The Canadian Pork Carcass Grading System and the 1992 National Carcass Cut Out

S.D.M. Jones
Agriculture and Agri-Food Canada
Lethbridge Research Centre, Lethbridge, AB. Canada T1J 4B1

Introduction

The need to be able to place an objective value on a pork carcass has been recognized by the Canadian swine industry for almost 30 years. Although grading of live pigs was developed as early as 1922, it was not until the 1960's that an effective value based system was introduced. In 1968, a national system was put in place which involved the measurement of fat thickness on the carcass mid-line with a ruler at two locations (shoulder and loin). The system was introduced following a national carcass cutout which provided a measure of the average cutout value within 10 lb increments of carcass weight. The average carcass was given an index of 100, with carcasses leaner or fatter than the average given indexes above or below 100 depending on their expected cutout value. Since the process of selling pigs was through a marketing board in each Province in Canada, the grade index was then linked to price. Producers then received a % premium over the average price for leaner carcasses (e.g. a 106 index carcass would received 6% more than average price). In addition, the system was set up for the highest grade indexes to be received in the optimum carcass weight range. The system has evolved over time, but the essential elements in terms of paying a higher price for carcasses in the optimum carcass weight range still remain. Major changes that have been introduced to pork carcass grading since 1968 are shown in Table 1.

As we move into the next century, it is clear that the swine industry is undergoing major changes. Production units are generally becoming much larger and more specialized, the number of processing plants have declined leading to single species plants with much higher capacity and primary markets continue to be North America with growth in the Pacific Rim countries and Mexico. To remain competitive with alternative protein sources, pork must not only have good value but also be safe to eat, lean, tender and of good visual appeal. Carcass assessment for value will therefore have to move beyond being in most cases a relatively simple measure of carcass composition (fat and lean content).

Progress in improving carcass composition

Several analyses have been conducted over the years to measure overall change in Canadian pork carcass composition (Fredeen 1984, Jones 1986 and Jones et al. 1992). Fredeen (1984) and Jones (1986) showed that about a 10 mm reduction in the combined total of shoulder and loin fat had been achieved over the period 1967 to 1981 (Table 2). The average carcass weight for all pigs increased by about 7 kg in the same period. Thus, the system was judged to be effective in improving carcass lean content through an approximate 10% reduction in fat thickness, while increasing carcass weight by the same factor. More recently, Jones et al. (1992) evaluated carcass compositional changes from 1978 through to 1992. It was found that the 1992 carcass had about 6% more lean meat than the analogous carcass of the same weight produced in 1978. These studies have shown collectively that the grading system has had its desired effects and not only has it provided a fair value for the carcass on an objective basis but it has also been successful in providing the market signals to continually improve the lean content of pork carcasses. It has been far more difficult to monitor change in composition in the US, since grading schedules differ and the results are not publicly disclosed.

Limitations of present system and requirements for the future

The present system of grading carcasses for lean yield only gives an overall approximation of carcass value. While this information is very useful is assigning a potential value to a carcass, the accuracy of a single probe technology is limiting along with its inability to provide further information.

The usefulness of a single point measurements on a pig carcass to predict carcass composition has been defined by quite a large number of research studies (for a review see Fisher 1990). Generally, the most precise estimates of pork carcass composition are provided by fat thickness measurements measured lateral to the dorsal mid-line (Table 3) and tend to reach a maximum value between the 12th rib and last rib. The amount of variation explained in carcass lean content by single point measures varies between 60-80% and the residual standard deviation (RSD) also varies between 1.5-3.0%, depending to some extent on the source population. Nevertheless, assuming an RSD of 2%, it would be expected that 65% of the pigs probed at a single point would have a predicted carcass lean content within 2% of the actual value determined by a cutout. While this degree of precision is quite satisfactory on groups of carcasses, it is of less value for the individual carcass. However, it is probably the best that a single point measurement of fat and lean thickness recorded by an electronic probe can achieve under commercial conditions.

The Danes recognized the limitations of single point measurements to predict carcass composition and developed the Carcass Classification Centre (CCC) in the late 1980's, which originally used 17 probes and a series of regression equations to predict carcass lean yield. The CCC was introduced in all Danish abattoirs as the standard method of grading pork carcasses. Results indicated that the CCC improved the precision for lean content prediction about 30% and opened the possibility to measure the lean content of the major primal cuts (Table 4). Improvements were made over the years to the CCC, the most important one being the use of neural network procedures for predicting carcass lean content. However, while the CCC is recognized as a major engineering feat, it still relied heavily on many mechanical parts which probably resulted in relatively high maintenance costs. It was also limited by the number of carcasses that could be graded per hour (e.g. 300-360) and was not adopted in North America.

Meat processors require more information than a simple estimate of carcass lean content and carcass weight to make optimum use of pork carcasses. It would be extremely valuable to have estimates of the lean content of each of the major primal cuts along with their expected size and weight on the slaughter floor, so that this information could be used to sort carcasses in the chiller, prior to cutting the following day. It would also be of considerable value to have an estimate of meat quality (e.g. meat colour, drip, marbling, tenderness), but this is a topic that will be covered in another presentation. It is clear that the present system of grading pork carcasses in North America (electronic probes) will not meet these requirements and alternative technologies have to be considered.

New Technologies for Grading Pork Carcasses

There are several promising developments underway at the present time that may in the future provide replacement systems for the present systems in use by industry. These include ultrasound, TOBEC (total body electrical conductivity) and video image analysis. While there are several more technologies that are under investigation such as computerized tomography or nuclear magnetic resonance, these are not thought at the present time to be contenders for grading hog carcasses at the commercial level.

Ultrasound

While ultrasound has been used by the swine industry for over 35 years, it has not been widely used for grading purposes. The Danes developed a hand held probe based on ultrasound (UFOM) but this did not find acceptance in the North American market. More recently, Goldenberg and Ananthanarayanan (1993) reported an automated carcass grading system for pork which used an ultrasonic probe to locate the grading point in Canada (3-4th last rib) and provide a series of fat measurements. The Danes in contrast have developed an ultrasonic based grading system for warm pork carcasses and marketed as the Autofom. The Autofom has been introduced commercially in one plant in North America and there are plans for at least one further installation.

The measurement principle of the Autofom system is a digitized three-dimensional scanning. The scanning pattern is provided by 16 ultrasonic transducers embedded in a fixed stainless steel transducer array. The Autofom is installed after the dehairing machine and can function at up to 1250 head per hour which fully meets current North American requirements. The Autofom has no moving parts and its maintenance is expected to be extremely low. It is fully automated and requires no human involvement since the gambrel pulls the carcass over the transducer array while the weight of the carcass ensures contact with the transducers. The Autofom produces an average of about 2,000 measurements per carcass consisting of subcutaneous fat measurements and for the loin, measurements of muscle thickness. The information can be used to optimize carcass break points and for sorting carcasses in the chiller. The present state of software development allows a calculation of carcass lean percentage and values for the weight and lean content of the major cuts (ham, loin, shoulder, belly). The Autofom has been stated to have a high degree of precision (Table 5).

TOBEC

Much of the work concerning the usefulness of TOBEC has been reviewed by Forrest (1995). TOBEC is based on the conductivity differential between lean and fat tissue. Carcasses are passed through an electromagnetic field and the amount of energy absorption detected is related to the conductive mass of the carcass or meat cuts. Since the fat-free mass is approximately 20 times more conductive than fat, the conductivity index is highly correlated with the lean tissue mass. Electromagnetic scanning is therefore a non-invasive technology capable of scanning carcasses, primal cuts or boxes of meat. Table 6 shows the precision of this technique for predicting carcass lean and the weight of lean in the major primal cuts.

Commercially TOBEC has been evaluated in processing plants and has operated at 300-400 carcasses per hour for about 18 months with few technical problems (Forrest 1995). It is probably most useful to scan warm pre-rigor carcasses where the temperature of the carcasses is relatively constant and the need to include carcass temperature in the prediction equations is not required. While TOBEC can provide a more precise estimate of carcass lean percentage and also provide estimates of the lean content of the cuts, it is not capable of providing dimensional data to optimize cutting procedures.

VIDEO IMAGE ANALYSIS

Video image analysis or VIA involves taking an image of the whole carcass by a video camera and then analyzing the image for dimensional measurements that have a relationship with carcass composition and value. The technique has been applied mainly to beef carcasses to date since muscle thickness measurements are thought to be more variable in beef cattle than in pigs. With the advent of specialized sire lines which are being marketed as terminal sires, it would be expected that muscle thickness will show increasing variation. To this end the measurements of items like ham shape and carcass length may be quite useful as additional information to probe measurements to increase accuracy in predicting carcass and cut lean content. In addition, images collected on the slaughter floor could also be used for the more efficient breaking of carcasses to optimize cut out value for the meat processor. There has to date been very little work done on using VIA to predict carcass composition in pig carcasses. However, it is clear that VIA could not be used as a stand alone system to grade pigs since the system would be unable to measure fat thickness off the mid-line of the carcass. VIA shows most promise as an add-on to existing probe type systems to provide additional information on cut size.

Conclusions

Pork carcass grading for carcass lean content using objective procedures has been adopted by a number of countries. In Canada, the system used has been shown on a national basis to encourage the marketing of leaner and more desirable carcasses. Probe based systems using a single measurement of fat depth and lean thickness have moderate precision for the determination of carcass value. Inherent variation within the system includes the effect of the operator and the limitations of how single point readings of carcass tissues relate to overall carcass lean content. The Danes have made the most progress in improving carcass evaluation equipment for grading purposes. The Carcass Classification Centre which is currently used in Denmark makes multiple probe measurements in each of the major primal cuts which allow carcass lean percentage and cut lean percentage to be calculated. The Autofom is a relatively new development and uses an array of transducers to provide a 3 dimensional scan along the whole of the pig's back. While there have been no studies released which document its precision for predicting carcass lean content, the extent of the information collected would suggest that it will have a similar precision to that of the Carcass Classification Centre. Other technologies such as electrical conductivity (TOBEC) and VIA appear to be some way away from industry acceptance. The final decisions for replacement of the electronic probes currently in use will be made on a cost effectiveness basis. For large processing plants the investment in an Autofom may show a much higher return on investment than in the smaller plants where the cost of the equipment would be applied against fewer hogs killed in any one year.

References

Fisher, A.V. 1990. New approaches to measuring fat in carcasses . In, "Reducing Fat in Meat Animals". Edited J.D. Wood and A.V. Fisher. Elsevier Applied Science, New York.

Forrest, J.C. 1995. New Techniques for estimate of carcass composition. In "Quality and Grading of Carcasses of Meat Animals". Edited S.D.M. Jones. CRC Press, Boca Raton, Fl.

Fredeen, H.T. 1994. Changes in the characteristics of commercial hog carcasses in Canada. Can. J. Anim. Sci. 64:569.

Goldenberg, A.A. and Ananthanarayanan. S. 1993. An approach to automation of pork grading. Food Res. Int. 27:191.

Jones, S.D.M. 1986. Changes in animal product composition and implications for animal production systems. Can. J. Anim. Sci. 66:23.

Jones, S.D.M., Tong, A.K.W., Robertson, W.M. and Skoczylas, P. 1992. National Pork Carcass Cutout Project. A comparison of the 1978 and 1992 cut-outs. Canadian Pork Council, Ottawa.

Table 1. Chronological sequence of main events in Canadian pork carcass grading (1968-1996)

1968 Index system introduced using a ruler to measure backfat at the mid-line at the shoulder and loin. Average carcass was given index of 100.
1982 System was modified to one single measurement of fat (loin).
1986 Electronic carcass grading introduced. Mandatory for plants slaughtering >1000 pigs per week. Two probes approved (Hennessey Grading Probe and Destron PG-100). Fat and lean measurement recorded between the 3/4th last ribs at 7 cm from the carcass mid-line.
1995 New prediction equations introduced for predicting carcass lean content.
1996 Pork carcass grading privatized and no longer a Federal responsibility.

Table 2. Fat depth measurements1 in Canadian pork carcasses
Carcass Weight Range kg
72.6-77.1
77.2-81.6
81.7-86.2
Year
1967
85
89
92
1974
81
84
87
1980
78
81
84
1981
76
79
83
1 Fat depth = sum of shoulder + loin fat in mm Jones. 1986

Table 3. Effect of location of measurement of fat depth on the precision for predicting pork carcass lean content.

Measuring siteResidual Standard
Deviation %
Shoulder

Loin

Last rib

3/4th last rib

5/6th last rib

2.67

2.60

2.41

2.26

2.31
Adapted from Fortin et al. 1984

Table 4. The precision of the Carcass Classification Centre for predicting carcass and cut lean percentages.

R2 Residual Standard
Deviation%
Carcass

Shoulder

Loin

Ham

0.82

0.76

0.85

0.81
1.63

1.77

2.27

1.73

Table 5. Precision of the Autofom for the prediction of Carcass and Cut Composition

R2
Residual Standard
Deviation
Carcass lean%

Shoulder1

Loin

Ham

Belly

0.80

0.90

0.90

0.85

0.75
1.80

1.50

1.70

1.75

1.80
1 Predictions in cuts are for weight of lean in lbs Data provided by SFK

Table 6. Precision of TOBEC for the prediction of carcass and cut composition

R2 Residual Standard
Duration
Carcass lean %

Ham1

Loin

Shoulder

0.86

0.83

0.86

0.85
2.05

0.64

0.60

0.51
1 Predications for cuts are for weight of lean (kg)