Development and evaluation of a model to predict herd-level milk fat concentration in Irish spring-calving grazing dairy herds

Description

Heffernan, Ruelle, and Dineen rebuild a model that predicts milk fat levels across Irish dairy herds. Their 2026 study serves spring-calving grazing herds, where milk fat swings with the grazing rotation. The original model spoke in rotations; the goal was a continuous day-of-year version that any calendar date can feed, without losing accuracy.

Rotation coefficients were mapped onto their calendar days and stretched into smooth daily values by linear extrapolation, then slotted into the original regression in place of rotation. On the original evaluation data the rebuilt model predicted strongly, explaining seventy-eight percent of variance with a concordance of 0.85. A national check followed, using 13,788 herd recordings from 5,165 herds collected across the 2023 season. Herd-level forecasting supports feeding and breeding decisions at scale.

Catalog record

Source: https://europepmc.org/articles/PMC13464197