Animal Variation and the Derivation of Optimal Shipping Strategies for Finishing Pigs22 October 2013
The complexity of animal performance and economic interactions on the options available for pig producers to market their pigs makes it imperative to use appropriate integrated growth models to maximise farm income, according to Neil S. Ferguson of Nutreco Canada Agrasearch. He explained the reasoning to the 2013 London Swine Conference.
Defining the optimum strategy to ship finisher pigs is essential for increasing farm profitability, and conversely if shipping management is inappropriately applied, loss of revenue can occur.
The complexity of performance and economic interactions and options available for pig producers to market their pigs makes it imperative to use appropriate integrated growth models to help determine the optimum (maximise farm income) solution.
The prediction of what shipping strategy should be adopted will depend on many factors but the essential components include understanding the consequences of animal variation and the main causes of variation, including social stressors, as well as the capacity to optimise over a range of production, nutrition and shipping scenarios.
The ability of an individual animal to cope with social stress is a major source of variation between animals within a population. Therefore, any introduction of animal variation into growth models must incorporate this genetic characteristic as well as the interactions between this ability to cope and the social stressor(s) most commonly observed in commercial conditions, namely stocking density, feeder space allowance and health challenges. The individual animal’s ability to cope to suboptimal conditions will influence its potential protein growth rate as well as the extent stocking density and feeder space allowance will affect the potential growth and voluntary feed intake.
A study of two different shipping strategies provides an example of how important it is to define the optimum marketing strategy based on the production system and the grading grid(s).
Most integrated pig growth models are concerned with the response of the ‘average’ individual within the herd (population) and assumes this is a good representation of the population. In most practical cases this assumption may hold true. However, it has been shown that the overall close-out (mean of the population) responses can differ significantly from the average individual response due to the variation in growth potential between individual animals within the herd (Curnow, 1973, Knap, 2000; Pomar et al., 2003; Brossard et al., 2009).
The extent of the difference between the average individual and the population mean response will depend on how large are the differences between individual genotypes within the population, the correlation between the genetic parameters defining the genotype, and the individual animal’s ability to cope with social stressors (Wellock et al., 2003 and 2004). The more individuals vary within a population (or the greater the weight variation), the more inappropriate it is to use the average individual response as a means of predicting the population mean response. For example, predicting nutrient requirements for a population based on the single deterministic response will introduce a bias against individuals with a higher nutrient requirement. These errors can be magnified during the optimisation process which is dependent on the herd nutrient responses.
Not only is the introduction of animal genetic variation essential for more accurate nutritional optimisation but according to Knap (1995), it also influences financial outcomes because of the effect animal variation has on the variation in production characteristics (feed intakes, growth rates, backfat, hot carcass weight, lean yield and MOFC. In the monthly Hormel Report (December 2005), they indicated that a reduction in shipping weight variation or having more pigs shipped at ideal weight, may result in a $5.70 per pig increase, and similarly in the Manitoba Pork Marketing Co-operative monthly report (January, 2008), they suggested a $2.00 per pig improvement.
Further reasons for considering between-animal variation in pig modelling are
- to predict more accurately the optimum strategy for shipping pigs to market to increase the proportion of “full-value” pigs per close-out, and
- to enhance production through better utilization of space and minimising performance failures.
There are other sources of variation (health, feed and physical environment) that influence the individual’s response and therefore the population response besides the genetic and social variation. However, for the purposes of this paper, these latter sources of variation will not be addressed, and attention will be given to the effect social interactions influence variation between individual animals and how between-animal variation may affect the derivation of optimal shipping strategies.
Drivers of Animal Variation
As described in previous papers (Ferguson, 2006) and other similar models (Knap 2000; Pomar et al., 2003; Wellock et al., 2003), the genetic potential growth of an individual pig can be defined by four components:
- potential rate of maturing (B),
- mature protein weight (Pm);
- inherent fatness or desired fat level relative to protein weight (LPm), and
- the amount of water relative to protein weight (WPm).
For the most part, these parameters are independent of each other (uncorrelated) and are assumed to be normally distributed, such that the genetic parameters of individuals can be randomly generated around the mean and standard deviation (Ferguson et al., 1997).
The exception is the strong correlation between Pm and B, but this will not be discussed further.
Initial size (body weight for a given age)
Individuals within a population are likely to have different body weights for a given age and therefore, different amounts of protein, lipid, water and ash.
Assuming a fixed starting age, initial bodyweight will vary according to the population mean weight and standard deviation. This variation at the start of the growing period will be a significant factor affecting the variation in bodyweight at slaughter.
Based on previous grower-finisher trials within Nutreco Canada, the coefficient of variation of bodyweights close to 25kg varied from six per cent to 17 per cent with an average of 11 per cent. Part of the variation in starting weight will be derived from the individual’s potential growth rate, and therefore start weight should be correlated with the genetic parameters (Wellock et al., 2004). Individuals with the highest growth potential will tend to have the highest start weight.
Earlier studies have clearly demonstrated that individual pigs within a pen interact differently with each other, and these interactions can affect individual performances (Tindsley and Lean, 1984). Data from Giroux et al., (2000) indicate that social interactions can account for nine per cent of the variation in average daily gain in growing pigs. Socially dominant individuals, often larger individuals, are less affected by social stresses and tend to perform better than their subordinates when exposed to suboptimal conditions, e.g. high stocking density, inadequate feeder space (Botermans, 1999).
Wellock et al. (2003) introduced a genetic parameter to facilitate the interaction between an individual pig and its social environment, and subsequent effects on performance. This parameter describes the ability of an individual pig to cope when exposed to suboptimal conditions (A2C) and would be used to define both when an individual becomes stressed and by how much the stressor decreases performance.
The mechanism by which A2C exerts its influence involves both the ability to attain potential growth and changes to feeding behaviour. Socially stressed animals (low A2C) will have a lower capacity to achieve potential (protein) growth as well as reduced accessibility to food and therefore their performance is expected to be reduced.
The evidence for these mechanisms is more circumstantial than exact, nevertheless it supports the idea that stressed animals do not attain their potential and their normal feeding behaviour is disrupted. (Nielsen et al., 1995; Turner et al., 2002; Anil et al., 2007).
This is certainly evident when comparing individual versus group housed pigs (Gonyou et al., 1992; Ferguson et al., 2001). It is likely that both time at the feeder and the rate of feeding will be adversely affected by socially stressed pigs (Nielsen, 1999; Turner et al., 2002).
Nielsen (1999) suggested that one of the reasons social constraints will adversely affect ad libitum feed intake is due to the desire to eat simultaneously with other pigs in the pen and therefore have to cope with increased competition for feeder space. The main social stressors that interact with A2C include stocking density, feeder space and health status (Figure 1).
Because of the strong correlation between size and dominance (Tindsley and Lean, 1984; D’Eath, 2002), it is reasonable to assume that there is a strong positive correlation between liveweight and A2C.
Turner et al. (2002) observed a greater reduction in growth rates in smaller pigs than larger pigs when grown under more stressful conditions (low feeder space and large group size). There is likely to be an increase in within-pen body weight variation when the level of stress increases. This was observed by Anil et al. (2007) when pigs kept in acceptable stocking density levels (>0.74 square metres per pig) had bodyweight standard deviations of 7.6 to 14.9kg compared to those pigs with less space (0.64 square metres per pig) of 11.7 to 16.6kg.
Although these differences were not statistically different, they do highlight the possibility that the weight differences between the small and large pigs in a pen may increase with higher levels of stress.
Tindsley and Lean (1984) noted that dominant pigs were generally the heaviest pigs and that the variation in ability to dominate may be responsible for liveweight variations. Pigs with a low ability to cope with stress will require more space and if this is limiting then they will be more severely affected than individuals with a higher A2C. Conversely, pigs with a very high ability to cope with stress will realise their potential growth at much lower space allowances and when their potential is constrained, the extent will be lower than pigs with a low A2C.
DeDecker et al. (2005) concluded from their study that increasing feeder space benefitted the smaller pigs more than the heavier pigs as they grew proportionally faster and became heavier than smaller pigs with less space.
Individual differences in voluntary feed intake can be explained partly by the variation in feeding behaviour associated with social ranking (Giroux et al., 2000). Feeding time and the rate of feeding change when there is limited feeding space but the extent of the change will depend on the social hierarchy (Nielsen, 1999; Turner et al., 2002). Therefore, individuals with a lower social standing will probably have less time at the feeder than the socially dominant pig(s).
A further source of variation between animals within a group, is related to their immunocompetence (Knap and Bishop, 2000; Flori et al., 2011). Clapperton et al. (2009) observed substantial genetic variation in immunity traits pigs in general, and therefore it can be assumed that individual pigs in a population will have different abilities to fight health challenges. It is expected that in a batch of pigs that on average show a low health status, there is likely to be more variation than a high health herd because there will still be some pigs showing a strong immunocompetency.
There is likely to be some interaction between A2C and health status, especially in better health conditions. Healthier individuals can cope better with stress than less healthy.
Optimisation of Feeding Strategy
Optimising nutritional strategies based on economic returns or animal performance, rather than least-cost formulation for a defined set of nutrient requirements is the most appropriate method for improving performance and profitability at the farm level.
Gous and Berhe (2006) define the criteria required for optimisation as:
- feed costs at defined nutrient levels
- animal responses to changing nutrient profiles
- fixed and variable costs associated with the production system and
- definition of revenue generating processes.
Figure 2 illustrates the relationship between animal biology, optimisation and animal variation defined in Watson 2.0.
The optimisation process is started by passing initial specifications (nutrients) to a feed formulator to determine the least-cost feeds (formulation) which, in turn, are fed to the animal biology component to produce a specific performance, including feed intake, growth, carcass characteristics. From the performance output, it is possible to generate the costs and revenue (economics) which is passed back to the optimizer to complete the cycle.
This process is repeated before identifying the ‘best’ solution to meet the optimisation objective.
Currently in Watson 2.0, the processes to be optimised include energy content, nutrient density, amino acid responses, carcass weights and feeding phases, while the objectives include maximising growth rates, margin over feed costs, net profit per pig and minimising feed:gain, cost per kg gain and nutrient excretion.
Within a batch of finishing pigs, there is sufficient between-animal variation in protein and fat deposition, feed intake and subsequent efficiency of nutrient utilisation to ensure differences in nutrient responses between the single average individual and the batch mean (Pomar et al., 2003; Brossard et al., 2009; de Lange et al., 2012).
However, the incremental cost of increasing the dietary amino acid level may not be offset by the increase in revenue generated from improved feed efficiency and/or higher carcass lean, resulting in differences in amino acid requirements between minimizing feed:gain and maximising margin over feed costs. As individual pigs will have different optimum performance and economic responses to amino acid intakes, it is necessary to incorporate this between-animal variation into the optimisation process.
Shipping Management Strategies
One of the most important drivers of profitability in pig production is the revenue per pig carcass, and therefore the higher, and less variable the income generated per batch of pigs, the more profitable the operation.
It is imperative for the producer to strive to optimise when individual pigs should be shipped, in order to maximise profitability. For this reason, animal variation will be introduced into Watson, from which it will be possible to determine how many pigs are at, or close to, the minimum ideal market weight and what are their performances and economics for each of the shipping weeks.
To illustrate the potential opportunities, a simple example is provided illustrating how many barrows and gilts could be shipped over a four-week period, having met producer-defined minimum shipping weights for each week of shipping.
For the purposes of this exercise, two shipping strategies were compared: 1) a fixed minimum shipping weight of 120kg, and 2) the minimum live weight for the first week of 120kg, followed thereafter by a minimum of 123kg. The target or ideal live weight was 125kg.
It is expected that first shipments will occur at Week 14, which will result in a significantly large proportion of the population (45 per cent) being shipped that week. As expected more barrows than gilts will be shipped earlier (Figure 3).
Shipping Management Optimisation
To determine the shipping strategy that will return the highest margin over feed costs (MOFC), it is necessary to define the range of either Weekly Shipping Proportions (per cent) or the Minimum Shipping Weight for each week.
The model will generate the performance and economic data for each individual pig, and then collate the data from those individual pigs that meet the minimum shipping weight for each of the shipping weeks.
For example, one combination could be: Ship 0 per cent pigs the first week, the heaviest six per cent in Week 2, the next 24 per cent heaviest in Week 3, the next 18 per cent heaviest pigs in Week 4, the next 18 per cent heaviest in Week 5, and the remaining 34 per cent in Week 6 (Figure 4). This would provide the solution for one shipping combination and therefore, it is necessary to repeat for all possible combinations before deriving the the weekly shipping strategy that produces the highest MOFC.
Case Study: Comparison of Conventional Weekly Shipping Versus an Alternative Shipping Strategy for Heavy Pigs
A trial was conducted to compare the biological and economical results of a conventional (CONV) versus an alternative (ALT) shipping strategy to a grid with a constant index up to a maximum of 112kg hot carcass weight (136kg shipping weight).
For the CONV treatment, shipping occurred weekly between weeks 13 and 17 with a target weight of 128-132kg (n=20 pens), while the ALT treatment involved shipping pigs only twice (two heaviest pigs from each pen (28 per cent) at week 13 and the remainder at the end of week 17; n=22 pens). The results are shown in Table 1.
Pigs on the ALT strategy had a significantly higher feed conversion ratio (+3 per cent or 0.07g per g), increased feed costs (+$6.80) and increased cost/kg gain (+$0.023/kg), resulting in a reduced MOFC (-$5.70/pig). The ALT strategy may reduce labour costs and increase the shipping weight at the end of 17 weeks by 2kg but there will be more variation in shipping weight.
It could be argued that under commercial conditions, the ALT strategy could have a larger throughput of pigs (i.e. higher stocking density) and therefore higher gross profit per year.
The economic outcomes of this study would be different if shipping to a grid that has a narrower ideal carcass weight range.
The point of this case study is to highlight the opportunity to define optimum shipping strategies based on the production constraints within the farm and the grading grids the pigs will receive payment from.
The complexity of animal performance and economic interactions on the options available for pig producers to market their pigs makes it imperative to use appropriate integrated growth models to help determine the optimum (maximise farm income) solution.
The consequences of not optimising the shipping strategy for a specific grid(s) can significantly reduce farm income.
To be able to define the optimum shipping strategy, it is necessary to understand how animal variation impacts when individual pigs are marketed and the optimisation process.
Gous and Berhe (2006) summarise the importance of incorporating animal variation into the optimisation process as follows: 'Models of individuals may be adequate for an understanding of the theory of growth and feed intake, as well as for ‘what-if’ scenario planning. However, for purposes of optimization, it is imperative to account for the variation inherent in the system if a realistic assessment of the population response is to be simulated.'
Managing between-animal variation and applying the optimum shipping strategy for a given slaughter plant(s) has the potential to improve farm profitability significantly.
Anil, L., S.S. Anil and J. Deen, 2007. Effects of allometric space allowance and weight group composition on grower-finisher pigs. Can J. Anim. Sci. 87:139-151.
Botermans, J.A.M. 1999. Feeding environment for growing-finishing pigs. Effects of competition for feeding frequency on performance, behavior, injuries, plasma cortisol and exocrine pancreatic secretion. Acta Univ. Agric. Sueciae – Agraria 192:1-45.
Brossard, L., J-Y. Dourmad, J. Rivest. and J. van Milgen, 2009. Modelling the variation in performance of a population of growing pigs as affected by lysine supply and feeding strategy. Animal 3:1114-1123.
Clapperton, M., A.B. Diack, O. Matika, E.J. Glass and C.D Gladney. 2009. Traits associated with innate and adaptive immunity in pigs: heritability and associations with preference under different health status conditions. Genet. Sel. Evol. 41:54-61.
Curnow, R.N. 1973. A smooth population response curve based on an abrupt threshold and plateau model for individuals. Biometrics 29:1-10.
D’Eath, R.B. 2002. Individual aggressiveness measured in a resident-intruder test predicts the persistence of aggressive behavior and weight gains of young pigs after mixing. Appl. Anim. Behav. Sci. 77:267-283.
DeDecker, J.M., M. Ellis, B.F. Wolter, S.E. Corrigan, G. Curtis. and R. Hollis. 2005. Effect of stocking rate on pig performance in wean-finish production system. Can. J. Anim. Sci. 85:1- 5.
De Lange, C.F.M, C.L. Levesque and B.J. Kerr. 2012 Amino acid nutrition and feed efficiency. In (J. Patience, Ed.): Feed Efficiency in Swine. Wageningen Academic Publishers, Netherlands. pp. 81-100.
Ferguson, N.S. 2006. Basic concepts describing animal growth and feed intake. In (.R.M. Gous, T. Morris and C. Fisher): Mechanistic modeling in Pig and Poultry Production. CAB International, UK. pp 22-53.
Ferguson, N.S., R.M. Gous and G.C. Emmans. 1997. Predicting the effects of animal variation on growth and food intake in growing pigs using simulation modeling. Anim. Sci. 64:513- 522.
Ferguson, N.S., G. Lavers and R.M. Gous. 2001. The effect of stocking density on the responses of growing pigs to dietary lysine. Anim. Sci. 73: 459-469.
Flori, L., Y. Gao, D. Laloe, G. Lemonnier, J-J. Leplat, A. Teillaud, A-M. Cossalter, J. Laffitte, P. Pinton, C. de Vauriex, M., Bouffaud, M-J. Mercat, F. Lefevre, I.P. Oswald, J-P. Bidanel and C. Rogel-Gaillard, 2011. Immunity traits in pigs: Substantial genetic variation and limited covariation. PLoS ONE 6:1-11.
Giroux, S., G.P.Martineau and S. Roberts, S. 2000. Relationships between individual behavioral traits and post-weaning growth in segregated early-weaned piglets. Appl. Anim. Behav. Sci. 70:41-48.
Gonyou, H.W., R.P. Chapple and G.R. Frank. 1992. Productivity, time budgets and social aspects of eating in pigs penned in groups of five or individually. Appl. Anim. Behav. Sci. 34: 291-301.
Gous, R.M and E.T. Berhe. 2006. Modeling populations for purposes of optimization. In (.R.M. Gous, T. Morris and C. Fisher): Mechanistic modeling in Pig and Poultry Production. CAB International, UK. pp 76-96.
Knap, P.W. 1995. Aspects of stochasticity: variation between animals. In(P.J. Moughan, M.W.A. Verstegen and M.I. Visser-Reyneveld): Modelling growth in the pig., Wageningen Press, Wageningen, pp 165-172.
Knap, P.W. 2000. Stochastic simulation of growth in pigs: relations between body composition and maintenance requirements as mediated through protein turn-over and thermoregulation. Anim. Sci. 71:11-30.
Knap, P.W. and S.C. Bishop. 2000. Relationships between genetic change and infectious disease in domestic livestock. In(W.G. Hill, S.C. Bishop, B. McGuirk, J. C. McKay, G. Simm and A.J. Webb, Ed). The Challenge in Animal Production, British Society of Animal Production Occasional Publications, Edinburgh. pp. 65 – 80.
Nielsen, B.L., A.B. Lawrence. and C.T. Whittemore. 1995. Effect of group size on feeding behavior, social behavior and performance of growing pigs using single-space feeders. Liv. Prod. Sci. 44: 73-85.
Nielsen, B.L. 1999. On the interpretation of feeding behaviour measures and the use of feeding rate as an indicator of social constraints. Appl. Anim. Behav. Sci. 63: 79-91.
Pomar, C. I. Kyriazakis, G.C. Emmans. and P.W. Knap. 2003. Modeling stochasticity: dealing with populations rather than individual pigs. J. Anim. Sci. 81(E Suppl. 2) E178-E186.
Tindsley, W.H. and I.J. Lean 1984.Effects of weight range at allocation on production and behavior in fattening pig groups. Appl. Anim. Behav. Sci. 12:79-92.
Turner, S.P., M. Dahlgren, D.S. Arey. and S.A.Edwards. 2002. Effect of social group size and initial live weight on feeder space requirements of growing pigs given food ad libitum. Anim. Sci. 75: 75-83.
Wellock, I.J., G.C. Emmans. and I. Kyriazakis. 2003. Predicting the consequences of social stressors on pig food intake and performance. J. Anim. Sci. 81: 2995-3007.
Wellock, I.J., G.C. Emmans. and I. Kyriazakis. 2004.. Modeling the effects of stressors on the performance of populations of pigs. J. Anim. Sci. 82: 2442-2450.
You can view other papers presented at the 2013 London Swine Conference by clicking here.