Adopting cover crops and buffer strips to reduce nonpoint source pollution: Understanding farmers’ perspectives in the US Northern Great Plains

Results and Discussion

Perception Difference on Water Quality Related Issues among States. Table 1 displays the average value for each water quality related issue, grouped by state, to compare producers’ perception differences at the state level. We also conducted Duncan’s multiple range tests for all listed issues, with different letters denoting statistically different results among the three states. Of the three states, South Dakota producers generally regarded the water quality related issues as more severe than the producers located in the other states. On the contrary, North Dakota producers provided significantly lower ratings on most of the water quality issues than those from South Dakota and Nebraska. On average, South Dakota producers provided ratings of greater than 1 (“slight problem”) for 7 out of 12 issues listed, while those from North Dakota only provided similar ratings for 1 issue. Nebraska farmers’ perceptions were in the middle, with 4 out of 12 issues rated as more than a slight problem.

The issues under the environment and recreation category are largely related to water appearance and associated water activities, and these issues could be readily discerned even at the early stage of water pollution. Most issues under the health and economy category, however, generally take a longer time to fully manifest, and therefore they are generally unobservable unless the water pollution is already a serious problem. As a result, we can see that the issues rated as more than a slight problem under the former category outnumbered those under the latter category for all the three states. The top three issues for all states include excessive aquatic plants and algal bloom, and both issues are caused by high levels of N in water (Strock et al. 2004; Suplee et al. 2009).

Nevertheless, even under the first category, no issue received a rating of greater than 1.5. This indicates that while most producers, especially those in South Dakota, started to realize water related issues in different aspects, no issue listed was considered as serious as a moderate problem in the surveyed states. This suggests water quality issues are not of paramount concerns in the Northern Great Plains. By contrast, a study conducted in Mashavera River Basin, Georgia, a hotspot of water quality deterioration, found that 44.6% of farmers perceived the rivers as polluted, 42.4% perceived limited swimming activities, and 63.8% perceived their health at risk due to water pollution (Withanachchi et al. 2018).

Water Pollution Sources and Agreeable Measures to Improve Water Quality. Producers were also asked about their perceptions on water pollution sources in their local areas. Of the five potential sources listed, the two leading pollution sources were fertilizers/pesticides and lawn care chemicals, with 28% producers viewing them either as a moderate problem or as a severe problem (figure 1). However, even though fertilizers and pesticides have been found as an important issue that leads to nonpoint source pollution (Dowd et al. 2008), 29% of the survey respondents failed to realize it as a pollution source in their local areas. Chemicals used to enhance crop production or grassland maintenance were considered as more important pollution sources than the septic systems and industry discharge, which is possibly linked to the predominant role of agriculture in our study region.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Farmers’ perceptions on water pollution sources in their local areas.

When being asked about their agreement with different types of measures to help improve water quality, nearly half of the producers showed either disagreement (31.8%) or strong disagreement (14.2%) toward paying taxes to protect local water quality (figure 2). In comparison, less than 20% of farmers disagreed (12.7%) and strongly disagreed (3.5%) to implement CPs if it involves additional costs from their pocket. When the option changes from CPs at some cost to CPs at no direct cost to farmers, the percentage of farmers who showed disagreement or strong disagreement further reduced to 8.1% and 0.2%, respectively. Such a contrast suggests that preventative measures that improve soil health and minimize water pollution from the source could garner more farmer support than curative measures that require tax payment to improve water quality. Additionally, when shifting from CPs at some cost to CPs at no direct cost to farmers, the percentage of farmers who showed agreement or strong agreement increased from 43% to 70%. This indicates that financial assistance provided for the conservation practices could potentially play an effective role for 27% out of 70% producers.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Farmers’ agreement with various efforts to improve water quality (CP = conservation practice).

Knowledge and Adoption Status of Cover Crops and Buffer Strips. The rest of this paper, therefore, will be devoted to studying knowledge and adoption status of CPs among farmers and factors potentially affecting farmers’ adoption decisions. Specifically, we studied two CPs—cover crops and buffer strips. As indicated in figure 3, very few respondents indicated that they had never heard of these two practices, and only 7.7% and 13.5% of the producers indicated these two practices were not applicable for their farms.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Farmers’ knowledge and usage status of cover crops and buffer strips.

Figure 3 shows that 40.2% of the respondents adopted cover crops and 36.4% adopted buffer strips. The finding that more farmers showed interest in cover crops than in buffer strips is potentially due to the soil and cash crop benefits attributable to cover crops as well as their wider applicability (Eanes et al. 2017). Additionally, 33.0% and 25.5% of the respondents identified themselves as knowledgeable nonusers of cover crops and buffer strips, respectively. This implies that among the current nonusers who believed the two CPs applicable for their farms, 63.4% and 50.9% identified themselves as knowledgeable for cover crop and buffer strip practices, respectively, and those people could be considered as the potential target groups to promote these practices.

It is noteworthy that the majority of nonusers indicated they “know how to use but do not use” the practice, which implies lack of knowledge is not the major barrier that prevents producers from adopting cover crops and buffer strips. For knowledgeable nonusers of cover crops (“know how to use but do not use”), 16.5%, 40.8%, and 76.0% showed agreement (either “agree” or “strongly agree”) toward tax payment, CPs at some cost, and CPs at no direct cost to farmers, respectively. Among the current cover crop users, the corresponding percentages for these three choices were 17.4%, 53.4%, and 70.6%, respectively. A comparison between the preferences of these two groups indicated that a higher percentage of current users indicated willingness to implement CPs at some cost (53.4%) when compared to knowledgeable nonusers (40.8%), while a higher proportion of knowledgeable nonusers expressed agreement for CPs involving no direct cost from farmers (76.0%) when compared to current users (70.6%). Similar differences in agreement levels exist between the knowledgeable nonusers and current users of buffer strips. For those who “know how to use but do not use” buffer strips, 15.4%, 45.7%, and 76.1% showed agreement (either “agree” or “strongly agree”) toward tax payment, CPs at some cost, and CPs at no direct cost to farmers. Among the current buffer strip users, the corresponding percentages for those three choices were 21.7%, 57.1%, and 73.0%, respectively. Such findings indicate that monetary concerns constitute a primary reason that hinders CP adoption among knowledgeable nonusers.

Descriptive Statistics. Table 2 demonstrates the descriptive statistics of the explanatory variables used in the cover crop and buffer strip regression models. Under the farmer characteristics category, our result showed that average age of the survey respondents was 62.8 years. On the highest education level achieved, 47.9% of the survey respondents had some college, college degree, or above, and the rest had either a high school degree or below. As expected with any farmer survey, we found 92.8% of our survey respondents were male.

Regarding farm characteristics, we found the average ownership score was 2.39, which indicated that majority of the respondents/producers had both owned and rented land in their operation. The area of farmland averaged 1,886 ac (763.2 ha). Similar to Wang et al. (2019), average farm size reported by the respondents for this survey was much higher than the average farm size based on the USDA National Agricultural Statistics Service (NASS) census data. We found that the majority of producers (68.9%) had livestock, which could help enhance cover crop profit by utilizing cover crops for grazing purposes. The adoption rates of CT and DCR were 79.7% and 77.3%, respectively, indicating that the majority of our survey respondents were current users of one or more CPs.

Eight variables were included under the information and perception category. Regarding information source, the mean value of 2.748 indicates that producers on average considered nearly three out of the nine listed information sources as important. The average rating for N is 2.075, indicating that regarding the role of N as a potential water pollutant, the survey respondents generally perceived it only as a slight problem in their local regions (table 2). Of all the respondents, only 27.6% viewed N pollution as a moderate or severe problem, while the rest perceived it as not a problem or slight problem. For producers who viewed N pollution as a moderate or severe problem, the adoption rates were 57.9% and 60.7% for cover crops and buffer strips, respectively. Adoption rates declined to 40.4% and 37.0% for cover crops and buffer strips, respectively, among those who viewed N pollution as not a problem or slight problem. This indicates that those who realize N pollution as a critical issue are much more likely to adopt CPs that prevent nutrient runoff into water bodies. Meanwhile, producers who do not view N as an important water pollutant could also implement adequate CPs, probably due to other reasons such as soil health concerns.

Table 2 shows the average values for the health and economic issues and the environment and recreation issues were 0.463 and 0.976, respectively, which indicates that the former was perceived as less severe than the latter. Of the six issues listed under both categories in table 1, the producers on average regarded less than one issue as a moderate or severe concern. The average ratings for follower attitude and profit priority were 2.513 and 2.906, respectively. This means that while producers demonstrated slight disagreement with both motivations on CP (i.e., cover crops) adoption, they showed more agreement with the economic profit motivation than the follower motivation. About 34.5% of producers perceived the high cost as a limiting factor for their adoption of buffer strips, yet more producers (44.8%) considered buffer strips as not suitable for their farm. Note that figure 3 demonstrates that only 13.5% reported buffer strips as not applicable for their farm, yet the percentage of farmers who reported buffer strips as not suitable (44.8% in table 2) is more than tripled the nonapplicability reporting. Table 2 shows that 54.7% of the respondents were from South Dakota and 19.5% from North Dakota, and the rest were from Nebraska.

Logit Model Regression. Table 3 presents logit model estimation results for cover crop and buffer strip adoption decisions. Coefficient and odds ratio for each explanatory variable are reported in table 3. After eliminating those observations either with missing values or deemed as nonapplicable for adopting cover crops or buffer strips, logistic regressions were performed using 276 and 250 observations for cover crop and buffer strip adoption, respectively. Log likelihood test was conducted, and the hypothesis that all coefficients are zero was rejected at 1% significance level for both models. The concordant percentages for the cover crop and buffer strip models were 74.6% and 70.3%, respectively, which indicates a reasonably good fit of the models.

As indicated in table 3, land ownership had a positive effect on cover crop adoption, which indicated that farmers are more likely to adopt cover crops on their owned land. This is consistent with an Alabama row crop study, which showed that farmers with more rented land are less likely to adopt cover crops because of a greater perceived risk to farm income on rental land (Bergtold et al. 2012). Short-term versus long-term considerations are important for cover crop adoption decisions (Bergtold et al. 2017). Producers are more likely to schedule sustainable management plans for a longer time on their owned land, and therefore they will be more likely to adopt CPs, such as cover crops, that generally take a longer term to manifest soil and economic benefits (Soule et al. 2000). The results showed that farmland size has a positive and significant impact on cover crops adoption, with a larger farming area associated with a greater likelihood of planting cover crops. Farm size is often a good indicator of gross revenue, and larger farms can benefit from economies of scale as they can spread the fixed costs associated with the planting and spraying machinery to more area (Fan et al. 2020). A national survey on early adopters of cover crops shows that growing cover crops is feasible on larger farms, and that the cover crop practice is becoming increasingly common in multiple cropping systems (Dunn et al. 2016).

Compared with those who did not own livestock, producers who owned were 3.119 times more likely to use cover crops. This is consistent with literature findings that livestock producers can incorporate cover crops with livestock production and receive short-term economic returns (MCCC 2015; Lazarus and Keller 2018; Wang et al. 2020b). Similarly, for adopters of CT and DCR practices, the likelihoods of using cover crops were 2.421 and 3.100 greater, respectively, than the nonadopters. The plausible reasons underlying this finding could be that farm CPs are complementary in nature and joint adoption increases crop yield and farm profit (Teklewold et al. 2013; Prokopy et al. 2019), and that producers who are conservation-minded are likely to try a combination of CPs to achieve maximum soil benefits (Schimmelpfennig and Ebel 2016; Bergtold et al. 2017). Similarly, Lee and McCann (2019) also found that producers using CT and DCR were more likely to utilize cover crops in soybean (Glycine max [L.] Merr.) production. Incorporating cover crops may be easier for producers who have already practiced CT because of available equipment. For instance, no-till producers already have the no-till drills and planters to handle cover crop residues, while additional expenses on relevant equipment may impede conventional tillage growers from planting cover crops (Grisso et al. 2009; Mirsky et al. 2013). In addition, producers planting a variety of crops may be more open to try cover crops because they have more relevant equipment and experience with different crops compared to the monoculture farmers (Singer et al. 2007; Arbuckle and Roesch-McNally 2015). Research findings show cover crops have been considered as critical and consistent components of DCR in the US Corn Belt (Singer et al. 2007).

Farmers who deemed more information sources as important in obtaining water quality information are more likely to adopt cover crops. More information about water quality enables producers to be more aware of water pollution (Eanes et al. 2017), which encourages them to take actions including adopting farm CPs in agricultural production (Prokopy et al. 2008; Arbuckle and Roesch-McNally 2015). Among the nine listed information sources, neighbors and friends may provide their opinions based on personal experience. Dealers, consultants, extension personnel, and associations may provide technical support and scientific evidence regarding how water quality problems can be dealt with in the area (Bergtold et al. 2012; Eanes et al. 2017; Prokopy et al. 2019). Government may provide cost shares to cover crop adopters, which is an important factor to account for when farmers make CP adoption decisions (Myers et al. 2019).

The number of issues farmers perceived as a moderate or severe problem under the health and economy category shows a positive effect on cover crop adoption. Specifically, when the perceived number of moderate and severe issues increase by 1, farmers are 1.552 times more likely to adopt cover crops. This suggests that producers who are more aware of the negative consequences of water pollution on health and economy related issues are more likely to plant cover crops. On the contrary, for those who show more agreement with CP adoption because others are practicing them, the likelihood of using cover crops is 0.637 times less likely. Currently, cover crop use is still a relatively new practice with low statewide adoption rates, e.g., the number of cover crop users only account for 10.3% of the total crop operations in South Dakota as of 2017 (USDA NASS 2017). This also echoes the low adoption rate in the semiarid Northern Great Plains (Wade et al. 2015). Therefore, most of the respondents probably know more nonadopters than adopters of cover crops. Consequently, producers with the follower attitude at this stage probably would more likely follow the conventional practice of no cover crops. In this case, one farmer’s adoption decision generates a positive externality, since it would likely motivate some neighbors with the follower mentality to adopt. Our finding in this regard supports the suggestions of previous research that government initiatives on improving adoption of CPs should target the early adopters (Reimer et al. 2012; Perry-Hill and Prokopy 2014).

Table 3 also presents the regression results of buffer strip adoption. Farmland shows a positive effect on buffer strip adoption as when farm size increases by 1,000 ac (404.7 ha), the odds of adopting buffer strips increase by 1.128. This indicates a large area encourages producers to use some land for buffers that improve local water quality. A synthesis of the literature shows that an average of 33 ft (10 m) buffer strip can provide an optimal trapping capability and reduce sediment transport (Liu et al. 2008). With a larger farming area, a producer is more willing to allocate a portion of the land to maintain an effective buffer width (Castelle et al. 1994; Rhodes et al. 2018).

Additionally, our results show that more information sources encourage farmers to practice buffer strips. Consistent with the adoption of cover crops and other CPs, adoption of buffer strips needs evidence from neighbors and friends, and some information can better facilitate producers to have access to technical and financial support (Ryan et al. 2003; Helmers et al. 2008; Rhodes et al. 2018). As a major fertilizer used for crop production, excessive N not utilized by crops has caused the nutrient to accumulate in water bodies (Dabney et al. 2006). We find that for producers who regard N as a more severe water pollution source, the likelihood of adopting buffer strips is 1.645 times greater. This indicates that as producers’ perceived environmental costs of agricultural pollutants increase, they would more likely use buffer strips. Evidence has shown that benefit perceptions and knowledge of environmental impacts can contribute to the adoption of riparian buffers by landowners and producers (Buckley et al. 2012; Rhodes et al. 2018). For example, a field survey in the Spring Creek watershed of central Pennsylvania also confirmed the belief that buffers that lead to positive outcomes would motivate landowners to adopt the riparian buffers (Armstrong and Stedman 2012).

Moreover, producers who placed more concerns on high expenses and suitability of buffer strips are 0.535 and 0.558 times less likely to adopt buffers, respectively. Extra costs or profit loss have been found to be significant barriers for CPs. Therefore, government subsidies are necessary to eliminate producers’ concern in this regard and to further promote buffer strip adoption. Producers’ concern with the suitability of buffer strips on their farm reveals the importance of technical support on buffer strip implementation in a site-specific manner. For example, Dosskey (2001) concluded that it remains unclear on the degree of pollution reduction to be expected from utilizing buffer strips in agricultural fields. Kreig et al. (2019) also mentioned that the effectiveness of filter strips in trapping water contaminants could be contingent on landscapes, site locations, and presence of tile drains.

Among the explanatory variables, the factors under farm characteristics and management category and those under information and perception category play a major role in cover crop adoption, while variables under information and perception category play a predominant role in buffer strip adoption. The primary reason that farm characteristics barely show any effect on buffer strip adoption could be that the benefits of buffer strips are entirely off-farm, and therefore the adoption decision is more contingent on the cost and perceived suitability. Similar to Wang et al. (2019), we find that information source and producer perceptions are the primary influencing factors on the adoption decisions of both CPs. However, both respondent characteristics and regional heterogeneity play no significant role in producers’ adoption decisions of cover crops and buffer strips.