A multi-disciplinary approach to improve the environmental performance of niche pork production systems and marketability of Heritage swine breeds

Final Report for LS08-211

Project Type: Research and Education
Funds awarded in 2008: $175,000.00
Projected End Date: 12/31/2012
Region: Southern
State: North Carolina
Principal Investigator:
Sang Hyon Oh
North Carolina A&T State University
Expand All

Project Information

Abstract:

This project investigated niche pork production systems that address market demands and natural resource conservation concerns. The result of this study will help determine the sustainability of outdoor systems utilizing a multi-disciplinary, systems-based research approach to better understand: 1) swine grazing systems that best maintain vegetative cover, 2) the relationship between stocking density, vegetation survival, soil loss and nutrient distribution, 3) the potential for different grass species to best withstand grazing by gestating sows, 4) heritage breeds that are best suited to deep-bedded hoop barn production systems, and 5) taste characteristics and consumer preferences for heritage breed pork products.

Project Objectives:

1. Determine how stocking rates affect vegetation survival, soil disturbance and nutrient distribution under different management schemes on perennial pasture.
2. Determine the effect of grazing management on nutrient loading and nutrient distribution in pastures.
3. Determine which of four vegetation types will continue to protect the soil when gestating sows have access for grazing and lounging during the gestation period.
4. Analyze the growth characteristics for heritage crossbreds to be raised in the hoop barn.
5. Compare meat quality of heritage crossbreds produced with different sire lines.

Introduction:

This project investigated niche pork production systems that address market demands and natural resource conservation concerns, with a specific focus on maximizing vegetative ground cover and nutrient distribution in pastures and understanding marketability of heritage breeds produced in alternative production systems. Several wholesale buyers are offering North Carolina growers twice the current market price to supply this demand. This represents an important opportunity for small-scale, limited resource farmers across the southeast. The majority of niche pork buyers require that pigs be raised outdoors on pasture or in deep-bedded hoop barn systems. In the case of North Carolina’s outdoor production systems, most operations are not sustainable if one evaluates nutrient loading, soil loss, and water quality impacts. Natural Resource Conservation Service (NRCS) planners recognize the near total lack of information available to guide them in developing conservation plans that addresses resource concerns on outdoor swine farms across the southeast. Lastly, consumer interest in heritage breed pork continues to rise however little is known about taste characteristics and production potential in alternative systems.

Conventional confinement pork production systems are often criticized for animal welfare and environmental impacts. In response, new markets have emerged for animal compassionate and outdoor pork production. Alternative production systems, including “pasture-based” and deep-bedded hoop barns have the potential to address these concerns as well as provide new opportunities for small, limited resource farmers. However, when pigs are allowed to exhibit their natural behaviors of rooting, wallowing and excreting waste in preferred areas, they can be highly destructive of vegetation and contribute to significant soil loss and nutrient build-up and run-off. Many limited resource farms appear to be overstocked and lack documented guidance on how to manage the animals & cropping/grazing management to prevent natural resource degradation. Furthermore, according to the ALBC, several pig breeds with potentially desirable consumer attributes are considered threatened or endangered. Limited research has been conducted to determine the suitability of heritage breeds to alternative production systems or to evaluate consumer preferences for meat products from heritage breeds.

This project will help determine the sustainability of outdoor pork production systems utilizing a multi-disciplinary, systems-based research approach to better understand: 1) swine grazing systems that best maintain vegetative cover, 2) the relationship between stocking density, rotation frequency, vegetation survival, soil loss and nutrient distribution, 3) the potential for different grass species to best withstand grazing by gestating sows, 4) types of heritage breeds that are best suited to deep-bedded hoop barn production systems, and 5) taste characteristics and consumer preferences for heritage breed pork products.

Cooperators

Click linked name(s) to expand
  • Dana Hanson
  • Morgan Morrow
  • Charles Raczkowski
  • Todd See

Research

Materials and methods:

Materials and Methods – Objective 1: Stocking Rates
The experiment was conducted for 91 days during summer 2008 and repeated in Spring 2009. A total of sixty purebred Yorkshire female and castrated pigs (18.4 and 118.5 kg initial and final BW, respectively) were randomly assigned to groups of five animals, in twelve paddocks sized (1351; 676; 450 and 338 m2) to equal the stocking rates under evaluation: 37, 74, 111 and 148 pigs/ha (15, 30, 45 and 60 pigs/a). Animals had ad libitum access to concentrate feed (16 % CP) and water. Wallows, shade, feeders and nipple waterers were provided at fixed locations within each paddock. The paddocks were managed under a continous management system. Ground cover was assessed weekly using a step-point technique with transect lines evenly spaced across the paddocks. The experimental design was a randomized complete block, with three field replicates. Data were log (log[x+10]) and square root ([x+10]1/2) transformed for percent bermudagrass cover (BGC), vegetation cover (VC), and bare soil (BS).

For gestating sows, a rotational grazing system experiment was performed during winter 2009 (January to March), and repeated during fall 2009 (September to October) and spring 2010 (April to May) to establish the effect of sows stocking rate on the ground cover of a dormant bermudagrass (Cynodon dactylon) stand. Each plot (0.4 ha) was divided into 9 sections. Twenty Yorkshire mature sows (avg BW: W09: 294; F09: 212; S10: 186 kg) were allocated to each of three stocking rates (10, 15 y 25 sows/ha) (4, 6 and 10 sows/a). Sows had permanent access to the central section, being defined as a heavy use area where shelter and water were provided, and to one of the other eight sections for one week. Sows were restrictedly fed concentrate daily in the morning (3 kg, 15% CP). Used conveyor belts were used as feeders and moved weekly to the section where the animals would graze. A step point procedure was performed weekly to estimate vegetative cover changes using evenly spaced transect lines.

Materials and Methods – Objective 2: Nutrient Monitoring
Management guidelines are needed which reflect a greater understanding of the impact of different swine stocking densities and movement patterns on nutrient loading and distribution. Soil composites samples were collected using a hand auger before the start and after animal removal at two plot locations (L) in gestating sow trials: heavy use area (HUA) or other sections (OS), and two depths (D): 15 cm (D1) and 30 cm (D2). Samples were analyzed at the NCDA soil laboratory, and standard procedures were implemented to measure nutrient concentrations. Statistical differences were established between SR for S (P = 0.03), Cu (P = 0.05), and Na (P = 0.02). Nutrient concentrations varied among L for Fe (P = 0.02). Except for P, nutrient concentrations were higher in D1. Following the removal of sows from the plots S, Cu and Na soil concentrations were higher with the highest SR.

A winter cereal rye/annual ryegrass (Secale cereale/Lolium perenne) mixture was sown into the pastures following removal of the pigs in fall 2009. The forage was harvested at early boot stage to 5cm stubble length in spring 2010 and followed by a sorghum-sudangrass hybrid no-till seeded into the plots and harvested at early boot stage to 5cm stubble in August and again at stem elongation in October 2010.

Soil samples were taken in spring and fall 2010 following the final harvest of each forage crop. Soil compaction and root numbers were determined at depths of 15, 30, 45, 60, 90 cm, and 15, 30, 60, 90 cm, respectively. A Delmi soil penetrometer was used to determine soil compaction and root numbers were determined by counting the roots in the cross-sections of harvested soil cores. The experimental design was a randomized complete block, with three field replicates. The experimental design was a randomized complete block with two field replicates. Data were log (log[x+10]) and square root ([x+10]1/2) transformed.

Materials and Methods – Objective 3: Survival of Vegetation Types during the Gestation Period
Because of their extensive rooting behavior, gestating sows raised outdoors cause the greatest disturbance to soil. Research is needed to determine which grass species planted in outdoor production systems can best protect soil integrity. A 4-paddock pasture system with four different types of grasses was used to evaluate grazing preferences as well as ground cover and plant survival. Four different grasses – the wild-type, infected Kentucky 31 Tall Fescue, the novel-type, infected Max Q Fescue, a multi-species pasture (Redtop, Kentucky bluegrass and Kentucky 31 fescue) and common bermudagrass as a control were established in sub-paddock areas, respectively. These are some of the grasses commonly found in the Piedmont area of NC. The first two are cool-season grasses that are well adapted to trampling and grazing, and bermudagrass is probably the most popular warm-season grass in the area. The inclusion of Max Q tall fescue responds to the need of knowing how this grass (a relatively new one in the market) responds under severe trampling and grazing conditions.

Six gestating sows used for this experiment were antibiotic free Yorkshire breed from CEFS, and they were bred with heritage breeds such as Berkshire, Large Black, and Tamworth (see Objective 4); once slaughtered, meat from these carcasses were used to evaluate meat quality of these progeny. Animals had continuous access to shade, water and feed in a common area with access to all four forage types at a stocking rate of 6 sows per acre. Pastures were used by sows during three gestation and breeding cycles for two years. Global positioning system was used to track the activity of each sow in the paddock and the ground cover was determined using a step-point technique with transect lines evenly spaced across the paddocks.

Materials and Methods – Objective 4: Growth of Heritage Crossbreds
The experiments were performed at the Center for Environmental Farming Systems (CEFS)/Cherry Research Station of the North Carolina Department of Agriculture and Consumer Services, and at the University Farm of North Carolina Agricultural and Technical State University (NCA&T). The CEFS Alternative Swine Unit is located in Goldsboro (latitude +35° 23′ 26.82″, longitude -78° 1′ 43.76″), and the NCA&T farm is in Greensboro (latitude +36° 4′ 16.63″, longitude -79° 43′ 33.02″). Both NC cities feature a humid subtropical climate (Köppen climate classification; Lohmann et al., 1993), with subtropical summer temperatures and mild winters, and average annual precipitation of about 110 cm.
The CEFS unit has been raising antibiotic-free Yorkshire sows in hoop structures for more than 10 years. For Trial 1, 24 gilts approximately 6 months of age were moved from CEFS to NCA&T. For Trial 2 and 3, fifty-four gilts total (the 24 at NCA&T; 30 at CEFS) were artificially bred with the semen of Berkshire (BY), Large Black (LBY), Tamworth (TY) or Yorkshire (YY; control) boars with 10 sows bred per sire breed at CEFS and 8 sows bred per sire breed at NCA&T. The same sows were used for each trial with sire breed randomly chosen each time. All sows were estrus synchronized with Matrix® as per the manufacturer directions prior to breeding. Sows were farrowed in a hoop structure. Pigs were castrated within a week and weaned when the youngest litter was 4 weeks old. Farrowing for Trials 1, 2 and 3 occurred in May and October 2009, and April, 2010, respectively. All pigs in each trial for each farm were raised as one group in a deep-bedded hoop (16 m x 32 m) from weaning through to finishing. The deep bedding, straw, corn stalks, or hay, was spread approximately 35–45 cm thick whenever needed and provided a comfortable environment for the animals, which allowed rooting and other natural behaviors.

Birth and weaning weight
For Trial 2 and 3 and at NCA&T only, body weights were recorded on the day after birth. Sire breed and trial were included in the statistical model for analysis. Weaning weights were recorded for all trials. PROC GLM was used to analyze the data (SAS 9.2), and location, sire breed, and trial were included in the statistical model as fixed effects. For all analyses, non-significant interactions were excluded from the final model. Least square means were estimated to compare the level of each effect with the pdiff option.

Body weight, Feed:Gain and ADG
Body weights were measured manually in both locations approximately every 30 days beginning at around 60 days of age and up to 240 days of age; weights were adjusted for common days of age (60, 90, 120, etc.) by multiplying actual ADG for each pig by the common day of age, rounding to the nearest 15 days of age for the common day of age (i.e. for pigs that were 70 days of age at weighing, weights were adjusted to 60 days of age). A FIRE (Feed Intake Recording Equipment, Osborne Industries Inc. Osborne, Kansas) system with 8 feeding stations was used at NCA&T. Body weights and daily feed intake for 106 finishing pigs were recorded from March to November, 2010 at NCA&T (from approximately 140 to 210 days of age for Trial 2 and 3) using the FIRE system, resulting in 101,394 observations. Data was eliminated before analysis if feed intake per visit was greater than 2 kg (Casey and Dekkers, 2001). The complete feed intake record for each pig was then evaluated for outliers by plotting feed intake by day and testing each feed intake observation with the Cook’s D test statistic. Outliers were removed based on the values from the equation of Cook’s D that were greater than 4/n, where n is the number of observations (Cook and Weisberg, 1982). After removal of outliers, 86,067 daily feed intake and body weight records were utilized in the subsequent analysis. This data was used to compare growth performance among breed types including growth pattern, feed:gain ratio, and ADG. Least square means of body weight were estimated with Proc Mixed in SAS 9.2 for fixed effects such as breed type and days of age within the sire breed. The differences within fixed effects were compared using least significant differences with a DIFF option.
For the adjusted body weight data collected manually in both locations, the same factors described above were included as fixed effects in the statistical model using Proc Mixed in SAS 9.2. The effects of trial, location, and interactions were not significant and were thus not included in the final model.

Materials and Methods – Objective 5: Carcass Traits & Meat Quality
From Objective 4, Trials taking place in Fall 2009, Spring 2010, and Fall 2010 were designated as Trial 1, 2, and 3, respectively. Sows farrowed in hoop structures and pigs were castrated within a week then weaned when the youngest litter was 4 weeks old. Weaned pigs were reared within deep-bedded hoop houses until harvest with ad libitum NRC based diets balanced for stage of production and free choice fresh water. The deep bedding, generally straw, corn stalks, or hay, was spread approximately 35–45 cm thick as needed to provide a comfortable environment for the animals which allowed rooting and other natural behaviors.

Carcass Data Collection
Carcass characteristics were measured using 104 randomly selected animals that were harvested at a USDA-inspected abattoir at approximately 200 days of age. At harvest, hot carcass weight (including the head) was collected prior to refrigeration. After refrigeration for 24 hours, carcass collection procedures followed NPPC guidelines (NPPC, 2000); back fat (BF) depth at 1st rib, 10th rib and last lumbar were collected as well as longissimus muscle area (LMA). The longissimus dorsi (LD) was collected from the right side loins. All the boneless loins were packed in ice and transported approximately one hour to the NCSU Processed Meat Laboratory for storage at 2?C until further analysis.

Pork Quality Measurements
Chops were cut 2.54-cm thick from each of the LD samples. Marbling score (1-10) and color score (1-6) was recorded (NPPC, 2000) and a LD chop was used to determine drip loss by placing a 100 g sample on a hook and hanging it in a plastic bag at 2?C for 48 hours. To determine ultimate pH, a chop sample was homogenized with a variable speed laboratory blender (Waring, New Hartford, CT) and deionized water added to a final dilution of 1:10. Samples were blended for 20 seconds and pH was determined using an Accumet Excel XL15 pH meter with glass tip probe (Thermo Fisher Scientific, Waltham, MA). Color of the LD was objectively evaluated by Minolta L*, a*, and b* measurements using a Minolta Chroma Meter (CR-200, Ramsey, New Jersey) using D65 illuminant and calibrated with a standard white plate. Minolta values were reported as the average color values from measurements conducted at three positions on the surface of each chop after a minimum of 20 minutes of the initial cut. Slice shear force was conducted to estimate tenderness (Shackelford et al., 2004).

Pork quality characteristics and sensory panel tests
Pork quality and sensory panel tests were conducted at North Carolina State University. The first sensory panel test was held with 109 consumers using pork from the pigs in Trial 1, and the second test was with 107 consumers using pork from the pigs in Trial 2 and 3. Consumers were recruited through a screener launched to an on-line database maintained by the Sensory Service Center with over 3,000 members. Consumers were compensated with a five dollar grocery store gift card for their participation. Consumers had to be at least 19 years old, no older than 59 years old, and consume pork at least a few times per year.
Chops for each test were thawed at 38o F and cooked on an Impinger conveyor oven set at 400o F. After being wrapped and labeled, chops were placed in a warming cabinet set to 170o F. Samples were replaced with fresh samples after 30 minutes. Samples were cut for consumers using a cutting mold, resulting in a 2×2×1 cm sample of pork. Samples were placed in labeled containers and presented to consumers with a fork, napkin, and cup of deionized water.
When consumers arrived for the taste test, they were instructed they would taste four samples of pork chops and would have to rank them in order of preference at the end of the test. Consumers answered demographic questions identical to what was asked on the recruitment screener. Consumers were then presented with their first sample and asked to indicate their overall liking on a 9 point Hedonic scale for which 1 = dislike extremely and 9 = like extremely. Consumers then indicated their overall flavor, juiciness and tenderness liking on the 9 point Hedonic scale. After all four samples were tasted, consumers ranked the samples in order of preference in which 1 = most preferred and 4 = least preferred.

Data analyses
For pork characteristics, location, trial, and sire were included in the statistical model as fixed effects using PROC GLM in SAS 9.2. Hot carcass weight was also included as a covariate in the model. For sensory panel data, scores were evaluated by ANOVA using SAS 9.2. Location, sire and interaction between them were included in the model as fixed effects. Overall ranking was evaluated by Friedman’s Rank Sum using XLSTAT. All statistics were calculated to 95% confidence.

Research results and discussion:

Results and Discussion – Objective 1: Stocking Rates
Orthogonal contrasts showed linear effects of stocking rate (SR) for bermudagrass ground cover (P = 0.0001), bare soil (P = 0.0001) and vegetation cover (P = 0.0001) in purebred Yorkshire female and castrated pigs (18.4 and 118.5 kg initial and final BW, respectively). Percent bermudagrass ground cover and vegetation cover decreased (87.4, 78.8, 73.8, 60.4 and 92.6, 84.8, 79.9, 67.6, respectively) whereas percent bare soil increased with increasing stocking rate (7.4, 15.2, 20.0 and 32.4). Daily gain was not influenced by stocking rate (avg: 0.89±0.02 kg/d). For the gestating sows, the SR affected LV (P = 0.09; 41.71a, 27.97b and 28.0b %, respectively, for 10, 15 and 25 sows/ha) whereas BS did not change (P = 0.31; 23.6, 32.7 and 34.4%, respectively, for 10, 15 and 25 sows/ha). Season had a pronounced effect on VGC components: LV W09: 7.2a, F09: 38.2b; S10: 52.3c%; P = 0.0005) and DR (W09: 67.2a, F09: 23.3b; S10: 17.9%b; P = 0.0002). Conversely, S had no effect on BS (W09: 22.5; F09:38.4; S10: 52.3%; P = 0.14). Statistical differences were established between SR for S (P = 0.03), Cu (P = 0.05), and Na (P = 0.02). Nutrient concentrations varied among L for Fe (P = 0.02). Except for P, nutrient concentrations were higher in D1. Following the removal of sows from the plots S, Cu and Na soil concentrations were higher with the highest SR.

Results and Discussion – Objective 2: Nutrient Monitoring
In spring, forage yield (avg: 5898 kg/ha; P = 0.72), soil compaction (avg: 78 kg/cm2; P = 0.4) and root numbers (avg: 2.8; P = 0.8) did not change with SR, but soil compaction (50, 87, 95, 77, 74 and 83 kg/cm2, respectively; P = 0.0001) and root numbers (avg: 8.1, 2.1, 0.59, 0.40, respectively; P = 0.0001) were affected by soil depth. Summer (avg: 6333 kg/ha; P = 0.08) and fall (avg: 783 kg/ha; P = 0.7) sorghum-sudangrass yield were not affected by SR. Fall root numbers were not affected by SR (avg: 2.9; P = 0.2) but strongly affected by soil depth (avg: 7.6. 2.5, 1.0 0.5, respectively; P = 0.0001). To maintain vegetative ground cover above 80% with continuous access to pasture, stocking rate must be kept below 74 hogs/ha during the finishing phase. Even though Ground cover was reduced with the heaviest sow stocking rate, under the conditions of these experiments, season had a more pronounced effect on vegetative ground cover than stocking rate. Following the removal of sows from the plots S, Cu and Na soil concentrations were higher with the highest SR (15 sows/ha). SR rate had no effect on subsequent forage yield. Conversely, soil compaction and root numbers were affected by soil depth, the latter being most prevalent in the top 15 cm.

Results and Discussion – Objective 3: Survival of Vegetation Types during the Gestation Period
Average staying times of gestating sows as well as survival of vegetation were not significantly different among different vegetation pasture areas. As a result, grass varieties do not significantly affect the rooting behavior of gestating sows among four different grasses – the wild-type, infected Kentucky 31 Tall Fescue, the novel-type, infected Max Q Fescue, a multi-species pasture (Redtop, Kentucky bluegrass and Kentucky 31 fescue) and common bermudagrass, while most time was spent in multi species, and rooting was exhibited in Bermuda grass pasture. Future trials will hopefully allow use of a more accurate tracking device as well as a larger sow number and more replications. Current data demonstrates that the sows frequently migrate towards the common area associated with Bermuda grass and the multispecies variety.

Results and Discussion – Objective 4: Growth of Heritage Crossbreds
Birth and weaning weight
For individual birth weights, there was an effect of breed (p<0.05) in which BY pigs were the lightest (p<0.05; 1.34 ± 0.03 kg), TY (1.49 ± 0.05 kg) and YY (1.42 ± 0.03 kg) pigs were the heaviest but similar to each other and LBY pigs were intermediate (1.40 ± 0.04 kg). However, litter birth weights were not influenced by sire breed, indicating that differences in individual birth weights were likely a reflection of litter size. However, because total number of litters per breed was limited, additional research would be beneficial. Individual (0.95 ± 0.02 and 1.87 ± 0.03 kg) and litter birth weights (7.86 ± 0.81 and 18.56 ± 0.98 kg) were significantly different (p<0.05) between Trials 2 and 3, respectively. The warmer weather typically seen in April (Trial 3) compared to October (Trial 2) may have contributed in part to the higher birth weights seen in Trial 3. Similarly, in a cooperative study with 999 litters from seven southern states (including North Carolina), pigs born in the warm season were heavier at birth than those born in the cool season (Coffey et al., 1994).
The effects of sire breed and trial were significant for individual weaning weights (p<0.05). The BY and LBY pigs were heavier than TY and YY pigs (7.76 ± 0.15, 8.07 ± 0.22, 6.40 ± 0.34 and 6.76 ± 0.16 kg, respectively). The IWW for the first trial (6.32 ± 0.24 kg) were significantly (P<0.05) lower than those in Trial 2 (7.89 ± 0.16 kg) and 3 (7.54 ± 0.15 kg). However, IWW were similar for CEFS (7.36 ± 0.18 kg) and NCA&amp;T (8.13 ± 0.13 kg).
For litter weaning weights, as with litter birth weights, there was no influence of sire breed, averaging 44.64 ± 3.97, 55.36 ± 6.55, 32.17 ± 10.36 and 38.13 ± 4.27 kg for BY, LBY, TY and YY, respectively. The LWW were lightest for Trial 1 (28.19 ± 6.45 kg), heavier for Trial 2 (42.61 ± 4.71 kg) than Trial 1 and were the heaviest for Trial 3 (56.93 ± 4.69 kg) (p<0.05). Location LWW were similar, averaging 38.28 ± 5.39 kg for CEFS and 46.87 ± 3.54 kg for NCA&amp;T. Gilts were used for Trial 1 and the same animals were used for Trial 2 and 3, so the age and parity of most of the females used naturally increased with trial, making it expected that weaning weights would also increase with trial. An increase in gain from birth to weaning over three parities was also noted for a collaborative study of 999 litters from seven southern U.S. states, including North Carolina (Coffey et al., 1994).

Body weight, ADG, and Feed:Gain
Tamworth sired pigs, included only in Trial 3, had fewer animal records than the other breed types, therefore, the estimates were very limited compared to other groups. Overall, the LBY group was heavier (p<0.05) than other groups for the FIRE system (Figure 1) and were also heavier (p<0.05) for adjusted manual weight measurements from 90 to 240 days of age (Table 3). However, average daily gain (from 140 to 210 days of age) was not influenced by breed type (0.79 ±0.03, 0.80±0.04, 0.97±0.09 and 0.82±0.03 kg/d for BY, LBY, TY and YY, respectively), and overall weight gain was not higher for LBY, indicating no actual benefit of LB sires over the other breeds (data for actual manual weights analyzed). As noted in the current study, no difference in ADG for breed type (Duroc sires with Tamworth, Tamworth x Landrace or Hampshire x Landrace sows) was noted in a previous outdoor versus confinement study conducted at North Carolina A&amp;T State University, though outdoor raised pigs grew 50% faster than confinement raised pigs (Talbott et al., 2003).
Overall weight gain from 60 to 240 days of age was lowest for TY (60.63±3.56 kg) and similar for the other breed types (72.51±1.48 kg for BY, 73.72±1.77 kg for LBY, and 71.00±1.69 for YY). Similarly, Tamworth x Berkshire pigs did not perform as well as Large Black x Berkshire pigs or other breed combinations for 112-day weight and age at market weight (Fahmy and Holtman, 1977). However, again, in the present study, the numer of observations for Tamworth pigs was low (Table 1). In contrast to the present study, a study in the UK (Wood et al., 2004) noted that average daily gain for a 12 week period starting at 9 weeks of age was lower for Berkshire and Tamworth purebreds than for Large White pigs (similar to the Yorkshire breed).
The estimated growth curves for the FIRE system data among breeding groups in this study were linear. This result can be found in previous studies. Taylor and Hazel (1955) have reported growth from 135 to 174 days of age reflected in linear lines, and quadratic growth curves have been reported with a coefficient of the second degree polynomial lower than -0.005 (not meaningful), indicating that growth could be considered to have occurred in a linear fashion (Quijandria and Robison, 1971).
Using the FIRE system, it was found that TY and YY pigs were more feed efficient, with lower feed conversions (p<0.05), than BY and LBY pigs based on the 106 animals from the FIRE system (3.39±0.63, 3.41±0.80, 2.16±1.19 and 2.46±1.07 for BY, LBY, TY and YY, respectively). However, more observations would be needed to make any firm conclusions about differences in feed efficiency for these breed types.
In Japan, purebred Berkshire pigs had similar ADG (0.78 kg/day) to that noted for the Berkshire sired pigs in the current study, but were not as feed efficient (4.87; Suzuki et al., 2003). However, in a North Carolina State University study using Berkshire-sired pigs from maternal-line females more similar to the animals in the current study, Berkshire-sired pigs had a feed conversion closer to that of the current study (3.70 vs 3.39; Hasty et al., 2002).
Few recent scientific studies been conducted comparing differences in growth patterns and productivity in outdoor systems, likely due in part to the relative difficulty in measuring feed intake and growth rates for pigs raised in outdoor systems compared to confinement systems. Honeyman and Harmon (2003) compared performance of finishing pigs in hoop structures during winter and summer in Iowa, with summer weather closer to North Carolina weather than winter weather. Crossbred pigs from terminal Duroc boars with white breeds of sows were used in the experiments, and average daily gains and feed efficiencies were similar for summer pigs compared to those in the current study (overall ADG for all breeds of approximately 0.85 kg and feed:gain of 2.86 averaged across breeds), with performance calculated to be greater than 0.83 kg/d and a feed conversion ratio of 2.87. So, overall, the pigs in the current study performed similarly to other pigs raised in a comparable production system.

Results and Discussion – Objective 5: Carcass Traits &amp; Meat Quality
Pork Quality Characteristics
Pork quality is an important factor for niche pork markets (Lammers et al., 2007). However, few studies have compared different breed types in alternative production systems. For the current study, the impact of location, trial and breed type on measured pork quality characteristics are noted in Table 3. Least square means and standard errors for each characteristic by location, trial and breed type are presented in Table 4. There was no effect of breed on hot carcass weight (which was used as a covariable in the statistical analyses), and the overall average for all animals was 87.7 ± 11.8 kg.

pH
A higher pH results in less drip loss and also increases dark color (most noticeable at a pH greater than 5.7). In general, the range of 5.6 to 5.9 would be considered normal in pork (Lammers et al., 2007). In this study, most samples were within the normal range except for Trial 3 which had samples with the lowest pH for all trials (Table 4). Though breed type did not influence pH, samples from CEFS were higher than those from NCA&amp;TSU and also had higher measured color indicators (L*, a*, b*), though pH values for both locations were within the normal range expected for pork (Table 4). Animals from CEFS were harvested at a younger (p < 0.05) age than those from NCA&amp;TSU which may have contributed to some location differences.

Color score, L*, a* and b*
Color score was assigned by a trained scientist while L*, a* and b* are instrument-measured values. Lightness (L*) is measured from 100, or white, to 0, or black; red-green, or redness, (a*) is measured with red being positive and green negative. Yellow-blue, or yellowness, (b*) is measured with yellow being positive and blue negative. Color scores averaged from 3.27 to 3.63, or an overall reddish pink (Table 4). Although the scientist-estimated color score was lower for YY than BY and LBY samples (TY were intermediate), the instrument measured scores did not support this difference (Table 4). As noted previously, differences for CEFS and NCA&amp;TSU in measured color values likely relate to the higher pH noted for CEFS pork. Redness (a*) was lower for the Spring trial (the only trial with no animals from CEFS represented), and lightness (L*) higher for Trial 2 and 3 (Trial 3 was the only one with TY pigs; Table 4).
It has been reported that pigs reared outdoor have darker meat color (Bee et al., 2004), but all of the pigs in this study were raised in a similar manner. Differences in pH and color have also been noted for different seasons (Gentry et al., 2002; Judge et al., 1959), and although a* was lower for Trial 2, the only trial with animals born in the Fall and harvested in the Spring, this trial was also the only trial for which no animals were harvested from CEFS. Research by others has noted differences in color due to pig breed type (Brewer et al., 2002; Judge et al., 1959; Skelley and Handlin, 1971), but no real important differences were noted for the present study.

Marbling Score
As with color scores, marbling scores were assigned by a trained scientist. The effect of location and trial on marbling score was quite significant, with CEFS pork nearly half that of pork from pigs at NCA&amp;TSU (perhaps due to the younger harvest age of the animals from CEFS), and Trial 2 scores were half or less than that of Trial 1 and 3 (Table 4). As noted for the present study, there was no effect of breed reported for marbling scores in a study using Duroc, Hampshire and Poland China pigs (Skelley and Handlin, 1971). In contrast, an earlier study with multiple breeds represented (Landrace, Poland China, Yorkshire, Hampshire, Spotted Poland China and Berkshire) indicated that there were breed effects on marbling score (Judge et al., 1959). Other studies have also noted breed differences for marbling (Wood et al., 1996), including differences for Large White, Meishan type and a synthetic line (Faucitano et al., 2005).

Backfat and Loin Muscle Area (LMA)
All three measures of backfat (last lumber, 10th rib and 1st rib) were higher for LBY than all other breed types (Table 4), which was not surprising based on visual appraisal of phenotype and the fact that the breed was traditionally known as a bacon-type breed. The TY pigs had generally lower backfat than all but YY pigs (Table 4), likely a result of the use of more modern Tamworth genetics for this study and perhaps because Tamworth were the youngest pigs at harvest (p < 0.05), even though harvest body weights were not influenced by location or breed. The effects of trial were not consistent for backfat (LLFT, BF10 and BF1), with Trial 1 highest for LLFT, Trial 2 lowest for BF1 and Trial 3 highest for BF10 (Table 4). The CEFS location was not represented in all trials and the number of LBY and TY pigs differed among trials (Table 1). Because these breed types had higher (LBY) and generally lower (TY) backfat measurements than the other breed types, it could be that the different numbers of those pigs in the study accounted for the differences among trials. Trial 1 had the greatest overall number of animals represented (Table 1) and had pigs with the largest LMA (Table 4). Again, based on visual appraisal of phenotype, it was not surprising that LBY pigs had smaller LMA than all but TY (the youngest) pigs. Several studies have indicated that backfat and LMA can be influenced by breed (Skelley and Handlin, 1971; Lo et al., 1992; Wood et al., 1996; Hiner et al., 2006 ).

Drip Loss and Slice Shear Force
Lower values of drip loss are better than higher levels. It is said that drip loss should not exceed 2.5% in general (Lammers et al., 2007). However, drip loss range for the present study was higher than 2.5%. Similarly, Lebret et al. (2006) reported drip losses of 3.3% and 5.7% on two and four days after harvest, respectively, in an outdoor study using Large White × Landrace crossbred pigs. There were no influences of breed type on drip loss in the present study. In contrast, drip loss was influenced by breed in a study in Germany in that Landrace/Pietrain animals with Duroc heritage had lower drip loss than those with Large White heritage (Morlein et al., 2007). Slice shear force was not influenced by breed type, though differences were noted for location and trial (Table 4). Other studies support the lack of an influence of breed on slice shear force (Edwards et al., 2003; Skelley and Handlin, 1971; Lo et al., 1992).

Sensory Panel Tests
Overall liking, overall flavor, juiciness and texture and preference rank were requested in the sensory panel tests. An effect of sire breed for some characteristics was noted for Test 1 (Table 5) but not Test 2 (Table 6). Overall, for the test with only three breeds (BY, LBY and YY), pork from YY was scored as having higher juiciness, texture, and overall liking, and, as would be expected given those scores, YY was also ranked as the most preferred (Table 5). Breed differences were also noted between Large White and Durocs for pork flavor intensity and overall liking scores (higher for Durocs), though tenderness was not different between the two breeds (Wood et al., 1996). Similarly, Duroc pigs had loins that were more tender and juicy than those from Yorkshire pigs (Hiner et al., 1965). In contrast to the present study, in 1971, Skelley and Handlin reported no differences in flavor, juiciness, tenderness and preference due to breed.

Participation Summary

Educational & Outreach Activities

Participation Summary

Education/outreach description:

Journal articles
Whitley, N., D. Hanson, W.. E. M. Morrow M. T. See and S.-H. Oh. 2012. Comparison of pork quality and sensory characteristics for Yorkshire crossbred pigs raised in hoop houses without antibiotics. Asian-Australasian Journal of Animal Sciences. Submitted.

Whitley, N., W. E. M. Morrow M. T. See and S.-H. Oh. 2012. Comparison of growth performance in antibiotic-free Yorkshire crossbreds sired by Berkshire, Large Black, and Tamworth breeds raised in hoop structures. Asian-Australasian Journal of Animal Sciences. Submitted.

Conference presentation
Ireland, S., B. Pope, T. Barrios, S.-H. Oh, and J. Green. 2010. Vegetation types to protect the soil when gestating sows have access for grazing. ASAS Southern meeting.

Oh, S.-H., D. Bautista, D. Hanson, M. Morrow, T. See. 2011. Comparison of pork characteristics of antibiotic free Yorkshire crossbreds raised in the hoop barn. ADSA-ASAS Joint Annual Meeting.

Oh, S.-H., D. Bautista, D. Hanson, N. Whitley, M. Morrow, T. See. 2012. Comparison of pork quality characteristics among Hereford, Tamworth and Large Black crossbred pigs raised in a hoop barn during the finishing phase. ASAS Southern Section Meeting.

Oh, S.-H., M. Dudley, J. Talton, B. Hardison, J. Gonzales, A. Meier, M. Morrow, and T. See. 2010. Growth characteristics of antibiotic free Yorkshire crossbreds raised in the hoop barn. ASAS Southern meeting.

Oh, S.-H., M. Morrow, T. See. 2011. Comparison of body weights in Berkshire and Large Black crossbreds produced by the use of antibiotic-free Yorkshire sows. ADSA-ASAS Joint Annual Meeting.

Pietrosemoli, S. and J. Green. Effects of stocking rate of mature sows on bermudagrass (Cynodon dactylon) ground cover during winter. Poster presented at the XXI Latin American Animal Production Society (ALPA) Meeting, October 18-23. San Juan, Puerto Rico. Extended abstract (3 pages) in MEMORIAS ALPA. Volumen 17. Suplemento I. 447-450. XXI Reunión Bienal. San Juan, Puerto Rico. 18-23 de Octubre de 2009.

Pietrosemoli, S., Green, J. and Vibart, R. Effects of stocking rate of weaned to finishing pigs on Bermudagrass ground cover. J. Anim. Sci. Vol. 87, E-Suppl. 2/J. Dairy Sci. Vol. 92, E-Suppl. 1: 449/2009. The 2009 American Society of Animal Science meeting, Montreal – Quebec, Canada.

Pietrosemoli, S; Guevara, J.C. and J. T. Green. Effects of sow stocking rates on soil nutrients in a bermudagrass (Cynodon dactylon) pasture. J. Anim. Sci. Vol. 89, E-Suppl. 3: 25/2011. The 2011 Southern section American Society of Animal Science Meeting. February 2011. Corpus Christi, Texas.

Pietrosemoli, S; Guevara, J.C. and J. T. Green. Effects of sow stocking rate and season on bermudagrass (Cynodon dactylon) ground cover. J. Anim. Sci. Vol. 89, E-Suppl. 1/J. Dairy Sci. Vol. 94 , E-Suppl. 1:289/2011. The 2011 American Society of Animal Science meeting, New Orleans, Louisiana.

Pietrosemoli, S; Guevara, J.C.; Cardona, J.; Maradiaga, W.; Lobo, A. and J. T. Green. Animal weight gain in a pastured hog production system. J. Anim. Sci. Vol. 88 , E-Suppl. 2/J. Dairy Sci. Vol. 93 , E-Suppl. 1: /2010. The 2010 American Society of Animal Science Meeting. July 2010. Denver, Colorado.

Renner, B.; Pietrosemoli, S.; Luginbuhl, J.M; Raczkowski, C.; Green, J.T. and J. Grossman. Effect of stocking rate on forage production, soil compaction and root numbers in a swine pasture system. Anim. Sci. Vol. 89, E-Suppl. 1/J. Dairy Sci. Vol. 94 , E-Suppl. 1:315/2011.

Master Science thesis.
RENNER, BART MICHAEL. The Effect of Stocking Rate on Nutrient Budgets in Outdoor Swine Management. (Under the direction of Dr. Jean-Marie Luginbuhl). Defended in November 2011.

Farmer and extension agent training.
1) August 2009. Livestock/Forage Agent training, Lake Wheeler Beef Unit, Raleigh, NC.
2) May 25, 2010. Workshop “Conservation Practices in Outdoor Hog Production”. CEFS. Goldsboro, NC.

Additional Engagement:
1) Hondurans Interns: We hosted 3 Hondurans students from May to August. 2009.
2) We hosted three Interns from North Carolina, Virginia and Uruguay from June to July in 2010. During the period of 2008 to 2010 we hosted 16 experimental area guided visit.

International Visiting scientists
Scientists from Uruguay, Costa Rica, Venezuela, Ecuador and Guadalupe have toured our experimental sites.

Project Outcomes

Project outcomes:

To develop a rational ground cover management, it is necessary to estimate potential hog stocking rates that can be maintained in an area during a specific period of time while limiting the occurrence of soil and ground cover deterioration. Over stocking can produce deterioration of the ground cover, whereas under stocking can result in less efficient utilization of the land area. Appropriate stocking rates and monitoring of ground cover conditions will help ensure that long term goals for natural resources are fulfilled.

It can be summarized that our results of experiments performed (Growing – Finishing stocking rate and Sows stocking rate) to evaluate the impact of hog stocking rates on Bermudagrass ground cover and soil deterioration as: Bermudagrass, with its rhizomes and stolons, has a great potential to provide sustainable cover within hog pastures. Vegetative ground cover in the Bermudagrass stand decreased as a result of animal activity, and paddocks with the higher stocking rates showed a faster decrease. If maintenance of ground cover is the main goal, under a continuous grazing system the stocking rates must be kept between 15 to 30 hogs/acre, with no more than two finishing cycles before an extended rest period is used. Sow stocking rate of less than 10 sows/acre must be implemented to maintain over 60 percent ground cover during 56-day periods in dormant and non dormant Bermuda growth seasons.

For growth of heritage crossbreds, overall no real benefit of one sire breed over another for growth performance was noted in this study, though the average performance was comparable to other hoop systems. Hoop environments could provide pigs with better conditions in specific seasons, and the high-demand niche market for humanely-raised pork means the potential exists for small-scale hog producers to profit using this system (Honeyman et al., 2006). Moreover, in comparison with confinement systems, alternative hoop production systems require far lower capital investments in buildings and equipment, which is a possible way for small-scale hog producers to remain in business.

In pork quality, although Large Black pigs seemed to have smaller loins and more backfat than the other breed types used, even given differences in harvest age, in general, few differences were noted for carcass traits due solely to breed. In addition, although trends indicated an overall preference for meat from purebred Yorkshire pigs in consumer sensory test panel 1, the other breeds had mean values that would still be considered acceptable. Therefore, the crossbred (and purebred) animals used in this study would all seem satisfactory to use in an outdoor system as far as carcass quality and meat sensory data.

Recommendations:

Areas needing additional study

Additional research is needed to evaluate the hypothesis that outdoor swine production based on perennial pastures with appropriate stocking rate managed under rational grazing systems will lead to: 1) diminished ground cover deterioration, 2) better distribution of manure/nutrients in the soil, 3) reduced need for animal feed, 4) a decrease in the amount of external nutrients deposited in the system, 5) increased soil biodiversity, and 6) improved animal welfare, health and performance.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.