J. Anim Sci. 2007. 85:1982-1989. doi:10.2527/jas.2006-408
© 2007 American Society of Animal Science
Prediction of nitrogen excretion in feces and urine of beef cattle offered diets containing grass silage1
T. Yan2,
J. P. Frost,
T. W. J. Keady3,
R. E. Agnew and
C. S. Mayne
Agri-Food and Biosciences Institute, Large Park, Hillsborough, Co Down BT26 6DR, UK
 |
Abstract
|
|---|
Data from 286 beef cattle, obtained in total diet digestibility assessments, were used to examine effects of dietary and animal factors on N excretion in feces and urine and to develop prediction equations for N excretion in beef cattle. The animals used were mainly from beef breeds, at various ages (from growth to finishing) and live BW (153 to 580 kg), and offered diets containing grass silage at production feeding levels. Dietary forage proportion ranged from 199 to 1,000 g/kg of DM and dietary CP concentration from 108 to 217 g/kg of DM. Linear and multiple regression techniques were used to examine relationships between the efficiency of N utilization and dietary and animal variables with the experimental effects removed. The statistical analysis indicated that N excretion was related positively (P < 0.001) to live BW and intakes of DM, N, and ME, and negatively (P < 0.001) to dietary forage proportion. The prediction equation for N excretion, developed using N intake alone, produced a large r2 (0.898) and a small SE (12.3). Addition of live BW and forage proportion as supporting predictors to this relationship only marginally increased R2 to 0.915 and reduced SE to 11.2. Nitrogen excretion was less well related to live BW (r2 = 0.771, SE = 18.5) than to N intake. Addition of N intake as a proportion of DMI or ME intake to the relationship between live BW and N excretion increased R2 to 0.824 and reduced SE to 16.2. The internal validation of these equations revealed that using N intake as the primary predictor produced a very accurate prediction of N excretion. In situations where data on N intake are not available, prediction equations based on live BW and dietary N concentration together can produce a relatively accurate assessment of N excretion. A number of mitigation strategies to reduce N excretion in feces and urine in beef cattle are discussed, including manipulation of dietary N concentration, diet quality, and level of feeding. The prediction equations and mitigation strategies developed in the current study provide an approach for beef producers to quantify N excretion against production and to develop their own mitigation strategies to reduce N excretion.
Key Words: beef cattle • live BW • nitrogen excretion • nitrogen intake • prediction
|
INTRODUCTION
|
|---|
In 1991, the European Union introduced the Nitrates Directives (European Community, 1991
), which aims to prevent the pollution of groundwater and surface water by nitrates arising from agricultural sources. The directive stipulates mandatory measures that must be included in an action program, one of which involves a limit on the amount of livestock manure (feces and urine) that may be applied to land each year, set at 170 kg of organic N (manure N) per ha. This limit will have very significant implications for stocking rates on livestock farms. Therefore, there is an increasing interest in developing approaches to predict N excretion and mitigation strategies to reduce N output in feces and urine in animal production. Recently, a number of prediction equations for N excretion in dairy cows have been published, using large datasets derived from total diet digestibility measurements (Wilkerson et al., 1997; Nennich et al., 2005; Yan et al., 2006). Yan et al. (2006) reported that dietary N concentration and milk production were the most important factors influencing the efficiency of N utilization for lactation in dairy cows. However, there is little comparable information available in the literature on prediction of N excretion and mitigation strategies to reduce N excretion in beef cattle. Information in this area would enable the beef industry to develop appropriate action programs to implement the European Union Nitrate Directives.
The objectives of the current study were to use digestibility data, derived from total diet digestibility assessments undertaken at the Agri-Food and Biosciences Institute of Northern Ireland, to examine effects of dietary and animal factors on the efficiency of utilization of dietary N and then to develop mitigation strategies and prediction equations for N excretion in feces and urine in beef cattle.
|
MATERIALS AND METHODS
|
|---|
Animals and Diets
All procedures involving animals in the experiments included in the current study were approved by Department of Health, Social Services and Public Safety of Northern Ireland.
The data set used in the current study was obtained from 286 beef cattle in 14 total diet digestibility studies (Table 1) undertaken at the Agri-Food and Biosciences Institute, formerly the Agricultural Research Institute of Northern Ireland, from 1984 to 2003. The animals used were of various ages (growth to finishing) and live BW (153 to 580 kg), and from different breeds (Holstein, Friesian, Aberdeen Angus, Simmental, Charolais, Limousin, Belgian Blue, and Blonde D’Aquitaine).
The animals were offered forage alone as a sole diet (n = 44) or a mixture of forage and concentrates (n = 242) at production feeding levels (AFRC, 1993). In the latter situation, the concentrates were offered as a complete diet with forage or were offered separately from forage. The forage proportion in mixed diets ranged from (g/kg of DM) 199 to 839, with a mean of 646 (SD = 165). The forage used was grass silage, with the exception of 8 animals, which were offered grass silage plus whole-plant wheat silage (n = 4) and grass silage plus corn silage (n = 4). These alternative forages accounted for less than 50% (DM basis) of the total forage consumed. The grass silages encompassed primary growth and first and second regrowth material. The grass was unwilted or wilted before ensiling and ensiled with or without application of silage additives. The concentrates used were based on barley, soybean meal, mineral and vitamin supplements, and by-products (corn gluten feed, molassed or unmolassed sugar-beet pulp, citrus pulp, or sugarcane molasses). The data on mean, SD, and range for animal and dietary variables are presented in Table 2.
Digestibility Measurements
Before commencing the digestibility studies, all cattle were housed in slatted accommodation, which was made of concrete with a 4-cm void between slats, offered the experimental diets for at least 20 d, and allowed free access to water. Cattle were then transferred to metabolism units and were tied in individual stalls with plastic mats for 8 d with total collection of feces and urine during the final 6 d. Feces and urine outputs were recorded and sampled daily as a proportion (5%) of total excretion of feces (by weight) and urine (by volume). The 6-d samples of feces and urine were separately mixed, and a representative sample taken for analysis as follows: feces samples were analyzed for oven DM, N, and GE concentrations and urine samples were analyzed for GE and N concentrations. During the 8 d in the metabolism units, total feed intake was recorded daily. The forage was sampled daily for determination of toluene DM, N, and GE concentrations. Samples of concentrates were collected daily and analyzed for oven DM, N, and GE concentrations. The details on methods adopted for analysis of diet, feces, and urine samples were as described by Mayne and Gordon (1984). Live BW was determined on the first and last day in the metabolism units. The data on mean, SD, and range for N intake and outputs and efficiencies of N utilization are presented in Table 2
.
Statistical Analyses
The correlation coefficients, presented in Table 3, were determined using the linear regression between N utilization variables and dietary and animal factors, without removing the experimental effect. Linear and multiple regression models were used to develop the prediction equations for N excretion (Tables 4 and 5). Two sets of variables were used to develop these equations; i.e., animal and dietary factors with or without total N intake because N intake may not be always available in commercial practice.
View this table:
[in this window]
[in a new window]
|
Table 4. Linear and multiple prediction equations for N excretion in feces and urine using live BW and N intake as the primary predictors1
|
|
View this table:
[in this window]
[in a new window]
|
Table 5. Internal validation of linear and multiple prediction equations for N excretion in feces and urine developed using two-thirds of the present data1
|
|
An internal evaluation was undertaken to validate the prediction equations for N excretion by dividing the whole data set into 2 subsets, one-third (n = 96) and two-thirds (n = 190) of the data. To obtain a similar range in N intake between 2 subdatasets, the whole dataset was divided into groups of 3 cows according to N intake from the highest to the lowest values, and within each group, 1 cow’s data were randomly selected for the one-third dataset and the remaining 2 cows’ data for the two-thirds dataset. Two-thirds of the data were used to develop relationships similar to those developed using the whole data set. These new equations were then evaluated using the remaining one-third of data and the mean-square prediction error (MSPE) technique (Eq. [A]). Mean prediction error (MPE, Eq. [B]), rather than MSPE, was used to describe the prediction accuracy, as follows:
 | [A] |
 | [B] |
where P or A is predicted or actual N excretion, A is the mean value of actual N excretion, and n is the number of pairs of values of P and A compared.
The preceding equations were fitted respectively to the following equations (Eq. [C] and [D]) to remove the experimental effect on the linear or the multiple relationship (Tables 4 and 5):
 | [C] |
 | [D] |
where ai represents the effect of experiments i for i = 1 to 14; x1, x2, ... and xn are the x-variables; and b1, b2, ... and bn are their regression coefficients. These equations each produce 3 optional fits. The first model to be fitted is an ordinary linear or multiple linear regression, ignoring the experimental effect. Next, the model is extended to include a common coefficient for each x-variable and a different constant for each experiment, thus giving a set of parallel relationships. Then, the final model has different constants and different regression coefficients for x-variables across the different experiments. The second model is selected, and a common constant was then obtained, with the error derived from each experiment being removed (Lawes Agricultural Trust, 2002).
All analysis for development of prediction equations for N excretion was performed using Genstat for Windows, sixth edition (Lawes Agricultural Trust, Rothamsted, UK).
|
RESULTS
|
|---|
Relationships Between N Utilization and Animal or Dietary Variables
The correlation coefficients (r, Table 3) were determined using linear relationships between N utilization and animal or dietary data without removing the experimental effects. Nitrogen intake, N excretion, and retained N were related positively (P < 0.001) with live BW, DMI, and energy intake (GE, DE, and ME), and negatively (P < 0.001) to dietary forage proportion; N intake was related positively (P < 0.001) with N excretion and retained N. Nitrogen intake as a proportion of DMI or GE intake was positively related to N intake, N excretion, and retained N (P < 0.05). However, these relationships had considerable differences in strength. For example, r values in the relationships of N excretion were highest with N intake, followed by DMI, energy intake (GE, DE, and ME), and live BW, low with dietary forage proportion; and the lowest with N intake as a proportion of DMI and GE intake.
Prediction Equations for N Excretion in Feces and Urine
Linear and multiple prediction equations for N excretion in feces and urine are presented in Table 4, and the relationships between manure N excretion and N intake and live BW are presented in Figure 1. All relationships are significant (P < 0.001), and each predictor had a significant effect on the relationship (P < 0.05). The r2 value in the linear relationship between N excretion and live BW is relatively high, and the SE value is relatively low (Eq. [1]). Addition of N intake:DMI ratio (Eq. [2]) or N intake:ME intake ratio (Eq. [3]) as a supporting predictor increased the R2 value and decreased SE value, and the improvement was greater with the inclusion of N intake:ME intake than N intake:DMI. However, the greater r2 value and lower SE for the relationship between N excretion and N intake (Eq. [4a]) indicates that N intake alone is a much better predictor of N excretion. On average, 77.5% of consumed N was excreted in feces and urine, as obtained in Eq. [4b] when the constant in Eq. [4a] was omitted. Addition of live BW and dietary forage proportion to Eq. [4a] only produced a small improvement on the relationship (Eq. [5] and [6]), although each supporting predictor had a significant effect (P < 0.05) on the relationship between N excretion and N intake. For example, including both variables along with N intake only increased R2 values from 0.898 to 0.915 and SE values reduced from 12.3 to 11.2 (Eq. [4a] vs. [6]).
Internal Validation of Prediction Equations for N excretion
The Eq. [i] to [iv] (Table 5), developed from 2-thirds of the present data set, are similar to those presented in Table 4, developed from the whole dataset. These new equations were validated (Table 6) using the remaining one-third of the dataset. The mean predicted N excretions from all 4 equations ([i] to [iv]) were close to actual data. The equation developed using live BW as the sole predictor (Eq. [i]) produced a relatively large MPE value and a low r2 value in the relationship between predicted and actual N excretion. When N intake:DMI ratio was added to Eq. [i], MPE was considerably reduced and r2 increased (Eq. [ii]). However, the best prediction accuracy was obtained using N intake as a sole or primary predictor. The equation developed using N intake as the sole predictor (Eq. [iii]) produced a small MPE value and a high r2 value in the relationship between predicted and actual N excretion in comparison with those using live BW as a primary predictor. Adding live BW as a supporting predictor to Eq. [iii] did not improve the MPE and r2 value (Eq. [iv]).
View this table:
[in this window]
[in a new window]
|
Table 6. Internal validation using one-third of the present data (n = 96) and equations (Table 5) developed from the remaining two-thirds of the data1
|
|
Residual plots (Table 6 and Figure 2) were also used to evaluate the prediction accuracy by plotting the predicted N excretion (x axis) against the corresponding residual N excretion of predicted minus actual (y axis). The residual plots from equations using live BW alone as the predictor (Eq. [i]) were scattered, but relatively uniformly distributed around the zero line with Eq. [ii] when N intake:DMI ratio was used together with live BW as predictors. However, the best distribution of the residual plots was derived from the prediction using N intake as a sole or primary predictor (Eq. [iii] and [iv]), with the vast majority of the residual plots distributed uniformly around the zero line. The SD and the range of the residual N excretion in Eq. [iii] and [iv], as presented in Table 6, were therefore much smaller than those in Eq. [i] and [ii]. Between Eq. [i] and [ii], the SD and the range of the residual N excretion were reduced with addition of N intake as a proportion of DMI. These results are in accordance with MPE and r2 in the relationship between predicted and actual N excretion as reported previously.
View larger version (13K):
[in this window]
[in a new window]
|
Figure 2. Relationships between predicted (x-axis) and residual (predicted minus actual, y-axis) N excretion (g/d) using one-third of the present data and Eq. [i] to [iv] in Table 5 developed from the remaining two-thirds of the data in beef cattle (NI = N intake, g/d).
|
|
|
DISCUSSION
|
|---|
Present Dataset
The majority of data in the present data set was obtained from beef cattle (Aberdeen Angus and Continental breeds), which were used to develop 2 sets of prediction equations for N excretion in feces and urine. One set of equations was developed using N intake as a primary predictor and another set using live BW and dietary N concentration as primary predictors because N intake may not always be available in practice. The present dataset is unique because there is little information available in the literature on developing prediction models for N excretion in beef cattle using a large dataset of digestibility measurements. Wilkerson et al. (1997) published a large N excretion dataset of nonbeef breed cattle (growing Holstein heifers and steers, n = 286). The average N excretion:N intake in their study was 5 percentage units greater than that in the current study (83 vs. 78%). This difference is likely to be attributed to the different breed of animals and the different type of forages used in the current study (grass silage) and Wilkerson et al. (1997, alfalfa, corn and grass silages and hay, cottonseed hulls, and fresh grass). Holstein cattle used by Wilkerson et al. (1997) are likely to have had a lower growth rate than beef cattle breed used in the current study (Steen, 1995). Moreover, the difference in the efficiency of N utilization between the current study and Wilkerson et al. (1997) can also be explained by the higher dietary CP concentration used in the latter (15.5 vs. 18.2%). This difference in dietary CP concentration can result in an increase in N excretion of 12 g/d for higher CP concentration diets in the study of Wilkerson et al. (1997) if using a live BW of 381 kg (the mean live BW in the current study) and the present equation [2] as presented in Table 4. The negative relationship between the efficiency of N utilization and dietary N concentration has also been reported in dairy cows (Tamminga, 1992; Yan et al., 2006). Therefore, the present dataset is unique for the beef industry in situations where grass silage is the primary source of forage.
Development of Prediction Equations for N Excretion
The current study examined the relationships between N excretion and a number of dietary and animal variables in beef cattle and found that the best fitting factor was N intake. Therefore, N intake was selected as a primary predictor for development of further prediction equations. Using N intake as a single predictor produced a very high r2 (0.90) and a very low SE (12.3). This r2 value is the same as that in the relationship between N excretion and N intake in lactating dairy cows (n = 564) as reported in our laboratory (Yan et al., 2006), and higher than that (0.78) reported by Kebreab et al. (2001) using a small dataset of lactating dairy cows. In the current study, adding live BW and forage proportion as supporting predictors only marginally improved the relationship between N intake and N excretion, although inclusion of each of these variables had a significant effect on the relationship. The r2 value slightly increased to 0.92 and SE reduced to 11.2. The r2 value of 0.92 using N intake, live BW, and forage proportion as predictors for N excretion is comparable to that of 0.94 obtained by Wilkerson et al. (1997) with growing and replacement cattle using similar predictors (DMI, dietary CP concentration, dietary NDF concentration, and live BW).
The present internal validation also demonstrated that using N intake only for prediction of N excretion in beef cattle resulted in a very good prediction accuracy in terms of mean, SD, and range of residual N excretion (predicted minus actual), and r2 value in the relationship between predicted and actual N excretion and mean prediction error. Adding live BW and dietary forage proportion as supporting predictors did not improve the prediction accuracy. Therefore, N intake is a very accurate predictor of N excretion in beef cattle. This result is similar to that obtained in our laboratory with lactating dairy cows (Yan et al., 2006).
However, N intake is calculated from dietary N concentration and DMI, which may not always be available in practice. This limitation restricts the ability to predict N excretion from N intake in beef cattle. Therefore in the current study a second set of prediction equations for N excretion was developed using live BW as a primary predictor because live BW is normally available in practice. Relating live BW to N excretion produced a greater r2 value (0.77) than that (0.23) in lactating dairy cows (Yan et al., 2006). Our r2 value was, however, lower than that (0.88) in beef cattle reported by Smith and Frost (2000), but the dataset used in their study was very small (n = 13). In the current study, adding N intake as a proportion of DMI or ME intake increased the r2 value to 0.82 and reduced SE value to 16.2. The internal validation demonstrated that including dietary N concentration as a supporting predictor to the relationship between live BW and N excretion greatly improved the prediction accuracy of N excretion. This is evidenced by a lower SD and range of residual N excretion (predicted minus actual) and mean prediction error, and a higher r2 value in the relationship between predicted and actual N excretion. These findings indicate that N excretion in beef cattle could be accurately predicted from live BW and dietary N concentration.
Mitigation Strategy to Reduce N Excretion
The current study evaluated the effects of a number of dietary and animal factors on the efficiency of N utilization in beef cattle (Table 7). The evaluation was undertaken by relating, respectively, N excretion as a proportion of live BW or N intake to a range of dietary and animal factors. The experimental effect on each relationship was removed using the technique as described previously. The most effective strategy to reduce N excretion is to manipulate dietary N concentration. For example, N excretion per unit of live BW is positively related to N intake as a proportion of DMI (Eq. [7]), GE intake (Eq. [8]), and ME intake (Eq. [9]). Increasing dietary N concentration by 1 g/kg of DM could increase N excretion per kg of live BW by 0.0089 g (Eq. [7]). Similarly, increasing N intake as a proportion of DM, GE, and ME intake significantly (P < 0.001) increased additional urine N excretion as a proportion of N intake (Eq. [10] to [12], r2 = 0.255 to 0.306), where additional urine N output was calculated as the difference between total urine N output and endogenous N excretion (AFRC, 1993). The increase in additional urine N output as a proportion of N intake could be proportionately 0.0076 g/g with an increase in dietary N concentration by 1 g/kg of DM (Eq. [10]). On the other hand, Yan et al. (2006) reported that dietary N concentration was negatively related to milk N output as a proportion of N intake, which indicated that the efficiency of N utilization for lactation increased with reductions in dietary N concentration. These findings clearly demonstrate that dietary N concentration could be used to control N pollution from beef cattle production. For dairy cows, it was suggested that N intake should be less than 400 g/d for average yielding cows (Kebreab et al., 2001), and dietary N concentration should not exceed 30 g/kg of DM (187.5 g of CP/kg of DM; Tamminga, 1992).
However, decreasing dietary N concentration below requirements, especially rumen degradable N content, can restrict microbial activity in the rumen, consequently reducing feed intake and growth in beef cattle (ARC, 1980). Therefore, animal productivity, health, and welfare should be considered when developing mitigation strategies because increasing animal productivity is also a mitigation approach to reduce N excretion per unit of production. Yan et al. (2006) reported that milk N:N intake ratio (g/kg) could be increased by proportionately 3.448 with each 1 kg/d increase in milk yield in dairy cows. Although there were no data on live BW gain in the current study, the level of feeding (as an indication of growth rate) was found to have a negative relationship with N excretion as a proportion of N intake (Eq. [13]). This indicated that increasing level of feeding increased the efficiency of N utilization. Increasing dietary quality (e.g., ME concentration) may be another approach to reduce N excretion in beef cattle. For example, the current study found that N excretion as a proportion of N intake could be reduced by proportionately 0.048 with each 1 MJ increase in ME per kg of DM of diet (Eq. [14]). The increase in dietary quality may give a better match in supplying fermentable N and OM to microbial organisms in the rumen, leaving less excess NH3 that is absorbed into blood and excreted in urine as urea. The current study also found that adult cattle could have a lower proportional N excretion than young cattle, because live BW was highly related to N excretion as a proportion of live BW (Eq. [15]). Increasing 1 kg live BW results in a decrease in N excretion as a proportion of live BW by 0.0003 g/kg.
In conclusion, beef production can result in excretion of high levels of N in feces and urine, which can pollute groundwater and surface water. The current study demonstrates that N excretion by beef cattle offered diets containing grass silage can be accurately predicted from N intake alone or from live BW and dietary N concentration. The equations developed in the current study therefore provide an approach for beef producers to quantify N excretion against production and consequently to develop their own mitigation strategies to reduce N excretion. The manipulation of dietary composition is an effective measure to increase the efficiency of N utilization and nutritional strategies, which could be used include use of low N content feeds or use of good quality diets to increase productivity.
|
Footnotes
|
|---|
1 The authors gratefully acknowledge funding from the Department of Agriculture and Rural Development and AgriSearch (via farmer levy funding). We also thank our colleagues for their assistance in the collation of the data for this study. 
3 Current address: Teagasc, Livestock Production Centre, Athenry, Co. Galway, Ireland.
2 Corresponding author: tianhai.yan{at}afbini.gov.uk
Received for publication June 27, 2006.
Accepted for publication May 1, 2007.
|
LITERATURE CITED
|
|---|
AFRC. 1993. Energy and Protein Requirements of Ruminants. CAB Int., Wallingford, Oxon, UK.
ARC. 1980. The Nutrient Requirements of Ruminant Livestock, Technical Review. CAB, Farnham Royal, UK.
European Community. 1991. Implementation of Nitrates Directive. http://ec.europa.eu/environment/water/water-nitrates/index_en.html Accessed Apr. 27, 2007.
Kebreab, E., J. France, D. E. Beever, and A. R. Castillo. 2001. Nitrogen pollution by dairy cows and its mitigation by dietary manipulation. Nitrogen Cycling in Agroecosystems 60:275–285.[CrossRef]
Lawes Agricultural Trust. 2002. Genstat for Windows. 6th ed. Rothamsted, UK.
Mayne, C. S., and F. J. Gordon. 1984. The effect of type of concentrate and level of concentrate feeding on milk production. Anim. Prod. 39:65–76.
Nennich, T. D., J. H. Harrison, L. M. VanWieringen, D. Meryer, A. J. Heinrichs, W. P. Weiss, N. R. St-Pierre, R. L. Kincaid, D. L. Davidson, and E. Block. 2005. Prediction of manure and nutrient excretion from dairy cattle. J. Dairy Sci. 88:3721–3733.[Abstract/Free Full Text]
Smith, K. A., and J. P. Frost. 2000. Nitrogen excretion by farm livestock with respect to land spreading requirements and controlling nitrogen losses to ground and surface waters. Part 1: cattle and sheep. Livest. Prod. Sci. 71:171–181.[CrossRef]
Steen, R. W. J. 1995. The effect of plane of nutrition and slaughter weight on growth and food efficiency in bulls, steers and heifers of three breed crosses. Livest. Prod. Sci. 42:1–11.[CrossRef]
Tamminga, S. 1992. Nutrition management of dairy cows as a contribution to pollution control. J. Dairy Sci. 75:345–357.[Abstract]
Wilkerson, V. A., D. R. Mertens, and D. P. Casper. 1997. Prediction of excretion of manure and nitrogen by Holstein cows. J. Dairy Sci. 80:3193–3204.[Abstract]
Yan, T., J. P. Frost, R. E. Agnew, R. C. Binnie, and C. S. Mayne. 2006. Relationships between manure nitrogen output and dietary and animal factors in lactating dairy cows. J. Dairy Sci. 89:3981–3991.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
X. Hao, M. B. Benke, D. J. Gibb, A. Stronks, G. Travis, and T. A. McAllister
Effects of Dried Distillers' Grains with Solubles (Wheat-Based) in Feedlot Cattle Diets on Feces and Manure Composition
J. Environ. Qual.,
June 23, 2009;
38(4):
1709 - 1718.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
INTRODUCTION
