Farm data integration – key to cattle management success (CowData)

• Forschungsstelle: BLW
• Projektnummer: 10.20_11

Schlüsselwörter (Deutsch)

Milchvieh, Informationsmanagementsystem für landwirtschaftliche Betriebe, Farm Management Systems, Entscheidungshilfesystem

Schlüsselwörter (Englisch)

dairy cattle, farm management information system, Farm Management Systems, decision support system

Schlüsselwörter (Französisch)

vaches laitières, système d'information sur la gestion agricole, systèmes de gestion agricole, système d'aide à la décision

Schlüsselwörter (Italienisch)

Bovini da latte, Sistema di gestione delle informazioni aziendali, Sistemi di gestione aziendale, Sistema di supporto alle decisioni

Kurzbeschreibung (Englisch)

The aim of this project is to demonstrate the advantages of a generic data platform. By means of case studies, we will showcase how existing sensor prototypes can be combined with additional on farm (sensor) data, to build new applications and improve decision support systems. The first case study will focus on grassland based production systems. Grassland based production systems often offer a possibility to use marginal land in LFAs (Less Favoured Area) that could otherwise not be used. The second case study will focus on behavioural monitoring. The aim of the second case study is to evaluate the possibilities of combining different information streams in order to develop algorithms for health and welfare monitoring. Besides demonstrating the advantages of a generic data platform by building new applications on the platform in the two case studies, the project also aims to disseminate the ‘success stories’ to different stakeholders and collect their feedback and input.

Projektziele (Englisch)

Dairy and beef cattle farming has become strongly technologized over the past years and applications that evaluate elements within the production system, such as rumination time, milk yield etc., are widely available. Large quantities of data are hereby generated fully- or semi-automated and give indications on improvement potentials within farm management. Following the principle “What you can measure, you can manage”, the project aims to use these data to provide farmers with easy-to-use decision support tools.
Today, farm management information system (FMIS) for dairy cattle exist, aiming to combine all available farm data. In practice, those commercial FMIS are often brand specific, which makes it difficult to integrate sensors from other brands contributing to the vender lock-in. Other commercial FMIS are more generic and connections between automatic milking systems (AMS) and other devices with executive functions are possible. However, integration of new developed techniques or sensors remains challenging. Yet, providing one generic data platform would have a lot of advantages - first of all for technology providers. They can use the data to build new algorithms and applications whether or not in combination with new developed sensors. Farmers will benefit from these new applications and will have the freedom to choose the most appropriate sensor, irrespectively from their milking system brand. This will enhance the implementation of new technology.
The aim of this project is to demonstrate the advantages of a generic data platform. By means of case studies,we will showcase how existing sensor prototypes can be combined with additional on farm (sensor) data, to build new applications and improve decision support systems.
The first case study will focus on grassland based production systems. Grassland based production systems often offer a possibility to use marginal land in LFAs (Less Favoured Area) that could otherwise not be used.These systems have increased in popularity amongst the cattle industry and the society. This is also partly due to the advantages of grazing for animal welfare and higher product quality. Although grassland based production systems and specifically grazing systems offer advantages, there are some deficiencies compared to conventionally housed cattle with concentrate based cattle production. These lie particularly in the limited knowledge on appropriate and sustainable grassland management to optimise the nutrient efficiency within these systems. Small and medium sized farms in the three countries participating in the call often use partial grazing models where it becomes essential to combine the indoor and outdoor environment. Farm Management Systems currently available focus on either the indoor or the outdoor environment. This project aims to provide a link between the two environments. Profound knowledge on the reliability and resilience of the various data sources, as well as their integrability in a decision support system, provides a foundation for the optimization of resource efficiency and thus of the entire production systems sustainability.
The second case study will focus on behavioural monitoring. Different sensors and commercial or prototyped technologies exist. All give some information on health or welfare for example rumination sensors, lameness detection sensors, activity sensors attached to milking clusters etc. The aim of the second case study is to evaluate the possibilities of combining different information streams of cow behaviour in order to develop algorithms for health and welfare monitoring. Hereby, the project will focus on the effects of parasite infections on cow behaviour as well as lameness in cattle.
Besides demonstrating the advantages of a generic data platform by building new applications on theplatform in the two case studies, the project also aims to disseminate the ‘success stories’ to different stakeholders and collect their feedback and input. Active communication with commercially available FMIS will be set up. An outdoor Farm Management System provider, and an indoor Farm Management System provider have expressed their interest in the project and agreed to participate in workshops and discussions.Critical points like open data standards and ownership will be discussed and options to incorporate more commercial sensors or FMIS will be explored.

AIM:
- Combining automatically and semi automatically collected data from the indoor and outdoor farm environment available in different countries into one generic data platform.
- Development of novel algorithms for decision support systems based on sensor data from various existing sensor applications in a holistic approach.
- Validation of the algorithms on research farms in three different countries.
- Development of at least one application with user-friendly interface and evaluation of the application regarding its usability to support cattle farmers, especially in efficient grazing management and health and welfare monitoring.

Abstract (Englisch)

The CowData project aimed to showcase how existing farm (sensor) data can be combined with additional sensor data to supply farmers with decision support tools.

The first trial aimed at the evaluation of an algorithm for Precision Pasture Management. Although we found that the algorithm developed under Irish grazing conditions was not fully applicable to the Swiss part-time rotational system, the field experiment allowed identifying those behavioural indicators, which are most sensitive to gradually decreasing herbage availability under the prevailing grazing conditions.

The second grazing trial aimed to investigate if the decreasing sward height during grazing has an influence on the behaviour of dairy cows. Cow behaviour in a rotational grazing system was found to be complex since it depends not only on sward height, but also on additional factors like amount of additional feed in the barn, sward composition or weather. Because of this complexity, we found that the behaviour of grazing dairy cows cannot be easily applied in a decision support tool for determining the appropriate time for new pasture allocation yet.

The third trial, investigated the effect of the cessation of milking at dry-off on dairy cow behaviour. based on the differences in lying behaviour between cows in the barn and at pasture, we assume that access to pasture buffered drying-off stress to some extent by allowing cows to take different lying positions and pursue their natural lying behaviour.