Wavelet analysis and sinusoidal modelling of environmental variables in a weaned piglet farm

Abstract

Characterizing the indoor environment of livestock buildings is particularly relevant because of its effects on production welfare and GHG emissions. However, the large number of variables involved and the complexity of the processes governing the evolution of indoor environment hinder the identification of the interrelation between environmental variables. This study analyzed outdoor temperature, animal activity and CO₂ and NH₃ concentrations in a weaned piglet building with mechanical ventilation during a complete breeding cycle. The analysis combined a traditional 3-harmonic cosine model with two innovative methods in this area, specifically, continuous wavelet transform (CWT) and wavelet coherence. The results reveal an inverse evolution of outdoor temperature and animal activity, with diurnal peaks, and of outdoor temperature and gas concentrations, with nocturnal peaks and a dominant period of 24 h. Outdoor temperature and CO₂ concentrations show a daily pattern with one distinct maximum and one distinct minimum, whereas animal activity and NH₃ concentrations present two maximum values. A particularly strong correlation is found between outdoor temperature and CO₂ concentrations, with an 11 h shift between series. The relationship between NH₃ concentrations and outdoor temperature is weaker, but the phase difference between both variables is in accordance with the shift for CO₂. The analysis reveals that the animals become active at dawn, 1h20’ before the increase in diurnal temperatures. Continuous Wavelet Transform and wavelet coherence enhance traditional harmonic analysis by providing a more detailed perspective of how the environmental study variables evolve and interrelate over time.