Investigating piglet crushing by the sow: A data mining approach

Gabriela Munhoz Morello, Purdue University

Abstract

Farrowing crates represent approximately 88 % of the pig farrowing systems in the U.S. (USDA`s NAHMS, 2006). This farrowing system is very advantageous for production efficiency and management; however, it presents several pig welfare issues that are subject to concern. Generally, alternatives to farrowing crates provide a little more flexibility to the sows to communicate better with their litters, as well as to perform more exercise, posture changes, and more natural behaviors during lactation. A major concern with the alternative farrowing systems is the higher pre-wean mortality, especially due to piglet overlay. High pre-wean mortality rates impact not only the producer`s profitability, but also the welfare of young piglets. Several studies demonstrate a high variability of crushing rate among sows submitted to the same physical environment. It is possible that factors other than the system’s physical configuration may be affecting the maternal behavior of sows and causing increased crushing. Therefore, the objective of this study was to investigate how environmental and management factors affect piglet crushing, as well as the duration and frequency of sow postures and posture changes. The study was conducted during May, 2013, through June 2014 in a 10,000 sow breeding to farrowing operation facility, located in northwestern Indiana, U.S.A. The present research focused on the first 48 h post-farrowing, which is when most of the crushing occurs. Data collection was concentrated in two farrowing rooms where sows were placed in 60 individual farrowing crates at each of a total of 27 experiment replicates. A total of 1287 sows and their litters were studied in this experiment. Production and management information were collected daily from the farm records, during the experiment replicates. Number of piglets born (alive, dead, mummy), number of liveborn piglets that died and the death causes were recorded daily. Birth assistance and manual checking of the sows cervix and vagina were also recorded daily. Sow behavior was recorded continuously for the 48 h post-partum for a sub-set of 59 sows (19 from spring, 11 from fall, 17 from summer and 12 from winter). Frequency and duration of posture changes were quantified from the behavior videos and evaluated as a function of the environmental variables measured. The procedure GLMSELECT was used on SAS 9.3 (SAS Institute, Cary NC) to select the most appropriate environmental variables affecting the duration and frequency of each sow posture and posture change. A data mining classification technique was used to identify patterns within the environment which led to piglet crushing. Approximately 12.9 ± 3.1 piglets were born alive per sow in the present study. Crushing by the sow was the main cause of mortality, accounting for 63 % of all piglet mortality causes. The second main identified cause of mortality was due to piglets being too weak and small which generally accounted for approximately 17 % of all piglet mortality within the 48-hour post-partum period, followed by splayed-leg which accounted for 10 % or less of all the piglet mortality. The statistical analysis indicated that environment also affected the duration and frequency of so posture changes. The increase in time exposure to T within 22 °C to 26 °C led to increased duration of a change from standing to lying sternally (P < 0.01). Increased exposure to RH of 50 % to 60 % tended to increase time taken to lie sternally (P < 0.10). Increased mean crate AV led to an increase in time taken to lie both sternally and laterally. Increased exposure to LI of 20 Lx to 40 Lx also led to increased time taken to lie down, whereas increased exposure to higher LI levels led to increased lying down events per hour. Surprisingly, there was an effect of number of fan energizing events on the behavior of sows, in that the more fans were turned on and off, the more sows changed their posture per hour (P < 0.01). Fan energizing events were partially correlated with LI and SI. Thus, it is possible that the LI and SI environments were suddenly being altered when fans were turned on and off, which possibly made sows more active. Increased sow posture changes per hour and decreased time taken to lie down have been previously associated with reduction in crushing rate. Therefore, the results of the present research indicate that changes within the sow microenvironment may allow for a substantial reduction in the crushing rate in farrowing facilities. Moreover, extreme conditions of T, RH, LI, AV, and mat T should be avoided not to increase crushing. (Abstract shortened by ProQuest.)

Degree

Ph.D.

Advisors

Richert, Purdue University.

Subject Area

Agriculture|Animal sciences|Agricultural engineering|Veterinary services

COinS