Sheep integration for diverse and resilient organic cotton systems

Objective 1

Exploring How Texas Organic Cotton Producers Make Decisions to Implement Sheep Cropping Techniques

We used a qualitative interview approach to explore how Texas organic cotton producers

make decisions about on-farm use of sheep cropping techniques. Qualitative interviews are often used to study a single case or situation that is small in scope and size but provides valuable evidence that could be used as a foundation to further study in a particular area. For our study, we interviewed five organic cotton farmers in West Texas. These five farmers are part of a small group of farmers who grow organic cotton in the state (~125 farmers). We answered two research questions in our study.

RQ1: What factors do Texas organic cotton farmers consider when making decisions about transitioning to sustainable sheep cropping techniques?

Table 1: List of Themes and Subthemes for RQ1

RQ2: What are information seeking behaviors of Texas organic cotton farmers when making decisions about adopting sheep cropping techniques?

Table 2: List of Themes and Subthemes for RQ2

Using Dimensions 4 and 6 of Neef and Neubert’s Participatory Research Framework to Describe Organic Cotton Farmers’ Participation in Research

We used an integrative literature review approach Torraco (2016) to describe the characteristics of how organic cotton farmers choose to participate in research and the benefits they perceive from participating in research projects. Using the interaction attributors of dimension four and benefits attributors of dimension six of Neef and Neubert’s (2011) participatory research framework, we identified descriptors of organic cotton farmers’ participation in research. We developed the research protocol using Torraco’s methods and refined it using Paré et al.’s (2016) guidelines. The sample consisted of articles published between 2006 and 2023 (N=26) that were gathered using Google Scholar, Web of Science, and Research Rabbit. The initial search used the terms “organic cotton,” “sheep cropping,” and “herbicides,” but did not provide a large sample. Thus, we added “livestock cropping.”

RQ1: How is Dimension 4 of Neef and Neubert’s participatory research framework described in the organic cotton farming and sheep cropping literature?

Table 3: Interaction Descriptors for Dimension Four of Neef and Neubert’s (2011) Participatory Research Framework

RQ2: How is Dimension 6 of Neef and Neubert’s participatory research framework described in the organic cotton farming and sheep cropping literature?

Table 4: Benefit Descriptors for Dimension Six of Neef and Neubert’s (2011) Participatory Research Framework

Sheep Integration for Diverse and Resilient Organic Cotton Systems: A Q-method Study

We used Q methodology to characterize cotton producers in regards their on-farm decisions. Q methodology is “a research technique, and associated set of theoretical and methodological concepts, originated and developed by William Stephenson, which focuses on the subjective or first-person viewpoints of its participants” (Watts & Stenner, 2012, p. 3). Q methodology is unique in that it incorporates similarities of each to explain rigorous research questions or objective and emphasizes “contextual human subjectivity,” while the statistical components (e.g., correlation and factor analysis) provide rigor (Leggette & Redwine, 2016, p. 58). Our sample includes 22 organic cotton farmers who would be among the population to consider integrating sheep into diverse and resilient organic cotton systems.

RQ1: Identify common viewpoints of organic sheep farmers as they relate to integrating sheep into diverse and resilient organic cotton systems

RQ2: Identify the interaction and benefit descriptors that organic sheep farmers agree are important and not important

Table 5: Q-set of Statements About Common Viewpoints of Organic Cotton Farmers Related to Integrating Sheep into Diverse and Resilient Organic Cotton Systems

Objective 2

Both growing seasons so far were extremely hot and dry in West Texas. In 2022, San Angelo experienced the hottest July on record (in 116 years), with an average daily high of 103.5 F and average temperature of 89.7 F. The area received only 3.37 inches of rain from January through July (71% below average for this period). Due to these conditions, all summer annual forage trials failed (as all sites were non-irrigated). In 2023, record high temperatures were observed in June (daily record of 115 F) and the area experienced the hottest August on record. The sheep-in-cotton trial was irrigated as frequently as possible throughout the season, enabling enough plant growth for crop and weed measurements, but water was inadequate for harvestable cotton production in both years.

In the first year of the sheep-weeding trial, grazing intensity had the greatest effect on total weed biomass (p = 0.036), weed canopy (p = 0.024), and the cotton to weed ratio (p = 0.074). The most intense grazing treatment (target 100% removal) at the 8-leaf initiation resulted in 89% and 49% greater weed biomass reduction than the least and moderately intense treatments, respectively. Similarly, the most intense grazing treatment at the mid-bloom initiation reduced weed canopy cover 33% and 80% more than the least and moderate intense grazing, respectively. Final weed biomass was less due to 8-leaf initiation with 100% weed removal (21 kg ha^-1) than mid-bloom initiation with 100% weed removal (77 kg ha^-1), 4-leaf initiation with 100% removal (180 kg ha^-1), as well as the weedy check (244 kg ha^-1) (Figure 1). Grazing initiation time had the greatest effect on total amount of sheep grazing minutes required for each plot (p < 0.0001). To achieve the same weed biomass removal as the 4-leaf grazing initiation, 8-leaf and mid-bloom initiations required 46% and 135% more sheep grazing minutes, respectively. As sheep were introduced into each plot, they all were not eating at the same time and/or throughout the entire time in the plot. The SGM unit allows for a more quantifiable metric to determine how long sheep should be grazing in the plot, rather than time (e.g., 30 minutes, 1 hour). Weed management treatments did not influence cotton canopy or final cotton biomass, likely due to a season of severe drought. Timing (cotton growth stage) of grazing initiation also did not influence cotton canopy, cotton biomass, weed biomass, and weed canopy to the same effect as grazing duration.

Figure 1: Sheep-weeding effect on ultimate weed biomass at the end of the 2022 season. Weed canopy coverage before and after grazing at each time and intensity in 2022.

Over both seasons (2022 and 2023) combined, grazing intensity effects were not significant within initiation timings, although a consistent relationship was observed with greater reductions of weed biomass at the later grazing timings, and in the weed free check, compared to the earlier grazing timings and the weedy check (Figure 2). 

Figure 2: Sheep grazing timing and intensity effects on ultimate weed biomass compared to a weedy check (WC) and weed-free check (WFC).

As grazing intensity was ultimately insignificant in our measured responses across years, a simplified analysis was conducted comparing only sheep grazing initiation timing to the weedy and weed-free checks. This showed that ultimate weed biomass was reduced in the sheep weeding treatments initiated at the 8-leaf growth stage and mid-bloom, although this reduction of weed biomass did not result in increased cotton biomass (Figure 3). Sheep spent considerably more time grazing weeds than cotton (Figure 4, 5), but as greater herbivory on the cotton occurred at earlier growth stages, we conclude that even slight early-season grazing damage to the cotton resulted in ultimately reduced/compromised yield potential. As the cotton matured, sheep grazed cotton only ~10% of the time compared to ~15% at earlier stages; however, by this point in the season the damage to cotton yield potential via weed competition had likely already occurred. 

Figure 3: Effects of sheep grazing initiation timing on ultimate weed vs. cotton biomass compared to the weedy check (WC) and weed-free check (WFC).

Figure 4: Proportions of sheep grazing time spent eating weeds vs. cotton relative to cotton growth stage.

Figure 5: Graduate student Mathew Stewart planting cotton for sheep grazing trials and weed population difference between grazed (right) and non-grazed(left) plot in 2022.

Lamb Conditioning and Grazing Preference 

Exposure and Aversion to Cotton

After day 3, intake of cotton decreased (Figure 1). From day 5 through day 8 the intake of cotton remained approximately 15%. This was largely a function of some of the lambs decreasing intake but continuing to consume some cotton. Those individuals were re-dosed with LiCl at the same rate until intake was zero for all individuals.  The last lamb was averted to cotton on day 8.

Seven of the 15 lambs were randomly allocated to Familiar treatment. These were also fed fresh cotton daily. Intake increased daily until day 5 of exposure (Figure 1). Thereafter, lambs consumed all of the cotton offered each day. Lambs in the naïve treatment were not exposed to cotton during this phase of the study.

Figure 1. Percent of cotton consumed (intake) during the conditioning phase of this study. “Averted” lambs were dosed with LiCl on day 3 of the study to create an aversion to cotton. “Familiar” lambs were not dosed.

Individual Day Cotton Bite Analysis

The number of bites of cotton was low for all treatments. Averted lambs took fewer bites of cotton. Conversely, familiar lambs selected cotton more frequently than averted lambs (Figure 2). The number of bites of cotton taken by naïve lambs was similar to familiar and averted lambs. Selection of cotton also differed by day (Figure 3). All lambs, regardless of treatment, selected more bites of cotton on day 1 and 4 of the study.

Figure 2.  The mean number of bites of cotton consumed across in each treatment (familiar, naïve, averted) measured per grazing period throughout the six-day trial.

Figure 3. The mean number of bites of cotton consumed across of all treatments by lambs over six days of grazing cotton plots.

The treatment by day interaction also differed for both selection of herbaceous weeds and cotton (Figure 4A and 4B). Regardless of treatment, lambs typically selected herbaceous weeds (grasses and forbs) and avoided cotton (Figure 4A and 4B). The mean number of bites of cotton varied across six-day trial (treatment by day interaction differed) (Figure 4B). When cotton was selected, familiar lambs selected cotton more frequently than averted lambs. Throughout the 6-day grazing trial, the selection of cotton for lambs averted to cotton remained near zero. On day 1, both naïve and familiar lambs consumed cotton. Selection of cotton decreased on day 2 and 3 followed by an increase on day 4. Selection of cotton declined after day 4 for both familiar and naïve lambs. Selection of cotton was not correlated with selection of herbaceous weeds (r² = 0.02).

Figure 4A. The number of bites of herbaceous weeds (grasses and forbs) consumed over six days for grazing on cotton plots

Figure 4B. The number of bites of cotton consumed over six days for grazing on cotton plots.

Weed Cover

After the grazing trial was completed nine of the grazed plots and three of the non-grazed plots (control) were measured for percent weed cover. After six days of grazing the plots the three treatment plots had similar cover of herbaceous weeds averaging 55 percent to 60 percent weed cover (Figure 5). The ungrazed control plots averaged 87.9 percent in weed cover.

Figure 5. The average percent (%) cover of herbaceous weed cover within the “control,” “familiar,” “naïve,” and “averted” treatment plots measured after the grazing trial.

Objective 3.

The ratio of grasses to forbs included in the mix had no main or interactive effect with any other variable. Cover % was influenced by the interaction of sampling date and cover type (Figure 1). Cover by volunteer forb species has increased from fall 2024 to 2025, while bare ground has decreased (Figure 1). No forbs that were included in the mixes germinated between fall 2024 – fall 2025. The primary volunteer forb species include kochia, Russian thistle, and bitter sneezeweed. Grass cover has been low (< 10%) at each sampling event (Figure 1). Of the three grass species included in the mixes (blue grama (BOGR), buffalograss (BODA), and western wheatgrass), two have been recorded in the plots so far and both are increasing in cover as the study progresses (Figure 2). Volunteer grass species include sandbur and windmill grass. Both of these species have decreased in cover since fall 2024 (Figure 2).

Figure 1: Grass cover percentage of three grass species during Fall, 2024 and Spring, Early Summer and Later Summer 2025.

Figure 2: Volunteer grass cover percentage o during Fall, 2024 and Spring, Early Summer and Later Summer 2025.

Biomass Production

Biomass production was influenced by the interaction of sampling date and functional group (grass vs. forb) (Figure 3). The ratio of grasses to forbs included in the mix had no main or interactive effect with any other variable. Dry matter from volunteer forbs increased from fall 2024 to late summer 2025. During almost all sampling periods forbs produced more biomass than grass. Total grass biomass was similar at all sampling points, but values indicate a numerical increase with time since planting. Similarly, grass species composition is changing as indicated by the cover data.

Figure 3:  Biomass production as influenced by the interaction of sampling date and functional groups (grass vs. forb).

Soil Bulk Density

In fall 2025, control areas and planted plots had similar bulk density (av. 1.52 g/cm³). There were no differences observed across different species mixes either.

Forage Nutrition

There was no species × month interaction for crude protein (P = 0.273) or ADF (P= 0.785) .

Sand dropseed had the greatest CP, while side oats grama had the least (Figure 1.1).

Side oats grama was the only species that didn't meet minimum maintenance requirements for dry beef cattle (7% CP).

CP content was greater in June than July, though it met requirements in both months (Figure 1.2).

Buffalo grass had significantly lower ADF than the other species (Figure 2.1).

There was an increase in ADF from June to July (Figure 2.2).

Ca:P ratio varied by species, and sand dropseed did not meet the ideal threshold.

• The lack of an interaction between species and month indicates all species had a similar seasonal response. Decreases in CP and increases in ADF from June to July reflect reduced quality with forage maturation. Buffalo grass had the best overall profile, likely due to its high leaf-to-stem ratio. Although sand dropseed had the greatest CP content, it did not have desired ADF and Ca:P values. This may be due to sand dropseed’s low leaf-to-stem ratio. Results suggest that protein supplementation may be necessary in late summer as CP is likely to decrease below requirements and energy is nearing the minimum threshold for lactation.

Winter Forage Trials

2023 Haygrazer Trials

2023 Yield Summary (Harvest 1)

2024 Haygrazer Trials

All varieties were planted at 500,000 live seed per acre with a small-plot cone drill on 7.5” spacing in 5 × 20-ft plots with four replications.

Objective 4

Publish agronomics, economics, and communication research results with a specific focus on public accessibility. The PDs published Extension publications, disseminated videos focusing on communicating shared values related to sheep-weeding/cropping and its role in environmental sustainability. The videos will follow The Center for Food Integrity’s recommendations for building trust and earning a social license (https://foodintegrity.org/trust-practices/trust-model/). The videos will be hosted on SARE’s Learning and Resources webpage and disseminated through Texas A&M AgriLife’s social media accounts, including those of the individual PDs professional program accounts.

Herbicide costs

One of the potential benefits of using grazing as a form of weed control is the savings achieved by through reduced herbicide use. Texas AgriLife Extension budgets give typical herbicide expenditures for cotton production. The following table, based on these budgets, gives the expected expenditure on herbicide per hectare for cotton production in the region where this study was conducted. If grazing is successfully used to completely replace herbicide use, this results in a total savings of $105.17 per hectare.

Table 1: Herbicide Cost per Hectare

Fencing

The most common pivot size is ¼ mile long and covers approximately 50 hectares. A square plot containing this pivot has a perimeter of two miles or 10,560 feet. The fencing used in this study cost $1.70 per foot. The expected life expectancy of this fencing is 7 to 10 years. This results in an expected cost of per hectare.    

Labor

A more complete analysis would consider the changes in labor costs associated with the different weed control methods. This would involve accounting for a decrease in labor use due to less chemical applications as well as the increased labor needed for animal management. These variables, however, were not tracked during the course of this project. For the purposes of this study it will be assumed that the amount of labor needed for using grazing as a means of weed control is not significantly different than the amount required with traditional methods.

Forage

The weeds consumed by sheep in this trial represent a valuable source of forage. For this study it was assumed that the plant matter had a similar nutritional content to Bermudagrass hay. Based on Texas Agrilife Extension budgets, this would put the value of forage at approximately $200 per ton. There was extensive variation in the study both between and within treatments. Weed removal by sheep was calculated in 2022 but not in 2023. This resulted in 4 observations for each treatment combination. The following table give the value of forage gained for these different combinations during the study. The row numbers correspond to grazing timing and the column numbers correspond to grazing intensity. For instance, grazing with timing 1 and intensity 2 resulted in utilization of $14.64 of forage. The average value of forage utilized across treatments was $72.19.

Table 2: Forage Values per Hectare.

Cotton Consumed

While the sheep in this study did selectively graze more weeds than cotton. They also consumed some cotton plants. In order to analyze the change in cotton production, the weed controls were used as a baseline. Each treatment within a block was compared to the weed control in that block. The decrease in cotton production was calculated as the difference between the weed control and treatment. The current price of $0.63 per pound of lint was used to translate this into financial cost. This methodology gave eight observations for each treatment combination. The average costs of these are given in the table below. The average loss in revenue due to cotton consumption across all timings and grazing intensities was $65.03.

Table 3: Cotton Losses per Hectare

Economic Summary

Combining the changes in these individual factors gives the expected net change resulting from switching from traditional methods to grazing as a form of weed control.

Table 4. Increase in revenue, cost and net change

The results of the study are promising with an average increase in revenue of  $70.09 across all treatments. The worst scenario occurred when sheep were grazed at the 4-leaf stage to 90% weed removal. This treatment resulted in a loss of $38.53 per hectare, compared with chemical control methods. On the other hand, The best outcome occurred when grazing was conducted at the 8-leaf stage to 70% weed removal. This resulted in an increase of $198.12 per hectare. In addition to the ecological benefits, It appears grazing as a means of weed control is also potentially economic advantageous.