Abstract

During the summer, a challenging thermal environment is known to cause a significant reduction in food intake, growth, milk production, reproduction and even death in cattle. In this study, we attempt to characterize the relationship of cattle body temperature with several environmental variables, such as air temperature, soil surface temperature, relative humidity, solar radiation, wind speed, incoming and outgoing short and long wave radiation. For these variables, the measurements taken over time are correlated. This places severe restrictions on the applicability of many conventional statistical methods that depend on the assumption of independent and identically distributed errors. In addition to these assumptions, there is serious collinearity among several weather variables and the variables are not stationary. Commonly used multiple regression models can be misleading when predictor variables are stochastic and issues of collinearity and non-stationary are ignored. In this paper, time series analysis is used as a tool to investigate the adequacy of classical regression models. Various aspects of dynamics of cattle body temperature and its relationship to environmental variables are discussed using the frequency and time domain analysis. Finally, we present a detailed approach for fitting cattle body temperature using a transfer function model with multiple environmental variables as inputs.

Creative Commons License


This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Share

COinS
 
Apr 19th, 10:30 AM

COMPARATIVE STUDY OF TIME SERIES AND MULTIPLE REGRESSION FOR MODELING DEPENDENCE OF CATTLE BODY TEMPERATURE ON ENVIRONMENTAL VARIABLES DURING HEAT STRESS

During the summer, a challenging thermal environment is known to cause a significant reduction in food intake, growth, milk production, reproduction and even death in cattle. In this study, we attempt to characterize the relationship of cattle body temperature with several environmental variables, such as air temperature, soil surface temperature, relative humidity, solar radiation, wind speed, incoming and outgoing short and long wave radiation. For these variables, the measurements taken over time are correlated. This places severe restrictions on the applicability of many conventional statistical methods that depend on the assumption of independent and identically distributed errors. In addition to these assumptions, there is serious collinearity among several weather variables and the variables are not stationary. Commonly used multiple regression models can be misleading when predictor variables are stochastic and issues of collinearity and non-stationary are ignored. In this paper, time series analysis is used as a tool to investigate the adequacy of classical regression models. Various aspects of dynamics of cattle body temperature and its relationship to environmental variables are discussed using the frequency and time domain analysis. Finally, we present a detailed approach for fitting cattle body temperature using a transfer function model with multiple environmental variables as inputs.