Modelling CO2 production of fast- and slow-growing broilers and quantifying litter contribution

Determination of the ventilation rate (VR) is essential to quantify gaseous emissions from broiler houses. When direct measurements are unavailable, the CO2 mass balance method is a common alternative. In practice, the CIGR model (published in 2002) is widely used for the estimation of the CO2 production. However, due to the genetic evolvements of modern broilers, it remains a question if the CIGR model needs to be updated. This study aimed to validate the CIGR model and explore other alternatives. Respiration chambers were used to quantify the CO2 production and litter contribution for fast-growing broilers (FGB, housed for 42 days) and slow-growing broilers (SGB, housed for 56 days) for two production rounds. FGB had a total CO2 production of 3.36 g h−1 bird−1, with litter accounting for 9.0% (0.28 g h−1 bird−1). SGB had a total CO2 production of 3.02 g h−1 bird−1, with litter contributing 5.8% (0.23 g h−1 bird−1). Friability and wetness scores of the litter were negatively correlated with litter CO2 production for SGB. The CIGR model overestimated total CO2 production by 6.2% for FGB and 19.9% for SGB, with Normalised Root Mean Squared Error (NRMSE) being 21.2% and 31.7%, respectively. To improve the accuracy of VR estimations, statistical Morgan–Mercer–Flodin (MMF) models for FGB and SGB, and physiological MMF models for both breeds were developed for total and bird CO2 production. All MMF models had R2 higher than 0.99, NRMSE lower than 5.3%, with random error accounting for more than 97.7% of the total error.