Predicting longevity-related traits in Swiss low-input and organic dairy cows from herdbook data of first versus second lactation

Dairy cows' longevity integrates multiple functional traits and influences dairy production sustainability. This study assessed the impact of productive lifespan on lifetime productivity in Swiss low-input and organic dairy farms, compared the predictability of longevity and lifetime production using first- versus second-lactation data, and evaluated whether lactation curve parameters (LCP) derived from test-day records improve predictions. Herdbook data of culled cows were analyzed using mixed-effects models to predict length of productive lifespan (LPL; days from first calving to culling), number of lactations until culling (MaxLN), lifetime milk production (LTP; kg energy-corrected milk, ECM), and average daily milk yield during productive lifespan (DMY_LPL; kg ECM). Records from first and second lactation were used in three datasets: a large dataset (n = 10,031 cows, 384 farms), a dataset with insemination records (n = 6,011 cows in first and 5,662 in second lactation, 372 farms), and a dataset with LCP (n = 6,048 cows in first and 6,735 in second lactation, 384 farms). Models based on second-lactation data with insemination records performed best for LPL and LTP, whereas MaxLN was best predicted using first-lactation data with insemination records. Predictability remained low for LPL (0.7%) and MaxLN (0.8%), and moderate for LTP (10.0%). Models including LCP best predicted DMY_LPL (64.9%). Predictive performance improved with additional factors, including somatic cell count, breed, calving interval, age at first calving, lactation persistency, milk yield traits, fat-to-protein ratio, and alpine pasturing. Based on second-lactation data, lifetime milk yield traits were reasonably predictable, whereas longevity prediction remained challenging.