Adsa Logo White Adsa Title White

Inconsistency of precision dairy monitoring technologies measuring the same behaviors.

B. W. Jones




Inconsistency of precision dairy monitoring technologies measuring the same behaviors.
B. W. Jones*1,2, I. C. Tsai3, Y. M. Chang4, J. M. Bewley5. 1Tarleton State University Stephenville, TX, 2Texas A&M AgriLife Research Stephenville, TX, 3DeLaval Lansing, MI, 4University of London London, UK, 5CowFocused Solutions Elizabethtown, KY.

The objective of this study was to evaluate agreement among precision dairy monitoring technologies (PDMT) measuring same behaviors. Holstein cows (n = 108) were fitted with PDMT 10d before their impending calving date. Six different PDMT, namely, IceQube (IQ; IceRobotics Ltd., Edinburgh, Scotland), AfiAct Pedometer Plus (AFI; Afimilk, Kibbutz Afikim, Israel), Track a Cow (TAC; ENGS Systems Innovative Dairy Solutions, Israel), Smartbow (SB; Smartbow GmbH, Jutogasse, Austria), CowScout S Leg (GEA; GEA Farm Technologies GmbH, B�nen, Germany) and Cow Manager SensOor (SEN; Agis Automatisering, Harmelen, Netherlands) were included in the study. Six different behaviors, lying time (hours/day), step (number/day), lying bouts (number/day), activity (hours/day), eating time (hours/day), and rumination (hours/day), were recorded. Before analysis, data was cleaned by removing potential outliers. Spearman's correlation was used to evaluate rank correlation among PDMT. All analyses were carried out in SAS (Cary, NC). For lying time, AFI, IQ and TAC show strong correlations (rs ranged from 0.79 to 0.90), and they were weakly correlated with SB (rs ranged from 0.33 to 0.43). For number of steps, GEA was moderately correlated with all other technologies (rs ranged from 0.50 to 0.61) and all other technologies were only weakly correlated (rs ranged from 0.39 to 0.47). Weak to moderate correlations among lying bouts measures existed (rs ranged from 0.42 to 0.66). Correlations between 2 activity measures and the 2 eating time measures were low (rs = 0.32 and 0.37, respectively). Moderate correlation for rumination time was observed between SEN and SB (rs = 0.61). Overall, there is limited agreement among most PDMT except for lying time. Although PDMT offers exciting opportunities to develop automated monitoring products to inform farm management, quality of captured data requires careful scrutiny.

Keywords: precision dairy monitoring technology, accelerometer.