Hierarchical Bayesian modeling of agrivoltaic dairy-calves under missing information

Livestock 5.0 provides a framework to optimize animal welfare and productivity amidst escalating climate change and current pressures to sustainably increase production through limited expansion of agricultural lands, while equilibrating the food-water-energy nexus. Directly tackling the food-water-energy nexus, agrivoltaics emerges as a promising solution for dual-land use for food and renewable energy production. In this paper, we evaluated the effectiveness of agrivoltaics for dairy-calves in a commercial farm through a joint hierarchical Bayesian model, developed to account for non-linear and time-varying relationships, pedigree structure, and missing information. Uncertainty from missing information was propagated to the posterior estimates. While weight at birth and calf sex were not associated with serum Brix, we found that serum Brix was positively associated with colostrum Brix, which allowed to estimate the probabilities of the serum Brix quality (poor, fair, good, or excellent) given colostrum Brix. Poor serum Brix quality was positively associated with disease event rate, with similar rates for the other quality classes. Serum Brix quality was not associated with calf weight, indicating that farm management overcame the negative effects of low serum Brix quality. Previous number of diseases was positively associated with disease event rate, while weight at birth was negatively associated with disease event rate. Agrivoltaics were not associated with disease event rate, indicating that agrivoltaics can be implemented as a safe, dual-use system for increased calf comfort an welfare.