Evaluation of milk production systems using a dynamic simulation model / Avaliação de sistemas de produção de leite por meio de modelo dinâmico de simulação

AUTOR(ES)
DATA DE PUBLICAÇÃO

2008

RESUMO

The Thesis is divided into four chapters. Chapter 1 aimed at selecting functions that mathematically represent lactation curves, providing a better fit to the observed data, to be used in simulation models of milk production systems. Data base was built with 6,459 recordings of daily milk production per cow obtained from monthly and quarterly milk controls between 2004 and 2007, coming from 472 complete lactations of animals from 10 different farms. The farms were divided into low, medium and high production levels according to average milk yield in liters per day per cow. Data were divided according to the lactation orders into first order, second order and third or the highest lactation order. Eight functions of lactation curves obtained from the literature were compared. The used goodness-of-fit measures were R, MSE, MSEP and BIC. The values of these measures obtained from each model were compared using the confidence interval, with 95% probability. For the group of farms with low average production, the Wilminks (1987) model gave better fit for lactations of first-order cows, whereas the Wood s (1967) model was better fitted for lactations of third-or-higher-order cows. The Wood s (1967) model provided better fit for all the lactation orders of farms with medium production. The Dijkstra et al.s (1997) model gave the best fits in all the lactation orders of the high production group. Chapter 2 aimed at developing user-friendly interface and easy-to-use software for calculating nutritional requirements, feed nutritional value, evaluation of rations and estimate of costs with dairy cattle feeding. Another objective was to evaluate the reliability of results generated by the software by comparing them with estimates obtained from the NRC software version 2001. The ZooCalc 1.0 software was developed using the Visual Basic 6.0 language, consisting of submodels of lactating cows, dry cows, heifers and calves, as well as a module for editing the composition and estimates of digestible fractions and dietary energy concentration. ZooCalc 1.0 is an application that can be used to calculate the nutritional requirements and feed evaluation for dairy cattle. ZooCalc 1.0 and NRC produce the same estimates for the digestible fractions and energy concentration of the tested feed. Both estimate the same values for NEm, NEl, NEc and MPc. Small differences were found for MPl because of the approximation errors produced by the NRC software. Small differences were also found for RDP, which might be due to approximation errors. There were small differences for the estimates of MPm and RUP, however, the cause is unclear. Chapter 3 aimed at developing and evaluating a dynamic simulation model, with stochastic components, to evaluate milk production systems. The simulation model was developed by using STELLA 9.0, ZooCalc 1.0 and electronic data sheets together. The physical structure of the herd was divided into 15 interlinked modules that represent categories or subdivisions of categories. Mathematical models obtained from the literature were used for body weight simulation of growing and lactating animals, as well as of milk production and milk composition. Electronic data sheets were used to calculate the nutritional requirements and dietary balance. After cost minimization, each ration was evaluated using ZooCalc 1.0. Some average responses of technical, economic and size indicators, the variation in the area used for silage corn production, the interval between calving and age at first insemination were obtained to evaluate whether the simulation model was working consistently with the expected. To assess the capacity of the model to generate simulated values that approach reality, the degree of linear association was measured between values generated by the model and values observed in practical situation. Sensibility analyses showed that the values produced by the simulation model were in agreement with the expected results, in response to variations in the area used for silage corn production, the interval between calving and the age at first insemination. There was correlation between the values generated by the simulation model and the values observed in practical situation, and the degree of linear association for the variables total number of animals, total number of cows and number of lactating cows can be considered high. Chapter 4 compared size, technical and economic indices of Holstein and Jersey herds in a confinement production system using the developed simulation model. It also evaluated the profitability of the activity by using different systems of allowance for milk price, because the percentage of fat and protein. The simulation model was the same described in Chapter 3. The comparisons were performed, initially, without considering the allowances for milk price based on the increase in the concentration of fat and protein. Later, the profitability of the activity was compared considering allowances of 1, 2, 3, 4 and 5 % for the milk price of the Jersey herd, for each unit of increase in the percentage of milk fat above the values obtained with the Holstein herd, and allowances of 2, 4, 6, 8 and 10 % for each unit of increase in the percentage of milk protein of the Jersey herd above the values obtained with the Holstein herd. Finally, the profitability of the activity was compared by using the allowance system based on quality proposed by Gimenes and Ponchio (2006) and the allowance system used by company A. The simulation time used for each comparison was 10 years and the herd was considered stabilized during the simulations. Confinement production systems were able to keep about 35.9% more Jersey animals than Holstein animals because of the lower roughage consumption per Jersey animal. Although the milk production per Jersey cow corresponded to approximately 69.0% of the milk production per Holstein cow, the annual milk production of the system using Jersey animals corresponded to 93.8% of the system with Holstein animals. When there is no allowance for percentage of fat and protein in the milk, although the Jersey breed provides gross income close to the Holstein (99.2%), the profit obtained with the Holstein breed is higher because of the lower production costs, mainly the ones related with concentrate feed. In order to the activity become more profitable with the Jersey breed, an allowance of 5.0% for milk price will be needed for each percentage unit of fat above the percentage obtained with the Holstein breed, or allowance of 10.0% in the milk price for each percentage unit of protein above the percentage obtained with the Holstein breed. With the use of the allowance system proposed by Gimenes and Ponchio (2006) the annual profit with the Jersey breed is higher than the profit obtained with the Holstein breed. However, with the application of the allowance system used by the company A, the Holstein breed becomes more profitable than the Jersey breed.

ASSUNTO(S)

modelos matemáticos simulação (computadores) animal nutrition mathematical models alimentação e rações feeds and feeding ruminante dairy cattle nutrição animal nutricao e alimentacao animal bovino de leite simulation (computers) ruminant

ACESSO AO ARTIGO

Documentos Relacionados