How much floor space do group housed sows need?

Originally appeared in National Hog Farmer, 2017

In the name of animal welfare, more hog producers are investigating the conversion from individual gestation stalls to group housing for pregnant sows. 

Sows in group housing usually require more floor space to maintain productivity and welfare compared with sows kept in individual gestation stalls. Reaching a balance of floor space allowance to maintain animal welfare without compromising sow inventory and total production from the barn can be tricky.

Researchers in this project, conducted at the University of Minnesota Southern Research and Outreach Center in Waseca between February 2015 and October 2016, set out to:

    • determine the minimal square footage needed to successfully house sows in pens with an electronic sow feeder (ESF)
    • evaluate the utility of housing sows in pens with lower space allowance after their social hierarchy is established to determine the efficient use of space and maintain welfare and performance of sows investigate effects of space allowance on performance, health, stress and behavior of gestating sows in an ESF system
    • evaluate method to monitor postural behavior of group-housed sows automatically using accelerometers

A total of 928 sows, ranging in parity from 1 to 9, from 20 breeding groups were used. All sows were group-housed during gestation in pens equipped with an ESF.

The ESF controlled sows’ individual feed intake by means of a radio-frequency identification ear tag. Gestating sows were managed as a static group in each pen.

Floor space allowances studied were 22 square feet, 20 square feet, 18 square feet, 16 square feet per sow, and 16 square feet per sow with more space (22 square feet) during the first week of mixing (22/16 square feet).

The floor space allowance treatments were achieved by adjusting pen size (858 to 948 square feet) and group size (42 to 51 sows).

Each floor space allowance was tested in four pens over 18 months. Sows were moved to ESF pens after being diagnosed pregnant at five weeks after mating. Sows remained in their ESF pens until Day 109 of gestation, when moved to individual farrowing stalls. Sows weaned their litters at about 18 days after farrowing. Sows that were rebred within one week after weaning a litter were considered to have completed the study.

To achieve Objectives 1 to 3, these factors were measured for all sows:

1. performance in gestation pens (body weight, backfat thickness and body condition at entering and leaving ESF pens, daily feed intake, the number of days that the sow did not eat or failed to consume their feed allotment)

2. days in gestation, farrowing rates and lactation performance (litter size, weight farrowed and weaned, pre-weaning mortality, and average daily feed intake during lactation)

3. days in lactation and wean-to-estrus interval (after weaning their subsequent litters)

Welfare indicators, including skin lesions, the number of wounds, prevalence of lameness and fighting among sows were monitored immediately after mixing, two days after mixing and seven days after mixing for all sows. Salivary samples were collected for cortisol analysis from eight focal sows in each pen (balanced for parity) before moving into ESF pens, two days after mixing in ESF pens and before moving to farrowing stalls.

Postural behaviors (lying lateral, lying on their sternum, standing and sitting) were measured in all pens through live observations during Week 3 and Week 9 after sows were moved into ESF pens. Live observations consisted of scan-sampling all sows in a pen at five-minute intervals for four hours during the observation days (two hours in the morning and two hours in the afternoon).

Objective 4

To achieve Objective 4, accelerometers (HOBO Pendant G Acceleration Data loggers, Onset Computer Corp., Bourne, Mass.) were fitted on a front leg and rear leg of eight focal sows in each pen during the same periods that live observations were conducted. Sow activities were recorded at 10-second intervals for three consecutive days.

Postural behaviors were calculated using the model generated in a validation trial. To validate accelerometers for detecting postural behaviors in pigs, an animal trial was conducted at the University of Minnesota West Central Research and Outreach Center in Morris.

Twelve pens of eight pigs weaned at 4 weeks of age were used. In each pen, four focal pigs were selected randomly for behavioral monitoring, which occurred at 5 weeks and 7 weeks of age. Each focal pig was fitted with a digital accelerometer (Onset Pendant G. Data Logger) on the rear leg, which recorded instant acceleration in three dimensions at 10-second intervals for 24 hours.

During the same period, behaviors of focal pigs were recorded continuously using digital cameras. Video recordings were viewed to register postural behaviors that focal pigs were performing continuously for two minutes. Acceleration data series corresponding to each behavior were extracted. Data were analyzed to determine the correct classification of postural behaviors using accelerometers.


The research concluded:

Objective 1 Floor space allowance did not affect weight, backfat thickness, condition scores or feed intake at any point during the study. No differences were detected in farrowing rates (95%, 92%, 94%, 94% and 95% for 16 square feet, 18 square feet, 20 square feet, 22 square feet and 22/16 square feet, respectively), completion rates (83%, 79%, 80%, 86% and 86%), live litter size farrowed (12.5, 12.7, 12.2, 12.3 and 12.5 pigs), litter size weaned (10.4, 10.5, 10.2, 10.2 and 10.6 pigs), litter weight farrowed and weaned, or wean-to-estrus interval among treatment groups.

No differences were observed in aggression among sows (total duration of fighting, frequency of fights and mean duration of each fight) and cortisol concentrations at any time point measured among treatment groups. Skin lesion scores for the body and vulva two days after mixing and before farrowing were similar across treatment groups. Prevalence of lameness two days after mixing was higher for sows allowed 22/16 square feet (9.5%) and 22 square feet (4.2%) than sows allowed 20 square feet (1.8%), 18 square feet (2.9%) and 16 square feet (1.5%), which may be associated with fighting to establish a dominance hierarchy during mixing in pens with larger open areas.

No difference was observed in prevalence of lameness before farrowing among treatment groups. Floor space treatments did not affect postural behaviors at Week 3 and Week 9 in ESF pens. Sows spent most of their time lying laterally (67% to 76% of total observation time across treatment groups) and lying on their sternum (16% to 24%), and less time on standing (2% to 4%) and sitting (under 1%).

As gestation progressed, sows spent more time lying laterally and less time lying on their sternum. Collectively, these results suggest that floor space allowance of 16 square feet per sow supports acceptable reproductive performance and welfare of gestating sows group-housed under conditions of the current study.

Objective 2 No difference was observed in any variables measured between sows with 16 square feet and sows with 22/16 square feet of floor space. Providing extra space during mixing did not benefit the welfare or performance of sows under conditions of the current study.

Objective 3 Increasing floor space allowance from 16 to 22 square feet did not affect the welfare or performance of sows. In general, there were no interactions between floor space allowance and sow parity category (Parity 1 and Parity 2, Parity 3 and Parity 4, Parity 5 and Parity 6, and Parity 7 to Parity 9) for variables measured, which suggests that sows, both young and old, responded to floor space treatments in a similar way. Providing extra space for the entire gestation period in the range tested did not benefit sows under conditions of the current study.

Objective 4 Accelerometers were validated for monitoring the posture of lying laterally and the posture of standing in pigs, with a correct classification rate of 93% for both postures in the current study. Accelerometers had difficulties classifying the posture of lying on their sternum with a correct classification rate of only 54%.

The posture of lying laterally was estimated for sows in group-housing systems using accelerometer data, which showed a pattern similar to the results obtained from live observations. However, using accelerometers to detect postural behaviors in loose-housed sows presented challenges to data recording and processing.

Researchers: Yuzhi Li, Shiquan Cui, Haifeng Zhang and Lee Johnston, University of Minnesota West Central Research and Outreach Center, Morris; and Sam Baidoo, University of Minnesota Southern Research and Outreach Center, Waseca. For further information, contact Li at [email protected]. Funding for this study was provided by the National Pork Board and the Pork Checkoff.