Abstract
Spatial modelling of water consumption in growing pigs can be a useful tool for
identifying high risk pens or sections in early detection of diseases and various
behavioural problems.
In this study a multivariate dynamic linear model (DLM) is developed based on data from simultaneous monitoring of water consumption across multiple pens in two separate herds. The two herds consist of a commercial finisher herd (Herd A) and a research farm with weaners (Herd B).
Parameters in the model can be defined individually at herd, section or pen level. This spatial distinction allows early warnings to be generated at pen level or merged at section or herd level to reduce the number of alarms. Information on which specific pens or sections are of higher risk of stress or diseases is communicated to the farmer and target work effort to pens at risk.
For Herd A, all model parameters defined at section level resulted in the best fit (MSE =13.85 litres2/hour). For Herd B, parameters defined at both pen and section level resulted in the best fit (MSE = 1.47 litres2/hour).
For both Herd A and Herd B, preliminary results support the spatial approach by
generating a reduced number of alarms when comparing section levels to pen levels.
This study is a part of an on-going project aiming to improve welfare and productivity in growing pigs using advanced ICT methods.
identifying high risk pens or sections in early detection of diseases and various
behavioural problems.
In this study a multivariate dynamic linear model (DLM) is developed based on data from simultaneous monitoring of water consumption across multiple pens in two separate herds. The two herds consist of a commercial finisher herd (Herd A) and a research farm with weaners (Herd B).
Parameters in the model can be defined individually at herd, section or pen level. This spatial distinction allows early warnings to be generated at pen level or merged at section or herd level to reduce the number of alarms. Information on which specific pens or sections are of higher risk of stress or diseases is communicated to the farmer and target work effort to pens at risk.
For Herd A, all model parameters defined at section level resulted in the best fit (MSE =13.85 litres2/hour). For Herd B, parameters defined at both pen and section level resulted in the best fit (MSE = 1.47 litres2/hour).
For both Herd A and Herd B, preliminary results support the spatial approach by
generating a reduced number of alarms when comparing section levels to pen levels.
This study is a part of an on-going project aiming to improve welfare and productivity in growing pigs using advanced ICT methods.
| Original language | English |
|---|---|
| Publication date | 2017 |
| Number of pages | 8 |
| Publication status | Published - 2017 |
| Event | European Conference on Precision Livestock Farming - Nantes, France Duration: 12 Sept 2017 → 14 Sept 2017 Conference number: 8 |
Conference
| Conference | European Conference on Precision Livestock Farming |
|---|---|
| Number | 8 |
| Country/Territory | France |
| City | Nantes |
| Period | 12/09/2017 → 14/09/2017 |