Introduction

Winter wheat (WW; Triticum aestivum L.) acreage in Ontario has increased 38.5% from 1941 to 2017; there were just over 370 000 ha of WW grown in 2017 (Hunt 1980; Kulasekera 2014; Mailvaganam 2018). At the same time, there has been a notable decline in wheat acreage being interseeded with red clover (RC; Trifolium pratense L.) in Ontario (Queen et al. 2009; Gaudin et al. 2013; Loucks et al. 2018). Growers struggle to establish consistent RC stands following WW, and an increasing number are opting instead for other cover crops such as oat (Avena sativa L.). It is likely that as little as 25% of WW acres in Ontario are being interseeded with RC (Johnson 2005).

RC is a short-lived perennial, growing for 1–2 yr in southern Ontario. It is shade tolerant, grows well in cool, moist conditions, and flowers approximately 9–10 wk after establishment. RC fits well as a relay cover crop alongside small-cereal crops when it is frost-seeded into the standing crop just before the spring thaw (Verhallen et al. 2003; Gaudin et al. 2013). This practice is generally understood to not affect the yield of the cereal crop (Gaudin et al. 2013).

Interseeding winter wheat with red clover (WWrc) provides a number of ecosystem services by increasing cropping system diversity and reducing risks of crop failure due to drought stress (Gaudin et al. 2015). This may be because RC can improve organic matter and aggregate stability (Carter and Kunelius 1993), as well as increase moisture retention and suppress weeds (Gaudin et al. 2013; Koehler-Cole et al. 2017), all of which can contribute to better water management. Being a nitrogen (N) fixing legume, RC increases organic soil-N pools (Schipanski and Drinkwater 2011). In studies of corn following RC, it was shown that RC can provide an N credit to the following corn crop and increase corn yields (Vyn et al. 1999, 2000; Gaudin et al. 2014). RC interseeded to WW as a best management practice has also been shown to contribute to better fertilizer N-use efficiency, reduced nitrate leaching in the non-growing season, and reduced N₂O emissions, thus requiring less fertilizer N than management that does not implement a RC cover crop (Jayasundara et al. 2007).

RC interseeded to WW tends to establish evenly, but subsequent plant desiccation and uneven recovery or death leads to nonuniformity that cannot be recovered after wheat harvest (Gaudin et al. 2013; Loucks et al. 2018). Light competition between the wheat canopy and RC can reduce clover biomass (Blaser et al. 2011), but this effect is thought to be less critical than moisture depletion by the wheat crop, which leads to plant desiccation and death and then nonuniformity (Cherr et al. 2006; Queen et al. 2009). Some have attempted to alter WW management — seeding rate, row spacing — to increase RC survival and biomass production, but with little success (Teich et al. 1993; Blaser et al. 2007).

When RC stands are not uniform, farmers no longer depend on RC to reduce the N requirements of the following corn crop or to control weeds and thus are forced to apply additional fertilizers and herbicides. Applying a full rate of fertilizer N on nonuniform RC stands is likely to result in excess N and therefore losses to the environment. Furthermore, increased input costs and low yields may negate the economic benefit of interseeding and may even result in a financial loss when using RC (Stute 2016).

Farmer adoption of the WWrc system and the agroecological benefits that the WWrc system offers are likely to increase if research into WWrc management leads to robust management solutions that can be extended to farmers; however research that is not informed by farmer knowledge and experience can lead to a disconnect and distrust of the scientific community by farmers (Wick et al. 2019). To address this, a survey was distributed in the summer of 2017 to a network of Ontario wheat growers. The objectives of the survey were to learn about the perceived advantages and disadvantages of WWrc, along with the RC and WW management practices of this group of farmers.

Materials and Methods

An online questionnaire was drafted with 41 questions exploring objective and subjective aspects of farmer experiences with the WWrc system (see  Supplementary data ¹). Qualtrics^® software (Qualtrics XM,  https://www.qualtrics.com/) was used to format the questionnaire, and the questionnaire received approval from Research and Ethics at the University of Guelph (Guelph, ON, Canada) before being distributed. A link to the online questionnaire was distributed through personal communication with growers, through grower association newsletters, and over social media.

All 199 survey responses were received between 20 June and 12 Sept. 2017. Of this group, when asked “What county is your home farm located in?”, 1 respondent identified that they farmed in Hungary, 2 farmed in Quebec, and 12 farmed in Ontario but did not name the county. Additionally, five respondents reported that they did not have prior experience with WWrc. These 20 respondents were not considered in the analysis to avoid skewed results, leaving a sample size of 179 respondents who previously or currently practiced WWrc. Descriptive statistics were analyzed in Microsoft Excel version 2016. Quotations are used to support themes and are preceded by an identification (ID) number that was assigned at the time of survey submission.

Results

Of the 179 respondents, 142 identified as currently practicing WWrc: 9 from central Ontario, 5 from eastern Ontario, 12 from Golden Horseshoe, 1 from northeastern Ontario, and 115 from southwestern Ontario — in total, 28 out of 49 Ontario counties were represented (Table 1). Next to the Temiskaming district, which only had one respondent, the Golden Horseshoe region had the lowest proportion of respondents (50%) identifying nonuniform stands as an issue for WWrc; central, eastern, and southwestern Ontario all had similar proportions of 78%, 80%, and 75%, respectively.

Table 1.

Survey responses according to region and reported prevalence of nonuniform red clover stands.

Of those currently practicing WWrc, the top three perceived advantages of the system were improved soil quality (97%, N = 138), yield increases for subsequent crops (76%, N = 108), and an N credit (75%, N = 107) (Fig. 1A). The top three perceived disadvantages were nonuniform stands (73%, N = 103), limitations to WW herbicide selection (43%, N = 61), and possible difficulties in combining WW when there is excessive RC growth (30%, N = 43) (Fig. 1B). The other 37 respondents were not currently practicing WWrc but had in the past, and for this group as well, the most commonly perceived disadvantage was nonuniform RC stands (70%, N = 26).

Fig. 1.

Perceived advantages (A) and disadvantages (B) of interseeding red clover to winter wheat (WWrc) identified by growers currently practicing WWrc. [Colour online.]

RC management

There were some dominant practices for RC management among respondents; the majority of growers used double-cut (68%), common seed (74%), seeded at a rate between 7 and 9 kg ha^-1 (70%), broadcasted (97%), and frost-seeded sometime in March or April (87%; Table 2). Fifty individuals, or 35% of the sample population practicing WWrc, used all these same practices; 40 of these (80%) identified nonuniform stands as an issue with WWrc.

Table 2.

Red clover management for those currently interseeding winter wheat with red clover and interactions with perceived uniformity issues.

The proportion of growers who identified nonuniformity as a challenge in WWrc did not change from common seed to certified seed (70%–68%; Table 2). A greater proportion of people planting a mix of single-cut and double-cut identified uniformity as a challenge, compared with those planting only single-cut or double-cut (91% compared with 72% and 71%, respectively). There were no discernable trends in how seeding time or seeding method affected perceived challenges with uniformity; however, as RC seeding rate increased, the proportion of people identifying nonuniformity as a challenge decreased from 86% at 7 kg ha^-1 to 25% at >11 kg ha^-1. Aggregating the growers’ seeding at a rate of 10 kg ha^-1 or greater, 53% reported challenges with nonuniform stands compared with 78% who seeded less than this.

WW management

There were also dominant practices in WW management; the majority of growers grew soft red wheat (79%) and direct seeded (i.e., no tillage, 74%) at 19 cm row spacing (75%; Table 3). The most popular seeding rates were either 157–168 kg ha^-1 (N = 25) or 0.57–0.65 million seeds ha^-1 (N = 49). The two most popular WW varieties were ‘Pioneer 25R40’ (N = 28) and ‘Dow Branson’ (N = 21).

Table 3.

Wheat management for those currently interseeding winter wheat with red clover and interaction with perceived uniformity issues.

As tillage increased, the proportion of people identifying nonuniform stands as an issue decreased from 77% under no tillage to 33% under conventional tillage (Table 3). Similarly, as row spacing increased from <15 to >20 cm, the proportion of people identifying nonuniformity as an issue decreased from 100% to 0%. There were conflicting trends in how WW seeding rate impacted nonuniformity between those reporting in kg ha^-1 and millions of seeds ha^-1.

Respondent feedback

Respondents were given the opportunity to list any management strategies or issues that they felt contributed to the success or lack of success of their RC stand; quotes related to these topics are categorized by ID numbers and themes and are listed in Table 4. A number of growers commented that a no-tillage system contributed to poor establishment (see quotes in Table 4, ID numbers 25, 78, 149, 154, 186, 189, 192). Explanations for this varied. Several respondents remarked that the amount of crop residue left on the surface affected RC seed–soil contact and thus stand uniformity (ID numbers 32, 117, 132, 154, 192). Some identified that forms of light tillage such as tine weeding (ID number 14) helped to incorporate seeds; others felt that it was necessary to plow down corn stalks (ID numbers 32, 117, 186). In some cases, this was accomplished because previous crops in the rotation required tillage, e.g., edible beans (ID numbers 100, 186, 189); however, other respondents attributed the nonuniform stands in no tillage to earthworm activity (ID numbers 25, 78). Several respondents identified competition between WW and RC as an issue (ID numbers 18, 58, 92, 98, 136, 155, 190). Limited moisture was thus identified as the source of competition between the two crops. Other respondents commented similarly on the necessity of adequate and timely rainfall for RC success (ID numbers 92, 162, 190). Some growers recommended wheat management strategies to make room for RC (ID numbers 76, 110, 173). Other growers seem resigned to RC success being entirely dependent on the weather (ID number 10).

Table 4.

Respondent’s short answers regarding issues and strategies contributing to red clover success in the winter wheat – red clover system.

Discussion

While the survey had an adequate response rate, its insights are limited by the uneven distribution of respondents in their regions and management decisions. There did not appear to be a regional association with RC stand uniformity, but 115 of the 142 growers currently practicing WWrc were from southwestern Ontario (Table 1) — though this does correspond with the dominant WW growing regions. As RC seeding rate increased, the proportion of people identifying nonuniformity as an issue decreased, but only 32 growers used a seeding rate ≥10 kg ha^-1 compared with 106 who used a seeding rate less than this (Table 2). Similarly, as tillage increased, the proportion of people identifying nonuniformity as an issue decreased, but only 35 growers used some form of tillage compared with 105 growers who did not till (Table 3). Growers planting wheat at 20 cm spacing or greater did not report any challenges with uniformity, but only four were represented by this management decision.

Nevertheless, these trends find further support from growers’ qualitative responses. Respondents believed there was an association between no tillage, crop residue left on the surface, and nonuniform stands (Table 4). As 97% of those currently practicing WWrc broadcast their RC seed, it is not difficult to imagine that crop residue could interfere with seed–soil contact resulting in reduced and uneven stands. Indeed, there is evidence to suggest that cover crops establish better when they are drilled as opposed to broadcast (Fisher et al. 2011). A few growers also commented on the increased earthworm activity in no tillage (Table 4), and this is supported by the literature (House and Parmelee 1985; Reeleder et al. 2006), as is the potential for earthworms to bury seeds (Bertrand et al. 2015). Additionally, in winter canola, the crown tends to be raised in no tillage compared with conventional tillage (Assefa et al. 2014). While this does not appear to affect canola winter survival, the ability of RC to survive summer moisture deficits could depend on conserving crown water content (Loucks et al. 2018), which might be compromised if the crown is raised. Whether RC grows large enough in a WW stand to consider crown placement an issue is unclear. Unfortunately, only one grower direct seeded their RC seed, and though this grower did not report challenges with nonuniform stands, this represents an insufficient sample size.

Growers also noted competition between the WW crop and the RC crop as a factor in RC stand uniformity (Table 4), and quantitative responses suggested that higher RC seeding and wider WW row spacing reduced the likelihood of nonuniform stands (Tables 2 and 3), possibly due to less competition. Other studies, however, have not found that adjusting WW row spacing effectively increases RC survival (Teich et al. 1993; Blaser et al. 2007). In the study by Blaser et al. (2007), the field was tilled with a tandem disc and a cultipacker before seeding WW; Teich et al. (1993) did not consider no tillage as it was still relatively uncommon at that time. If tillage is one of the main management factors influencing RC stand uniformity, this may override other management efforts. It has been noted before that the rise of nonuniform RC stands seems to correspond with the arrival of direct seeding technology (Gaudin et al. 2013), but a study by Blaser et al. (2012) did not find any difference in RC biomass between tilled and no-tillage treatments. More research on the effect of tillage is warranted.

A few final qualifications are necessary. First, weather in any given year will interact with a management practice to influence RC stand uniformity, but this year-to-year interaction could not be captured through the survey. Secondly, the conflicting impacts of WW seeding rates on RC stand uniformity may be due to the WW seeding date, which has a dramatic impact on the optimal seeding rate. Thirdly, the analysis assumed that the management practice under question is truly associated with RC stand uniformity; however, it may be that several management practices (or none) contribute to RC stand uniformity and the relationship is merely coincidental. Indeed, it may be that stand uniformity has less to do with management and more to do with weather (Westra 2015), as suggested by some of the respondents (Table 4). Nevertheless, survey responses showed some indication that using tillage, wider WW row spacing, and higher RC seeding rates may improve RC stand uniformity. Due to limited sample sizes with these management practices, more research is needed to formally evaluate them.