Modeling heat stress effects on first service to conception rates in Canadian Holstein dairy cattle

Heat stress (HS) can result in decreased production and poor fertility performance in dairy cattle. There is limited understanding of the point at which heat load begins to affect fertility, causing a major challenge for the industry. The temperature-humidity index (THI) is a metric commonly used to represent the realized heat load on livestock, as it incorporates both ambient temperature and humidity percentage. The objectives of this study were to estimate the threshold of THI at which the interval of days from first service to conception (FSTC) begins to increase due to HS, evaluate the effect of estrus synchronization on that threshold, and identify geographic regions of concern for Canadian dairy farming. Data comprised 2,033,928 FSTC records on 1,239,053 Canadian Holsteins in parities 1, 2, and 3, collected between November 2008 and April 2022. Hourly ambient temperature and relative humidity were extracted from the NASA Prediction of Worldwide Energy Resources (POWER) database (https://power.larc.nasa.gov/; accessed Jan. 7, 2025), and daily THI was calculated by averaging hourly THI across the full 24-h day. Daily THI was averaged across 3 temporal windows relevant for follicular growth and zygotic survival: 3 d before to 2 d after first insemination, 7 d before to 2 d after first insemination, and 10 d before to 2 d after first insemination. The FSTC phenotype was adjusted using a single-trait repeatability model to account for known sources of environmental and genetic effects on the trait. Segmented linear-linear regressions were fit to identify the point at which the residual FSTC began to increase in response to HS effects. Thresholds varied between THI 65 and 68.6 across parities and temporal windows. The window spanning 3 d before to 2 d after first insemination showed the highest overall accuracy, and the average threshold for this window (THI 66) was selected as the representative onset of HS effects on FSTC. These results suggest that shorter windows may be sufficient to evaluate HS effects on fertility. Timed artificial insemination (AI) protocols were found to have a greater sensitivity to HS than heat detection methods, displaying a difference in thresholds (timed AI: 64; heat detection: 72.8). However, heat detection inseminations were found to have a higher rate of increase in FSTC above group-specific thresholds (timed AI: 0.44; heat detection: 3.27). Analysis of weather data indicated that Ontario is a province of concern, with average summer temperatures exceeding THI 66. Additionally, Ontario had an average of 50% of 24-h day with THI above 66 as well as 50% of summer days with average THI above 66. All provinces, except for British Columbia, had an observed hottest season within the past 4 years within the dataset. Results of this study suggest an increase in FSTC above THI 66 with a significant effect of timed AI protocol on HS response. A genetic analysis is needed to evaluate a possible genotype by environment interaction and the corresponding re-ranking of sires.