Heat stress assessments in dairy cattle based on milk MIR spectral data

The effect of heat stress (HS) on milk production traits is well documented. However, HS effect on the detailed milk
composition has been much less studied. First findings in different environments (e.g. Belgium, Tunisia) reported
responses in MIR spectra to HS in dairy cows. This study investigated the effect of HS on milk production traits
and MIR spectrum which reflects the detailed composition of milk samples in the German federal state of BadenWürttemberg (BW). Milk test-day records for the period 2012 to 2019 were merged with temperature and humidity data provided by public weather stations in BW. Values for the daily average temperature-humidity index (THI) were calculated. Statistical analysis was made in R using ‘glmnet’ and spectral data was first standardized, then pre-processed by first derivative. Models relying in THI index were used to compare between HS effect on milk traits (milk yield, fat, protein contents and fatty acids) and 212 points of standardized spectra. Fixed effects were breeds, parity, milking moment, DIM classes, age at calving. The average 3-day lag average THI value was used as a covariate. In this study, the monthly average THI values ranged between 36±7.3 and 65±5.2 from January until August, respectively. The average daily THI was 50±8.6, 49±8.1, 64±5.4, and 37±7.5 during the autumn, spring, summer and winter periods, respectively. The average THI values indicated that dairy cows are exposed to HS during the summer period in BW.
The results showed that there are differences in MIR spectra recorded during HS and thermoneutral conditions; specific wavenumbers of the MIR spectrum were reacting differently. Further analyses will be required to identify potential MIR milk-based phenotypes that could be used in this context for herd management and breeding for resilience to HS.