Prediction of urinary and fecal nitrogen excretion by beef cattle.

Prediction of urinary and fecal nitrogen excretion by beef cattle R. L. Dong, G. Y. Zhao, L. L. Chai and K. A. Beauchemin J ANIM SCI published online August 22, 2014

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://www.journalofanimalscience.org/content/early/2014/08/22/jas.2014-8000

www.asas.org

Downloaded from www.journalofanimalscience.org at Umeå University on August 26, 2014

LRH: Dong et al. RRH: Prediction of nitrogen excretion in cattle

Prediction of urinary and fecal nitrogen excretion by beef cattle1

R. L. Dong,*† G. Y. Zhao,* L. L. Chai, † and K. A. Beauchemin†2

*College of Animal Science and Technology, China Agricultural University, State Key Laboratory of Animal Nutrition, Beijing, 100193, P.R. China; †Agriculture and Agri-Food Canada, Lethbridge Research Centre, Lethbridge, Alberta Canada T1J 4B1

1

This project was supported by Agriculture and Agri-Food Canada. The authors wish to thank

Luis Tedeschi, Texas A & M University, for his helpful comments on an earlier draft of the manuscript. 2

Corresponding author: [email protected]

Received April 28, 2014. Accepted August 3, 2014.

1

Downloaded from www.journalofanimalscience.org at Umeå University on August 26, 2014 Published Online First on August 22, 2014 as doi:10.2527/jas.2014-8000

ABSTRACT: An analysis of predicting urinary and fecal N excretion from beef cattle was conducted using a dataset summarizing 49 published studies representing 180 treatment means for 869 animals. Variables included in the dataset were initial BW (kg), DMI (kg/d), dietary CP content (% of DM), N intake (g/d), apparent total tract N digestibility (%), and urinary and fecal N excretion (g/d). Correlation analysis examined relationships between animal and dietary variables and N excretion. A mixed model regression analysis was used to develop equations to predict N excretion in urine and feces and the proportion of urinary N in total N excretion as a function of various animal and dietary variables. Of the single animal and dietary variables, N intake was the best predictor of N excretion in urine and feces, whereas apparent total tract N digestibility was best to predict the proportion of urinary N in total N excretion. Low prediction errors and evaluation of the equations using cross-validation indicated the prediction equations were accurate and robust. Urinary and fecal N excretion can be accurately and precisely predicted by N intake, whereas the proportion of urinary N in total N excretion was best predicted solely using apparent total tract N digestibility.

Key words: apparent total tract N digestibility, beef cattle, crude protein, nitrogen excretion, nitrogen intake, prediction

2


INTRODUCTION

Ammonia (NH3) emissions are a major air quality concern at regional, national, and global levels (Burgos et al., 2007). Agricultural sources and livestock farming in particular, especially intensive cattle operations, are large contributors to NH3 emissions (Bussink and Oenema, 1998; Groot Koerkamp et al., 1998). Ammonia is volatilized from animal waste. Dietary CP intake and the digestibility of CP influence NH3 emissions by affecting N excretion (Cole et al., 2005; Todd et al., 2013). The majority of NH3 emitted is produced from microbial hydrolysis of urinary urea to ammonium by the urease enzyme (Mobley et al., 1995). Beef cattle excrete more N in urine (approximately 60 to 80%) than in feces (approximately 20 to 40%) (Varel, 1997), and given that urinary N is more volatile than fecal N (Todd et al., 2013), there is interest in predicting route (urine and feces) of N excretion. Nitrogen excretion has been extensively studied in dairy cattle and numerous prediction equations have been published (Wilkerson et al., 1997; Jonker et al., 1998; Nennich et al., 2005; Nennich et al., 2006; Zhai et al., 2007; Spek et al., 2013). In beef cattle, several equations have been developed that predict total N excretion from the chemical composition of the diet and a description of the animal (Guo et al., 2004; Guo and Zoccarato, 2005; Yan et al., 2007); however, these are based on limited data. Only Waldrip et al. (2013) predicted urinary and fecal N excretion separately. However, those equations were developed from a limited number of studies, the majority of which were managed similarly to commercial practices (e.g., fed high concentrate diets with ionophores and given growth promoting implants), thus predictions of N excretion for beef cattle fed various diets needs further investigation.

3


The objective of this study was to develop equations to predict urinary and fecal N excretion from growing and finishing beef cattle fed a range of diets and to assess the accuracy of these equations.

MATERIALS AND METHODS

The project involved 3 phases: (1) a comprehensive database was constructed, (2) new prediction equations were developed using the database, and (3) the new equations were evaluated using cross-validation (Picard and Cook, 1984).

The Database

A database was constructed from 50 published studies (Appendix 1) conducted using beef cattle. The initial database included 206 observations from 872 cattle in research trials or commercial feedyards, but, following removal of outliers, only 180 observations from beef cattle in 49 studies were included in the database used for final analyses. The studies were identified using databases of published literature and keywords such as beef cattle, crude protein, digestibility, feedlot, nitrogen, nitrogen excretion, and so forth. Only studies in which the cattle production system was designed for meat rather than milk production were selected. The study was incorporated into the database if it supplied information on initial BW (kg), DMI (kg/d), dietary CP content (% of DM), N intake (g/d), and N excretion in urine and feces (g/d). Studies that specifically evaluated as treatments the effects of implants or hormones on N metabolism were not included in the database. A summary of the studies is presented in Table 1. 4


Within the dataset, the data were divided into 3 dietary CP concentration levels: (1) diets with CP content ≤ 9% of DM (low CP); (2) diets with CP content between 9 and 15% of DM (moderate CP); and (3) diets with high CP content > 15% of DM (high CP). When not reported, variables within a study were calculated from other variables provided. For example, when not reported, dietary CP content of the diet was calculated from DMI and N intake, and the proportion of urinary N in total N excretion was calculated by dividing urinary N excretion by total N excretion. The beef cattle used were of various ages (growing and finishing) and BW (101 to 626 kg) and from different breeds (Holstein, Crossbred, Nellore, Angus, Hereford, Angus-Hereford, Belgian white-blue, and Charolais). Diets fed to the cattle differed extensively. The majority of the studies were conducted in North America. The studies conducted in the United States with feedlot cattle fed high grain diets were based on processed corn (cracked, dry-rolled, steam-flaked, high moisture, and ground) (Bunting et al., 1989; Hill and Utley, 1989; Hill and West, 1991; Lapierre et al., 1992; Smith et al., 1992; Hill et al., 1996; Knaus et al., 1998, 2001, 2002; Bierman et al., 1999; Greenwood et al., 2001; Cole et al., 2003, 2006, 2011; Archibeque et al., 2006, 2007; Spiehs and Varel, 2009; Vasconcelos et al., 2009; Brake et al., 2010; Burciaga-Robles et al., 2010; Arias et al., 2012; Luebbe et al., 2012) accounting for 60 to 90% of DM, or sorghum (dry-rolled, steam-flaked) accounting for 77% of DM (Theurer et al., 2002). The forage-based (56 to 100% of DM) diets contained a variety of forages, including switchgrass hay, tallgrass-prairie hay, corn silage, alfalfa hay, gamagrass, tall fescue hay, and sudan hay (Archibeque et al., 2001, 2002; Wickersham et al., 2008 a,b; Taylor-Edwards et al., 2009; Waggoner et al., 2009a; Alvarez Almora et al., 2012; Drewnoski and Poore, 2012). The studies conducted in Canada with feedlot finishing diets were composed of 50 to 91% barley grain (dry-rolled and ground; Walter et al., 5


2012; Hünerberg et al., 2013b; Koenig and Beauchemin, 2013b) or barley grain supplemented with corn or wheat dried distillers grains plus solubles (Walter et al., 2012; Hünerberg et al., 2013a,b). The forage-based diets consisted of barley silage accounting for 54 to 55% of DM and 38 to 44% of DM as barley grain-based concentrates (dry-rolled, steam-rolled, and ground) (Hünerberg et al., 2013a; Koenig and Beauchemin, 2013a). In the Brazilian studies, the cattle were fed corn silage diets with concentrates containing ground corn and soybean meal (Gandra et al., 2011) or corn starch and cottonseed meal (Véras et al., 2007). Studies from western Europe offered silage (grass and corn alone or as mixtures) alone or with concentrates containing barley (dry-rolled), corn (steam-flaked), wheat (cracked), soybean meal, rapeseed meal, molassed sugar beet pulp, and fish meal (Rouzbehan et al., 1996; Fiems et al., 1997; Browne et al., 2005; Valkeners et al., 2008). In Southern European countries, the cattle consumed barley straw (chopped coarsely) supplemented with corn (ground), barley (ground), sunflower meal, and tapioca (Devant et al., 2000). In Asian countries, Japanese studies used high concentrate diets (up to 80% of DM) that contained corn, rye, wheat, wheat bran, corn gluten feed, soybean meal, rapeseed meal, beet pulp, and alfalfa meal along with mixed hay (20% of DM) (Mwenya et al., 2005). In India, a wheat straw-based diet was used along with barley, wheat bran, groundnut oil cake, and mustard oil cake (Giri et al., 2000). Rice straw was fed in Thailand, along with rice bran, coconut meal, and palm kernel meal (Huyen et al., 2012). The publications reported whole animal N balance from metabolism studies with individual animals as the experimental unit or from feedlot studies with penned cattle. There were differences among studies in the method of determining urinary and fecal N output. Some studies collected total urinary and fecal output directly (e.g., Fiems et al., 1997; Wickersham et al., 2008 a,b; Koenig and Beauchemin, 2013a,b); whereas other studies used markers to estimate 6


fecal N output (e.g., Cole et al., 2006; Gandra et al., 2011; Luebbe et al., 2012). In studies where urine was not collected, urinary N excretion was estimated in most cases as the difference between N intake and the sum of fecal N and retained N (e.g., Cole et al., 2003, 2006; Cole and Todd, 2009), although sometimes urine volume was estimated from creatinine (e.g., Devant et al., 2000).

Development of Prediction Equations

Correlation analysis was carried out to determine the best independent variables for predicting N excretion. Multiple regression analysis was conducted using initial BW (kg), DMI (kg/d), dietary CP content (% of DM), and N intake (g/d) as independent variables and N output (g/d) in urine or feces as the dependent variable. Similarly, multiple regression was conducted using the same independent variables plus apparent total tract N digestibility (%) and the proportion of urinary N in total N excretion as dependent variable. The MIXED regression model procedure (SAS Inst. Inc., Cary, NC) was used for these analyses. The independent variables included in the final equations were selected using a stepwise approach by manually removing non-significant variables (P > 0.05). Collinearities between variables were evaluated by variance inflation factors which are measures of how highly correlated each independent variable is with the other predictors in the equation. Large values of variance inflation factors (> 2.5) for a predictor imply large inflation of SE of regression coefficients due to this variable being in the equation. When multi-collinearity was identified the variable with the lowest correlation coefficient in predicting N excretion was removed. Univariate regression relationships with dietary CP content (% of DM) or N intake (g/d) as independent variable and N 7


output in urine and feces (g/d) or the proportion of urinary N in total N excretion as dependent variable were also developed using the MIXED regression model procedure (SAS Inst. Inc., Cary, NC). Study was included in the models as a random effect to account for differences among studies. Outliers in the dataset were identified as extreme values using Mahalanobis outlier analysis in JMP Statistical Software (SAS Institute, Cary, NC). Studentized residuals were then plotted vs. predicted values from the mixed regression analysis for urinary N, fecal N, and the proportion of urinary N. If the studentized residual for a particular observation was outside the range of -2.5 to 2.5 and the observation had been identified as an outlier, the treatment mean was removed from the dataset. In total, 26 observations were identified as outliers, leaving 49 studies and 180 observations for each variable in the final analysis. Equations were then compared based on root mean square error (RMSE) and Akaike’s information criterion (AIC). The equation with the smallest RMSE and AIC was chosen as the best fit model.

Validation of Prediction Equations

The robustness of the prediction equations was evaluated using cross-validation (Picard and Cook, 1984). The data were split randomly into 4 subsets with all data from a particular study in the same subset. Each subset was in turn left out and equations to predict N excretion in urine and feces and the proportion of urinary N in total N excretion were developed based on the remaining 3 subsets. These equations had the same variables as the newly developed final equations. Each equation was then used to predict the N excretion for the omitted subset. The 8


procedure was repeated for all subsets. Model performance was evaluated as the difference between predicted and observed values using univariate linear regression and indices including RMSE, mean absolute error (MAE), mean bias error (MBE), and index of agreement (IA) (Willmott, 1982). The RMSE, MAE, and MBE were calculated based on the mean difference between observed and equation-predicted values, which were calculated as: MAE = MBE =

𝑛

�𝑖=1 |P𝑖 −O𝑖 | 𝑛

n

�𝑖=1(P𝑖 −O𝑖 ) n

RMSE = [n−1 ∑𝑛𝑖=1(P𝑖 − O𝑖 )2 ]0.5

where Pi and Oi are the individual equation-predicted value and observed value, respectively, and n is the number of observations. Mean prediction error (MPE) was used to describe the overall prediction accuracy (Yan et al., 2007) and was determined by RMSE divided by the observed mean (Ellis et al., 2009). It is assumed that greater IA and r2 values and smaller RMSE, MAE, MBE, and MPE values represent a better fit equation.

RESULTS AND DISCUSSION

Animal Care and Use Committee approval was not required for this study because the data were obtained from published literature.

The Database

9


A summary of the studies comprising the dataset used for predicting N excretion in urine and feces and the proportion of urinary N in total N excretion is presented in Table 2 with data for mean, SD, and range for BW of animals, DMI, dietary CP content, N intake, apparent total tract N digestibility, and N excretion. There was a large range in initial BW of the beef cattle in the dataset, ranging from 101 to 626 kg with an average BW of 329 kg. Thus, the studies were conducted using cattle that ranged widely in maturity and represented growing, finishing and mature cattle. However, it should be noted that because treatment means rather than individual animal data were used, animal variability was not considered in the analysis. Dietary CP content in the dataset averaged 13.3% and ranged from 5.5 to 23.5%. The lowest CP content was for a diet of low quality prairie hay used in a study by Wickersham et al. (2008b) whereas the greatest CP content was for a diet containing a barley silage-based diet supplemented with barley grain and wheat dried distillers grains plus solubles (Hünerberg et al., 2013a). Average DMI was 6.62 kg/d, ranging from 2.92 to 10.70 kg/d. There was no difference in DMI between the high and moderate CP level groups (P = 0.89) or between the moderate and low CP level groups (P = 0.54) or between the high and low CP level groups (P = 0.63). As noted by Imaizumi et al. (2010), there are no conclusive effects of dietary CP content on DMI, but DMI is often lower when the CP content of the diet is below the minimum requirements for microbial growth in the rumen due to reduced ruminal fiber digestibility (Colmenero and Broderick, 2006). The lack of difference in DMI among the CP groupings was probably a reflection of confounding caused by differences in BW, dietary ingredients, and other variables. Apparent total tract N digestibility (Table 2) averaged 67.5% and varied from 46.6 to 86.9%, with the high dietary CP group being greater than the moderate dietary CP group (P < 10


0.01), which was in turn greater than the low dietary CP group (P < 0.01). Increasing dietary CP content increased apparent total tract N digestibility likely as a result of dilution of endogenous fecal N. Daily N intake (Table 2) ranged from 52 to 350 g/d, averaging 141 g/d. As expected, the greater the CP concentration in the diet, the more N consumed by cattle (Noftsger and St-Pierre, 2003). Nitrogen intake was positively correlated with CP content (r = 0.68; P < 0.001, data not shown). As intended, there were differences (P < 0.01) among the dietary CP groups for average N intake. The dataset showed a large range in urinary N excretion (Table 2) varying from 13.7 to 201.3 g/d, averaging 60.4 g/d. Just as N intake varied among the 3 dietary CP groupings, so did urinary N excretion (P < 0.01): high CP group > moderate CP group > low CP group (Table 2). Total N excretion ranged from 41.5 to 303.0 g/d, with the high dietary CP grouping > moderate dietary CP grouping > low dietary CP grouping (P < 0.01). Urinary N excretion increased with increasing CP concentration because increased dietary CP content and N intake generally lead to substantial increases in urinary N loss (Castillo et al., 2000). Daily fecal N excretion (Table 2) averaged 43.7 g/d, with a large range from 15.1 to 101.9 g/d. The least fecal N excretion was from Angus-Hereford crossbred heifers fed 1.43 × maintenance energy intake with a diet low in N (Smith et al., 1992), whereas the greatest fecal N excretion was from beef heifers fed a barley silage-based diet supplemented with barley grain and wheat dried distillers grains plus solubles (Hünerberg et al., 2013a). Daily fecal N excretion from cattle in the greatest CP level group differed from that of animals in the moderate CP level group (P < 0.001) or in the low CP level group (P < 0.001), but there was no difference between the moderate and low CP level groups (P = 0.25). 11


The proportion of urinary N in total N excretion (Table 2) averaged 0.55 with a range of 0.26 to 0.85. The proportion of urinary N in total N excretion from cattle in the high or moderate CP groups differed from that of animals in the low CP groups (P < 0.001). To our knowledge, Waldrip et al. (2013) is the only study to publish equations that separately predict N excretion in urine and feces for beef cattle. The equations developed by Waldrip et al. (2013) were based on 12 studies and 47 dietary treatments, representing mainly feedlot finishing cattle typical of North America. When these equations were evaluated using our larger dataset representing a wide range of production systems and an assortment of diets, prediction accuracy and precision were substantially lower than when applied to the original data, as would be expected (data not shown). Use of their equations should be restricted to feedlot cattle fed diets similar to those used in their development database (i.e., high grain feedlot diets). For that reason, we used the dataset to develop prediction equations that would be useful for cattle fed a broad range of diets.

Relationships between N Excretion in Urine and Feces and Dietary or Animal Variables

Pearson correlation coefficients were determined between N excretion and animal or dietary variables (Table 3). Urinary N excretion, total N excretion, and the proportion of urinary N in total N excretion were related positively (P < 0.001) to N intake, DMI, dietary CP content, apparent total tract N digestibility, and initial BW. Fecal N excretion was related positively (P < 0.001) to all variables, except apparent total tract N digestibility (P = 0.76). Relationships between N excretion in urine and feces and N intake (P < 0.001) are consistent with the studies of Yan et al. (2007) and Todd et al. (2013). The correlation coefficient for the proportion of 12


urinary N in total N excretion was greatest with apparent total N digestibility, followed by dietary CP content and N intake, whereas its relationship with initial BW and DMI was low. Because our objective was to develop more accurate N prediction equations, the significant variables were selected as predictors in equation development.

Prediction Equations for N Excretion in Urine

Mixed model regression equations for predicting urinary N excretion are presented in Table 4 (Eq. [1] to [2]). Univariate regression relationships between urinary N excretion and dietary CP content (Eq. [1]) or N intake (Eq. [2]) were established for the dataset. In both equations each variable had a significant effect (P < 0.05) on the relationship. Using N intake as the independent variable resulted in a better fit equation (AIC = 1254, Eq. [2]) compared with using CP content alone (AIC = 1343, Eq. [1]). Comparison of predicted values produced with Eq. [2] to observed values for urinary N excretion is shown in Fig. 1. There was no significant mean or slope bias. Regression analysis indicates good agreement (P < 0.001) between predictions and observations. The r2 value of 0.98 for this equation indicates great accuracy in predicting urinary N excretion. When urinary N excretion was singly regressed against N intake (Eq. [2]), the RMSE and AIC were relatively low (Table 4). Using dietary CP content as a predictor to estimate urinary N excretion (Eq. [1]) resulted in greater RMSE and AIC, and thus was less accurate than using N intake. The positive relationship between daily N intake and urinary total N excretion has been reported previously (Archibeque et al., 2001; Marini and Van Amburgh, 2003; Brake et al., 13


2010). Waldrip et al. (2013) showed urinary N excretion in beef cattle could be better predicted as a function of daily N intake (r2 = 0.82) than as a function of both N intake and DMI. The univariate regression of urinary N excretion on N intake in the present study had a slope similar to that reported by Waldrip et al. (2013), but the intercept was less (21.18 vs. 14.12).

Prediction Equations for N Excretion in Feces

Mixed model regression equations for predicting fecal N excretion are presented in Table 4 (Eq. [3] to [4]). Univariate linear relationships between fecal N excretion and dietary CP content (Eq. [3]) or N intake (Eq. [4]) were established. Each predictor had a significant effect (P < 0.05) on the relationship. Using N intake alone as an independent variable resulted in a better fit equation (AIC = 1088, Eq. [4]) compared with using CP content alone (AIC = 1178, Eq. [3]). Comparison of values predicted with Eq. [4] to observed values for fecal N excretion is shown in Fig. 2. Neither the mean (-0.0008 ± 0.2005) nor the slope bias (0.0151 ± 0.0150) was significant. Regression analysis indicates good agreement (P < 0.001) between predictions and observations. The r2 value of 0.96 for this equation indicates great accuracy in predicting fecal N excretion. For the univariate regression relationships between fecal N excretion and N intake (Eq. [4]), RMSE and AIC were relatively low. When fecal N excretion was singly regressed against dietary CP content (Eq. [3]), RMSE and AIC value were both greater than when the regression was based on N intake, indicating that predictions based on N intake would be preferred. Vasconcelos et al. (2009) reported a linear relationship between total fecal N and N intake in beef cattle. Waldrip et al. (2013) found that fecal N excretion could be modeled as a function of N intake using the equation: Fecal N, g/d = 0.154 (N intake, g/d) + 24.28. That 14


equation has a slope nearly identical to that in Eq. [4], but the intercept of the model of Waldrip et al. (2013) is about 1.5 times that of Eq. [4].

Prediction Equations for the Proportion of Urinary N in Total N Excretion

Mixed model regression equations for predicting the proportion of urinary N in total N excretion are presented in Table 4 (Eq. [5] to [7]). Univariate regression relationships between the proportion of urinary N in total N excretion and dietary CP content (Eq. [5]) or N intake (Eq. [6]) were established. Multivariate relationships between the proportion of urinary N in total N excretion and initial BW, N intake, DMI, dietary CP content, and apparent total tract N digestibility were also established, but only apparent total tract N digestibility (Eq. [7]) was found to be significant during the stepwise process. Each predictor in the equation had an effect (P < 0.05) on the relationship. Using apparent total tract N digestibility as an independent variable resulted in a better fit equation (AIC = -640, Eq. [7]) compared with using CP content (AIC = -532, Eq. [5]) or N intake (AIC = -497, Eq. [6]). Comparison of predicted values produced with Eq. [7] to observed values for the proportion of urinary N in total N excretion is shown in Fig. 3. There was no significant mean or slope bias. Regression analysis indicates good agreement (P < 0.001) between predictions and observations. The r2 value of 0.96 for this equation indicates great accuracy in predicting the proportion of urinary N in total N excretion. When the proportion of urinary N in total N excretion was regressed against apparent total tract N digestibility (Eq. [7]), AIC (-640) and RMSE (0.022) were lower than when dietary CP content (Eq. [5]) or N intake (Eq. [6]) were used as the sole predictor; thus, use of dietary CP content and N intake resulted in less accurate predictions than using apparent total tract N 15


digestibility. Hence, apparent total tract N digestibility proved to be the best predictor to estimate the proportion of urinary N in total N excretion compared with using dietary CP content or N intake alone. This finding is similar to that of Waldrip et al. (2013). They attempted to predict the partitioning of total excreted N into urinary N as a function of N intake or CP content; however, regression analysis indicated that the r2 values for such models were relatively low. Apparent total tract N digestibility was selected as the best predictor of proportion of N excretion in urine in our present study probably because apparent total tract N digestibility is related to dietary CP content. Greater dietary CP content and N digestibility would lead to a greater amount of N being absorbed as ammonia from the rumen or amino acids from the small intestine, and the excess N above requirements would be excreted in urine, thereby increasing the proportion of N excreted in urine.

Validation of Prediction Equations for N Excretion in Urine and Feces and the Proportion of Urinary N in Total N Excretion

The performance of the newly developed equations for predicting N excretion in urine and feces and the proportion of urinary N in total N excretion was validated using crossvalidation presented in Table 5. Between the equations of urinary N excretion, Eq. [2] offered the better prediction. The accuracy of predictions of urinary N excretion based on CP content (Eq. [1]) was low; the r2 and IA were lower and RMSE, MAE, MBE, and MPE greater than for using N intake in the urinary N excretion equation. Thus, it is recommended that Eq. [1], based solely on dietary CP content, be used to predict urinary N excretion only when N intake estimates are not available. 16


To predict fecal N excretion, Eq. [4] offered the better prediction when evaluated. The r2 and IA were greater while RMSE, MAE, and MPE were lower compared with Eq. [3]. For Eq. [4] the r2 value of 0.64 was slightly greater than the value (0.61) of Waldrip et al. (2013; Eq. 9 in their publication), but RMSE of 6.9 g/d observed by Waldrip et al. (2013) was somewhat less than our RMSE of 8.3 g/d. Predictions based on dietary CP content using Eq. [3] were not adequate for fecal N excretion by beef cattle, similar to the findings of Waldrip et al. (2013). Thus, use of Eq. [3] to predict fecal N excretion is not recommended. Given that urinary N excretion accounts for a large percentage in the total N excretion, equations for predicting the proportion of urinary N in total N excretion were also developed. To predict the proportion of urinary N in total N excretion, Eq. [7] offered the best prediction of the proportion of urinary N in total N excretion. Using apparent total tract N digestibility (Eq. [7]) to predict the proportion of urinary N in total N excretion in the present study improved the accuracy of prediction compared with using N intake (our Eq. 6 and Eq. 10 of Waldrip et al., 2013) or CP content (our Eq. 5 and Eq. 13 of Waldrip et al., 2013). Observed and predicted values produced from Eq. [7] agreed better (r2 = 0.68, IA = 0.87) than did values from Waldrip’s Eq. 10 and 13, and thus were more accurate. Predictions made by Eq. [5] based on CP content and Eq. [6] based on N intake resulted in lower r2 and IA but greater RMSE, MAE, and MPE compared with using apparent total tract N digestibility alone and thus were less accurate. So, Eq. [5] and Eq. [6] are not recommended due to inaccurate prediction.

Conclusions

17


This study used treatment means from a wide range of beef cattle studies to examine relationships between N excretion and a number of animal and dietary variables and evaluated the effects of these variables on N excretion of beef cattle fed a broad range of diets. Of all the single animal and dietary variables, N intake was the primary predictor of N excretion in urine and feces whereas apparent total tract N digestibility was the primary predictor of the proportion of urinary N in total N excretion. Other variables only marginally improved model predictions. Evaluation of the equations using cross-validation with the complete dataset indicated that the predictions of N excretion in urine and feces and the proportion of urinary N in total N excretion were accurate. Results indicated urinary N can be accurately and precisely predicted from N intake (Eq. [2]) and fecal N excretion can be best predicted using N intake (Eq. [4]). The proportion of urinary N in total N excretion was best predicted solely using apparent total tract N digestibility (Eq. [7]).

LITERATURE CITED

Archibeque, S. L., J. C. Burns, and G. B. Huntington. 2001. Urea flux in beef steers: effects of forage species and nitrogen fertilization. J. Anim. Sci. 79:1937-1943. Brake, D. W., E. C. Titgemeyer, M. L. Jones, and D. E. Anderson. 2010. Effect of nitrogen supplementation on urea kinetics and microbial use of recycled urea in steers consuming corn-based diets. J. Anim. Sci. 88:2729-2740. Burgos, S. A., J. G. Fadel, and E. J. DePeters. 2007. Prediction of ammonia emission from dairy cattle manure based on milk urea nitrogen: relation of milk urea nitrogen to urine urea nitrogen excretion. J. Dairy Sci. 90:5499-5508. Bussink, D. W., and O. Oenema. 1998. Ammonia volatilization from dairy farming systems in temperate areas: a review. Nutr. Cycl. Agroecosys. 51:19-33.

18


Castillo, A. R., E. Kebreab, D. E. Beever, and J. France. 2000. A review of efficiency of nitrogen utilisation in lactating dairy cows and its relationship with environmental pollution. J. Anim. Sci. 9:1-32. Cole, N. A., R. N. Clark, R. W. Todd, C. R. Richardson, A. Gueye, L. W. Greene, and K. McBride. 2005. Influence of dietary crude protein concentration and source on potential ammonia emissions from beef cattle manure. J. Amin. Sci. 83:722-731. Colmenero, J. J. O., and G. A. Broderick. 2006. Effect of dietary crude protein concentration on milk production and nitrogen utilization in lactating dairy cows. J. Dairy Sci. 89:17041712. Ellis, J. L., E. Kebreab, N. E. Odongo, K. Beauchemin, S. McGinn, J. D. Nkrumah, S. S. Moore, R. Christopherson, G. K. Murdoch, B. W. McBride, E. K. Okine, and J. France. 2009. Modeling methane production from beef cattle using linear and nonlinear approaches. J. Amin. Sci. 87:1334-1345. Groot Koerkamp, P. W. G., J. H. M. Metz, G. H. Uenk, V. R. Phillips, M. R. Holden, R. W. Sneath, J. L. Short, R. P. P. White, J. Hartung, J. Seedorf, M. Schröder, K. H. Linkert, S. Pedersen, H. Takai, J. O. Johnsen, and C. M. Wathes. 1998. Concentrations and emissions of ammonia in livestock buildings in northern europe. J. Agr. Eng. Res. 70:7995. Guo, K., A. Mimosi, R. Fortina, and I. Zoccarato. 2004. A computer model to predict the nitrogen excretion in growing-finishing cattle. Livest. Prod. Sci. 88:273-284. Guo, K., and I. Zoccarato. 2005. A dynamic model to predict the nitrogen excretion in growingfinishing cattle. Ecol. Model. 187:219-231. Hünerberg, M., S. M. McGinn, K. A. Beauchemin, E. K. Okine, O. M. Harstad, and T. A. McAllister. 2013a. Effect of dried distillers grains plus solubles on enteric methane emissions and nitrogen excretion from growing beef cattle. J. Anim. Sci. 91:2846-2857. Imaizumi, H., F. A. P. Santos, C. M. M. Bittar, P. S. Correia, and J. C. Martinez. 2010. Diet crude protein content and sources for lactating dairy cattle. Sci. Agric. 67:16-22. Jonker, J. S., R. A. Kohn, and R. A. Erdman. 1998. Using milk urea nitrogen to predict nitrogen excretion and utilization efficiency in lactating dairy cows. J. Dairy Sci. 81: 2681-2692. Marini, J. C., and M. E. Van Amburgh. 2003. Nitrogen metabolism and recycling in Holstein heifers. J. Amin. Sci. 81:545-552. Mobley, H. L., M. D. Island, and R. P. Hausinger. 1995. Molecular biology of microbial ureases. Microbiol. Rev. 59:451-480. 19


Nennich, T. D., J. H. Harrison, L. M. VanWieringen, D. Meyer, 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. Nennich, T. D., J. H. Harrison, L. M. VanWieringen, N. R. St-Pierre, R. L. Kincaid, M. A. Wattiaux, D. L. Davidson, and E. Block. 2006. Prediction and evaluation of urine and urinary nitrogen and mineral excretion from dairy cattle. J. Dairy Sci. 89:353-364. Noftsger, S., and N. R. St-Pierre. 2003. Supplementation of methionine and selection of highly digestible rumen undegradable protein to improve nitrogen efficiency for milk production. J. Dairy Sci. 86:958-969. Picard, R. R., and R. D. Cook. 1984. Cross-validation of regression models. J. Am. Stat. Assoc. 79:575-583. Smith, S. B., R. L. Prior, L. J. Koong, and H. J. Mersmann. 1992. Nitrogen and lipid metabolism in heifers fed at increasing levels of intake. J. Amin. Sci. 70:152-160. Todd, R. W., N. A. Cole, H. M. Waldrip, and R. M. Aiken. 2013. Arrhenius equation for modeling feedyard ammonia emissions using temperature and diet crude protein. J. Environ. Qual. 42:666-671. Varel, V. H. 1997. Use of urease inhibitors to control nitrogen loss from livestock waste. Bioresour. Technol. 62:11-17. Waldrip, H. M., R. W. Todd, and N. A. Cole. 2013. Prediction of nitrogen excretion by beef cattle: A meta-analysis. J. Anim. Sci. 91:4290-4302. Wickersham, T. A., E. C. Titgemeyer, R. C. Cochran, E. E. Wickersham, and E. S. Moore. 2008b. Effect of frequency and amount of rumen-degradable intake protein supplementation on urea kinetics and microbial use of recycled urea in steers consuming low-quality forage. J. Anim. Sci. 86:3089-3099. Wilkerson, V. A., D. R. Mertens, and D. P. Casper. 1997. Prediction of excretion of manure and nitrogen by Holstein dairy cattle. J. Dairy Sci. 80:3193-3204. Willmott, C. J. 1982. Some comments on the evaluation of model performance. Bulletin of the American Meteorological Society 63:1309-1313. Yan, T., J. P. Frost, T. W. J. Keady, R. E. Agnew, and C. S. Mayne. 2007. Prediction of nitrogen excretion in feces and urine of beef cattle offered diets containing grass silage. J. Amin. Sci. 85:1982-1989. Zhai, S., J. Liu, Y. Wu, and J. Ye. 2007. Predicting urinary nitrogen excretion by milk urea nitrogen in lactating Chinese Holstein cows. Anim. Sci. J. 78:395-399. 20


Appendix 1 Alvarez Almora, E. G., G. B. Huntington, and J. C. Burns. 2012. Effects of supplemental urea sources and feeding frequency on ruminal fermentation, fiber digestion, and nitrogen balance in beef steers. Anim. Feed Sci. Technol. 171:136-145. Archibeque, S. L., J. C. Burns, and G. B. Huntington. 2001. Urea flux in beef steers: effects of forage species and nitrogen fertilization. J. Anim. Sci. 79:1937-1943. Archibeque, S. L., J. C. Burns, and G. B. Huntington. 2002. Nitrogen metabolism of beef steers fed endophyte-free tall fescue hay: effects of ruminally protected methionine supplementation. J. Anim. Sci. 80:1344-1351. Archibeque, S. L., H. C. Freetly, N. A. Cole, and C. L. Ferrell. 2007. The influence of oscillating dietary protein concentrations on finishing cattle. II. Nutrient retention and ammonia emissions. J. Anim. Sci. 85:1496-1503. Archibeque, S. L., D. N. Miller, H. C. Freetly, and C. L. Ferrell. 2006. Feeding high-moisture corn instead of dry-rolled corn reduces odorous compound production in manure of finishing beef cattle without decreasing performance. J. Anim. Sci. 84:1767-1777. Arias, R. P., L. J. Unruh-Snyder, E. J. Scholljegerdes, A. N. Baird, K. D. Johnson, D. Buckmaster, R. P. Lemenager, and S. L. Lake. 2012. Effects of feeding corn modified wet distillers grain plus solubles co-ensiled with direct-cut forage on feedlot performance, carcass characteristics, and diet digestibility of finishing steers. J. Anim. Sci. 90:35743583. Bierman, S., G. E. Erickson, T. J. Klopfenstein, R. A. Stock, and D. H. Shain. 1999. Evaluation of nitrogen and organic matter balance in the feedlot as affected by level and source of dietary fiber. J. Anim. Sci. 77:1645-1653. Brake, D. W., E. C. Titgemeyer, M. L. Jones, and D. E. Anderson. 2010. Effect of nitrogen supplementation on urea kinetics and microbial use of recycled urea in steers consuming corn-based diets. J. Anim. Sci. 88:2729-2740. Browne, E. M., D. T. Juniper, M. J. Bryant, and D. E. Beever. 2005. Apparent digestibility and nitrogen utilisation of diets based on maize and grass silage fed to beef steers. Anim. Feed Sci. Technol. 119:55-68. Bunting, L. D., J. A. Boling, and C. T. MacKown. 1989. Effect of dietary protein level on nitrogen metabolism in the growing bovine: I. nitrogen recycling and intestinal protein supply in calves. J. Anim. Sci. 67:810-819.

21


Burciaga-Robles, L. O., C. R. Krehbiel, D. L. Step, B. P. Holland, C. J. Richards, M. A. Montelongo, A. W. Confer, and R. W. Fulton. 2010. Effects of exposure to calves persistently infected with bovine viral diarrhea virus type 1b and Mannheimia haemolytica challenge on animal performance, nitrogen balance, and visceral organ mass in beef steers. J. Anim. Sci. 88:2179-2188. Cole, N. A., P. J. Defoor, M. L. Galyean, G. C. Duff, and J. F. Gleghorn. 2006. Effects of phasefeeding of crude protein on performance, carcass characteristics, serum urea nitrogen concentrations, and manure nitrogen of finishing beef steers. J. Anim. Sci. 84:3421-3432. Cole, N. A., L. W. Greene, F. T. McCollum, T. Montgomery, and K. McBride. 2003. Influence of oscillating dietary crude protein concentration on performance, acid-base balance, and nitrogen excretion of steers. J. Anim. Sci. 81:2660-2668. Cole, N. A., K. McCuistion, L. W. Greene, and F. T. McCollum. 2011. Effects of concentration and source of wet distillers grains on digestibility of steam-flaked corn-based diets fed to finishing steers Prof. Anim. Sci. 27:302-311. Cole, N. A., and R. W. Todd. 2009. Nitrogen and phosphorus balance of beef cattle feedyards. In: Proceedings of the Texas Animal Manure Management Issues Conference, September 29-30, Round Rock, Texas:17-24. Devant, M., A. Ferret, J. Gasa, S. Calsamiglia, and R. Casals. 2000. Effects of protein concentration and degradability on performance, ruminal fermentation, and nitrogen metabolism in rapidly growing heifers fed high-concentrate diets from 100 to 230 kg body weight. J. Anim. Sci. 78:1667-1676. Drewnoski, M. E., and M. H. Poore. 2012. Effects of supplementation frequency on ruminal fermentation and digestion by steers fed medium-quality hay and supplemented with a soybean hull and corn gluten feed blend. J. Anim. Sci. 90:881-891. Eisemann, J. H., A. C. Hammond, T. S. Rumsey, and D. E. Bauman. 1989. Nitrogen and protein metabolism and metabolites in plasma and urine of beef steers treated with somatotropin. J. Anim. Sci. 67:105-115. Fiems, L. O., B. G. Cottyn, C. V. Boucqué, D. F. Bogaerts, C. Van Eenaeme, and J. M. Vanacker. 1997. Effect of beef type, body weight and dietary protein content on voluntary feed intake, digestibility, blood and urine metabolites and nitrogen retention. J. Anim. Physiol. Anim. Nutr. 77:1-9. Gandra, J. R., J. E. Freitas Jr, R. V. Barletta, M. M. Filho, L. U. Gimenes, F. G. Vilela, P. S. Baruselli, and F. P. Rennó. 2011. Productive performance, nutrient digestion and metabolism of Holstein (Bos taurus) and Nellore (Bos taurus indicus) cattle and Mediterranean Buffaloes (Bubalis bubalis) fed with corn-silage based diets. J. Livest. Sci. 140:283-291. 22


Giri, S. S., A. Sahoo, and N. N. Pathak. 2000. Feed intake, digestibility, plane of nutrition and live weight gain by crossbred growing bulls fed on grainless diets containing different nitrogen sources. Anim. Feed Sci. Technol. 83:195-203. Greenwood, R. H., E. C. Titgemeyer, G. L. Stokka, J. S. Drouillard, and C. A. Löest. 2001. Effects of L-carnitine on nitrogen retention and blood metabolites of growing steers and performance of finishing steers. J. Anim. Sci. 79:254-260. Hill, G. M., G. L. Newton, M. N. Streeter, W. W. Hanna, P. R. Utley, and M. J. Mathis. 1996. Digestibility and utilization of pearl millet diets fed to finishing beef cattle. J. Anim. Sci. 74:1728-1735. Hill, G. M., and P. R. Utley. 1989. Digestibility, protein metabolism and ruminal degradation of beagle 82 triticale and kline barley fed in corn-Based cattle diets. J. Anim. Sci. 67:17931804. Hill, G. M., and J. W. West. 1991. Rumen protected fat in kline barley or corn diets for beef cattle: digestibility, physiological, and feedlot responses. J. Anim. Sci. 69:3376-3388. Hünerberg, M., S. M. McGinn, K. A. Beauchemin, E. K. Okine, O. M. Harstad, and T. A. McAllister. 2013a. Effect of dried distillers grains plus solubles on enteric methane emissions and nitrogen excretion from growing beef cattle. J. Anim. Sci. 91:2846-2857. Hünerberg, M., S. M. McGinn, K. A. Beauchemin, E. K. Okine, O. M. Harstad, and T. A. McAllister. 2013b. Effect of dried distillers’ grains with solubles on enteric methane emissions and nitrogen excretion from finishing beef cattle. Can. J. Anim. Sci. 93:373385. Huyen, N. T., M. Wanapat, and C. Navanukraw. 2012. Effect of Mulberry leaf pellet (MUP) supplementation on rumen fermentation and nutrient digestibility in beef cattle fed on rice straw-based diets. Anim. Feed Sci. Technol. 175:8-15. Knaus, W. F., D. H. Beermann, P. J. Guiroy, M. L. Boehm, and D. G. Fox. 2001. Optimization of rate and efficiency of dietary nitrogen utilization through the use of animal byproducts and(or) urea and their effects on nutrient digestion in Holstein steers. J. Anim. Sci. 79:753-760. Knaus, W. F., D. H. Beermann, T. F. Robinson, D. G. Fox, and K. D. Finnerty. 1998. Effects of a dietary mixture of meat and bone meal, feather meal, blood meal, and fish meal on nitrogen utilization in finishing Holstein steers. J. Anim. Sci. 76:1481-1487. Knaus, W. F., D. H. Beermann, L. O. Tedeschi, M. Czajkowski, D. G. Fox, and J. B. Russell. 2002. Effects of urea, isolated soybean protein and blood meal on growing steers fed a corn-based diet. Anim. Feed Sci. Technol. 102:3-14. 23


Koenig, K. M., and K. A. Beauchemin. 2013a. Nitrogen metabolism and route of excretion in beef feedlot cattle fed barley-based backgrounding diets varying in protein concentration and rumen degradability. J. Anim. Sci. 91:2295-2309. Koenig, K. M., and K. A. Beauchemin. 2013b. Nitrogen metabolism and route of excretion in beef feedlot cattle fed barley-based finishing diets varying in protein concentration and rumen degradability. J. Anim. Sci. 91:2310-2320. Lapierre, H., H. F. Tyrrell, C. K. Reynolds, T. H. Elsasser, P. Gaudreau, and P. Brazeau. 1992. Effects of growth hormone-releasing factor and feed intake on energy metabolism in growing beef steers: whole-body energy and nitrogen metabolism. J. Anim. Sci. 70:764772. Luebbe, M. K., J. M. Patterson, K. H. Jenkins, E. K. Buttrey, T. C. Davis, B. E. Clark, F. T. McCollum, N. A. Cole, and J. C. MacDonald. 2012. Wet distillers grains plus solubles concentration in steam-flaked-corn-based diets: Effects on feedlot cattle performance, carcass characteristics, nutrient digestibility, and ruminal fermentation characteristics. J. Anim. Sci. 90:1589-1602. Mwenya, B., C. Sar, B. Santoso, T. Kobayashi, R. Morikawa, K. Takaura, K. Umetsu, S. Kogawa, K. Kimura, H. Mizukoshi, and J. Takahashi. 2005. Comparing the effects of β14 galacto-oligosaccharides and l-cysteine to monensin on energy and nitrogen utilization in steers fed a very high concentrate diet. Anim. Feed Sci. Technol. 118:19-30. Rouzbehan, Y., H. Galbraith, J. H. Topps, and J. Rooke. 1996. Response of growing steers to diets containing big bale silage and supplements of molassed sugar beet pulp with and without white fish meal. Anim. Feed Sci. Technol. 62:151-162. Smith, S. B., R. L. Prior, L. J. Koong, and H. J. Mersmann. 1992. Nitrogen and lipid metabolism in heifers fed at increasing levels of intake. J. Anim. Sci. 70:152-160. Spiehs, M. J., and V. H. Varel. 2009. Nutrient excretion and odorant production in manure from cattle fed corn wet distillers grains with solubles. J. Anim. Sci. 87:2977-2984. Taylor-Edwards, C. C., N. A. Elam, S. E. Kitts, K. R. McLeod, D. E. Axe, E. S. Vanzant, N. B. Kristensen, and D. L. Harmon. 2009. Influence of slow-release urea on nitrogen balance and portal-drained visceral nutrient flux in beef steers. J. Anim. Sci. 87:209-221. Theurer, C. B., G. B. Huntington, J. T. Huber, R. S. Swingle, and J. A. Moore. 2002. Net absorption and utilization of nitrogenous compounds across ruminal, intestinal, and hepatic tissues of growing beef steers fed dry-rolled or steam-flaked sorghum grain. J. Anim. Sci. 80:525-532.

24


Valkeners, D., A. Théwis, M. Van Laere, and Y. Beckers. 2008. Effect of rumen-degradable protein balance deficit on voluntary intake, microbial protein synthesis, and nitrogen metabolism in growing double-muscled Belgian Blue bulls fed corn silage-based diet. J. Anim. Sci. 86:680-690. Vasconcelos, J. T., N. A. Cole, K. W. McBride, A. Gueye, M. L. Galyean, C. R. Richardson, and L. W. Greene. 2009. Effects of dietary crude protein and supplemental urea levels on nitrogen and phosphorus utilization by feedlot cattle. J. Anim. Sci. 87:1174-1183. Véras, R. M. L., S. d. C. Valadares Filho, R. F. D. Valadares, L. N. Rennó, P. V. R. Paulino, and M. A. d. Souza. 2007. Balanço de compostos nitrogenados e estimativa das exigências de proteína de mantença de bovinos Nelore de três condições sexuais. Rev. Bras. Zootecn. 36:1212-1217. Waggoner, J. W., C. A. Löest, C. P. Mathis, D. M. Hallford, and M. K. Petersen. 2009a. Effects of rumen-protected methionine supplementation and bacterial lipopolysaccharide infusion on nitrogen metabolism and hormonal responses of growing beef steers. J. Anim. Sci. 87:681-692. Waggoner, J. W., C. A. Löest, J. L. Turner, C. P. Mathis, and D. M. Hallford. 2009b. Effects of dietary protein and bacterial lipopolysaccharide infusion on nitrogen metabolism and hormonal responses of growing beef steers. J. Anim. Sci. 87:3656-3668. Walter, L. J., T. A. McAllister, W. Z. Yang, K. A. Beauchemin, M. He, and J. J. McKinnon. 2012. Comparison of wheat or corn dried distillers grains with solubles on rumen fermentation and nutrient digestibility by feedlot heifers. J. Anim. Sci. 90:1291-1300. Wickersham, T. A., E. C. Titgemeyer, R. C. Cochran, E. E. Wickersham, and D. P. Gnad. 2008a. Effect of rumen-degradable intake protein supplementation on urea kinetics and microbial use of recycled urea in steers consuming low-quality forage. J. Anim. Sci. 86: 3079-3088. Wickersham, T. A., E. C. Titgemeyer, R. C. Cochran, E. E. Wickersham, and E. S. Moore. 2008b. Effect of frequency and amount of rumen-degradable intake protein supplementation on urea kinetics and microbial use of recycled urea in steers consuming low-quality forage. J. Anim. Sci. 86:3089-3099. Wray, M. I., W. M. Beeson, and T. W. Perry. 1980. Effect of soybean, feather and hair meal protein on dry matter, energy and nitrogen utilization by growing steers. J. Anim. Sci. 50: 581-589.

25


Table 1. General review of the studies used to predict N excretion Source1 Wray et al., 1980 Bunting et al., 1989 Eisemann et al., 1989 Hill and Utley, 1989 Hill and West, 1991 Lapierre et al., 1992 Smith et al., 1992 Hill et al., 1996 Rouzbehan et al., 1996 Fiems et al., 1997 Knaus et al., 1998 Devant et al., 2000 Giri et al., 2000 Archibeque et al., 2001 Greenwood et al., 2001 Knaus et al.,

Country

Experimental Design (n)2

USA

Switchback (12)

USA USA

Completely randomized (8) Balanced single reversal (12)

Diet composition3 Whole corn, cane molasses, and ground corn cobs Corn, soybean meal, molasses, and cottonseed hulls Cracked corn, soybean meal, molasses, cottonseed hulls, and wheat straw Cracked corn, triticale, barley, and peanut hulls Cracked corn, soybean meal, barley, and peanut hulls

USA

Latin square (6)

USA

Latin square (6)

USA

Split plot (6)

Cracked corn, soybean meal, and alfalfa hay

Completely randomized (25) Latin square (6); cross-over (8) Completely randomized block (24)

Concentrate (corn, molasses, soybean meal, barley, sugar beet pulp) with grass silage Cracked corn, sorghum, pearl millet, soybean meal, and peanut hulls

USA USA UK

Bale silage and concentrate

Belgium

Cross-over (12)

Maize silage and concentrate (sugarbeet pulp, soybean meal, and coconut cake)

USA

Latin square (4)

Ground corn, soybean meal, and grass hay

Completely randomized (20) Completely randomized block (25)

Concentrates (corn, barley, tapioca, sunflower and soybean meal) with barley straw

Spain India USA USA USA

Latin square (8) Incomplete Latin square (7) Latin square (4)

Wheat straw with concentrate Gamagrass and switchgrass hay with cornbased vitamin and mineral supplement Rolled corn, molasses, blood meal, corn gluten meal, soybean meal, and alfalfa Cracked corn and soybean, meat and bone,

26


Forage level in diet

Breed

Initial BW, kg

Method to estimate N4

20%

Hereford, Angus

241 to 340

*

30%

Angus

234 avg

ƪ

41%

Hereford × Angus

290 to 300

*

23%

Crossbred

229 avg

*

21%

British

236 to 266

*

18%

Hereford × Angus

326 to 349

*

20%

Angus × Hereford

265 avg

*

14 to 15%

British

213 to 251

*

85 to 100%

Crossbred

285 to 290

*

50%

Belgian White/blue (normal, doublemuscled)

430 to 632

*

9%

Holstein

407 avg

*

6.5%

Crossbred

96 to 106

£

54 to 58%

Crossbred

305 to 310

*

90%

Angus

202 to 232

*

10%

Angus × cross

216 avg

*

14 to 16%

Holstein

259 avg

*

2001 Archibeque et al., 2002 Knaus et al., 2002 Theurer et al., 2002 Cole et al., 2003 Browne et al., 2005

feather, blood meal with grass hay USA

Latin square (8)

Fescue hays and corn based supplement

USA

Latin square (4)

Corn with mixed grass hay

USA

Cross-over (7)

USA

Completely randomized (27)

UK

Cross-over (7)

Mwenya et al., 2005

Japan

Latin square (4)

Archibeque et al., 2006

USA

Cross-over (8)

Cole et al., 2006

USA

Completely randomized blocked (318)

USA

Latin square (8)

Brazil

Latin square (12)

Belgium

Latin square (6)

USA

Latin square (5)

USA

Latin square (5)

Archibeque et al., 2007 Véras et al., 2007 Valkeners et al., 2008 Wickersham et al., 2008a Wickersham et al., 2008b Cole and Todd, 2009 Spiehs and Varel, 2009 TaylorEdwards et al., 2009 Vasconcelos et al., 2009

USA USA

Completely randomized (10) Completely randomized block (24)

90%

Angus

232 to 274

*

27 to 29%

Holstein

251 avg

*

15%

Hereford × Angus

329 to 367

*

10%

Crossbred

400 to 410

§

79 to 83%

Holstein

390 to 432

*

20%

Holstein

280 to 302

*

10%

Charolais

425 to 469

*

10.50%

Crossbred

310 to 320

§

10%

Charolais × cross

294 to 336

*

75%

Nellore

255 to 285

ƚ

60%

Belgian Blue (double muscled)

292 to 316

*

Grass-prairie hay and protein supplement

96 to 98%

Angus × Hereford

259 to 297

*

Grass-prairie hay and RDP

94 to 100%

Angus × Hereford

323 to 409

*

6 to 12%

Crossbred

427 to 443

ǁ

Dry-rolled corn, soybean meal, and CDGS with alfalfa hay

11%

Crossbred

437 to 469

*

Sorghum grain, cottonseed meal, cane molasses, and alfalfa hay Steam-flaked corn, cottonseed hulls, and Sudan hay Grass or corn silage or a mixture of the two (1:2 or 2:1) with concentrate Concentrate (corn, rye, wheat, wheat bran, corn gluten, and soybean meal ) with mixed hay Dry-rolled or high moisture corn, soybean meal, and corn silage Steam-flaked corn, molasses, fat, urea, conttonseed meal, and alfalfa hay Dry-rolled corn, soybean meal, molasses, and corn silage Corn silage with concentrate (corn starch, urea, and cottonseed meal) Corn silage, sugar beet pulp, soybean meal, linseed expeller, barley, corn, and urea

Steam-flaked corn and alfalfa hay

USA

Cross-over (10)

Corn silage with corn-based vitamin and mineral supplement and urea

88%

Holstein, Angus

236 to 367

*

USA

Completely randomized (54)

Steam-flaked corn, molasses, fat, cottonseed meal, urea, and alfalfa

10%

Crossbred

315 to 353

ƚ

27


Waggoner et al., 2009a

USA

Waggoner et al., 2009b

USA

Brake et al., 2010 BurciagaRobles et al., 2010 Cole et al., 2011 Gandra et al., 2011 Arias et al., 2012 Alvarez Almora et al., 2012 Drewnoski and Poore, 2012 Huyen et al., 2012 Luebbe et al., 2012 Walter et al., 2012 Hünerberg et al., 2013a Hünerberg et al., 2013b Koenig and Beauchemin, 2013a Koenig and Beauchemin, 2013b

USA USA USA Brazil

Completely randomized block (20) Completely randomized block (24) Latin square (6) Completely randomized block (24) Latin square (10) Completely randomized (10)

Alfalfa hay, corn silage, sorghum, sudan hay with cracked corn, molasses and soybean meal

57%

Angus × crossbred

256 to 268

*

Wheat, corn, molasses, urea, soybean hulls, sorghum silage, and alfalfa hay

48%

Angus × crossbred

247 to 253

*

Dry-rolled corn, cane molasses, DDGS, urea, and alfalfa hay

10%

British

211 to 277

*

Corn dent, dried corn distillers grain, and alfalfa hay

6%

Angus × crossbred

283 to 345

*

Steam-falked corn, cottonseed meal, urea, CDGS, NMDGS, and cottonseed hulls

8%

Crossbred

237 to 267

*

75.63%

Nellore

408 avg

ǂ

Corn silage with concentrate

USA

Latin square (4)

Cracked corn, WDGS, and DDGS with coensiled, haylage, and corn silage

12 to 35%

Crossbred

525 to 587

*

USA

Cross-over (8)

Switchgrass hay with supplements (soybean hulls, corn, and soybean meal)

60%

Crossbred

239 to 284

*

USA

Latin square (6)

Fescue hay with a soybean hull and corn gluten feed blend

56 to 100%

Angus

344 to 380

*

Thailand

Latin square (4)

Rice straw with concentrate

76%

Crossbred

405 to 435

*

USA

Latin square (6)

10%

Crossbred

463 to 499

ƚ

Canada

Latin square (5)

6%

Hereford

414 to 426

*

Canada

Latin square (8)

55%

Crossbred

354 to 424

*

Canada

Latin square (16)

8%

Crossbred

488 to 570

*

Canada

Latin square (4)

Barley silage, barley grain, and vitamin and mineral supplement

54%

Angus

464 to 494

*

Canada

Latin square (4)

barley grain, vitamin and mineral supplement, and barley silage

9%

Angus

546 to 582

*

Dry-rolled corn, steam-flaked corn, molasses, NMDGS, and alfalfa hay Barley, WDDGS, and CDDGS with barley silage Barley silage with barley grain, CDDGS, and WDDGS Barley grain, CDDGS, and WDDGS with barley silage

28


1

In chronological order.

2

n = number of cattle included in this study.

3

CDGS, wet corn distillers grains with solubles; NMDGS, wet corn and sorghum distillers grains with solubles; WDGS, wet

distillers grains with solubles; DDGS, dried distillers grains with solubles; CDDGS, corn dried distillers grains with solubles; WDDGS, wheat dried distillers grains with solubles. 4

* = Collected total urine and total feces; ƪ = Collected total urine and partial feces. Feces production was estimated with external

marker; £ = Collected partial urine and partial feces. Urine volume was estimated from creatinine. Feces production was estimated with external marker; § = Collected partial feces and estimated feces production with an internal marker. Urinary N was estimated as difference between N intake and fecal and retained N; ƚ = Collected total urine and partial feces. Total feces production was estimated with an internal marker; ǁ = Urinary N was estimated as difference between N intake and fecal and retained N. Feces production and fecal N excretion were calculated form nutrient intake and digestibility; ǂ = Collected partial urine and partial feces. Total urinary volume was estimated by dividing daily excretion of creatinine by creatinine concentration in urine. Feces production was estimated with an external marker.

29


Table 2. Summary of animal and dietary variables and N intake and output for the dataset used to predict N excretion Item BW, kg

CP, % of DM

DMI, kg/d

TTND, %

N intake, g/d

Urinary N, g/d

Fecal N, g/d

Total N excretion, g/d

Urinary N /total N excretion

Grouping

Mean

SD

Minimum

Maximum

CV

n

Complete dataset CP ≤ 0.09 0.09 < CP ≤ 0.15 0.15 < CP Complete dataset CP ≤ 0.09 0.09 < CP ≤ 0.15 0.15 < CP Complete dataset CP ≤ 0.09 0.09 < CP ≤ 0.15 0.15 < CP Complete dataset CP ≤ 0.09 0.09 < CP ≤ 0.15 0.15 < CP

329 326 322 353 13.3 7.5c 12.6b 17.6a 6.62 6.94 6.59 6.64 67.5 54.4c 67.6b 70.7a

100 61 93 130 3.0 1.3 1.4 2.4 1.8 1.3 1.7 2.3 7.0 5.3 6.5 4.6

101 217 101 101 5.5 5.5 9.1 15.2 2.92 4.14 2.92 3.47 46.6 46.6 50.2 60.6

626 420 626 626 23.5 9.0 15.0 23.5 10.70 8.70 10.70 10.12 86.9 65.3 86.9 82.5

0.31 0.19 0.29 0.37 0.23 0.17 0.11 0.14 0.28 0.19 0.26 0.34 0.10 0.10 0.10 0.06

180 11 131 38 180 11 131 38 180 11 131 38 180 11 131 38

Complete dataset

141

55

52

350

0.39

180

CP ≤ 0.09 0.09 < CP ≤ 0.15 0.15 < CP

c

83 134b 185a

20 42 72

57 52 86

124 240 350

0.24 0.31 0.39

11 131 38

Complete dataset

60.4

34.9

13.7

201.3

0.58

180

CP ≤ 0.09 0.09 < CP ≤ 0.15 0.15 < CP

c

26.5 57.0b 82.2a

10.0 28.4 46.5

14.0 13.7 30.0

48.9 181.8 201.3

0.38 0.50 0.57

11 131 38

Complete dataset

43.7

13.9

15.1

101.9

0.32

180

CP ≤ 0.09 0.09 < CP ≤ 0.15 0.15 < CP

36.8b 41.5b 53.2a

5.5 11.1 19.1

27.7 15.1 21.5

43.9 71.1 101.9

0.15 0.27 0.36

11 131 38

Complete dataset

104.2

44.5

41.5

303.0

0.43

180

CP ≤ 0.09 0.09 < CP ≤ 0.15 0.15 < CP

63.3c 98.5b 135.4a

14.2 34.3 61.4

41.7 41.5 69.2

92.0 213.2 303.0

0.22 0.35 0.45

11 131 38

Complete dataset

0.55

0.11

0.26

0.85

0.20

180

CP ≤ 0.09 0.09 < CP ≤ 0.15 0.15 < CP

b

0.06 0.11 0.10

0.34 0.26 0.33

0.53 0.85 0.74

0.15 0.19 0.17

11 131 38

0.41 0.55a 0.59a

n = number of observations; TTND = apparent total tract N digestibility; a-cMeans within an item with different superscripts differ (P < 0.01). 30


Table 3. Pearson correlation coefficients (r) for relationships between N output variables and dietary and animal factors for the dataset (n = 180) Parameter

Urinary N excretion

Fecal N excretion

Urinary N/Total N

Total N

N intake

0.878

0.806

0.488

0.941

DMI

0.652

0.667

0.283

0.720

Diet CP

0.645

0.501

0.510

0.663

TTND

0.624

0.023

0.834

0.498

Initial BW

0.537

0.553

0.283

0.594

Note: TTND = apparent total tract N digestibility; all values were significant at P ≤ 0.001 except for TTND and fecal N excretion (P = 0.76). Table 4. Prediction of N excretion in urine, feces and the proportion of urine N in total N excretion using N intake, dietary CP concentration, or apparent total tract N digestibility for the dataset (n = 180) Equation number

Independent variable1

Intercept

SE

Slope

SE

RMSE2

AIC3

Equations for Urinary N, g/d [1]

CP

-22.00

3.84

6.04

0.41

4.91

1343

[2]

NI

-14.12

2.71

0.51

0.02

4.07

1254

Equations for Fecal N, g/d [3]

CP

19.68

2.66

1.81

0.24

3.45

1178

[4]

NI

15.82

1.88

0.20

0.01

2.68

1088

Equations for Proportion of Urinary N in Total N excretion [5]

CP

0.328

0.03

0.016

0.002

0.027

-532

[6]

NI

0.402

0.023

0.001

0.0001

0.031

-497

31


[7] 1

TTND

-0.162

0.04

0.010

0.0006

0.022

-640

NI = N intake (g/d), CP = dietary crude protein (% of DM), TTND = apparent total tract N

digestibility (%). 2

RMSE = root mean square error (g/d).

3

AIC = Akaike’s information criterion, a measure of the relative goodness of fit of the model; a

smaller value means a better model fit.

32


Table 5. Evaluation of equations developed to predict N excretion in beef cattle using cross-validation with the complete dataset

N excretion1

Regression2

Difference indices3

Equation number O

SD

P

SD

Slope

SE

Intercept

SE

r2

RMSE

MAE

MBE

MPE

IA

Equations for Urinary N, g/d [1]

60

35

59

19

0.34

0.03

38.87

2.24

0.38

27.5

20.64

-1.20

0.45

0.71

[2]

60

35

58

28

0.71

0.03

15.24

2.04

0.77

17.1

12.10

-2.32

0.28

0.92

Equations for Fecal N, g/d [3]

44

14

44

6

0.19

0.03

35.96

1.30

0.20

12.4

9.84

0.44

0.28

0.54

[4]

44

14

44

11

0.63

0.04

16.03

1.61

0.64

8.3

6.23

-0.09

0.19

0.88

Equations for Proportion of Urinary N in Total N excretion [5]

0.55

0.11

0.55

0.05

0.24

0.03

0.41

0.02

0.25

0.09

0.07

-0.01

0.17

0.61

[6]

0.55

0.11

0.55

0.06

0.25

0.04

0.41

0.02

0.22

0.10

0.08

-0.01

0.17

0.63

[7]

0.55

0.11

0.54

0.08

0.57

0.03

0.23

0.02

0.68

0.06

0.05

-0.01

0.12

0.87

1

O, means of observed value; P, means of predicted value.

2

Significant at P < 0.001. r2, coefficient of the determination for linear regression between predicted and observed value. 33


3

RMSE = root mean square error (g/d); MAE = mean absolute error (g/d); MBE = mean bias error (g/d); MPE = RMSE/O; IA =

index of agreement.

34


Figure Captions

Figure 1. Relationship between (a) predicted and observed urinary nitrogen excretion (g/d) using prediction Eq. [2], and (b) centered predicted values and residuals (observed - predicted) of Eq. [2]. Predicted values were centered by subtracting the mean of all predicted values from each predicted value.

Figure 2. Relationship between (a) predicted and observed fecal nitrogen excretion (g/d) using prediction Eq. [4], and (b) centered predicted values and residuals (observed - predicted) of Eq. [4]. Predicted values were centered by subtracting the mean of all predicted values from each predicted value.

Figure 3. Relationship between (a) predicted and observed proportion of urinary N in total N excretion using prediction Eq. [7], and (b) centered predicted values and residuals (observed - predicted) of Eq. [7]. Predicted values were centered by subtracting the mean of all predicted values from each predicted value (TTND = apparent total tract N digestibility).

Figure 1. (a)

250

Regression for Eq. [2]: Urinary N = 0.51(N intake) - 14.12 y = 0.97 (±0.01)x + 1.64 (±0.72) y=x r2 = 0.98

Predicted urinary N, g/d

200

P < 0.001 150

100

50

0 0

50

100

150

200

250

Observed urinary N, g/d

(b) 35


Residual (Observed-Predicted)

15 y = 0.0052 (±0.0089)x - 0.0006 (±0.3045) r² = 0.0019, P = 0.56 10 5 0 -140 -120 -100 -80 -60 -40 -20

0

20

40

60

80 100 120 140

-5 -10 -15

Centered predicted urinary N, g/d

Figure 2. (a)

120

Regression for Eq. [4]: Fecal N = 0.20(N intake) + 15.82 y = 0.95 (±0.01)x + 2.27 (±0.64) y=x r2 = 0.96

Predicted fecal N, g/d

100

P < 0.001 80 60 40 20 0 0

20

40

60

80

100

120

Observed fecal N, g/d

(b) 36



15 y = 0.0151 (±0.0150)x - 0.0008 (±0.2005) r² = 0.0056, P = 0.32 10 5 0

-60

-50

-40

-30

-20

-10

0

10

20

30

40

50

-5 -10 -15

Centered predicted fecal N, g/d

Figure 3. (a)

Predicted urinary N/total N excretion

0.9 0.8 0.7

Regression for Eq. [7]: Urinary N/total N excretion = 0.010(TTND) - 0.162

y=x

y = 0.95 (±0.015)x + 0.03 (±0.008) r2 = 0.96 P < 0.001

0.6 0.5 0.4 0.3 0.2 0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Observed urinary N/total N excretion

(b) 37


60


0.10 y = 0.0143 (±0.0158)x - 0.00008 (±0.002) 0.08 r² = 0.0046, P = 0.37 0.06 0.04 0.02 0.00 -0.4

-0.3

-0.2

-0.1

-0.02

0.0

0.1

0.2

0.3

-0.04 -0.06 -0.08 -0.10

Centered predicted urinary N/total N excretion

38


0.4

Diurnal variations of fecal and urinary iodine excretion in rats.

Nitrogen dioxide exposure and urinary excretion of hydroxyproline and desmosine.

Influence of dry matter and nitrogen intakes on fecal nitrogen losses in cattle.

Prediction of manure nitrogen and organic matter excretion for young Holstein cattle fed on grass silage-based diets.

Genomic prediction in Japanese Black cattle: application of a single-step approach to beef cattle.

Twenty-four hour urinary excretion of vitamins, minerals and nitrogen by Eskimos.

Protein requirements of man: obligatory urinary and fecal nitrogen losses in elderly women.

Fecal Polyomavirus Excretion in Infancy.

Urinary creatinine and creatine excretion in pigs in relation to body weight and nitrogen balance.

Urinary calcium excretion by neomycin-treated dogs.

Prediction of the DST results in depressives by means of urinary-free cortisol excretion, dexamethasone levels, and age.

Effect of phase-feeding crude protein on performance and carcass characteristics of crossbred beef bulls: an application to reduce nitrogen compounds in beef cattle diets.

Change of urinary nitrogen excretion in response to low-protein diets in adults.

Mineral balance and urinary and fecal mineral excretion profile of broilers housed in thermoneutral and heat-distressed environments.

Selection signatures in Canchim beef cattle.

Genetic disorders in beef cattle: a review.

Labile nitrogen reserves and plasma nitrogen fractions in growing cattle.

Urinary growth hormone excretion.

[Prediction of ovulation by urinary LH surge].

Urinary excretion of furazabol metabolite.

Chlorophyllin-enhanced excretion of urinary and fecal mutagens in rats given 2-amino-3-methylimidazo[4,5-f]quinoline.

Comparison of alternative approaches to single-trait genomic prediction using genotyped and non-genotyped Hanwoo beef cattle.

Identification and Characterization of Cefotaxime Resistant Bacteria in Beef Cattle.

Significance of liver and erythrocyte arginase in beef cattle.