Loss of nutrients from agricultural fields has resulted in a greater amount of nutrients in surface and ground water resources globally (Tilman et al., 2011). Flow of nutrients through terrestrial and aquatic systems is also linked with many of the United Nation's Sustainable Development Goals (SDGs) (Ladha et al., 2020) (also see https://www.un.org/sustainabledevelopment/sustainable-development-goals/). Due to runoff from agricultural nutrient applications, impaired water quality has become a major concern with its associated ecological and environmental challenges (Alagele et al., 2019). In addition, excessive tillage practices have also led to soil and water degradation globally (Tilman et al., 2002; Smith et al., 2019). Implementation of strategies for reducing the use of these harmful practices while sustaining agricultural productivity at high levels is a significant priority (Tilman et al., 2011).

As states across the U.S. Midwest have struggled to meet nutrient loss reduction goals from agriculture through traditional individual-focused programs and approaches (Secchi and Mcdonald, 2019), watershed management groups have come to be seen as a promising strategy to catalyze effective action (Church and Prokopy, 2017). Because agriculture is the predominant source of nonpoint source pollution in the region and the main driver of harmful algal blooms (HABS) (Brooks et al., 2016), gulf hypoxia (Rabotyagov et al., 2014; Lee et al., 2018), and impaired waterways (Alagele et al., 2019), watershed management activities that actively engage farmers are of particular interest (Morton and Brown, 2011). Watershed management, which includes water resource utilization control, water pollution control, and economic policies, is an effective means of dealing with these issues at the watershed scale (Heathcote, 2009). A strong global consensus has begun to evolve that effective land and water management must start at the watershed level, and that land and water management actions must be taken in the context of watersheds and the human communities in them (Heathcote, 2009; Ramsar Convention Secretariat, 2010).

Thus, watershed management groups that strive to engage farmers have become increasingly central to water quality improvement efforts across the Midwest (Fishers Farmers Partnership, 2020; Indiana Watershed Initiative, 2020; Iowa Agriculture Water Alliance, 2020; Minnesota Association of Soil Water Conservation Districts, 2020; Natural Resources Conservation Service, 2020). Owing to the complexity of watersheds, uncertainty is one of the key factors influencing watershed management programs as successful management depends upon changes in human behaviors (Floress et al., 2015). These individual behaviors are influenced by a variety of social, economic, institutional, psychological, and biophysical attributes (Floress et al., 2015). Understanding how these factors may influence farmers' decisions to join watershed management groups is important as little is known about farmers' decisions to join such groups.

Although little research on the effectiveness of engaging farmers in such approaches exists, the few studies that have examined the impacts of farmer involvement in watershed groups point to improvements in soil and water conservation outcomes (Morton, 2008; McGuire et al., 2013; Church and Prokopy, 2017; Lee et al., 2018). Lee et al. (2018) found that farmers' active involvement in watershed management (WM) activities had strong direct and indirect relationships with their use of cover crops, a highly effective and heavily promoted nutrient loss reduction practice for farmland (Christianson et al., 2018). At the same time, however, the study found that only 27% of surveyed farmers were involved in organized watershed management activities. As recruiting farmers into watershed groups requires time, effort, money, and other resources (Church and Prokopy, 2017), improved understanding of predictors of watershed management involvement could lead to enhanced farmer participation and achievement of nutrient loss reduction goals. Given that farmer involvement in watershed management appears to lead to improved conservation outcomes, this study asks the question: what factors influence farmers' engagement in watershed management activities?

Of 4,534 valid responses, 1,250 farmers (27.57%) reported that they were involved in watershed management activities in their area (Table 5). Our results show that 12 out of 21 variables were statistically significant at the p < 0.05 level or lower. The MacFadden, Cox and Snell, and Nagelkerke pseudo-R², three common measures of goodness of fit, were 0.15, 0.26, and 0.29, respectively, indicating that the model explains a substantial amount of variance in the dependent variable. The binomial regression model correctly classified 74% of observations, indicating a relatively good fit of the model to the data set (Table 5).

The variables measuring farmers' awareness of (Awareness) and attitudes toward (Attitudes) the NRS were positive and significant (Table 5). The odds ratio statistic indicates that a one-unit increase in the 5-point awareness and attitude scales corresponded to an increase in odds of WM engagement of 16% and 35%, respectively.

One of the two perceived capacity variables had a significant negative effect on farmers' participation in WM. The variable measuring perceived lack of agronomic capacity (Capacity.Agron) was significant and negatively related to farmers' participation in WM, with a one-unit increase in the scale (indicating higher perceived agronomic barriers to nutrient loss reduction) corresponding to a 13% decrease in odds of participation in WM (Table 5).

Among farmers' demographic and farm characteristics, age (Age) and the 8-category measure of gross farm sales (GrossSales) were negatively related to farmers' engagement in WM. The odds ratio statistic indicated that a one-unit increase in age corresponds to a decrease in odds of engagement in WM by 1%, and one unit increase in gross farm sales corresponds to a 6% decrease in odds of engagement in WM (Table 5). The strongest predictor among the farmer and farm characteristics was the variable measuring presence of a stream, lake, or similar water body. Farmers who farmed land bordered a water body (WaterBorder) were 2.02 times more likely to report participation in WM. Other farm characteristics-crop acres, pasture acres, and percent rented cropland-were not significant.

Three out of four information channels through which farmers had learned about the NRS were significant (Table 5). Learning about the NRS through public (Info.Pub) or private (Info.Priv) information sources were positively significant, with a one-unit increase in the number of the public sector and private sector sources associated with a 22% and 27% increase in odds of participation in watershed management activities. The third information channel, the farm press, was significantly and negatively related, with farmers who checked that source being 30% less likely to report engagement in watershed management activities. Popular press sources (Pop.Press) was not significant.

Among the four variables measuring the influence of different actors on nutrient management decisions, influence of public sector actors such as the Natural Resources Conservation Service (Infl.Pub) was positively related to WM. In contrast, the influence of agricultural on-farm research and demonstration groups (Infl.On-farm) was negative and significant, with a one-unit increase in Infl.Pub, resulting in a 30% increase in odds and a one-unit increase in Infl.On-farm resulting in a 16% decrease in odds (Table 5). Private sector entity (Infl.Priv) and family and landlords (Infl.Fam) influence were not significantly related to WM involvement.

The cost-share and technical assistance variable was the strongest predictor of involvement in watershed management activities from among the network-related variables. The coefficient for whether or not a farmer received cost-share or technical assistance to help with conservation practices (CS.TA) was positive and significant. The corresponding odds ratio statistic indicates that farmers who received cost share or technical assistance were more than twice (2.51) as likely to report participation in WM (Table 5).

Establishment of watershed management groups that involve farmers has become a more common strategy in Midwestern water quality improvement efforts (Fishers Farmers Partnership, 2020; Indiana Watershed Initiative, 2020; Iowa Agriculture Water Alliance, 2020; Minnesota Association of Soil Water Conservation Districts, 2020; Natural Resources Conservation Service, 2020). This has raised interest in improving our understanding of predictors of farmers' participation in such groups. Our analysis identified 12 variables that appear to positively or negatively and significantly influence Iowa farmers' engagement in watershed management activities.

Results suggest that the type of information source from which farmers learned about the Iowa Nutrient Reduction Strategy was related to WM behavior. Much effort has gone into NRS-related outreach, with public sector actors such as university extension targeting both individual farmers and the private sector entities such as agricultural retailers that farmers rely on for agronomic advice (INRS, 2020). The results showing that farmers who learned about the NRS through public and private sector sources of information were more likely to participate in WM suggest that those efforts may have directly or indirectly influenced some farmers' decisions to join organized WM efforts. On the other hand, farmers who cited the farm press as an information source were less likely to participate in watershed management activities. A possible explanation for this may be that the farm press tends to focus articles on productivity-related themes (Walter, 1996). It is possible that coverage of conservation practices skews toward on-farm production-related practices rather than off-farm collective approaches to soil and water conservation such as watershed management. Rust et al.'s (2021) findings that farmers did not view the farm press as a credible source of information about sustainable agricultural practices, perhaps because of perceived bias toward agribusiness, support this interpretation. This may point to a need for proponents of watershed management groups to develop communication and outreach strategies that bring their work to farm press outlets' attention.

A second major finding identified associations between entities that influence farmers' nutrient management decisions and WM involvement. Farmers who attributed higher levels of influence to public sector entities such as NRCS were more likely to report WM action. Combined with the finding that farmers who had received cost-share or technical assistance were twice as likely to be involved in WM activities, this offers evidence that public-sector groups' intensive efforts to promote conservation action positively impact farmer participation in WM. These findings align with Davenport and Seekamp's (2013) emphasis on relational and formal networks and support Church et al.'s (2019) recommendation that conservation agency involvement in watershed projects, especially combined with the promotion of cost-share and technical assistance, can play a key role in encouraging farmers' adoption of conservation practices to reduce impaired water quality.

One surprising result was the small negative relationship between influence of organizations that facilitate on-farm research and demonstration and WM involvement. Given the major role that such organizations play in promoting farmer conservation practice adoption, this result was perplexing. It may be that the specific groups that comprise the factor—Practical Farmers of Iowa, Iowa Learning Farms, and the Iowa Soybean Association—tend to focus their on-farm production and conservation practices research and demonstration work on individual farms and farmers rather than group-based watershed management. While these groups certainly support involvement in watershed groups, it could be that farmers who cite them as influential may be more focused on on-farm practices rather than off-farm, collective activities.

The strong positive relationship between farmers' awareness of the NRS and attitudes toward the strategy and its water quality improvement objectives and their engagement in watershed management action is encouraging. A central tenet of the Iowa NRS is that efforts to increase awareness and change attitudes will lead to behavior change. These results, which align with our conceptual frameworks and previous research (Prokopy et al., 2019), suggest that, at least in terms of farmer involvement in WM, the focus on shifting awareness and attitudes may be effective.

Another result that we wish to highlight is the negative relationship between the agronomic capacity variable and involvement in watershed management. Results show that farmers who reported a lower perceived capacity to reduce nutrient loss in their cropping systems were less likely to engage in WM activities. This finding is important because watershed groups that involve farmers often focus on helping them to surmount their perceived capacity barriers through peer-to-peer learning and other assistance (Morton, 2008; McGuire et al., 2013). Thus, this result indicates that watershed management groups and other stakeholders should continue to focus on increasing farmer confidence in their capacity to address nutrient loss, and specifically take self-efficacy challenges into account in farmer WM recruitment efforts. These findings and recommendations align with those suggested in previous studies (Arbuckle and Roesch-McNally, 2015; Burnett et al., 2018; Lee et al., 2018).

Another negative predictor of involvement in WM activities that requires discussion is gross farm sales. Farm sales is a proxy of farm size, so the result indicates that larger-scale farmers may be less likely to become involved in watershed management. Because larger-scale operations farm a disproportionate amount of land relative to their numbers (USDA ERS, 2020), their engagement in soil and water conservation efforts is critically important. Efforts to bring more larger-scale farmers into WM and similar conservation efforts should be increased.

Finally, the result indicating that farmers whose farm operations were bordered or bisected by streams and other water bodies were twice as likely to be involved in WM has major implications for outreach. This finding suggests that watershed planning efforts that employ GIS, remote sensing, modeling, and innovative tools such as the Agricultural Conservation Planning Framework to facilitate watershed-level conservation planning and action (Ranjan et al., 2019) may find such farmers to be more receptive to targeted outreach.

Despite major financial and other investments in individual-focused traditional programs and approaches, agricultural production practices across the U.S. Midwest still lead to major environmental challenges, including harmful algal blooms (Brooks et al., 2016), gulf hypoxia (Rabotyagov et al., 2014), biodiversity loss, and impaired waterways (Alagele et al., 2019), and nutrient loss reduction goals are far from met (Secchi and Mcdonald, 2019). As individual-focused programs have faltered, voluntary collective actions such as watershed management groups have become increasingly central to water quality improvement efforts across the Midwest (Church and Prokopy, 2017). In most agricultural regions, farmers manage the majority of the land, and how they manage it largely determines watershed health. Because farmers' engagement in watershed management appears to directly or indirectly affect the adoption of key practices (McGuire et al., 2013; Lee et al., 2018), results presented in this paper can serve to inform ongoing and increasing efforts to involve farmers in WM activities.

Our findings have broad relevance given the increasing role that watershed management planning and action plays in working toward improved water quality, which it turn makes a vital contribution to social, economic and ecological benefits and services. Diverse countries including the U.S. have agreed upon goals like the United Nations Sustainable Development Goals (SDGs) in an integrated manner (see https://www.un.org/sustainabledevelopment/sustainable-development-goals/). Policy and decision-makers are increasingly looking for policy options that will help them achieve these agreed upon goals (McElwee et al., 2020). It is suggested that many land challenges such as clean water and sanitation (SDG 6), Life under water (SDG 14) and Life on land (SDG 15) can be met with a range of response options readily available, such as reducing the alteration of natural ecosystems and increasing adoption of conservation practices that reduce nutrient loss to surface water (McElwee et al., 2020). Our study, by identifying factors associated with farmers' participation in watershed management activities, can also assist decision-makers to craft policy and goals setting for achieving different SDGs including SDGs 6, 14, and 15.

In summary, this research points to several key levers to help increase farmers' involvement in watershed management activities. Specifically, it highlights the importance of engaging with the information sources and influential actors that can be related to farmer proclivity to take part in WM, as well as the critical role that awareness of water quality issues and attitudes toward amelioration efforts can play. Extension and outreach efforts should continue to focus on raising awareness and attitudes and further align information sources and influential actors, including watershed groups themselves, on working to recruit farmers into watershed management groups and collective efforts to improve water quality.

The datasets presented in this article are not readily available because our Institutional Review Board human subjects research protections make it difficult to share primary survey data. Requests to access the datasets should be directed to J. Gordon Arbuckle, arbuckle@iastate.edu.

This research involving human participants was reviewed and approved by the Iowa State University, Institutional Review Board (IRB).

SU and JA conceived the idea for the manuscript and wrote the manuscript. JA designed the survey and collected the data. SU conducted the data analysis. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer PB declared a past co-authorship with one of the authors SU to the handling editor.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.