fractionation of compost and manure samples used in the bioassay.{dagger}

 
Organic P accounted for 8% of TP in the composted manures and 19% of TP in the non-composted manures, reflecting greater mineralization of Po during composting. While H2O- and NaHCO3–extractable Po fractions have traditionally been considered labile, recent techniques using enzymatic hydrolysis suggest that up to 50% of Po in H2O extracts and as much as 90% of Po in NaHCO3 extracts may be enzymatically nonhydrolyzable (He and Honeycutt, 2001).

Manure and compost analyses by most commercial laboratories include major nutrients such as total N, TP, and total K rather than extractable nutrients. However, drawing conclusions based on TP in isolation can be misleading, as demonstrated by our P fractionation results. For example, Composts 3 and 6 had very similar TP concentrations (10.7 and 10.4 g kg–1, respectively) and inorganic P proportions (92 and 97% of TP). However, the individual Pi fractions in each of these composts were quite different. Thus, while Compost 3 had 35% of TP as PiH2O, Compost 6 had only 9% of its TP in this labile fraction. Conversely, Compost 3 had only 27% of its TP as PiHCl (recalcitrant Pi) compared with 52% in this form in Compost 6. If limiting the potential for environmental transport in runoff were the goal, then Compost 6 would be favored over Compost 3.

On the other hand, at first glance, Composts 5 and 6 appear to have very different TP concentrations (4.7 g kg–1 vs. 10.4 g kg–1). However, labile P values (sum of water- and NaHCO3–extractable Pi and Po) were similar (2.74 vs. 3.43 g kg–1) for the two composts. Thus Compost 5, which had only 45% of the TP of Compost 6 had in fact 80% of the labile P of Compost 6. Therefore, land application of Compost 5 with its low level of TP, instead of Compost 6, may offer a false sense of security if limiting environmentally mobile P was the aim.

Biomass Yield, Phosphorus Uptake, and Apparent Phosphorus Recovery
Cumulative biomass yield during the first three growth cycles did not differ significantly among treatments, although it was slightly lower for the control (Fig. 1). At all subsequent harvests, however, plants from the unamended control treatment had significantly lower cumulative biomass yield. Cumulative biomass yields for composted and non-composted manure treatments were similar at all harvest dates, and yields also did not differ significantly between the organic amendments and the fertilizer during the first six harvests. However, the fertilizer treatment gave significantly lower cumulative biomass yields at subsequent harvests. All amendment treatments significantly improved total cumulative biomass yield over the 363-d period compared to the control (Table 5). There was no significant difference in the total cumulative biomass yield between the composted manure (overall mean 37.9 g pot–1), non-composted manure (37.8 g pot–1), and fertilizer (31.1 g pot–1) treatments, indicating that P availability from the amendments was not a limiting factor for optimal plant growth. Plant available P (or modified Kelowna-extractable P, MKP) was 20 mg kg–1 or higher at all harvests in amended soils, except on Days 313 and 363 (Harvests 9 and 10) when the fertilizer treatment resulted in lower values (Fig. 2). The low MKP in fertilizer-treated soils at these later harvests may explain the observed decline in cumulative biomass yield mentioned above. For optimum canola yield (one harvest) under field conditions in Alberta, a threshold of 60 mg kg–1 MKP in the 0- to 0.15-m soil depth is recommended (Alberta Agriculture, Food and Rural Development, 2003). Although MKP levels were below this threshold in our amended soils, deficiency was not expected since the crops were harvested just 30 d after planting.


Figure 1
View larger version (21K):
[in this window]
[in a new window]
  		Fig. 1. Cumulative canola biomass at each of the 10 harvests in the bioassay. Vertical bars represent standard errors.

 

View this table:
[in this window]
[in a new window]
  		Table 5. Cumulative biomass, P uptake, and apparent P recovery during the bioassay.{dagger}

 

Figure 2
View larger version (25K):
[in this window]
[in a new window]
  		Fig. 2. Modified Kelowna-extractable soil P measured at each of the 10 harvests in the bioassay. Vertical bars represent standard errors.

 
Cumulative P uptake was lowest for the control treatment at all but the first harvest (Fig. 3). Organic amendments gave the highest (and similar) CPU values, although these were not significantly higher than those for the fertilizer during the first five harvests. There was a nearly linear increase in CPU for the organic amendments at successive harvests, reflecting the high levels of MKP throughout the bioassay and the corresponding near-linear increase in cumulative biomass yield. For the control and fertilizer treatments, however, the trends showed diminishing increments in the CPU, reflecting the corresponding changes in cumulative biomass yield and MKP. Slow mineralization of the organic P pool explains the observed sustained supply of MKP from the organic amendments, in contrast to inorganic fertilizer P, which is immediately available but whose availability declines with time as observed in this study. This is consistent with previous reports of sustained P availability from swine manure versus declining availability of fertilizer P (Loria and Sawyer, 2005). Also, much more TP was applied with composted and non-composted manure than with fertilizer in our study (Table 5). As evident from Fig. 3, only a third of the fertilizer P (i.e., 20 mg kg–1) remained in the soil after 10 harvests, compared with 74 mg kg–1 for composted manure and 60 mg kg–1 for non-composted manure.


Figure 3
View larger version (22K):
[in this window]
[in a new window]
  		Fig. 3. Cumulative phosphorus uptake (CPU) by canola plants measured at each of the 10 harvests in the bioassay. Vertical bars represent standard errors.

 
Total CPU over the 363-d period ranged from 23.8 mg kg–1 soil for the control to 73.9 mg kg–1 amended soil for the composted manure treatments (Fig. 3). Similar P uptake values have been reported for a controlled environment chamber experiment in which white clover (Trifolium repens L.) was grown in a clayey calcareous Eutrochrept amended with composts made from solid kitchen and garden wastes (Sinaj et al., 2002). In that study, compost application increased plant P uptake from 31.4 mg kg–1 in the unamended control to 37.9 to 42.7 mg kg–1 soil in compost-amended soils. In the present study, composted and non-composted manures resulted in similar overall CPU values and increased the CPU compared to the control treatment (Table 5). These results are consistent with those of Eghball and Power (1999), who reported similar P uptake by corn from composted and non-composted cattle manure, and Sikora and Enkiri (2005), who indicated that composting had little effect on P availability. In our study, however, the similarities in CPU between composted and non-composted manures were despite the fact that TP concentration was, on average, 57% higher in composted than non-composted manures. The overall CPU was significantly higher for composted manures than fertilizer. This was consistent with the much higher total N and total P amounts applied with composted manure (348 mg kg–1 and 170 mg kg–1, respectively) than with the fertilizer (100 and 60 mg kg–1 soil, respectively). More significantly, however, these results suggest that, although N and P availability from the organic amendments is much slower than that from inorganic fertilizers, amendment rates applied under conditions similar to those in this study may result in excessive bioavailability of the nutrients. In the absence of adequate plant uptake, this may increase the risk of bioavailable P loss to aquatic systems.

Since the organic amendments had similar CPU values and more P was applied with composted- than with non-composted manure, APR was significantly higher for non-composted (33%) than composted manure (24%) (Table 5). However, there was no significant difference in APR between the organic amendments and fertilizer (27%). Similar results have been reported by Sikora and Enkiri (2005), who demonstrated in a controlled environment chamber study that composted poultry manure was as effective as triple superphosphate fertilizer in supplying P to fescue (Festuca arundinacea Schreb.) in a P-deficient soil. The difference in APR between plants treated with composted manure and those treated with non-composted manure could be due to the manure having a higher percentage (59%) of its TP in the labile (i.e., PH2O + PHCO3) form than composted manure (48.4%).

Our results suggest no significant effect of composting on beef cattle manure P plant availability. This was despite the substantially lower C to P ratio in composted (geometric mean = 25) than in non-composted manures (75). These findings are consistent with those of Dao et al. (2001), who found little change in water and Mehlich 3–extractable P content in poultry manure following composting.

Predicting Cumulative Phosphorus Uptake
Of the 22 amendment (composted and non-composted manures) chemical properties (predictors) included in the preliminary model for predicting ACPU, 14 had VIP > 0.8 (Table 6). Total PH2O, PiH2O, and TP were the most important variables, with VIP > 1.3. Based on PLS coefficient values, all but two (N to P and C to P ratios) of the predictors were positively correlated with ACPU. The inverse relationship with N to P and C to P ratios was expected and was due to the strong positive correlation between ACPU and TP. Cross-validation results based on the 14 significant (VIP > 0.8) predictors indicated that eight PLS components gave the minimum PRESS (root mean PRESS = 0.36). However, the eight components did not explain significantly more variation than just one component (root mean PRESS = 0.43).

Further statistical tests indicated that a one-factor model (root mean PRESS = 0.44) with only TPH2O and TP as predictors could most adequately predict ACPU from the amendments. The model accounted for 87% of the variability in independent variables and 81% of the variability in CPU. The following equation, containing "pseudo"-regression coefficients as described above (see Statistical Analysis), was derived to estimate ACPU from amendment TPH2O and TP concentrations:

Formula 5[5]
where ACPU, TPH2O, and TP are in units of g kg–1. The model had a good predictive power, as evidenced by the excellent relationship (r2 = 0.83) between predicted values, obtained by cross-validation, and observed ACPU data (Fig. 4).


Figure 4
View larger version (12K):
[in this window]
[in a new window]
  		Fig. 4. Predicted amendment-derived cumulative phosphorus uptake (ACPU) values (mg kg–1 amendment) from the one-factor partial least squares (PLS) model versus observed values for composted and non-composted manures used in the bioassay.

 
Our results have important implications for the management of composted or non-composted manure in cropping systems. When applied in excess of crop requirements, these organic amendments can result in a significant buildup of P in soil (Eghball and Power, 1999). A key issue in land application as a management option for high P organic amendments is defining agronomic rates that are environmentally acceptable. Predictive models, which use amendment chemical properties that are easily measurable in the laboratory, can help estimate amendment P availability for plant uptake following soil application of the amendment.

Because of the limited data used in our study (eight composted and two non-composted manure samples), the model may be valid only for the range of chemical concentrations recorded in this study. However, the predictive power of the model was reasonable in light of the small data set, highlighting the capability of PLS regression in handling such data. Nevertheless, the data were obtained under constant temperature and optimum soil moisture conditions (field capacity). The model, therefore, may not reflect the short-term effects of wetting and drying events and temperature fluctuations. Recent research findings highlight the importance of changes in soil water potential and temperature in field estimates of P mineralization (Grierson et al., 1999). Therefore, caution is needed when using models developed under controlled environment conditions.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Our results demonstrate that agronomically and environmentally important attributes such as ACPU associated with composted and non-composted manure application can be predicted based on P fractions and an optimization procedure involving PLS regression analysis. In this study, PLS regression provided an effective method for predicting ACPU using only two P fractions (PH2O and TP) that can be measured using simple laboratory procedures. The one-factor model accounted for 81% of the variation in ACPU and had good predictive power for this parameter. This model has important implications on the assessment of environmental fate and risk of manure P applied to cropland. Accurate prediction of P availability and plant P recovery may help tailor manure and compost applications to plant needs and minimize the buildup of bioavailable P, which can contribute to eutrophication of sensitive aquatic systems. Further research is needed to formulate such predictive models under field conditions where temperature fluctuations and wetting and drying cycles are important considerations. Additional research is also needed on different manure types. Avoiding P accumulation in the soil is also an important aspect in decreasing the risk of P loss since biochemical processes will ultimately release some of the P that is not bioavailable in the short term.


   ACKNOWLEDGMENTS
 
We thank M. Abul Kashem for sequential phosphorus extractions and Brian P. Handerek for technical assistance. Porcupine Corral Cleaning Ltd., Picture Butte, AB, and EcoAg Initiatives Inc., High River, AB, kindly provided some of the composts used in this study.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Lethbridge Research Centre Contribution no. 38705055.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zvomuya, F.
Right arrow Articles by Olson, B. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zvomuya, F.
Right arrow Articles by Olson, B. M.
Agricola
Right arrow Articles by Zvomuya, F.
Right arrow Articles by Olson, B. M.
Related Collections
Right arrow