From Data to Decision: HPDA and HPC Reveal the Value of Cooling in Dairy Farming
Cypriot dairy farms face rising challenges as heatwaves increasingly threaten milk production and farm profitability. Estimating the economic value of installing cooling systems is complicated by biological variation among cows and the influence of weather and management practices. Traditional analysis methods struggle to separate these factors, leaving farmers uncertain about the return on investment.
To tackle this, the NCC Cyprus collaborated with the innovative farm consultancy Vettaky and the Kouloubris dairy farm in order to assess the effect of the installation of a cooling system in the field. A Bayesian hierarchical model powered by High-Performance Computing, was developed. The model integrates years of milk yield and weather data and accounts for individual cow variability while inferring the impact of the cooling system on the total productivity of the farm. By performing full Bayesian inference at scale, HPC infrastructure enabled robust, data-driven results that isolate the effects of heat stress and quantify the benefits of intervention with statistical confidence.
The analysis revealed a clear productivity gain: Milk losses under heat stress dropped by approximately 0.31 tonnes per cow per year which corresponds to a median return on investment for the farm of around €7,000, excluding running costs. This provides farmers with concrete evidence to support climate adaptation investments while improving productivity and resilience. The approach can be scaled to other farms and regions, offering a powerful, HPC-enabled tool to strengthen the dairy sector against the growing challenges of climate change.
The figure depicts the posterior probability density of the marginal effect of the cooling system. The analyses showed that the installation cost is well below the 95th percentile indicating a clear positive outcome for the farm.