Forecasting reproductive performance in Holstein heifers and cows in a hot environment: a time-series analysis.

Abstract

This study aimed to predict the pregnancy rate (PR) and number of services per pregnancy (SP) in a large high-input dairy herd in a prolonged high ambient temperature zone. Also, the impact of climatic conditions on reproductive performance was assessed. An autoregressive integrated moving average (ARIMA) model was used in data fitting to predict future monthly PR and SP using data from 2014 to 2020. The highest predicted PR for cows was in January (35.3%; 95% CI = 30.5-40.1), and the lowest was in August (12.5%; 95% CI = 7.5-17.6). Temperature-humidity index (THI) and PR were significantly negatively correlated in the same month (r = 0.7) and 2.5 months earlier and 2.5, 5, and 7.5 months later. The predicted highest SP for cows was in September (6.2; 95% CI = 4.8-7.7) and the lowest for March (2.8; 95% CI = 1.3-4.2). The predicted highest PR in heifers was in January (62.2%; CI = 51.6-72.9) and the lowest in May (52.3%; 37.9-66.7). The cross-correlation between THI and PR in heifers was not significantly correlated in the same month, but significantly negative correlations occurred 5, 7.5, and 10 months earlier. SP in heifers were related to seasonality, with the predicted maximum SP occurring in May (1.9; CI = 1.2-2.6) and the minimum in February (1.6; CI = 1.0-2.2). It was concluded that weather strongly influenced the monthly reproductive performance rhythms of Holstein cows and heifers. Also, ARIMA models robustly forecasted reproductive outcomes of dairy cows and heifers in a hot desert climate.