PERFORMANCE TREND MEASUREMENT POSSIBILITIES IN AUTOMATIC MANAGEMENT SYSTEMS.

Dairy farmers increasingly are employing precision farming practices and computer software that enables them to manage large herds at the individual animal level. In South Africa, dairy farmers have been adopting similar strategies with a trend towards larger production units and the incorporation of automatic milking systems (AMSs). The software used in automatic systems can record production, reproduction and health parameters daily. The systems record all variables and movements of individual animals, from the day of birth to the day the cow exits the herd. Production levels, milk flow rate, milk composition, milk conductivity, milking time, body weight and activity levels are recorded routinely. A centrally managed software system combines these variables to identify and sort animals with sub-optimal production levels, animals with possible signs of heat and infectious or metabolic ailments. Birth dates, insemination dates and calving dates are documented and saved, and applied for monitoring fertility traits. The technology assists dairy breeders to scrutinize milk yield, fertility and health records of animals, enabling informed management decisions and has the potential to increase selection accuracy. 

Milk recording in the country has declined considerably and therefore from an industry point of view data and statistics to monitor performance and other progress are becoming inadequate and non-representative. AMS data offers a solution because of the many variables recorded, but the majority of AMS producers do not participate in national recording since they can obtain the necessary information themselves for facilitating day-to-day management from the system. Extraction of the data however does not include time trends which are necessary to evaluate progress in different critical variables. Therefore, the aim of the study referenced below was to perform a production analysis with the primary objective of constructing a template for extracting and analysing herd performance data which can be used by all farmers using AMSs in South Africa. 

Two large herd dairy producers representing respectively a total mixed ration (TMR) system, and a pasture-based production system participated in the study. By extracting retrospective animal records from multiple years, comprehensive data tables were constructed for different production analyses. The data was extracted from the AfiFarm herd management software from S.A.E Afikim, Kibbutz, Israel. Analyses included time-trend evaluation of herd numbers, mean production and reproduction performance at the heifer and cow level, distribution of exit reasons and assessing the relationship between the genetic merit of sires and the mean performance of progeny.

Findings in the study confirmed that automatic management systems permit extraction and analyses of multiple variables imperative to dairy management at the herd and cow level. The software used in these systems has the potential to serve as a platform to add a vast number of dairy cow performance records for future analyses. 

This study was the first production analyses in both a TMR and pasture-based AMS in South Africa. Results provided insight into the possibilities for extracting data from automatic software and should serve as a platform for administering future research on documented animal records. A production analyses performed on historic data can provide added value to dairy farmers by tracking the performance of animals and management decisions, which can be continuously updated by comparing previous results with findings from present-day data.  

Recommendation: Farmers should be informed of the potential of the “large data” at their disposal. Mutual trust between the MPO, scientists, consultants and farmers with AMS can facilitate that a large proportion of farmers using AMS can provide access to their data to the benefit of themselves and the industry. Benchmarks for economically important production, reproduction and health traits can be established for farmers with similar systems, i.e. similarities in geographical location, feeding systems employed, herd size and breeds used. By comparing results, herds that are unable to reach benchmark threshold values for traits measured can be identified.  

Reference: 

A.U. Gresse, 2019.Alternative approaches for analyses of production performance from automatic milking systems in South Africa. MSc Agric. Thesis, Department of Animal and Wildlife Sciences, University of Pretoria.