Avaliação dos comportamentos de pastejo e suas relações com caracteres individuais dos bovinos de corte
Carregando...
Data
2016-03-22
Autores
Aguilar, Natalia Maria Alejandra [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
O objetivo principal foi identificar relações entre comportamento durante o pastejo e temperamento dos bovinos de corte, com objetivos específicos de: predizer o comportamento a partir de informação de dispositivos de precisão e, verificar a influência do temperamento sobre o comportamento durante o pastejo de bovinos mantidos em pastagens tropicais suplementadas (Brachiaria brizantha - cv. Marandu). Inicialmente comparou-se dados comportamentais de observação visual dos bovinos com dados dos colares GSP com sensores de inclinação (3300 LR, Loteck®), determinando por meio de árvores de decisão os limiares de predição e classificação dos comportamentos. Quarenta e oito tourinhos Nelore (PV = 231 ± 19,6 kg; de 10 - 12 meses), foram recriados em 14 piquetes recebendo um dos um dos seis planos nutricionais, resultantes de três alturas de pastagens (15, 25 ou 35 cm) com dos tipos de suplementos (minerais ou proteína), durante as águas. Doze animais por ciclo levaram aleatoriamente um colar GPS durante 24 horas por 7 dias. Foram encontradas associações entre dados do colar e observações visuais, com variação do 75,11% explicada pelos componentes principais (PC1 e PC2). A variável do colar cabeça baixa (%), mostrou alta correlação (P <0,0001) com pastejo (GRA, r = 0.76) e com rumia (RUM, r = - 0.78); entretanto variáveis do colar: X-act, Y-act, somatória de ambos e distância percorrida, tiveram baixa associação com comportamentos observados. O modelo de árvore de decisão de menor erro de classificação (13,98%), foi escolhido para prever o comportamento animal durante intervalos não observados; apresentando alta precisão para GRA (90%) e RUM (76%) e menor para não-pastejo (Not-gra, 59%). Determinando-se a previsão média total de tempo diário para GRA (38,8%), RUM (44,2%) e Não-gra (17%). Não foram encontradas diferenças (P> 0,05) dos comportamentos previstos entre os diferentes planos nutricionais. Sugerindo que, colares GPS com sensores de inclinação fornecem informações aceitáveis para prever e classificar os comportamentos de bovinos em pastagens. Seguidamente, avaliou-se a associação do temperamento com a produtividade (GMD e PVfinal) e o comportamento dos bovinos (tempo de GRA, RUM, Not-gra e caminhada diária). Inicialmente 126 novilhos foram mantidos em 18 piquetes das mesmas características ao estudo anterior. Todos os animais foram pesados individualmente no início e final do estudo, sendo avaliado o temperamento por escore de reatividade visual (RS), enquanto o animal foi mantido na balança e velocidade de saída (FS, m/s) após de sair da balança. Para avaliar o efeito do temperamento sobre o comportamento utilizou-se os tempos previstos dos comportamentos descritos previamente. O temperamento mostrou baixa relação de PVfinal com RS (rs = 0,18; P = 0,05), entretanto não teve efeitos sobre GMD (P> 0,05). Encontrou-se efeito sobre os tempos previstos dos comportamentos (P <0,0001) Not-gra avaliado por ambos métodos (FS e RS); RUM avaliado por FS; GRA com apenas tendência negativa (P = 0.07) avaliado por FS. Com base nestes resultados, demostrou-se que o temperamento bovino, avaliado pelas respostas ao manejo, não representa diretamente o comportamento durante o pastejo de bovinos mantidos em pastagem suplementadas, sugerindo que estas características provavelmente sejam independentes.
The main aim was to identify the possible relationship between behaviour during the grazing and beef cattle temperament, specific aims were: predict the behaviour from information given by precision devices, and verify the influence of temperament on the behaviour cattle during grazing in supplemented tropical pastures (Brachiaria - Marandu cv.). Initially, compared cattle behaviour visual observed and data from collars GSP with built-in tilt sensors (LR 3300, Loteck®), it was determined by decision trees the thresholds for the prediction and classification of behaviour cattle. Forty-eight young bulls Nellore (BW = 231 ± 19.6 kg; 10 to 12 mo-old) were breeding in 14 paddocks that received one of the six nutritional plans, resulting from three heights of pastures (15, 25 or 35 cm) and the types of supplements (minerals or protein), on season wet. Twelve animals randomly per cycle wear a GPS collar for 24 hours for 7 days. Found associations among the data behavioural visual observations and from GPS collars, where75.11% explained by the principal components (PC1 and PC2). The variable head down (%) showed high correlation (P < 0.0001) with grazing (GRA, r = 0.76) and ruminating (RUM, r = - 0.78); but the others variables by collar: X-act, Y-act, sum of both and distance traveled, had a low correlation with observed behaviour. The decision tree model of lower misclassification (13.98%), was chosen to predict animal behaviour during intervals unobserved. Shown high precision with GRA (90%) and RUM (76%) and lower with not-grazing (Not-gra, 59%). The determining average daily time predict to GRA (38.8%), RUM (44.2%) and Not-gra (17%). There were no differences (P > 0.05) of behaviour referred among different nutritional plans. Results suggested, that GPS collars with tilt sensors provide acceptable information to predict and classify the cattle behaviour in pastures. Followed, was assessment the association of temperament with productivity (ADG and BWfinal) and behaviour (GRA, RUM, Not-gra and daily distance walked). Initially 126 steers were kept in 18 paddocks of the same characteristics the previous study. All animals were individually weighing at the beginning and end of the study, while assess temperament by two usual test: visual reactivity score (RS) while the animal was kept inside the squeeze chute; and flight speed (FS, m.s-1) velocity to exit the squeeze chute. Furthermore, to evaluate the effect of temperament on the behaviour we used the of predicted behaviors time described previously. The temperament shows low correlation BWfinal with RS (rs = 0.18; P = 0.05), but had no effect on ADG (P > 0.05). Was found effect on the distance walked, the behavior predicted time (P < 0.0001) where Not-gra assessed by both methods (FS and RS); RUM assessed by FS and GRA with only negative tendency (P = 0.07) assessed by FS. Based on these results, it demonstrated that the cattle temperament, measured by responses to management, is not directly the behavior during the grazing of cattle kept on pasture supplemented, suggesting that these characteristics are likely to be independent.
The main aim was to identify the possible relationship between behaviour during the grazing and beef cattle temperament, specific aims were: predict the behaviour from information given by precision devices, and verify the influence of temperament on the behaviour cattle during grazing in supplemented tropical pastures (Brachiaria - Marandu cv.). Initially, compared cattle behaviour visual observed and data from collars GSP with built-in tilt sensors (LR 3300, Loteck®), it was determined by decision trees the thresholds for the prediction and classification of behaviour cattle. Forty-eight young bulls Nellore (BW = 231 ± 19.6 kg; 10 to 12 mo-old) were breeding in 14 paddocks that received one of the six nutritional plans, resulting from three heights of pastures (15, 25 or 35 cm) and the types of supplements (minerals or protein), on season wet. Twelve animals randomly per cycle wear a GPS collar for 24 hours for 7 days. Found associations among the data behavioural visual observations and from GPS collars, where75.11% explained by the principal components (PC1 and PC2). The variable head down (%) showed high correlation (P < 0.0001) with grazing (GRA, r = 0.76) and ruminating (RUM, r = - 0.78); but the others variables by collar: X-act, Y-act, sum of both and distance traveled, had a low correlation with observed behaviour. The decision tree model of lower misclassification (13.98%), was chosen to predict animal behaviour during intervals unobserved. Shown high precision with GRA (90%) and RUM (76%) and lower with not-grazing (Not-gra, 59%). The determining average daily time predict to GRA (38.8%), RUM (44.2%) and Not-gra (17%). There were no differences (P > 0.05) of behaviour referred among different nutritional plans. Results suggested, that GPS collars with tilt sensors provide acceptable information to predict and classify the cattle behaviour in pastures. Followed, was assessment the association of temperament with productivity (ADG and BWfinal) and behaviour (GRA, RUM, Not-gra and daily distance walked). Initially 126 steers were kept in 18 paddocks of the same characteristics the previous study. All animals were individually weighing at the beginning and end of the study, while assess temperament by two usual test: visual reactivity score (RS) while the animal was kept inside the squeeze chute; and flight speed (FS, m.s-1) velocity to exit the squeeze chute. Furthermore, to evaluate the effect of temperament on the behaviour we used the of predicted behaviors time described previously. The temperament shows low correlation BWfinal with RS (rs = 0.18; P = 0.05), but had no effect on ADG (P > 0.05). Was found effect on the distance walked, the behavior predicted time (P < 0.0001) where Not-gra assessed by both methods (FS and RS); RUM assessed by FS and GRA with only negative tendency (P = 0.07) assessed by FS. Based on these results, it demonstrated that the cattle temperament, measured by responses to management, is not directly the behavior during the grazing of cattle kept on pasture supplemented, suggesting that these characteristics are likely to be independent.
Descrição
Palavras-chave
Árvores de decisão, Etologia, GPS, Raça Nelore, Temperamento, Decision trees, Ethology, Nellore breed, Temperament