AgBioForum
Volume 18 // Number 1 // Article 12
PDF Comment on this article Issue contents Previous article
Farmer Experience with Weed Resistance to Herbicides in Cotton Production
The University of Tennessee, Knoxville
University of Arkansas at Monticello
Louisiana State University AgriCenter
Mississippi State University, Delta Research and Extension Center
Texas A&M Research and Education Center
Texas A&M AgriLife Extension Center
Cotton Incorporated
A mail survey of 2,500 potential cotton farmers in 13 southern cotton-producing states was conducted in 2012 to assess the temporal and geographic extent of weed resistance to herbicides in cotton production, appraise changes in production practices after the emergence of herbicide-resistant weeds, evaluate the effectiveness of those changes in managing resistant weeds, and ascertain the influence of herbicide-resistant weeds on cotton weed-control costs. Over two-thirds of the farmers surveyed reported herbicide-resistant weeds on their farms. Pigweed and horseweed were the dominant resistant weed problems, accounting for 61% and 25% of the responses, respectively. Newly observed infestations of pigweed and horseweed peaked in 2008-2009 and have declined thereafter. Farmers relied extensively on labor-intensive and mechanical/chemical practices to control resistant weeds. The proportion of farmers in the sample who indicated they had total weed control costs of $50 or more per acre nearly doubled with the emergence of herbicide-resistant weeds on their farm.
Key words: weed resistance, herbicides, pigweed, horseweed, labor-intensive practices.
Introduction

No-tillage and conservation tillage practices in cotton (Gossypium hirsutum) production have provided substantial benefits to farmers and the environment (Givens et al., 2009; Price et al., 2011). Glyphosate-resistant cotton has been an important factor in the increased adoption of no-tillage and conservation tillage practices (Price et al., 2011; Roberts, English, Gao, & Larson, 2006). However, substantial problems related to weed resistance to glyphosate have developed in cotton and other crops and have created a significant threat to the progress made in the adoption of conservation tillage practices (Price et al., 2011). Cotton growers have experienced more problems with weed resistance given cotton’s slower emergence after planting and fewer registered herbicides when compared with other major row crops (Laws, 2010; Norsworthy, Griffith, Scott, Smith, & Oliver, 2008).

Anecdotal reports suggest cotton farmers are abandoning crops due to weed problems, using tillage to control weeds where no-tillage practices had been used previously, relying more on older residual herbicides and sprayer technologies, using non-glyphosate-resistant crops, and devoting more effort to hand hoeing and other labor-intensive practices to manage weeds (Baldwin, 2011; Robinson, 2010; Smith, 2010). Herbicide-resistant (HR) weeds have increased weed control and other production costs and reduced crop yields in some cases (Brandon, 2011; Givens et al., 2011; Rowland, Murry, & Verhalen, 1999; Smith, 2010). Documenting the extent of weed resistance to herbicides in cotton production and the effectiveness of the practices farmers are using to combat the problem can provide guidance in designing strategies to address the problem.

Weed resistance has become a serious problem in crop production because farmers have widely adopted herbicide-tolerant seeds with a concomitant reduction in the use of integrated weed-management practices (Harrington et al., 2009). Previous research surveying farmers about HR weeds has mostly focused on farmer awareness and perceptions about the problem and farmer use of selected weed-control practices to manage herbicide resistance. Surveys of farmers have found that most of them are acutely aware of the potential for weeds to develop resistance to herbicides (Foresman & Glasgow, 2008; Givens et al., 2011; Johnson et al., 2009; Llewellyn, Lindner, Pannell, & Powles, 2002; Prince et al., 2012a). Prior survey research also indicates that farmers have relied on an assortment of chemical and cultural practices to manage weed resistance to herbicides on their farms (Foresman & Glasgow, 2008; Frisvold, Hurley, & Mitchell, 2009; Givens et al., 2011; Johnson & Gibson, 2006; Prince et al., 2012a). Surveys of farmers have also reported that some farmers used more tillage and labor to control HR weeds (Frisvold et al., 2009; Prince et al., 2012a; Johnson & Gibson, 2006).

Notwithstanding the findings of the aforementioned studies, information about when farmers first experienced weed-resistance problems and the geographic extent of those problems has not been documented through a survey of farmers. In addition, the literature on weed resistance has not assessed farmer perceptions of how changes in weed-management practices have affected weed-control costs. The research objectives of this analysis were to 1) assess the temporal and geographic extent of weed resistance to herbicides in cotton production for cotton producers; 2) appraise changes in cotton-farmer production practices due to herbicide resistance and the effectiveness of those practices; and 3) evaluate how changes in production practices to manage weed resistance have influenced cotton weed control costs. To achieve the aforementioned research objectives, a mail survey of cotton farmers in 13 cotton-producing states was conducted in 2012 to provide data about farmer experiences with and adaptations to HR weeds.

Methods and Data

The population of interest was the set of active cotton producers in 13 cotton-producing states: Alabama, Arkansas, Florida, Georgia, Louisiana, Mississippi, Missouri, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, and Virginia. The list frame of 13,894 cotton producers for the 2010 marketing year was furnished by the Cotton Board in Memphis, Tennessee (J. Reeves, personal communication, November, 2011).

A random sample of 2,500 cotton farmers (17.9% of the list frame), weighted geographically in relation to the documented cases of weed resistance in four historical outbreak epicenters—Lauderdale County, TN; Nash and Edgecombe Counties, NC; and Macon County, GA—was selected from the list frame to identify the sample of farmers to receive surveys. The sample size corresponds with a 99% confidence interval with a 2% margin of error (Lohr, 1999). The first documented cases of glyphosate-resistant horseweed (Conyza canadensis) in cotton were in Lauderdale County in west Tennessee in 2000 (Hayes, Mueller, Willis, & Montgomery, 2002; Steckel, 2006) and in Nash and Edecombe Counties in North Carolina in 2003 (Yancy, 2003). The first confirmed case of glyphosate-resistant pigweed (Palmer amaranth) in cotton was in 2005 in central Georgia in Macon County (Culpepper et al., 2006). Glyphosate-resistant pigweed has since been confirmed in Alabama, Arkansas, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, and Virginia (Culpepper, Whitaker, MacRae, & York, 2008; Duzy, Price, & Balkom, 2011; Roberson, 2011).

A cluster-based sampling strategy was used as the survey sampling design, whereby counties were considered to be primary sampling units. Counties and cotton farmers in a county were jointly and randomly selected using Tillé’s (1996) method, an unequal probability sampling without replacement algorithm. The probability sampling weights assigned to each county were based on three criteria: 1) the proportion of cotton acres in a county to total cotton acres in the 13-state region in 2010 (US Department of Agriculture [USDA], National Agricultural Statistics Service [NASS], 2011); 2) the Euclidean distance of the most dense cotton-growing area in a county to the closest one of the four epicenters where weed resistance was initially documented (Lauderdale, TN; Nash and Edgecombe, NC; and Macon, GA); and 3) the number of cotton producers in each county. This sampling strategy ensures that counties producing relatively more cotton were more likely to be selected, with the likelihood of selection discounted by the distance to the nearest original outbreak area. Figures 1 and 2 highlight the geographic distribution of the selection probabilities. Darker shaded counties were assigned higher selection probabilities.

Figure 1. Distribution of pigweed resistance to herbicides reported by farmers in 13 cotton-producing states.
Note: Pigweed resistance to herbicides was identified and reported by one or more farmers in the county. Counties not sampled refer to counties in the list frame that can be selected. The probability of counties selected to receive survey questionnaire were based on three criteria: 1) the proportion of cotton acres in a county to total cotton acres in the 13-state region in 2010 (USDA NASS, 2011); 2) the straight-line distance of the most dense cotton-growing area in a county to the closest one of the four epicenters where weed resistance was initially documented (Lauderdale County, TN; Nash and Edgecombe, NC; and Macon, GA); and 3) number of cotton producers in each county. This sampling strategy ensures that counties that produced relatively more cotton were more likely to be selected, but the likelihood of selection is discounted by their distance from the original outbreak area.



Figure 2. Distribution of horseweed resistance to herbicides reported by farmers in 13 cotton-producing states.
Horseweed resistance to herbicides was identified and reported by one or more farmers in the county. Counties not sampled refer to counties in the list frame that can be selected. The probability of counties selected to receive survey questionnaire were based on three criteria: 1) the proportion of cotton acres in a county to total cotton acres in the 13-state region in 2010 (USDA NASS, 2011); 2) the straight-line distance of the most dense cotton-growing area in a county to the closest one of the four epicenters where weed resistance was initially documented (Lauderdale County, TN, Nash and Edgecombe, NC, and Macon, GA); and 3) number of cotton producers in each county. This sampling strategy ensures that counties that produced relatively more cotton were more likely to be selected, but the likelihood of selection is discounted by their distance from the original outbreak area.

Following Dillman’s (1978) mail survey procedures, the questionnaire, a postage-paid return envelope, and a cover letter explaining the purpose of the survey were sent to each randomly selected cotton producer. A post card was first sent to each cotton producer to inform them of the survey questionnaire coming in a week. The initial mailing of the questionnaire was on February 15, 2012, and a reminder post card was sent one week later. A follow-up mailing was sent on March 7, 2012 to cotton producers who did not respond to previous inquiries. The second mailing included a letter restating the importance of the survey, the questionnaire, and a postage-paid return envelope. Finally, the third reminder post card was sent one week later on March 15, 2012.

Of the 2,500 addresses that were randomly selected, 329 individuals responded to the survey and two were undeliverable due to no forwarding address. Among the 329 responses, 20 individuals declined participation because they had retired or no longer produced cotton, leaving 309 usable responses. Assuming all remaining non-respondents are active cotton farmers, the total number of cotton farmers surveyed was 2,478 (2,498 − 20) and the survey response rate was 12.47% for the 13-state region (309 ÷ 2,478 = 0.1247 × 100 = 12.47%). Although 309 cotton farmers responded to the survey, some respondents did not answer all survey questions.

To achieve the first research objective, the survey asked farmers when they first heard about weed resistance in their area (i.e., the year), whether HR weeds had been identified on their farm, and what problems they had with herbicide resistance, including weed species identified, cotton area affected, and year first identified on their farm. Responses were analyzed using ArcView (Esri, Redlands, CA) and Stata (StataCorp, College Station, Texas) to assess the temporal and geographic extent of weed resistance to herbicides in cotton production. Frequency counts, percentages, and cumulative percentages of when farmers first heard about weed resistance and reported pigweed and horseweed on their farms were summarized for 1999 through 2011. Mean and total area affected by pigweed and horseweed were also summarized. Spatial distributions of resistant pigweed and horseweed were superimposed on the maps of the selected counties (Figures 1 and 2).

To complete the second research objective, we asked cotton farmers who indicated problems with herbicide resistance about adaptations they made in their cotton-production practices, their perceptions about the effectiveness of those practices, and whom they relied on to develop a plan to manage resistance. Farmers were asked about how their use of tillage practices and HR crops changed with weed resistance. Specifically, farmers were asked to estimate their cotton area in no tillage, reduced tillage, and conventional tillage before and after the emergence of HR weeds on their farm. Similarly, farmers were asked to estimate the cotton areas before and after resistant weeds on their farm that were planted in conventional (i.e., non-genetically modified), Liberty Link, Widestrike, and Roundup Ready/Flex cotton. Cotton areas reported by farmers for each of the aforementioned tillage practices and cotton cultivars were aggregated to test for statistical differences before and after resistant weeds. The null hypothesis was that the aggregated crop area for a practice remained the same before and after herbicide resistance. For each practice, a 95% confidence interval was estimated for the aggregated cotton area managed before resistance and then compared for statistical difference with the aggregated cotton area after resistance by a t-test (Cochran, 1977; Stuart & Ord, 1994).

Farmers reporting weed-resistance problems were also asked about other production practices they used to manage herbicide resistance. Farmer responses to the 20 practices were organized into three categories: labor, mechanical/chemical, and cultural practices. These categories were used to evaluate the relative proportions of farmers using the categories to manage weed resistance on their farms. Farmers were also asked to rate on a Likert scale the effectiveness of their management practices (1 = “not effective” to 5 = “very effective”). Finally, farmers were asked to indicate who helped them develop a plan to manage weed resistance—extension/university personnel, crop scouts/consultants, chemical/fertilizer dealers, custom applicators, other farmers, and other sources. Farmers were instructed to check all applicable categories.

To accomplish the third research objective, we asked farmers to indicate their best estimate of total weed-control costs per acre before and after resistance on their farms by checking one of the following cost categories: $0-$49, $50-$99, $100-$149, $150-$199, $200-$249, $250-$299, and ≥$300 per acre. Total weed-control costs were defined broadly in the survey to include “herbicide chemicals and application costs, tillage, labor, hand hoeing, scouting, etc.” Proportions of farmers in the before and after weed resistance samples were calculated for each cost category. A sample proportion test was performed to test for statistical difference in the proportions of farmers before and after weed resistance (Snedecor & Cochran, 1967). The null hypothesis for a particular category was H0: mb = ma, where mb and ma are the proportions of farmers in the weed-cost category before and after weed resistance, respectively. To test the proportion difference, a z-score statistic was calculated as z = (mbma)/SE, where SE = {m(1 − m)[(1/nb)+(1/na)]}0.5 is the standard error, m = (mbnb + mana)/(nb + na), and nb and na represent the number of farmers in the before and after resistance samples, respectively. Potential differences in the proportions of farmers reporting weed-control costs in the before and after resistance groupings were evaluated for the $0-$49, ≥$50, $0-$99, and ≥$100 categories.

Results

Farmers indicated hearing about weed resistance as early as 1980 (n=264, Table 1). However, the number of cotton farmers who first heard about weed resistance increased rapidly after the earliest confirmed reports of glyphosate-resistant horseweed and pigweed in Georgia, North Carolina, and Tennessee in 2000-2005. A previous survey found that only 44% of farmers were aware of glyphosate-resistant weeds in their states in 2005 (Givens et al., 2011), suggesting greater awareness in later years. More than three quarters (79%) of sampled farmers had heard about herbicide resistance in their area by 2009. New instances of farmers hearing about herbicide resistance peaked in 2008 and 2009 and declined thereafter. By 2011, 89% of the sample had learned of weed resistance to herbicides in their area. Another survey in 22 states found similar results, indicating that southern cotton, corn (Zea mays), and soybean (Glycine max) growers are becoming more aware of glyphosate resistance in their area (Prince et al., 2012a).

Table 1. Year when farmers first heard about weed resistance in their area.
Year Number % Cumulative %
Never 29 10
Checked year, but no specific year 5 2
1980-1999 10 3 3
2000 9 3 6
2001 2 1 7
2002 4 1 8
2003 7 2 11
2004 15 5 16
2005 27 9 25
2006 26 9 34
2007 22 7 41
2008 56 19 60
2009 57 19 79
2010 18 6 85
2011 11 4 89
Total 298 100  

A total of 307 farmers responded to the question about whether HR weeds had been identified on the farm. Among these farmers, 213 (69%) indicated that HR weeds had been identified on their farms. Growers in the South more frequently reported glyphosate-resistant weed infestations on their farms (Prince et al., 2012a). In Prince et al.’s (2012a) 2010 survey, only 32% reported experiencing glyphosate-resistant weeds on their farms, the majority of which were in the South (53%). A 2006 survey found that 39% of the 64% who experienced glyphosate resistance were southern growers (Foresman & Glasgow, 2008). Givens et al. (2011) found that 15% had such experience on their own farms in 2005, suggesting that more farmers are experiencing and/or noticing HR weeds as time progresses.

Farmers in the sample indicated that cotton, soybean, and corn were the main crops impacted by weed-resistance problems. The 197 farmers who reported that HR weeds were a problem on their farms identified a total of 349 instances of weed problems on one or more crops. Pigweed was the dominant problem weed (61%), while horseweed was the next most troubling (24%). The remaining instances of problem weeds were ragweed (Ambrosia; 5%) and other weeds (10%).

Spatial distributions of reported instances of resistant pigweed and horseweed were superimposed on the maps of county probabilities. Figure 1 reports the distribution of HR pigweed identified by responding farmers. Pigweed resistance to herbicides tended to be clustered around the epicenter counties where resistance was first reported in Lauderdale County, TN, Edgecombe and Nash Counties, NC, and Macon County, GA. Limited instances of pigweed resistance to herbicides were also reported in Texas. The reporting of pigweed resistance to herbicides by cotton farmers in Texas is consistent with recent popular-press reports of HR pigweed appearing in that state (e.g., Smith, 2012). Figure 2 shows the distribution of horseweed herbicide resistance identified by survey respondents. Reported horseweed resistance to herbicides tended to be clustered around the epicenter counties where resistance was first reported in Lauderdale County, TN, and Edgecombe and Nash Counties, NC. Limited instances of horseweed were reported in Texas. Fewer horseweed cases were reported in Georgia and South Carolina.

The 197 farmers who reported herbicide-resistance problems indicated one or more herbicides that were ineffective in controlling weeds (n=349, Table 2). Answers were classified by herbicide mode of action (Baumann, Dotray, & Prostko, 1999). Problems with weed resistance to amino acid synthesis herbicides occurred at the highest frequency (318 times). Thus, more than 90% were related to amino acid synthesis herbicides. Glyphosate made up the vast majority of responses classified into the Amino Acid Synthesis Herbicides category. Pigweed and morning glory were reported to be the problematic weeds after glyphosate-resistant cotton adoption (Kruger et al., 2009). Crop area in the sample affected by pigweed and horseweed resistance is presented in Table 3.

Table 2. Herbicide categories that farmers reported having problems with weed resistance in the sample.
Herbicide category Respondents a Total answers a
Number b %b Number %
Amino acid synthesis herbicides 191 97 318 91.1
ALS herbicides 13 7 13 3.7
Seedling growth-inhibitor herbicides 3 2 3 0.9
Cell membrane disrupters and organic and arsenicals 2 1 2 0.6
Lipid synthesis-inhibitor herbicides 2 1 2 0.6
Photo-inhibitors herbicides 1 1 2 0.6
Growth-regulator herbicides 1 1 1 0.3
Other 6 3 8 2.3
Total answers 349 100
a 197 farmers gave a total of 349 answers of weed resistance to one or more herbicides on their farms.
b Numbers of respondents do not sum to 197 across herbicide categories, and percentages of respondents do not sum to 100% because each respondent can indicate more than one herbicide category.

Table 3. Crop acres affected by pigweed and horseweed in the sample.
Weed/crop Number Mean acres of crops affected Total acres of crops affected
Pigweed
Cotton 174 608 105,795
Corn 25 564 14,104
Wheat 5 726 3,632
Peanuts 19 154 2,917
Soybean 76 515 39,145
Total 299
Horseweed
Cotton 81 757 61,283
Corn 19 366 6,945
Wheat 1 92 92
Peanuts 5 144 719
Soybean 38 498 18,922
Total 144
Note: 197 farmers gave 299 answers for crop acres affected by pigweed and 144 answers for crop acres affected by horseweed.

Figure 3 summarizes the frequency and cumulative frequency of the year pigweed was first identified on cotton farms as reported by respondents (n=171). The highest frequency of pigweed first identified on a farm was in 2009 with 49 instances (24%), followed by 2008 with 44 cases (22%), 2010 with 32 cases (16%), 2007 with 22 cases (11%), and 2011 with 16 cases (8%). The number of pigweed resistance problems increased rapidly after the first confirmation of resistant pigweed in Georgia in 2005. Instances of pigweed resistance problems first identified on cotton farms peaked in 2008-2009 and declined thereafter.

Figure 3. Cumulative distribution of when pigweed was first identified on cotton farms (171 farmers).
Previous cumulative % refers to cumulative percentage of farmers who reported the year when pigweed was first identified on their farms before the current year. Newly observed % refers to percentage of farmers who reported the year when pigweed was first identified on their farms for the current year.

Figure 4 summarizes the frequency and cumulative frequency of the year horseweed was first identified on cotton farms (n=84). Similar to the responses for pigweed, the highest frequency of new cases of HR horseweed on cotton farms occurred in 2009 with 16 instances (19%), followed by 2008 with 14 instances (17%), 2006 with 10 instances (12%), and 2004 with 9 instances (11%). Similar to the pattern for HR pigweed, the number of new instances of horseweed resistance problems increased rapidly after initial reports in Tennessee and North Carolina in 2000-2003. New instances of HR horseweed problems peaked in 2008-2009 and declined thereafter.

Figure 4. Cumulative distribution of when horseweed was first identified on cotton farms (84 farmers).
Previous cumulative % refers to cumulative percentage of farmers who reported the year when pigweed was first identified on their farms before the current year. Newly observed % refers to percentage of farmers who reported the year when pigweed was first identified on their farms for the current year.

Figure 5 compares cotton acres produced with no tillage, reduced tillage and conventional tillage before and after resistance. For no-till cotton, the point estimate of aggregate cotton acres after resistance falls below the 95% confidence interval of estimated aggregate cotton acres before resistance. For reduced till and tillage cotton, however, the point estimate of aggregate cotton acres after weed resistance was within the 95% confidence interval of estimated aggregate cotton acres before weed resistance. Results suggest that cotton farmers significantly decreased cotton acres in no-till after resistance, but did not change the acres produced with reduced till or conventional tillage (Figure 5). Price et al. (2011) concluded that the conservation tillage (including no-tillage) practice was at risk following the emergence and spread of glyphosate-resistant Palmer amaranth because farmers revert back to tilling fields because tillage can control resistant weeds by disrupting weed seed germination (Shrestha, Lanini, Wright, Vargas, & Mitchell, 2006; Steckel, Sprague, Stoller, Wax, & Simmons, 2007).

Figure 5. Change in total cotton acres in alternative tillage practices reported by farmers in the sample due to weed resistance to herbicides.
Note: The vertical bars within the “before resistance” category of total cotton acres for each tillage practice in the sample of farmers represent a 95% confidence interval for comparison with the “after resistance” category of total cotton acres for each tillage practice. If the point estimator for total cotton acres after resistance falls outside the 95% confidence interval of estimated total cotton acres before resistance, estimated total cotton acres after resistance were significantly different from the estimated total cotton acres before resistance. If the point estimator for total cotton acres after resistance falls in the 95% confidence interval of estimated total cotton acres before resistance, estimated total cotton acres after resistance were not significantly different from the estimated total cotton acres before resistance.

Figure 6 compares cotton acres in three HR varieties and conventional seed before and after resistance. For Liberty Link and Widestrike cotton, the point estimate for aggregate cotton acres after weed resistance was above the 95% confidence interval of estimated aggregate cotton acres before resistance. For Roundup Ready/Flex, the point estimate for summed acres after weed resistance falls below the 95% confidence interval of estimated aggregate cotton acres before resistance. For conventional cotton, the point estimate of aggregate cotton acres after resistance was within the 95% confidence interval of estimated aggregate cotton acres before resistance. Results indicate that farmers in the sample significantly increased the use of Widestrike and Liberty Link cotton but decreased the use of Roundup Ready/Flex cotton after resistance. Farmers did not change their acres of conventional cotton. Increasing use of Widestrike and Liberty Link cotton but decreasing use of Roundup Ready/Flex cotton suggests that farmers were using other types of weed-resistant cotton so they could use different chemicals for which the weeds were not resistant.

Figure 6. Change in total herbicide-resistant cotton acres in alternative tillage practices reported by farmers in the sample due to weed resistance to herbicides.
Note: RR=Roundup Ready
The vertical bars within the “before resistance” category of total cotton acres for each use of HR cotton crop in the sample of farmers represent a 95% confidence interval for comparison with the “after resistance” category of total cotton acres for each use of HR cotton crop. If the point estimator for total cotton acres after resistance falls outside the 95% confidence interval of estimated total cotton acres before resistance, estimated total cotton acres after resistance were significantly different from the estimated total cotton acres before resistance. If the point estimator for total cotton acres after resistance falls in the 95% confidence interval of estimated total cotton acres before resistance, estimated total cotton acres after resistance were not significantly different from the estimated total cotton acres before resistance.

Of the 213 respondents who identified HR weeds on their farms, 197 indicated they adopted one or more of 20 management practices to control resistant weeds (Table 4). The 197 farmers provided a total of 1,154 answers for the practices they used, or an average of 5.6 practices per farmer. Responses were classified into three categories: labor, mechanical/chemical, and cultural practices. Results suggest that farmers relied heavily on labor-intensive practices (41% of the 1,156 answers). Among the individual labor practices, the practice used by the largest number of farmers was hand hoed or pulled weeds (177 of 197 respondents, or 90%), followed by increased field scouting (107 of 197 respondents, or 54%), cleaning harvesting equipment (93 of 197 respondents, or 47%), and hand sprayer to spot-spray weeds (78 of 197 respondents, or 40%). The finding that farmers extensively relied on labor-intensive practices is consistent with Frisvold et al.’s study (2009). They found that of the best management practices that were widely practiced, more than half were labor intensive.

Table 4. Production practices used by cotton farmers to manage weed resistance.
Practices to manage weed resistance by categories Farmers
Number %
Labor intensive totals 475 41
Hand hoed or pulled weeds in field 177 90
Increased field scouting 107 54
Cleaned harvest equipment 93 47
Hand sprayer to spot-spray weeds 78 40
Collaborated with neighbors to control weeds 15 8
GPS weed mapping 1 1
Others 4 2
Mechanical/chemical totals 479 42
Changed in-season herbicide program/chemistry 136 69
Hooded sprayer/post-directed herbicides 129 65
Fall tillage after harvest to kill growing weeds 42 21
Wick applicator/weed wiper 37 19
Fall residual herbicide program 24 12
Variable rate spray technology 19 10
Summer fallowed fields 5 3
Others 13 4
Cultural totals 200 17
Controlled weeds in field boarders/ditches 92 47
Winter cover crop to suppress weeds 71 36
More crop rotations 68 35
Narrower row spacing 15 8
Abandoned part or all of a crop in a field 10 5
Burned field stubble/field boarders/ditches 10 5
Higher plant population 6 3
Others 2 1
Note: 197 farmers gave a total of 1,154 answers for management practices used to control HR weeds on their farms.

Cotton producers also extensively relied on mechanical/chemical methods to control weed resistance, which made up 42% of the 1,156 answers. The chemical practice used by the largest number of farmers was a change in in-season herbicide program/chemistry programs (136 of 197 respondents, or 69%). The mechanical practice used by the largest number of farmers was hooded sprayers/post-directed herbicides (129 of 197 respondents, or 65%), followed by fall tillage after harvest to kill growing weeds (42 of 197 respondents, or 21%). Cotton farmers relied to a lesser extent on cultural practices, making up 17% of the 1,156 answers. The cultural practice used by most farmers was controlling weeds in field boarders and ditches (92 of 197 respondents, or 47%), followed by winter cover-crop planting to suppress weeds (71 of 197 respondents, or 36%) and more crop rotations (68 of 197 respondents, or 35%). Johnson et al. (2009) found that the most effective strategy for most farmers to control glyphosate resistance was to use chemicals, followed by using correct label rates. Our results suggest that farmers used various combinations of labor, mechanical/chemical, and cultural practices to manage weed resistance on their farms.

The 200 farmers who reported HR weeds on their farms indicated one or more sources of information used to develop weed-resistance management plans, and they provided 387 answers about the sources of information used (Table 5). Cotton producers used, on average, 1.9 information sources to aid in developing a plan to manage weed resistance. Of these 200 farmers, 59% relied on fertilizer dealers, 53% used extension personnel, 40% used crop scouts and consultants, and 27% used other farmers. Of the 387 answers reported by farmers, fertilizer dealers (30%) and extension personnel (27%) were used the most to develop weed-resistance management plans. Crop scouts and consultants were the third-most cited information source (20%), followed by other farmers (14%).

Table 5. Information sources used by farmers to develop plans to manage weed resistance.
Source Number a Percent of respondents b Percent of answers b
Chemical/fertilizer dealer 117 59 30
Extension/university personnel 105 53 27
Crop scout/consultant 79 40 20
Other farmers 53 27 14
Other 25 13 6
Custom applicator 8 4 2
a Number of respondents is the same as number of answers for each information source.
b 200 respondents gave 387 answers because each respondent could check more than one information source.

Table 6 summarizes the 5-point-scale ratings of management-practice effectiveness used to control HR weeds. Among the 213 farmers who identified HR weeds on their farm, 186 farmers rated the effectiveness of their management practices; 17% rated their strategies as very effective, 46% rated them as effective; 30% rated efforts as neutral; and 6% rated their practices as not very effective. Only 2% rated their efforts as ineffective. Thus, 63% of the 186 responding farmers rated their altered weed-management practices as either effective or very effective. Similarly, of the growers Prince et al. (2012b) found experiencing glyphosate-resistant weeds, 67% reported that their modifications were very effective or somewhat effective in controlling HR weeds. The most effective practices in their study (rated 9 or 10 on scale of 1 [least effective] to 10 [most effective]) were chemical and cultural methods—using correct label rate (62%), rotating crops (37%), and rotating herbicide chemistries (34%).

Table 6. Effectiveness of management practices to control herbicide-resistant weeds.
Effectiveness of management practices Number %
Not effective 3 2
Not very effective 12 6
Neutral 55 30
Effective 85 46
Very effective 31 17
Total 186 100
Note: 186 farmers rated the effectiveness of management practices.

A total of 166 farmers identified the total weed-control cost categories that best fit their farm situation before and after the onset of weed resistance on their farms (Figure 7). Total weed-control costs were defined broadly in the survey to include “herbicide chemicals and application costs, tillage, labor, hand hoeing, scouting, etc.” Results from the proportional difference tests indicated that the proportion of farmers in the $0-$49 cost category was significantly smaller after resistance (mb = 0.52, ma = 0.08, nb = 87, na = 13, z-score = 3.00, p-value < 0.01) and the proportion of farmers in the ≥$50 cost categories was significantly larger after resistance (mb = 0.48, ma = 0.92, nb = 79, na = 154, z-score = -7.65, p-value < 0.01). About half (48%) of the respondents reported total weed-control costs ≥$50 per acre before HR weeds first appeared on their farms, while 92% of the respondents indicated they had weed-control costs ≥$50 per acre after weed resistance.

Figure 7. Total weed-control costs for cotton reported by farmers with and without weed resistance on their farms.

Similarly, results from the proportional difference tests indicated that the proportion of farmers in the $0-$99 cost categories was significantly lower after resistance (mb = 0.81, ma = 0.44, nb = 135, na = 73, z-score = 5.46, p-value < 0.01), and the proportion of farmers in the ≥$100 cost categories was significantly higher after resistance (mb = 0.19, ma = 0.56, nb = 31, na = 93, z-score = -3.61, p-value < 0.01). About one-fifth (19%) of the respondents had total weed-control costs of $100 or more before resistance, but 56% of the respondents reported weed-control costs ≥$100 after resistance. In summary, our results suggest that a larger proportion of farmers in the sample reported an increase in weed-control costs after the emergence of HR weeds. The percentage of farmers in the sample who indicated they had total weed-control costs ≥$50 per acre nearly doubled with the emergence of HR weeds on their farms.

Conclusions

More than two-thirds of the farmers surveyed experienced HR weeds on their farms. Pigweed was the dominant weed problem, followed by horseweed, ragweed, and other non-specific weeds. Newly observed infestations of pigweed and horseweed peaked in 2008-2009 and declined thereafter. The geographic and temporal distributions of specific HR weeds may be useful for university/extension personnel, crop consultants, and fertilizer dealers to communicate with farmers about how to more effectively manage HR pigweed and horseweed.

Although most farmers were aware that widespread use of herbicide-tolerant seeds encouraged the emergence of resistant weeds, some farmers tried alternative herbicide-tolerant seeds to prevent resistant weeds. This research will enhance opportunities for university/extension personnel, crop consultants, and fertilizer dealers to provide more accurate information to help farmers make decisions about avoiding and/or managing HR weeds.

Labor-intensive practices are the most frequently used practices to control weed resistance to herbicides, followed by cultural practices and then chemical and mechanical/tillage methods. At times, labor-intensive practices to manage resistant weeds can influence farmers to abandon cotton production due to intensive labor requirements (Smith, 2010). Therefore, university/extension personnel, crop consultants, and fertilizer dealers could help some farmers by providing alternative integrated management practices that best fit different production systems.

Weed-control costs significantly increased after the emergence of HR weeds. The percentage of farmers in the sample who indicated they had total weed-control costs of $50 or more per acre nearly doubled with the emergence of HR weeds on their farms. Future research should evaluate which management practices are most cost-effective in adapting to HR weed infestations and how to effectively assist farmers to adopt those practices.

References

Baldwin, F. (2011). Going back in time for weed control. Delta Farm Press. Available on the World Wide Web: http://deltafarmpress.com/soybeans/going-back-time-weed-control.

Baumann, P.A., Dotray, P.A., & Prostko, E.P. (1999). Herbicides: How they work and the symptoms they cause (Texas Agricultural Extension Service Publication B-6081). College Station: Texas A&M University.

Brandon, H. (2011). Resistant weeds increase complexity of herbicide programs. Delta Farm Press. Available on the World Wide Web: http://deltafarmpress.com/soybeans/resistant-weeds-increase-complexity-herbicide-programs.

Cochran, W.G. (1977). Sampling techniques (3rd edition). New York: Wiley.

Culpepper, A.S., Grey, T.L., Vencill, W.K., Kichler, J.M., Webster, T.M., Brown, S.M., et al. (2006). Glyphosate resistant Palmer amaranth (Amaranthus palmeri) confirmed in Georgia. Weed Science, 54, 620-626.

Culpepper, A.S., Whitaker, J.R., MacRae, A.W., & York, A.C. (2008). Distribution of glyphosate-resistant Palmer amaranth (Amaranthus palmeri) in Georgia and North Carolina during 2005 and 2006. Journal of Cotton Science, 12, 306-310.

Dillman, D.A. (1978). Mail and telephone surveys: The total design method. New York: Wiley & Sons.

Duzy, L.M., Price, A.J., & Balkcom, K.S. (2011). Estimating the net returns of managing pigweed in cotton. In S. Boyd et al. (Eds.), Proceedings of the 2011 National Cotton Council Beltwide Cotton Conference (pp. 336-339), January 4-7, 2011, Atlanta, GA.

Foresman, C., & Glasgow, L. (2008). US grower perceptions and experiences with glyphosate-resistant weeds. Pest Management Science, 64, 388-391.

Frisvold, G.B., Hurley, T.M., & Mitchell, P.D. (2009). Adoption of best management practices to control weed resistance by corn, cotton, and soybean growers. AgBioForum, 12(3&4), 370-381. Available on the World Wide Web: http://www.agbioforum.org.

Givens, W.A., Shaw, D.R., Kruger, G.R., Johnson, W.G., Weller, S.C., Young, B.G., et al. (2009). Survey of tillage trends following the adoption of glyphosate-resistant crops. Weed Technology, 23,150-155.

Givens, W.A., Shaw, D.R., Newman, M.E., Weller, S.C., Young, B.G., Wilson, R.G., et al. (2011). Benchmark study on glyphosate-resistant cropping systems in the United States. Part 3: Grower awareness, information sources, experiences and management practices regarding glyphosate-resistant weeds. Pest Management Science, 67, 758-770.

Harrington, J., Byrne, P.F., Peairs, F.B., Nissen, S.J., Westra, P., Ellsworth, P.C., et al. (2009). Perceived consequences of herbicide-tolerant and insect-resistant crops on integrated pest management strategies in the western United States: Results of an online survey. AgBioForum, 12(3&4), 412-421. Available on the World Wide Web: http://www.agbioforum.org.

Hayes, R.M., Mueller, T.C., Willis, J.B., & Montgomery, R.F. (2002). Glyphosate-resistant horseweed and factors influencing its control. In Proceedings of the Southern Weed Science Society (pp. 119-120), January 28-30, 2002, Atlanta, GA.

Johnson, W.G., & Gibson, K.D. (2006). Glyphosate-resistant weeds and resistance management strategies: An Indiana grower perspective. Weed Technology, 20(3), 768-772.

Johnson, W.G., Owen, M.D.K., Kruger, G.R., Young, B.G., Shaw, D.R., Wilson, R.G., et al. (2009). U.S. farmer awareness of glyphosate-resistant weeds and resistance management strategies. Weed Technology, 23(2), 308-312.

Kruger, G.R., Johnson, W.G., Weller, S.C., Owen, M.D.K., Shaw, D.R., Wilcut, J.W., et al. (2009). U.S. grower views on problematic weeds and changes in weed pressure in glyphosate-resistant corn, cotton, and soybean cropping systems.Weed Technolology,23, 162-166.

Laws, F. (2010). Cotton: 2 strikes in weed resistance. Southeast Farm Press. Available on the World Wide Web: http://southeastfarmpress.com/cotton/cotton-2-strikes-weed-resistance.

Llewellyn, R.S., Lindner, R.K., Pannell, D.J., & Powles, S.B. (2002). Resistance and the herbicide resource: Perceptions of Western Australian grain growers. Crop Protection, 21(10), 1067-1075.

Lohr, S.L. (1999). Sampling: Design and analysis. Pacific Grove, CA: Doxbury Press.

Norsworthy, J.K., Griffith, G.M., Scott, R.C., Smith, K.L., & Oliver, L.R. (2008). Confirmation and control of glyphosate resistant Palmer amaranth (Amaranthus palmeri) in Arkansas. Weed Technology, 22(1), 108-113.

Price, A.J., Balkcom, K.S., Culpepper, S.A., Kelton, J.A., Nichols, R.L., & Schomberg, H. (2011). Glyphosate resistant Palmer amaranth: A threat to conservation tillage. The Journal of Soil and Water Conservation, 66(4), 265-275.

Prince, J.M., Shaw, D.R., Givens, W.A., Newman, M.E., Owen, M.D.K., Weller, S.C., et al. (2012a). Benchmark study: II. A 2010 survey to assess grower awareness of and attitudes toward glyphosate resistance. Weed Technology, 26(3), 531-535.

Prince, J.M., Shaw, D.R., Givens, W.A., Newman, M.E., Owen, M.D.K., Weller, S.C., et al. (2012b). Benchmark study: IV. Survey of grower practices for managing glyphosate-resistant weed populations. Weed Technology, 26(3), 543-548.

Robinson, E. (2010). Old technology coming out of the closet. Delta Farm Press. Available on the World Wide Web: http://deltafarmpress.com/management/old-technology-coming-out-closet.

Roberson, R. (2011). Glyphosate resistant pigweed arrives in Virginia. Southeast Farm Press, Available on the World Wide Web: http://southeastfarmpress.com/cotton/glyphosate-resistant-pigweed-arrives-virginia.

Roberts, R.K., English, B.C., Gao, Q., & Larson, J.A. (2006). Simultaneous adoption of herbicide-resistant and conservation-tillage cotton technologies. Journal of Agricultural and Applied Economics, 38, 629-643.

Rowland, M.W., Murry, D.S., & Verhalen, L.M. (1999). Full-season Palmer amaranth (Amaranthus palmeri) interference with cotton (Gossypium hirsutum). Weed Science, 47(2), 305-309.

Shrestha, A., Lanini, T., Wright, S., Vargas, R., & Mitchell, J. (2006). Conservation tillage and weed management (ANR Publication 8200). Oakland: University of California, Division of Agriculture and Natural Resources (ANR).

Smith, J.T. (2012). Pigweed resistance spreading in Texas. The Farmer-Stockman. Available on the World Wide Web: http://farmprogress.com/library.aspx/pigweed-resistance-spreading-texas-41/48/2051.

Smith, R. (2010). Resistant pigweed can run you out of cotton business. Southeast Farm Press. Available on the World Wide Web: http://southeastfarmpress.com/cotton/resistant-pigweed-can-run-you-out-cotton-business.

Snedecor, G.W., & Cochran, W.G. (1967). Statistical methods (6th edition). Ames: Iowa State University Press.

Stuart, A., & Ord, J.K. (1994). Kendall’s advanced theory of statistics: Distribution theory, Volume I (6th edition). London: Arnold.

Steckel, L.E., Sprague, C.L., Stoller, E.W., Wax, L.M., & Simmons, F.W. (2007). Tillage, cropping system, and soil depth effects on common waterhemp (Amaranthus rudis) seed-bank persistence. Weed Science, 55, 235-239.

Steckel, L. (2006). Horseweed (Fact Sheet 06-0034). Knoxville, TN: University of Tennessee Extension.

Tillé, Y. (1996). An elimination procedure for unequal probability sampling without replacement. Biometrika, 83, 238-241.

US Department of Agriculture (USDA), National Agricultural Statistics Service (NASS). (2011). Quick stats [database]. Available on the World Wide Web: http://quickstats.nass.usda.gov/.

Yancy Jr., C.H. (2003). Glyphosate resistant horseweed causing concern in North Carolina. Southeast Farm Press. Available on the World Wide Web: http://southeastfarmpress.com/glyphosate-resistant-horseweed-causing-concern-north-carolina.


Suggested citation: Zhou, X., Larson, J.A., Lambert, D.M., Roberts, R.K., English, B.C., Bryant, K.J., Mishra, A.K., Falconer, L.L., Hogan Jr., R.J., Johnson, J.L., & Reeves, J.M. (2015). Farmer experience with weed resistance to herbicides in cotton production. AgBioForum, 18(1), 114-125. Available on the World Wide Web: http://www.agbioforum.org.
© 2015 AgBioForum | Design and support provided by Express Academic Services | Contact ABF: editor@agbioforum.org