Optimizing Electrical and Mechanical Palmer Amaranth Control and Reducing Seed Production and Viability

Final report for GS19-205

Project Type: Graduate Student
Funds awarded in 2019: $16,498.00
Projected End Date: 02/28/2022
Grant Recipient: North Carolina State University
Region: Southern
State: North Carolina
Graduate Student:
Major Professor:
Expand All

Project Information

Summary:

Palmer amaranth (Amaranthus palmeri) is the most troublesome weed in U.S. crops. In an effort to reduce herbicide use, tillage and labor costs to control Palmer amaranth, commercially available electrical and mechanical control methods will be evaluated to optimize use efficiency. Electrical and mechanical weed control eliminate weeds that are taller in height than the crop canopy. Mechanical and electrical weed control has the potential to reduce the amount of between-row cultivation, herbicide application, and hand labor needed to achieve weed control, as well as the ability to selectively manage weeds that extend above the crop canopy to preserve the biodiversity of other small non-weedy plants that can support a high diversity of insect species. Additionally, it is hypothesized that seed production and viability can be reduced from electric treatments to sexually mature plants. Sweet potato, peanut, and cucumber are economically important crops in North Carolina that each are easily outcompeted for light by Palmer amaranth, which can grow to 2 m tall. Therefore, studies were conducted in each crop to optimize the use efficiency of each method of control. Another study was conducted to evaluate the effect that electrical applications have on Palmer amaranth pre-dispersal seed viability.  An economic analysis evaluated each control method and the results will be disseminated to growers and the scientific community.  We are still working on creating an extension publication to provide results to growers and extension agents.  

Project Objectives:
  • Develop recommendations needed for electric weed control and mechanical weed removal to optimize crop yield and quality.
  • Develop a management program to reduce weed seed number and viability of seeds added to the soil seed bank.
  • Assess the cost to benefit economics for using the electric method of weed control and mechanical weed puller and disseminate research-based knowledge and recommendations through grower meetings, technical extension articles, and refereed journal publications.

Research

Materials and methods:

Palmer Amaranth Management

Field studies were initiated at the Horticultural Crops Research Station (35.023°N, 78.280°W) near Clinton, NC in 2020 and 2021 in field with historically high Palmer amaranth densities (50 to 100 plants m-2). ‘Maxi pick’ pickling cucumber, ‘Walton’ Virginia-type peanut, and ‘Covington’ sweetpotato were planted into independent studies. Cucumbers were overseeded and thinned to a 15 cm in-row spacing, peanuts were seeded to a 6 cm in-row spacing, and nonrooted sweetpotato cuttings were transplanted to a 30 cm in-row spacing in July of each year. Due to COVID-19 delaying the delivery of the tractor needed to operate The Weed Zapper, the planting timings were later than recommended for peanuts and cucumber in North Carolina. The soil was an Orangeburg loamy sand (fine-loamy, kaolinitic, thermic Typic Kandiudults) with a 6 pH and 1% organic matter for sweetpotato and cucumber in 2020, a Norfolk loamy sand (fine-loamy, kaolinitic, thermic Typic Kandiudults) with a 5.5 pH and 0.5% organic matter for peanut in 2020, and a Norfolk loamy sand with a 6 pH and 0.5% organic matter in 2021. Soil fertility was maintained at levels sufficient for each crop and overhead irrigation was applied as needed.

            The experimental design for each study was a randomized complete block with treatments replicated four times. Cucumber and peanut plots consisted of four rows, where the center two rows were planted and used for data collection, and the outside two rows were not planted. Sweetpotato plots consisted of 5 rows where the first four rows were planted and treated, only the center two rows were used for data collection, and the fifth row was not planted. Treatments were arranged in a three by four factorial where the first factor included a treatment method of electrical (The Weed Zapper Annihilator 6R30, Old School manufacturing, LLC, Sedalia, MO), mechanical (Bourquin Organic Weed Puller / Roguing machine - Posi Pull, Bourquin Design and Manufacturing, Inc., Colby, Kansas), or hand Palmer amaranth control, and the second factor consisted of treatments applied when Palmer amaranth was approximately 0.3, 0.6, 0.9, or 1.2 m above the crop canopy. The Weed Zapper was set to the first pass broadleaf setting and powered by a Case IH Puma 220 (190 power take off (PTO) horse-power) (CNH Industrial America LLC, Racine, WI) at 2,200 engine rpm and 3.5 km h-1. The Weed Puller was applied at 1.5 km h-1 with the wheels rotating at 25 rpm. After each initial treatment, additional treatments were applied if successive weeds reached the assigned height. A nontreated plot was included for comparison. The nonplanted rows within a plot were used to ensure the adjacent plots did not contact the 15’ Weed Zapper electrode. All crops were planted into 1 m wide rows because of available tractor tire spacing. Interrow weeds were controlled using cultivation, and clethodim 135 g ai ha-1 (Select Max, Valent U.S.A. Corporation, Walnut Creek, CA) was applied to the whole study areas for grass control.

Effects of treatments on Palmer amaranth control were evaluated using a scale of 0% (no treatment effect) to 100% (plant death). Percent crop death was calculated by dividing the difference between the number of plants counted before the first treatment application and the number counted at the end of the growing season by the number of plants counted before the first treatment application. Sweetpotato crop stand could not be accurately counted at the end of the growing season due to between-row cultivation burying vines. Peanuts were dug and inverted 16 WAP. Because of a late peanut planting and limited equipment and personnel availability in 2020 due to COVID-19 restrictions, peanuts pod counts were used to estimate the yield response. Sweetpotato storage roots were harvested using a chain-digger 16 WAP; hand sorted into canner (>2.5 to 4.4 cm diam), no. 1 (>4.4 to 8.9 cm), and jumbo (>8.9 cm) grades and weighed. Total yield was calculated as the sum of all grades.

            Data were assessed for homogeneity of variance by examining residual plots. Arsine square-root transformations were required for Palmer amaranth control data. Back-transformed least square means were presented. ANOVA was conducted using PROC GlIMMIX (SAS version 9.4, SAS Institute, Cary, NC). For Palmer amaranth control data, fixed effects included application method, application timing, and their interactions, whereas study, all interactions including study, and replication nested within study were considered random effects. For crop data, fixed effects included year, application method, application timing, and their interactions, whereas replication nested within year was considered a random effect. When significant interactions were present, a slice statement was used to evaluate simple effects within the interaction. Least square means were separated according to Tukey-Kramer HSD at a significance level of a = 0.05.

 

Seed Production and Viability

Raised beds spaced 1 m apart were prepared in July 2020 and 2021 at the Horticultural Crops Research Station near Clinton, NC. The soil was an Orangeburg loamy sand (fine-loamy, kaolinitic, thermic Typic Kandiudults) with a 6 pH and 1% organic matter in 2020, and a Norfolk loamy sand with a 6 pH and 0.5% organic matter in 2021. Natural populations of Palmer amaranth were thinned to approximately 50 plants m-2 using between-row cultivation. The study was arranged in a randomized complete block design with 4 replications. Plots consisted of 4 rows each 3 m long. Treatments consisted of electricity applied to Palmer amaranth at first visible inflorescence, two wk after first visible inflorescence (WAI), or four WAI using The Weed Zapper Annihilator 6R30 powered by a Case IH Puma 220 at 2,200 engine rpm and 3.5 km h-1. In addition, check plots, which were treated without electricity, were included for comparison.

Five gynoecious plants per plot were harvested immediately after application. In addition, plots from the first application stage were harvested 2 and 4 WAI, and plots from the second treatment stage were harvested 4 WAI. At the last treatment timing, plants in the nonelectrically treated check were large and at reproductive maturity; thus, plots previously treated and the check were harvested before the last treatments were applied to avoid seed shatter. After the 4 WAI treatment, the nonelectrically treated check was harvested again for comparison.

Plots were threshed using a mill (Thomas Scientific, Swedesboro, NJ), and seeds were separated from floral material using a vertical air column separator and stored at 4 C and 25% relative humidity. Seed production was estimated by multiplying the number of seeds in 1 g subsamples from each plot by the total plot weight. Four months after being placed in cold storage, 30 mature (black) seeds per experimental unit were placed into 10 cm diam petri dishes containing filter paper moistened with 10 ml water, sealed with parafilm, and placed back into cold storage for 4 wk to overcome dormancy. Then the petri dishes were placed into a germination chamber at a 16 h photoperiod set to 35/25C d/night and 100% relative humidity for 3 d. Germination was counted, then seedlings (radicle and hypocotyl) were imaged using a flatbed scanner (Expression 10000 XL; Epson America, Long Beach, CA). Seedlings were arranged to avoid overlapping during scanning. Root measurement image analysis software (WinRHIZO 2019a; Regent Instruments, Quebec, QC, Canada) was used to measure total seedling length. Average seedling length was calculated by dividing the total length of seedlings by the number of germinated seeds in the experimental unit. Seed viability was assessed on 30 seeds per plot using a crush test (Sawma and Mohler 2002).

Residual plots were assessed for normality and homogeneity of variance. Data were subjected to ANOVA using PROC MIXED. Fixed effects included plant section, treatment timing, and their interaction, and replication was treated as a random effect. Dunnett’s test was used to compare treatments to the respective nonelectrically treated check (α = 0.05).

Research results and discussion:

Palmer Amaranth Control

Treatment method by application height interactions were present for each evaluation timing (P < 0.0001), thus, the interactions were assessed by application height (Table 1). After treatment, Palmer amaranth that were not tall enough to be controlled and plants that were not fully controlled continued to grow, requiring four weekly and two biweekly treatments for the 0.3 and 0.6 m application heights, respectively, though only one treatment was required for the 0.9 and 1.2 m application heights. Hand weed removal consistently resulted in optimal weed control. Four WAT, the electrical applications controlled Palmer amaranth at least 27% more than the mechanical applications when applied at the 0.3 and 0.6 m timings. At the 0.9 and 1.2 m application timings 4 WAT, electrical and mechanical applications controlled Palmer amaranth by at most 87%. When weeds were small, the weed puller would only grab and uproot weeds that had greater separation in height from the crop, whereas the electrical applications could control weeds that were closer to the crop in height. Once the weeds became larger, the electrical applications were less effective because of the increased weed biomass that needed to be treated.

 

Crop Death

Palmer amaranth densities were high (50 to 100 plants m-2), thus cucumber were actively damaged or uprooted with the hand removal plots only maintaining 50% stand after treatment (Table 2). As the cucumber were shaded by the Palmer amaranth, plants responded with etiolation causing long plants that grew upwards with the weeds. The tendrils grabbed onto the Palmer amaranth further increasing difficulty of weed removal. In a typical field situation, because of the absence of herbicides in these high densities of Palmer amaranth, this cucumber crop would likely be a complete loss as the cost to remove weeds with the care needed to prevent crop damage would be high. The use of mechanical and electrical weed control further increased the plant death by at least 14% compared to the hand removal.

            Peanut plant death was greatest when the weed puller was used for Palmer amaranth control. However, the weed zapper and hand removal each caused less than 7% peanut plant death on average across treatment heights. When treatments were applied at the 0.3 and 0.6 m weed heights, peanut plant death was 4%, averaged across treatment methods. At the 1.2 m treatment height, 20% peanut plant death was observed.

 

Yield

Cucumber yield was suboptimal due to the timing of planting, resulting in heat and water stress and disease. Cucumber plots had less than 0.3 fruit per plant in the highest yielding plots; thus, data were not presented. Due to a significant treatment method by application height interaction for peanut pod count and total sweetpotato yield (P < 0.05), data were assessed by application height (Table 3). At the 0.6 and 0.9 m treatment heights, hand removal resulted in greatest pod counts. At the 0.3 m height, the P-value of 0.2 indicated no differences between treatments, though the hand-rogued plot yielded higher on average than the mechanical or electrical treatments. The greatest total sweetpotato yield was achieved by hand removing Palmer amaranth prior to reaching 0.9 m in height. Though hand removal often resulted in the greatest peanut pod count and total sweetpotato yield, the weed puller and weed zapper resulted in similar yield to the hand rogued plots depending on the treatment timing. With additional research to provide insight into the optimal applications, there is potential for the weed zapper and the weed puller to be used as alternatives for hand-removal in peanut and sweetpotato. These results are contrary to those reported by Simard et al. (2019), who observed sub-optimal control from using the weed puller in soybean. The difference in results is likely due to the taller crop height of soybean compared to peanut and sweetpotato.

            Visible injury could not be consistently rated due to Palmer amaranth interference, though up to 20% injury was observed after treatment using the weed zapper in sweetpotato, depending on environmental conditions and treatment timing (data not shown). More research is needed to assess the negative impacts that electrical treatment can have on sweetpotato. Injury that was observed was often located near Palmer amaranth that had been treated. It is likely that electricity traveled through sweetpotato that was close to or contacting the weed at the timing of treatment. The impact of treatment injury on sweetpotato yield could not be determined due to the interference of Palmer amaranth.

 

Seed Production and Viability

The nonelectrically treated control produced 11,000 seeds per plant 4 weeks after first visible inflorescence, which is less than the typical 200,000 to 600,000, because plant densities were high (approximately 50 plants m-2). Treatments at varying reproductive maturities did not reduce the seed production immediately after treatment at a 95% confidence level (Table 4). However, after treatment, plants primarily died and ceased maturation, reducing seed production at 4 weeks after first visible inflorescence by 93% and 70% when treated 0 and 2 weeks after first visible inflorescence, respectively. Treatments did not have a negative effect on germination or seedling length, indicating that the quality of seeds produced were unchanged by electrical treatment in the present study based on the variables that were recorded.

 

Economics

Each of these studies were conducted on a research station with historically high (50 to 100 plants m-2) Palmer amaranth densities. Thus, these studies were conducted under weed pressures that were expected to be worse than the average farm. These numbers will be used for a rough comparison of economics with the disclaimer that actual data will vary depending on weed densities and costs specific to individual circumstances. On average, to hand remove the weed densities in each study required 40, 81, 53, and 53 h ha-1 for removal at the 0.3, 0.6, 0.9, and 1.2 m application timings, respectively. The adverse effect wage rate for employing H-2A employees in North Carolina is $13.15 h-1 (USDOL-ETA 2021); however, this does not include additional cost of transportation, meals, housing, and fees. Using $13.15 h-1, for discussion sake, would equate to $530, $1060, $690, and $690 ha-1 for removal at the 0.3, 0.6, 0.9, and 1.2 m application timings, respectively, in the present study. Comparatively, a four-row (4.3 m wide) electrical or mechanical weed control implement could cover approximately 2 ha h-1 at 4.8 km h-1, not factoring in the time required to turn around and the end of rows and travel between fields. Assuming the same wage rate, the cost of labor to operate the implements would be approximately $26.30 ha-1. This leaves approximately a $500 ha-1 difference from the cheapest recorded hand removal cost to cover the cost of equipment, fuel, and maintenance.

            The cost of hand weed removal is directly related to the density of weeds present. In the present study, weed densities were much greater than would be expected in a grower field. In a survey of sweetpotato growers in North Carolina in 2018, the cost of hand weed removal was $62 to $370 ha-1, with a mean of $150 ha-1 (Smith and Moore, unpublished data). The survey included primarily conventional cropping systems, though a few organic sweetpotato systems were included in the surveyed group. Using this more relevant number of $150 ha-1 would leave a difference of approximately $120 ha-1 between hand removal and electrical or mechanical control to cover the operation cost of the control implements. In 2019, the estimated cost of operation of a 190 PTO horsepower tractor was $116 h-1 (Lattz and Schnitkey 2019). The cost of operation will depend on power demand of the implement being used, which will be greater when using the weed zapper than the weed puller. However, this estimated figure will be used for discussion for both implements. At 4.8 km h-1, the cost of operating the tractor with a 4.3 m wide implement would be $58 ha-1. This would leave approximately $62 ha-1 to cover the cost of the equipment and maintenance.

            The cost of electric and mechanical weed control implements will vary depending on the source, model, and year. But for discussion sake, the cost of the 4.3 m electrical weed control implement is approximately $56,000. With this figure, it would require 903 ha treated with the electrical implement, respectively, rather than hand removal to pay for the cost of the weed zapper. However, we acknowledge that the cost of hand labor would likely be higher than used for this discussion due to additional costs and the cost of equipment maintenance not being included. Thus, this discussion should be used for informational purposes only, and the numbers should be adjusted to fit individual circumstances. The cost of herbicides may be lower than the cost of controlling weeds with the electrical weed zapper, and, if used as pre-emergence control, would offer greater yields as there would be less total competition with the crop. However, the electrical and mechanical implements can be useful for cropping systems that are grown for markets that require the absence of synthetic herbicides, or where herbicide resistance causes escapes.

 

Table 1. The influence of treatment method and application height on Palmer amaranth control.a

 

Palmer amaranth control

 

 

1 WATbc

 

2 WAT

 

3 WAT

 

4 WAT

 

 

0.3 m

 

0.3 m

0.6 m

 

0.3 m

0.6 m

0.9 m

 

0.3 m

0.6 m

0.9 m

1.2 m

Treatment method

 

–––––––––––––––––––––––––––––– %d –––––––––––––––––––––––––––––

Hand

 

96a

 

 

96a

 

98a

 

 

99a

 

96a

 

99a

 

 

97a

 

95a

 

99a

 

99a

 

Mechanical

 

46c

 

 

60c

 

65b

 

 

64c

 

55c

 

86b

 

 

64c

 

64b

 

84b

 

82b

 

Electrical

 

72b

 

 

84b

 

76b

 

 

89b

 

72b

 

82b

 

 

91b

 

95a

 

85b

 

87b

 

aData pooled across years and crops.

bMeans within a column for dependent variables followed by the same letter are not significantly different according to Tukey-Kramer HSD at a significance level of a = 0.05.

cAbbreviations: WAT, wk after treatment application.

dRating scale: 0%, no treatment effect; 100%, plant death.

Table 2. The influence of treatment method and application height on cucumber and peanut plant death.a

 

 

Crop Plant Deathb

 

 

Cucumber

Peanut

Treatment method

 

––––––––––– %c –––––––––––

Hand

 

50b

 

7b

 

Mechanical

 

67a

 

27a

 

Electrical

 

64a

 

1b

 

Application Height

 

 

 

 

 

0.3

 

59

 

4b

 

0.6

 

57

 

4b

 

0.9

 

65

 

17ab

 

1.2

 

59

 

20a

 

aData pooled across years.

bMeans within a column for dependent variables followed by the same letter are not significantly different according to Tukey-Kramer HSD at a significance level of a = 0.05. Means within a column not followed by a letter are not significantly different according to a nonsignificant F statistic (P > 0.05).

cRating scale: 0%, no treatment effect; 100%, plant death.

Table 3. The influence of treatment method and application height on yield of ‘Walton’ peanut and ‘Covington’ sweetpotato.a

 

Yieldb

 

 

Peanut

 

Sweetpotato

 

 

0.3 m

0.6 m

0.9 m

1.2 m

 

0.3 m

0.6 m

0.9 m

1.2 m

Treatment method

 

––––– Pod per plant –––––

 

–––––– 1000 kg ha ––––––

Nontreated

 

3.6

 

7.5

Hand

 

26.7

 

27.9a

 

24.3a

 

15.6

 

 

8.8a

 

8.5a

 

5.8

 

6.4a

 

Mechanical

 

22.2

 

15.b

 

15.3b

 

19.5

 

 

6.0b

 

6.6b

 

6.3

 

4.4b

 

Electrical

 

21.0

 

17.1b

 

15.3b

 

13.2

 

 

6.2b

 

5.3b

 

5.5

 

4.9ab

 

aData pooled across years.

bMeans within a column for dependent variables followed by the same letter are not significantly different according to Tukey-Kramer HSD at a significance level of a = 0.05. Means within a column not followed by a letter are not significantly different according to a nonsignificant F statistic (P > 0.05).

eAbbreviations: WAT, wk after treatment application; NIS, nonionic surfactant.

fRating scale: 0%, no treatment effect; 100%, plant death.

 

Table 4. The influence of electrical treatment on Palmer amaranth progeny seed production, germination, and seedling length.a

 

 

 

Seed production

Germination

Average seedling length

Total seedling length

 

 

 

0 WAIb

2 WAI

4 WAI

0 WAI

2 WAI

4 WAI

0 WAI

2 WAI

4 WAI

0 WAI

2 WAI

4 WAI

 

Application timing

 

–––––– seed per plant ––––––

–––––––– % –––––––––

––––––– cm –––––––––

–––––––––– cm ––––––––

 

Nonelectrically treated

 

100

 –

2,500

 –

11,000

 –

42

 –

75

 –

80

 –

2.2

 –

3.2

 –

3.4

 –

18

 –

74

 –

81

 –

 

 

0 WAI

 

70

 

160

 ***

770

 ***

26

 

61

 

75

 

2.1

 

4

 

4.0

 

20

 

61

 

78

 

 

 

2 WAI

 

 

 

1,900

 

3,300

 ***

 

 

62

 

85

 

 

 

3.6

 

3.7

 

 

 

70

 

101

 ***

 

 

Nonelectrically treated

 

 

 

 

 

9,200

 –

 

 

 

 

84

 –

 

 

 

 

3.9

 –

 

 

 

 

97

 –

 

 

4 WAI

 

 

 

 

 

8,500

 

 

 

 

 

79

 

 

 

 

 

4.0

 

 

 

 

 

94

 

 

                                                                   

aMeans followed by *** are statistically different from the nonelectrically treated check according to Dunnett’s Test (α = 0.05).

bAbbreviation: WAI, wk after first visible inflorescence.

Participation Summary

Educational & Outreach Activities

1 Curricula, factsheets or educational tools
1 Journal articles
2 On-farm demonstrations
1 Online trainings
2 Tours
4 Webinars / talks / presentations
2 Workshop field days

Participation Summary:

150 Farmers participated
25 Ag professionals participated
Education/outreach description:

Presentations were given at sweetpotato field days, sweetpotato research meetings with growers, extension agents, and other industry professionals, national, regional, and local annual weed science meetings. A video was recorded for the 2020 virtual sweetpotato field day and the url is listed here.   https://youtu.be/mkpR9QB_RMI         The following abstracts, awards, and journal articles are related to this project.  

  • Moore LD, Jennings KM, Monks DW, Leon RG, Boyette MD, Jordan DL (2022) Influence of herbicides on germination and quality of Palmer amaranth (Amaranthus palmeri) seed. Page 190 In Proceedings of the Southern Weed Science Society (SWSS) 75th annual meeting. Austin, TX: SWSS
  • Moore LD, Jennings KM, Monks DW, Boyette MD, Jordan DL, Leon RG (2021) Evaluating electrical and mechanical methods for Palmer amaranth (Amaranthus palmeri) control. Page 10 In Proceedings of the SWSS 74th annual meeting. Virtual: SWSS
  • Moore LD, Jennings KM, Monks DW, Leon RG, Jordan DL, Boyette MD (2021) Evaluating electrical and mechanical methods for Palmer amaranth (Amaranthus palmeri) control. Page 98 In Proceedings of the  Northeaster Plant, Pest, and Soils Conference (NEPPSC). Virtual: NEPPSC
  • Honorable mention. Presentation. Northeastern Plant, Pest, and Soils Conference (NEPPSC). 2021
  • Moore LD, Jennings KM, Monks DW, Boyette MD, Leon, RG, Jordan DL, Ippolito SJ, Blankenship CD, and Chang P (2022) Evaluation of electrical and mechanical Palmer amaranth (Amaranthus palmeri) management in cucumber, peanut, and sweetpotato.  Weed Technology (in review)

Project Outcomes

100 Farmers reporting change in knowledge, attitudes, skills and/or awareness
1 Grant received that built upon this project
2 New working collaborations
Project outcomes:

We anticipate that several growers both conventional and organic will purchase the weed zapper and/or the weed puller for weed management in their sweetpotato crop.  

Knowledge Gained:

While only 0 sweetpotato farmers have adopted use of the weed zapper or weed puller we anticipate there will be many farmers who adopt these technologies.  A video was made of the weed zapper removing weeds in sweetpotato.  This video was requested by the executive director of the NC Peanut Growers Association.  Peanuts are a low growing crop with a similar architecture to sweetpotato so we anticipate some peanut growers may also purchase this technology.  From our farm visit and our research project we learned that the weed zapper and the weed puller offered a sustainable way to prevent weeds from reducing yield.  They could be an economical and agronomic sustainable part of a weed management program.   They did not appear to increase soil erosion or reduce soil health.  Through this project we were able to learn to use advancing technology for weed management, provide graduate student training, and provide technology demonstration to growers. 

Recommendations:

We appreciate the funds received to conduct this research as it allowed us to begin our in depth research with the Weed Zapper.  We will continue to evaluate the Weed Zapper to determine if it can be an alternative to herbicides to control cover crops prior to crimping.  Additional studies will determine the effect on weed control in other crops and in weed management systems in sweetpotato and other crops.  

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.