Project Summary

Project Title: Wool Education and Research – Montana State University and Texas A&M University

Primary Project Contacts:

Montana State University Dr. Rodney Kott Dr. Lisa Surber

Texas A&M University Dr. John Walker Mr. Faron Pfeiffer

Applicant Entity: Montana State University

Summary:

The American wool industry has changed significantly in the last 25 years. Years ago, domestic processing and use of wool far exceeded exports. Currently, approximately 50% of the US wool clip is exported. Adapting to the requirements of a global market is increasingly challenging to the American producer. Many growers have limited exposure to marketing expectations. Expanding avenues of domestic use include next-to-skin garments and socks. These uses require very specific wool quality needs. There is growing interest in using OFDA technology to build market specific lines of wool for these micron requirements in order to capitalize on the extremely attractive price premiums currently offered for micron specific lots of wool. Additionally, consumer interest in wool products demand domestic wool processors and companies to be able to source verify wool. The goal of this proposal is to provide leadership and technical assistance to American sheep producers on wool production, preparation and marketing to help improve the quality, marketing efficiency and competitiveness of US wool both internationally and domestically. The objectives are: (1) provide educational outreach programs for producers, stakeholders and end users of American wool;

(2) develop a sheep community of practice within eXtension that offers a wool education component;

(3) develop, evaluate and implement systems to apply OFDA technology for objectively measuring wool and to develop criteria that incorporate objective measurement of wool into a sheep selection program. Wool education has been identified as a major need within the sheep industry because of the decreased number of states within the US with a university sheep and wool program. Montana State University and Texas A&M University are the only remaining university wool labs in the US. Funding will allow these universities to expand their outreach efforts in order to meet the national needs of sheep and wool producers.

 

It has been a pleasure working with NSIIC on the joint project entitled “Wool Education and Research-Montana State University and Texas A&M University” with Texas A&M on a project which will hopefully benefit the Sheep and Wool Industry.  I have outlined the accomplishments and tasks associated with each objective. 

 

Summary:

 

Objective 1. To provide educational outreach programs for producers, stakeholders and end users of American wool. There is a renewed interest in utilizing more domestic wools to produce American-made wool products. Companies such as Rambler’s Way, Ibex, and Crescent Sock Company are particularly interested in source verifying wool back to the American farmer and rancher. The inclusion of the super-wash facility at Chargeurs has made these types of markets possible. Many of these companies are currently utilizing imported wool and reluctant to switch because of concerns regarding the availability of the type of wool needed in the quantities needed. Also, our marketing structure does not always sup-port this type of marketing efforts. The type of wool needed is available however they are not often separated at the grower level. This proposal would provide educational outreach to all segments of the wool industry. Wool education has been identified as a major need within the sheep industry because of the decreased number of states within the US with a university sheep and wool program.  Montana State University and Texas A&M University are the only remaining university wool labs in the US. Funding will allow these universities to expand their outreach efforts in order to meet the national needs of sheep and wool producers, stakeholders and end users of American wool.

 

Scientists associated with the “Wool Laboratories” at Montana and Texas have made numerous presentations at state, regional and national meetings associated with wool quality, wool preparation, selection for wool traits and the potential out of state use of these laboratories.  Presentations have been made in North Dakota, Idaho, South Dakota and Wyoming.  The number of out of state wool samples received and questions each lab has fielded has increased substantially.  In addition, numerous discussions have been initiated with manufactures and or growers in the potential for use of US wool in a branded US product.

 

Objective 2. To develop a sheep CoP within eXtension that offers a wool education component. The purpose of the proposed CoP is to transfer knowledge regarding sheep production and products to producers, stakeholders, and consumers. For such advancements to successfully transfer to the industry, a centrally located peer reviewed pool of educational resources is needed. Because no CoP for sheep currently exists, the proposed CoP will cover a wide variety of topics of interest to the sheep community (e.g., producers, stakeholders, industry, consumers), including (among others): Sheep Management Practices; Sheep Breeds; Sheep Purpose (wool versus meat versus dual-purpose); Markets; History; International Perspectives (how the industry differs in other countries); Showmanship; Wool; Meat Products (available products, how to prepare, etc.); Producer Profiles; Feed Efficiency; Genetic Selection; Prolificacy; Seasonality; Shearing; Feeding/Nutrition; Feedlot; Grazing/Pasture; Organic Sheep Production; Natural Sheep Production; Health; and Veterinary Care. The CoP in sheep and wool will provide a resource for extension educators and industry advisors as ASI moves forward with “Grow our Flock.” Additionally, the creation of a CoP for sheep will provide professional development opportunities for extension agents and specialists.

 

Sheep CoP within eXtension was created and is currently active http://www.extension.org/sheep .  An initial “work” meeting was held at the ASI meeting in January 2013 followed by a program during the convention.  The basic structure of the web-site and ask-a-expert was developed.  Also a plan was developed to recruit members of the sheep CoP to help develop content.  There are currently 48 members. In 2013, 33 “Ask an Expert” questions were answered.  At the 2014 ASI convention, an eXtension roundtable was held to review progress of the Sheep CoP. 

 

Objective 3. To develop, evaluate and implement systems to apply OFDA and Near Infrared Spectroscopy (NIRS) technology for objectively measuring wool and to develop criteria that incorporate objective measurement of wool into a sheep selection program. For the past few years, buyers have placed the burden of determining quality wool and grade lines on the producer. Although the U.S. has given some assistance in educating producers on wool quality improvement, a method for economically testing wool at the producer location needs to be implemented. Optical fiber diameter analyzer (OFDA) technology would allow the U.S. to become more competitive in the world wool market.  Quality improvement can be done with rapid results for fiber diameter, staple length and yield. The ability to test wool on-farm allows the U.S. to have a larger quantity of more uniform wool, less waste, and therefore, the ability to export more pounds of higher quality clips. There is growing worldwide interest in using new technology to build specific lines of wool. With extremely attractive price premiums currently offered for finer lots of wool, many producers and researchers are exploring the use of the OFDA technology in sorting wool by micron lines. It is important to be aware that wool characteristics are highly heritable. Through testing and selection, producers are able to develop sheep with desirable wool characteristics. As the ability of producers to measure different wool characteristics at the farm becomes available it becomes essential that they understand the relationship between these new wool quality measurements and overall sheep productivity. NIRS will be used in a pilot project to determine the feasibility of using this technology for source verification of wool. NIRS has been use for source verification of other agriculture commodities.

 

Yield and average fiber diameter (AFD) are the two most important attributes of wool for determining its value. In contrast to AFD, the determination of yield is time consuming and expensive. Current wet chemistry methods for determination of yield require duplicate subsamples of greasy wool cores to be washed, dried, and weighed after which subsamples of the cleansed cores are further gravimetrically analyzed for residual grease, inorganic ash, and vegetable matter. The cost of wool yield determination has provided a motivation for the development of alternative methods to analyze this important wool metric. The use of near-infrared spectroscopy (NIR), a rapid non-destructive method for the determination of the chemical composition of agricultural materials, for predicting wool yield has been the ultimate goal of multiple research efforts since the mid-1970s.

 

Three objectives were pursued in the current research namely: 1) determine the potential of NIR for the prediction of yield for commercial wool samples; 2) determine the potential for NIR for the prediction of yield for evaluating the merit of individual animals; and 3) determine the potential for NIR to discriminate between international and domestic wools. Because NIR has previously shown good potential for the determination of AFD and because a single spectral measurement can be used to predict multiple constituents this metric was also evaluated for the first two objectives.

 

Methods

For objective 1, 1485 core samples from wool bales representing a broad cross-section of U.S. wool production were supplied by Yocom-McColl Testing Laboratories, Denver CO along with yield and AFD data as determined by International Wool Testing Organization (IWTO) standard methodology were scanned.  For objective 2, 1092 individual animal fleece cores collected from rams that participated in the Texas A&M AgriLife Research Rambouillet Ram Central Performance Test conducted between 2000 – 2006 with yield and AFD data were scanned. For objective 3, 17 international and 16 domestic wool samples of known origin and provided by Bollman Industries were scanned.

 

A near-infrared reflectance spectrometer (Feed and Forage Analyzer Model6500M, Foss North America, Eden Prairie, MN) fitted with a transport mechanism and using a customized sample holder (scanning area = 82 cm2) was used to obtain duplicate spectra (at 2 nm intervals in the range 400 to 2498 nm). For objective 1 and 2 if the Mahalanobis distance between duplicate scans exceeded 1 the samples were eliminated, otherwise duplicate spectra were average before further analysis. The select procedure of WiniSI II software (version 1.04, Infrasoft International, Port Matilda, P A) and constituent data for yield and AFD was used to subset samples from Objective 1 and 2 into calibration and validation data sets. The calibration data sets were structured to be the minimum data set that represent the spectral diversity of the entire data and have a rectangular distribution of the constituents. A small modification to these specifications occurred so that the calibration data sets adequately represented the distributional tails of the constituent data. This procedure resulted in the calibration set containing about one-third of the samples and the validation set containing the remaining two-thirds of the samples.

 

Results

 

Validation of the NIR calibrations for average fiber diameter (AFD) and clean wool fiber present (CWFP) are shown in Table 1. The AFD r2 value for commercial samples is considered usable in most applications, including quality assurance, and the ratio error range (RER) is considered fair and applicable for screening (Williams 2001).  The lower values for these statistics for the AFD of the ram test data set are primarily a result of the smaller range and standard deviation of that data set (Fig. 1). The greater concern for the ram test validation is the deviation of the slope (0.91) from the ideal value of 1. Validation for CWFP was similar for both the commercial samples and the ram test. The r2 and RER values are considered fair and acceptable for screening and some approximate calibrations. Slopes and bias estimates are considered good to excellent for CWFP. The RMSE reported in this study for AFD and CWFP are similar to previously reported results (Hammersley and Townsend 2004). Thus the 95% confidence intervals (CI) for AFD and CWFP are about ±1.6 μ and ±2.8%, respectively. For comparison the 95% CI for the OFDA2000 is about ±0.48 μ and for wool base is about ±1.8%. Thus the loss in precision using NIR compared to standard test methods is about ±1 μ for AFD and ±1% for yield.

The hope for this study was that improved instrumentation, calibration algorithms, and a large calibration/validation data set would result in improved precision and accuracy of the NIR calibrations as compared to previous studies. However, this was not the case. Because NIR has previously been extensively evaluated by large wool testing organizations in Australia and New Zealand and found not to meet the standards for commercial trade and because the results of this study did not demonstrate significant improvement over previously published results further investigations to develop a standard NIR test for yield determination of greasy wool to be used in commercial trade do not appear warranted.

 

Table 1. Validation statistics for calibration of near-infrared spectra of greasy wool for the prediction of average fiber diameter (AFD) and clean wool fiber present (CWFP).

Sample Set

Calibration  N

Validation

N

r2

RMSE

Slope

Bias

RER

 

AFD

Commercial

592

1195

0.92

0.80

0.97

-0.01

19.3

Ram Test

436

660

0.61

0.73

0.91

-0.07

9.3

 

CWFP

Commercial

668

1103

0.84

1.68

0.96

0.06

13.2

Ram Test

417

675

0.86

1.70

0.99

0.05

15.5

N

r2

 

RMSE

Bias

RER

number of samples in data set

simple coefficient of determination, i.e., percentage variation in lab measurements accounted for by NIR predictions

Root Mean Square Error

Lab Average – NIR Prediction Average

Range in Lab Values ÷ RMSE

                 

 

 Ram test spectra were further evaluated to determine if NIR had potential as a technique for individual animal selection. For this investigation spectra from the calibrations were developed from individual fleeces from a Rambouillet ram central performance test conducted in 2000, 2001 and 2003 and validated with spectra from fleeces from the same test collected in 2004, 2005 and 2006 either with or without 10% of the validation year fleeces added to the calibrations (Table 2). The results indicated that using a previously developed calibration to predict future samples resulted in validations equivalent to the first study for the prediction of AFD when a subsample of 10% of the fleeces were included in the calibration. However, for CWFP the addition of 10% of samples from the sample set that was being predicted was still less than when sample selection procedures insured that the validation data set was adequately represented in calibration data set.

 

Figure  SEQ Figure \* ARABIC 1. NIR predicted compared to lab value for average fiber diameter of wool sample from commercial and ram test samples.

Table 2. Comparison of validation statistics using 2000, 2001, and 2003 ram test samples to develop calibrations to validate with 2004, 2005, and 2006 either with or without 10% of samples from the validation set included in the calibration. Samples used in calibration were not included in validation.

Percentage of Validation Samples in Calibration

r2

RMSE

Slope

 

AFD

0%

0.62

0.9

0.85

10%

0.64

0.8

0.92

 

CWFP

0%

0.66

3.4

0.86

10%

0.77

2.1

0.95

 

Classification of international and domestic wool was accomplished by developing discriminant functions using the spectra from international and domestic wool. The discriminant functions were calculated using the spectra from 4 replicate scans of each sample and were validated with the averaged spectra from these 4 scans. Discriminant functions were developed for both greasy and lab scoured wool. This validation was clearly not independent and the results only indicate the potential for NIR to classify wools based on origin. Figure 2 shows there was no overlap in the classification of two groups of wool.  It also shows that washed wool could be better differentiated than greasy wool. Finally, the domestic wools were more homogeneous than the international wools, which is not surprising considering that the international wools originated from 4 different continents.  These wools were also measured on the OFDA2000 to determine if there were differences in their physical properties. The only metric that differed between international and domestic wool was standard deviation of along staple fiber diameter, which was greater for international wools (P < 0.03).

Figure 2. Classification of greasy and lab scoured wool using discriminant functions to differentiate the wools based on spectral characteristics.

 Again, it has been a privilege to work with NSIIC and Texas A&M on this project as we continue to promote and assist the US sheep industry.  If you would like additional information or a more detailed report, please let me know.

 

Home