Predicting lifetime production and longevity of organic dairy cows from 1st or 2nd lactation data

Longevity of dairy cows is a key trait combining all functional traits and is decisive for sustainability of dairy production at economic, environmental and ethical level. We used herdbook data of culled Swiss dairy cows to predict longevity (days) and average lifetime daily milk yield (LT_DMY, kg milk) of dairy cows in low-input dairy farms from data easily available for the farmer. First, we assessed the suitability of 1st vs 2nd lactation data to predict longevity and LT_DMY without (n=10,031 cows, 384 farms) and with information on number of inseminations (n=6,011 cows, 372 farms). Second, we tested if lactation curve parameter estimates (LCPs) derived from test day records can be successfully used to predict LT_DMY and longevity (n=1,632 cows, 321 farms). Finally, we investigated breed differences between local dual-purpose breeds and pronounced dairy type breeds in a subset of mixed herds (n=1,796 cows, 72 farms). Models based on 2nd lactation data were consistently better across all traits investigated. Although estimation of LCPs was only possible with sufficient reliability for about 16% of the cows, models including LCPs performed best in predicting LT_DMY with a mean predictability of 73.3%. By contrast, longevity models performed best when using insemination data, but mean predictability only reached 4.6%. Somatic cell count, breed, calving interval, age at first calving, lactation curve persistency, fat protein ratio and information on alpine pasturing were additional traits improving predictions of both traits. Investigation of breed differences in mixed herds revealed lower LT_DMY in the local breeds Simmental and Original Braunvieh compared to Swiss Fleckvieh and Holstein cows. Original Braunvieh lived longer than Holstein (1,949±70 SE vs 1,709±54 SE days, P=0.046), while the local Simmental cows (1,681±61 SE days) did not. We conclude that it seems possible to develop models for TL_DMY, while reliable prediction of longevity remains challenging using information at farmer’s hands