Improving fertility of dairy cows on farm remains difficult, as it is a complex trait which depends on a plethora of physiological, management and external (e.g. semen quality) factors. Even small improvements in the fertility (e.g. in terms of conception rate) potentially make a big difference for cow longevity, lifetime production and as such, sustainability. The availability of (1) new sensor data to get a better grip on physiological status of the cows, and (2) new systems to accurately predict the best insemination window based on milk progesterone, offers new opportunities to analyse and understand fertility performance, and subsequently improve this trait. Before external and physiological factors that impact conception success and are derived from these new sensor data can be investigated, it is crucial that we have certainty on whether the correct insemination time was applied. This study uses an extensive dataset from 5 modern dairy farms with a Herd Navigator systems, which farmers use to determine the optimal moment of insemination upon detection of a drop in milk progesterone preceding luteolysis. Despite that in many cases this identification moment goes well, factors such as sampling rate and milking frequency impact performance of the system, and therefore fertility. More specifically, the use of an algorithm for which a time lag is introduced that depends on both absolute progesterone concentration, rate of change of progesterone during luteolysis and a fixed threshold leads to inconsistencies in detection of luteolysis when sampling rate of milking frequency deviate. In this study, we analysed the factors influencing correct identification of the insemination window, to better separate cases in which conception failed due to human or system errors and cases in which the cow physiology caused poor fertility performance. To this end, we analysed different traits related to insemination and sampling frequency of the Herd Navigator and looked at their effect on success of conception across the 5 farms. We found that in farms with a better fertility performance, the time between the true luteolysis and the insemination time was more consistent. Additionally, we found that the sampling frequency in cycles with deviating characteristics differed and impacted the moment of insemination applied by the farmers. Lastly, we saw that in farms with a lower milking frequency, prediction of the best window for insemination was less reliable and therefore in these farms, the progesterone based milk sampler worked less well. From this analysis, we can make a better selection of the cycles in which insemination time was correct, and thus conception success depended on other factors. This is a first step towards improved understanding of fertility performance on farm, and as such, of a more sustainable dairy sector.

Early predictions of cows’ probability of survival to different lactations would help farmers in making successful management and breeding decisions. For this purpose, this research explored the adoption of joint models for longitudinal and survival data in the dairy field. An algorithm jointly modelled daily first-lactation sensor data (milk yield, body weight, rumination time) and survival data (i.e., time to culling) from 6 Holstein dairy farms. The algorithm was set to predict survival to the beginning of the second and third lactations (i.e., second and third calving) from sensor observations of the first 60, 150, and 240 days in milk of cows’ first lactation. Using 3-time-repeated 3-fold cross-validation, the performance was evaluated in terms of Area Under the Curve and expected error of prediction. Across the different scenarios and farms, the former varied between 45% and 76%, while the latter was between 3.5% and 26%. Significant results were obtained in terms of expected error of prediction, meaning that the method provided survival probabilities in line with the observed events in the datasets (i.e., culling). Furthermore, the performances were stable among farms. These features may justify further research on the use of joint models to predict the survival of dairy cattle.