Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A farm systems approach to the adoption of sustainable nitrogen management practices in California

A farm systems approach to the adoption of sustainable nitrogen management practices in California Improving nitrogen (N) fertilizer management in agricultural systems is critical to meeting environmental goals while maintaining economically viable and productive food systems. This paper applies a farm systems framework to analyze how adoption of N management practices is related to different farming operation characteristics and the extent to which ferti- lizer, soil and irrigation practices are related to each other. We develop a multivariate probit regression model to analyze the interdependency of these adoption behaviors from 966 farmers across three watersheds and diverse cropping systems in the Central Valley of California. Our analysis demonstrates that farmers adopt varying combinations or portfolios of practices, with the most common portfolio featuring a core set of fertilizer-focused practices. Irrigation infrastructure is an especially important farm operation characteristic for encouraging adoption of innovative practice portfolios that integrate water and fertilizer management. These findings highlight the ability for a farm systems approach to improve our understanding of farmer decision-making across diverse agricultural landscapes. Keywords Agricultural decision-making · Nitrogen management · Farmer adoption · Farm systems · Multivariate probit regression Abbreviations NUE Nitrogen use efficiency BMP Best management practice QAP Quadratic assignment procedure CGSP Colusa Glenn Subwatershed Program SJDWQC San Joaquin County and Delta Water Quality ESJWQC East San Joaquin Water Quality Coalition Coalition ET Evapotranspiration ILRP Irrigated Lands Regulatory Program MVP Multivariate probit regression N Nitrogen * Jessica Rudnick California Sea Grant, University of California San Diego, jrudnick@ucsd.edu 9500 Gilman Dr., #0232, La Jolla, CA 92093-0232, USA Mark Lubell Department of Environmental Science and Policy, University mnlubell@ucdavis.edu of California Davis, One Shields Drive, Davis, CA 995616, USA Sat Darshan S. Khalsa sdskhalsa@ucdavis.edu Department of Plant Sciences, University of California Davis, One Shields Drive, Davis, CA 995616, USA Stephanie Tatge statge@ucdavis.edu The Freshwater Trust, 1717 I St., Suite A, Sacramento, CA 95811, USA Liza Wood belwood@ucdavis.edu Department of Agricultural and Resource Economics, University of California Berkeley, 100 Academic Hall, Molly Sears Berkeley, CA 94720, USA molly_vandop@berkeley.edu Patrick H. Brown phbrown@ucdavis.edu Vol.:(0123456789) 1 3 784 J. Rudnick et al. To address these problems, we apply a ‘farm systems’ Introduction framework that envisions farming operations as complex systems in which multiple parts or subsystems of the farm Nitrogen (N) fertilizers are a dominant input in industri- are interrelated and interdependent, such that change in alized agricultural systems, significantly enhancing crop one part of the system is likely to induce change in another growth and yields, while also generating one of the most (Giller 2013; Shaner et al. 2019). The concept of farm sys- challenging sources of current environmental pollution tems has emerged over time in agricultural development (Good and Beatty 2011; Osmond et al. 2015; Kanter et al. and agroecology literatures, usually with an emphasis on 2020). Excess fertilizer not taken up by the crop is suscep- the importance of interdisciplinary research and recogni- tible to leaching, runoff and volatilization, leading to con- tion of the multiple biophysical, ecological, social and eco- tamination of drinking water resources, ecosystem dam- nomic factors at play in agricultural systems (Norman 1980; ages and release of nitrous oxide, a potent greenhouse gas van Rooyen 1984; Bawden 1995; Snapp and Pound 2008; (U.S. EPA 2017; Harter et al. 2012; Tomich et al. 2016). Giller 2013; Shaner et al. 2019). However, this integrated While agricultural research and extension has devoted sig- approach remains rather nascent in applied farmer behavior nificant attention to developing farm management strate- and adoption research (Church et al. 2020). This paper aims gies for improving N use efficiency (NUE), understanding to reconcile this gap by applying a farm systems framework the factors influencing farmers’ adoption of these prac- to evaluate farmers’ adoption of a suite of N management tices remains an active area of agricultural social science practices, paying careful attention to both the interdepend- research (Reimer et al. 2017). Moreover, a growing body ency between individual management practices and how of ecological modelling research suggests that it will be practice portfolios vary across farm types. necessary in most circumstances to simultaneously imple- This study is grounded in the empirical context of the ment multiple best management practices, in order to see Central Valley of California, where N management has the desired improvements in ecological and social out- become a key focus of the state given extensive nitrate pol- comes (Bosch et al. 2013; Teshager et al. 2017; McLellan lution in groundwater resources and associated threats to et al. 2018). Yet, every practice will not necessarily be effi- drinking water (Harter et al. 2012). The diversity in agro- cient or effective under the ecological or operational con- nomic, economic and ecological factors across the region ditions of every farm, thus farmers must ultimately be able allow us to test two core hypotheses. First, farmers adopt to determine the practices that best fit their unique context portfolios of practices that reflect interdependencies, with and tailor a portfolio of practices across their farm that practices in the same farm management area (e.g. ferti- work together synergistically to improve N management. lizer, soil, irrigation) more likely to be co-adopted. Second, This paper develops a conceptual and analytical frame- the benefits and costs of practices, and thus the portfolio work that encompasses two problems related to the com- adopted, vary across heterogeneous agronomic, economic plexity of N management and vexed agricultural policies and ecological conditions that shape different farm opera- that incentivize or mandate practice adoption. First, man- tions. While these hypotheses do not exhaust the possible agement practices across a farm are interdependent and implications of a farm systems framework, they are impor- therefore it is important to analyze the mix or portfolio tant initial ideas applying the idea of farm systems to farmer of practices farmers use to meet their management goals. adoption on N management. Thus, research must move beyond analyzing only a single The remainder of this article is organized as follows: we practice at a time or counting multiple practices in ways situate our farm systems study on N management within that do not account for interdependencies. the context of broader adoption literature and develop our Second, not every practice or portfolio of practices will two core hypotheses related to N management in Califor- be effective under the ecological or operational conditions nia. We then provide more details on our study context and of every farm. Thus policies that encourage widespread research design, which relies on survey data from 966 farm- adoption of a specific practice or as many practices as ers in three watersheds of California’s Central Valley. We possible, do not adequately account for different farming describe a statistical analysis method, called multivariate contexts. The heterogeneity of practice benefits and costs probit regression, which allows us to estimate the probability across different agro-ecological contexts contributes to a farmer adopts different individual practices, accounting the inconsistent empirical results in terms of what farm for interdependence among those practices. This statistical operation and operator factors predict practice adoption framework allows us to better capture these interdepend- (Knowler and Bradshaw 2007; Prokopy et al. 2008, 2019; ences and go beyond analyses that look at practices in isola- Baumgart-Getz et al. 2012; Wauters and Mathijs 2014; tion or as a simple sum or index. The results focus on which Ulrich-Schad et al. 2017; Liu et al. 2018). portfolios of practices are likely to be co-adopted, and how farm operation characteristics, especially related to irrigation 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 785 systems, impact the portfolio adoption decisions of differ - the balance between crop N demand and N losses (Denny ent farms. The conclusion discusses theoretical and policy et al. 2019; Snyder 2017). Thus, effective N management implications of applying the farm systems perspective. will require a portfolio of management practices that influ- ence different management areas of the farm, tailored to the combination of ecological, agronomic and economic dimen- Theory and hypotheses: linking farm sions of a specific operation, all of which can influence how systems to nitrogen management practices a management practice works. We draw on this literature to analyze two hypotheses: The concept of farm systems research has played an impor- adoption decisions on multiple practices are interdependent tant role across agricultural development and agroecology with one another (H1) and adoption of portfolios of practices fields (Fresco and Westphal 1988; Snapp and Pound 2008; vary across farming operation characteristics (H2). These Giller 2013; Shaner et al. 2019), with the farm system being hypotheses are tested in the empirical context of California, defined as “a complex interaction of soils, water sources, which provides a diverse agricultural landscape with produc- crops, livestock, labour, and other resources and character- tion of both annual and perennial commodities, farms rang- istics within an environmental setting which the farm family ing in scale and structure from small acreage, family-owned manages in accordance with its preferences, capabilities and to very large acreage, corporate operations, and a complex available technologies” (Shaner et al. 1982). Farm systems hydrologic system where farmers irrigate using both surface approaches emphasize interdisciplinarity to understand and groundwater and a variety of irrigation technologies. interdependence between different parts of a farm and dif- We evaluate farmer’s adoption decisions of eight different N ferent characteristics of the farmer, and integrate farmers management practices across three farm management areas: throughout the research process to translate results to exten- fertilizer, soil, and irrigation (See Table 1 for descriptions sion, outreach, policy and management goals (van Rooyen of practices). 1984). While some recent work on sustainable agriculture has Hypothesis 1: interdependence of practice adoption explored integrated mental models and systems thinking (Levy et  al. 2018; Halbrendt et  al. 2014; Hoffman et  al. Interdependence across system components is one of the 2014), the farms system perspective has not been widely most fundamental premises of any systems approach. Within embraced in the applied social science research on farmer a farm system, interdependence is present across different decision-making and conservation practice adoption aspects of the farm and the management practices employed. (Church et al. 2020). Instead, much of the research on best In the context of N management specifically, N availability management practice (BMP) adoption emphasizes theories and N losses depend on applications of N fertilizer to crops, of decision making on a single action, like Diffusion of Inno- irrigation management which may introduce additional N vation (Rogers 2003) or Theory of Planned Behavior (Ajzen from nitrate-contaminated groundwater or push N through 1989). Often these studies set the dependent variable as the the soil profile, and soil management which may introduce adoption of one practice or an index composed of multiple organic N sources (e.g. compost, cover crops) or influence practices, and independent variables are some mix of farmer soil properties that relate to nutrient retention (Khalsa et al. demographics, attitudes, perceptions and farm operation 2018). Furthermore, ecosystem modelling suggests portfo- characteristics (Prokopy et al. 2019; Ranjan et al. 2019). We lios of practices will be necessary in order to reduce N losses aim to apply the farm systems framework to understand how enough to have a positive impact on water quality (Bosch farmers consider relationships between practices and how et al. 2013; Teshager et al. 2017; Denny et al. 2019). the specific characteristics of a practice influence its adopt- Portfolio approaches have been applied to measure the ability on different types of farms (Reimer et al. 2012a). relationships between adoption decisions to estimate which N management is a clear case demanding the application practices are complementary (i.e. work in congruence with of a farm systems approach. A substantial body of agro- one another to enhance efficacy), conflictual (i.e. result in nomic research has developed tools to approximate crop N worse outcomes when used together) or substitutable (i.e. demand and recommend best management practices that redundant; one can replace the use of the other) with oth- reduce N losses (Snyder 2017; Khalsa and Brown 2019). ers in contributing to a management goal (McAllister et al. For example, the “4Rs” conceptual framework (“Right 2009; Teklewold et al. 2013; Kassie et al. 2015; Zulfiqar rate”, “Right time”, “Right place”, and “Right source”) et  al. 2016; Murendo et al. 2016; Koppmair et al. 2017). offers farmers and farm advisors a decision support tool From a decision-making perspective, co-adoption of multi- to adjust fertilizer applications (Mikkelsen 2011). More ple practices may be linked through multiple mechanisms, recently, research has highlighted how other aspects of the such as information sources that jointly recommend related farm, especially soil and irrigation management, influence practices (Tucker and Napier 2002), farmer experimentation 1 3 786 J. Rudnick et al. Table 1 Descriptions of N management practices evaluated in study, including the management area of the farm with which the practice is typi- cally associated Practice name Description Farm management area Leaf testing Test crop leaf for crop nutrient status to determine if plant is up-taking enough nutrients Fertilizer Split application Divide fertilizer applications into smaller doses and apply in different applications at needed Fertilizer times in season Soil testing Test soil for residual nitrate at beginning of season and adjust fertilizer application rate as Fertilizer appropriate Irrigation well N testing Test irrigation water in wells for nitrate content and adjust fertilizer application rates as needed Fertilizer Cover crops Plant cover crops to help hold moisture and nutrients in the soil; provides an organic source of Soil nitrogen that breaks down more slowly over time Moisture probe Test soil water content to determine depth of soil saturation and more precisely control irriga- Irrigation tion to give crop just enough water, which still retaining fertilizer in root zone Pressure bomb Determine plant-water stress and adjust irrigation scheduling as appropriate, including when Irrigation fertilizer is applied so that fertilizer stays in root zone ET-based irrigation scheduling Use evapotranspiration (ET) data to determine plant water losses, and calculate how much Irrigation water needs to be replaced with irrigation. Appropriately place fertilizer in the irrigation set so that fertilizer stays in root zone with multiple practices (Pannell et al. 2006), the adoption of potential disconnect between fertilizer application and water one practice lowering the perceived risk or cost of another management is exacerbated further by modern, large scale practice (Feder 1982), technical knowledge ‘spillovers’ that farm operations where often there are specialized teams of contribute to a better understanding of an additional prac- employees and consultants that oversee different areas of the tice (Conley and Udry 2010), co-dependence or use of the operation, with little interaction (Kling and Mackie 2019). same farm infrastructure (Hanson et al. 2009), or increased With these considerations, we hypothesize (H1) that fer- sophistication that drives the farmer to use information from tilizer management practices will be at the core of practice one practice to amend their use of another practice (Aubert adoption portfolios (i.e. higher adoption and co-adoption et al. 2012). with each other), with irrigation management practices at While our data does not allow us to test these mechanisms the periphery (i.e. lower adoption and co-adoption rates). explicitly, we draw on our interdisciplinary understand- ing of N management in California to hypothesize which Hypothesis 2: practice portfolios fit farming practices may be more frequently co-adopted. The relevant operation characteristics N management practices for which we measure adoption (see Table  1) were identified by University of California The complexity of the ecological and agronomic processes Cooperative Extension farm advisors and experts in nutri- that shape N management make it such that there is no pana- ent management. Direct N fertilizer application and moni- cea, or ‘one-size-fits-all’ approach to the selection of appro- toring practices have traditionally been a strong emphasis priate management practices. The portfolio of management of extension, including the 4R’s principles and monitoring practices responds to site-specific farm characteristics that nutrient availability in the plant-soil system. As a result, we influence operational compatibility and economic feasibil - predict farmers have adopted these practices at higher rates ity. This site-specificity is not unique to N management; for and have better knowledge of their interdependencies. In example, a study in different regions of China found that contrast, attention toward irrigation and soil management conservation management practices broadly have differ - practices has been more recent, as improved understandings ent impacts on crop yields and conservation goals under and technologies to monitor nutrient movement through- different climate conditions and cropping systems (Zheng out the agro-ecosystem have been developed (Coates et al. et  al. 2014). The results suggest that farmers must adapt 2005; Khalsa and Brown 2017; Schellenberg et al. 2009; their selection of management practices, and subsequently Fernández and Brown 2013). The extent to which irriga- practice portfolios, to ‘fit’ their specific operations to achieve tion practices influence N losses also depends heavily on the intended results. farm’s water source, soil type and irrigation infrastructure Unlike Midwest agricultural landscapes dominated by in place, complicating the ability for farm advisers to make staple field crops (USDA Midwest Climate Hub 2017), general practice recommendations around irrigation. This California features a diverse agricultural landscape which 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 787 provides an excellent research system for analyzing how Research design portfolios of practices may vary in their fit with different types of farms (Lopus et al. 2010; Shaffer 2013). Farm size Research context: nitrogen management is one of the most consistent predictors of practice adoption in California’s Central Valley (Prokopy et al. 2019), based on the argument that larger farms have more financial capital and economics of scale, California ranks as the most economically valuable agricul- which reduces barriers to practice adoption associated with tural state in the United States by annual crop cash sales. cost, time to return on investment, and risk (Feder and The state boasts more than 400 commodity crops grown Umali 1993; Ghadim et al. 2005; Kipling et al. 2019). Cali- across 77,000 farms and ranches on 25 million acres of land, fornia farm operations also feature crops across the “crop spread along a 500 mile longitudinal gradient (California hierarchy”, from low-value annuals to high-value perenni- Department of Food and Agriculture 2018). The Mediter- als (Blank 2001). High-value perennial crop systems like ranean climate is ideal for perennial and annual crops in fruit and nut orchards, are increasingly replacing low-value most areas of the state, yet creates reliance on irrigation annuals in California (Howitt et al. 2008), and receive a and a highly engineered water system. Top commodities higher return on investment for implementing practices that include dairy, grapes, almonds, berries, livestock, lettuce, increase efficiency or have yield gains. walnuts, tomatoes, pistachios and citrus. Farms also vary The California context also offers the opportunity to study widely in scale and structure—from small and mid-sized the effect of irrigation infrastructure on practice adoption. family-owned operations to very large, multi-commodity The state’s year-round growing season and Mediterranean international corporations (California Department of Food climate (i.e. dry summers, wet winters) creates reliance on and Agriculture 2018). irrigation, encouraging many combinations of irrigation sys- Importantly for our focus on N management, California tems and water sources, from gravity-fed flood irrigation to is one of the first states in the U.S. to implement an agri- systems with pressurized drip or sprinkler infrastructure, cultural non-point source pollution regulatory program, the fed by surface or groundwater. Irrigation systems can have Irrigated Lands Regulatory Program (ILRP). The ILRP is a large influence over N leaching (Letey and Vaughan 2013), implemented through local entities known as “Water Quality and different systems may be more or less compatible with Coalitions” and includes mandatory elements around report- different practices. Drip irrigation systems have been widely ing use of best management practices and an N budget, as adopted in California’s perennial nut crops and high value well as attendance at one educational meeting per year, held annual crops, particularly on farms with sandy soils that rely in each Water Quality Coalition (Central Valley Regional on groundwater (Taylor and Zilberman 2017). Some N man- Water Quality Control Board 2020; for more information, agement practices are implemented more easily through drip see Online Appendix). The N management practices we irrigation systems, like split application where the farmer study in this paper are consistent with those tracked as part can deliver fertilizer sets through the drip irrigation, provid- of the ILRP mandatory reporting. This policy landscape ing a low-labor way to distribute fertilizer throughout the offers a unique context to study the potential effects of gov - season. Other practices have co-evolved with the diffusion ernance on farmer decision-making, compared to well-docu- of drip irrigation, like adjusting irrigation rates according to mented studies evaluating practice adoption under voluntary evapotranspiration (ET) estimates of crop water needs (Han- policy settings (Reimer et al. 2018; Hillis et al. 2018). son et al. 2009; Taylor and Zilberman 2017). As a result, farmers working in operations with pressurized irrigation infrastructure may find their operations more compatible Survey and data collection with the aforementioned practices, whereas farmers working with gravity fed irrigation systems may show a propensity This paper employs data collected through a mail survey toward a different suite of practices. conducted in 2018 across the Central Valley of California. In considering how management practices may vary due The project integrated stakeholder feedback throughout to underlying differences in farm attributes, we hypoth- the research process and included multiple phases of inter- esize (H2) that individual practices and practice portfolios views, focus groups, and preliminary survey data collec- will be distinguishable across different farm types, based tion that both informed our survey design and dissemina- on crop type, farm size, irrigation system and water source tion strategy, and helped in interpreting results. An external characteristics. advisory committee, representing policymakers, farmers, directors of the Water Quality Coalitions, and nationwide researchers and extension specialists also provided survey 1 3 788 J. Rudnick et al. review. Institutional Review Board approval for the study Department of Agriculture 2018). We removed all obvious was obtained through the University of California Davis. non-agricultural entries (e.g. golf courses or public lands The survey was distributed to farmer members from using pesticides) from the mailing lists. This list contained three Water Quality Coalitions: the Colusa Glenn Sub- our best estimate of all eligible farmers who would report to watershed Program (CGSP), the San Joaquin County and SJDWQC or ESJWQC under the ILRP. The survey was sent Delta Water Quality Coalition (SJDWQC), and the East San to all farmers in SJDWQC region (n = 2322) and to a random Joaquin Water Quality Coalition (ESJWQC) (see Fig. 1). sample of 33% of farmers in ESJWQC region (n = 1243), Together, these Coalitions covered over 900,000 acres of due to the size of the Coalition. In aggregate, this totaled irrigated cropland and approximately 7500 individual farm- 4994 surveys mailed across all three regions. ing operations in 2017. These regions represent a longitu- We followed a four-wave mailing process using a modi- dinal transect of the Central Valley that captures a range of fied Tailored Design Method, which included a cover letter agricultural, ecological and socio-political dimensions. The and survey, followed by a reminder postcard, then second most important crop types in these regions include almonds, letter and survey, and final reminder postcard (Dillman et al. walnuts, grapes, tomatoes, sunflowers, pistachios, alfalfa, 2008). In CGSP, the Coalition permitted us access to join corn and wheat. Rice is also a top production crop in the our survey response data to their anonymized mandatory CGSP region, but was removed from our sampling frame reporting data on management practices adopted on each since our management practices focus on non-flooded crop- field. These data allowed us to compare survey-reported ping systems. practice adoption rates with adoption rates reported on man- Mailing addresses were provided by CGSP, to which datory paperwork by farmers who did not respond to the the survey was sent to all members (n = 1471). In SJD- survey, offering opportunity to evaluate the self-selection WQC and ESJWQC, mailing addresses were obtained from bias that is prevalent in survey-based research. We found county Agricultural Commissioner offices, who maintain that adoption rates did not differ between survey respondents publically-available databases of all commercial farming and non-respondents, thus indicating our survey respondents operations in compliance with Pesticide Use Reporting were representative of adoption behavior occurring across requirements. In these regions, organic farmer addresses the watershed (see Online Appendix Table A1). were also obtained through the U.S. Department of Agri- culture Organic INTEGRITY Database (United States Fig. 1 Map of Central Valley of California, highlighting three Water Quality Coalitions where survey was distributed 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 789 combined in a composite variable; all alpha scores were > Variable measurement 0.70, a widely-accepted cut-off to indicate internal validity (Santos and Reynaldo 1999). See Online Appendix Table A2 The survey questionnaire included 30 questions covering a range of topics related to farmers’ views on N management, for information on survey questions and composite variables. Finally, farmer demographic variables included a binary including their adoption of eight different N management practices on their largest parcel of their most important crop variable for college education, a categorical variable for income class, and a continuous variable for years in agri- in the 2017 crop year, measured as a binary variable. In this paper, we treat these eight N management practices as the culture. Binary variables were included to distinguish the three Water Quality Coalitions, with the baseline as farmers simultaneous adoption outcome variables in a multivariate probit regression model. who didn’t identify with the three Coalitions of focus or left the question unanswered. The binary measurement of these practices is a data limi- tation of this paper, especially for practices that are applied Survey respondents in a temporally (e.g. multiple times/year versus every other year) or spatially (e.g. full operation versus particular fields) We received a total of 966 partial and full survey responses heterogeneous fashion. Given the known heterogeneity of operations on our mailing lists (e.g. operations varied back (CGSP: n = 377, SJDWQC: n = 312, ESJWQC: n = 183), constituting an average response rate of 20% from a single crop up to 14 unique crops, and from < 1 acre to > 20,000 acres), we were constrained to developing a (CGSP: 30.7% SJDWQC: 14.4%, ESJWQC: 15.4%). We removed 101 responses from farmers reporting on irrigated survey tool that was general enough to fit every possible respondent. Furthermore, the ILRP collects practice adop- pasture, bringing our useable number of respondents to 865. Response rates were adjusted for the possible non-eligible tion data in a binary fashion as well. Thus, aligning our data structure with that of the regulatory program allows us the addresses included in our original mailing lists (Ameri- can Association for Public Opinion Research 2016). Our best opportunity for data comparisons and to draw policy- relevant conclusions. response rate is on par with recent surveys using similar designs and regarding similar topics (Denny et al. 2019; The survey measured farm operation characteristic vari- ables of interest: most important crop type (aggregated into Wilson et al. 2014; Arbuckle and Rosman 2014). All survey data was digitized for analysis. a binary variable: perennial/annual crop), farm size (log- transformed), primary irrigation type (aggregated into a Respondents are fairly representative of the full farm- ing populations in the surveyed regions, when compared to binary variable: pressurized systems- drip, micro-sprinkler, versus gravity-fed systems- flood, furrow, border strip) and USDA 2012 Census of Agriculture data (see Online Appen- dix Tables A3–A5). The average farm size of our respond- water source (aggregated into two binary variables: access to surface water versus groundwater, and access to both water ents is 355 acres (minimum < 1 acre, maximum ~ 12,000 acres). In aggregate, our survey respondents manage 329,800 sources versus single source). We also measured a number of behavioral variables typi- acres of land across the Central Valley, approximately 35% of the acreage of the study area. Seventy-nine percent of cally used in agricultural adoption research to include as controls (Prokopy et al. 2019). These included information- respondents own their land; 80% of respondents are male; 84% of respondents identify as White or Caucasian, 4% as related variables measuring access to information from three perspectives: a tally of the total number N manage- Hispanic or Latino and 3% as Asian or Asian American. Sixty-one percent of respondents have at least some college ment information sources, a binary variable for the use of Certified Crop Advisers to create N budgets (“consultants”), education. On average, respondents have 35 years of farm- ing experience, and the median gross farm income bracket and a binary variable for the completion of a Self-Certifi- cation course, which is a voluntary educational component is $100,000–$200,000. Respondents listed all crops and acreage they cultivate, of the ILRP which allows farmers to self-certify their own N budgets. Socio-behavioral concepts included problem though we only asked about practice adoption on their most important crop, as self-identified on the survey. Sixty-four awareness (“acceptance of agricultural N sources”), envi- ronmental values (“conservation motivation”), and perceived percent of farmers report only growing one crop, 27% two to four crops, and 4% have five or more crops. Eighty-five per - behavioral control (“self-efficacy”) (Reimer et al. 2012b). These latent variables were constructed using exploratory cent of respondents indicate a perennial crop as their most important. Seventy-four percent of respondents have pres- factor analysis to combine multiple survey question items measured on five-point Likert scales (Costello and Osborne surized irrigation systems on their most important parcels. Forty percent of respondents rely on groundwater only, 44% 2005), which improves reliability (McIver and Carmines 1981; DeVellis 2003; Santos 1999). Cronbach alpha scores have access to surface water (only) through riparian rights or irrigation district delivery water, and 16% have both surface were used to verify internal consistency between the items 1 3 790 J. Rudnick et al. and groundwater access (See Online Appendix Table A6 for Jenkins (2003), our MVP models are estimated using the all descriptive statistics). Geweke–Hajivassiliou–Keane (GHK) simulator in Stata 16. Multivariate normal probabilities are calculated at each Analysis approach: multivariate probit iteration of the simulation. Simulation bias is minimized by for estimating interdependencies increasing the number of random draws from the simulator, to at least as large as the square root of the sample size; we Much of the existing adoption literature uses a standard ran the model with 35 random draws (Cappellari and Jenkins quantitative approach of estimating some type of linear 2003). We also tested for any ordering effects in the depend- model with an individual practice, or count of practices, as ent variables by running the model with multiple different the dependent variable, and multiple predictor variables to orders for the practice dependent variables; results were con- test hypotheses about drivers of adoption (Prokopy et al. sistent across all runs. As an additional robustness check, we 2019). Here, we need an empirical model that simultane- fit individual univariate probit regression models for each of ously estimates farm and farmer variable influences on the the eight practices, which produces very similar coefficient adoption of multiple practices, and how those practices are estimates (See Online Appendix for additional discussion related to each other. To accomplish this goal, we employ a on robustness and Table A7 for univariate probit results). multivariate probit (MVP) model that allows estimation of To test H1, we evaluate the MVP variance–covariance multiple binary probit regression models (in our case 8 mod- matrix alongside a co-occurrence matrix. The co-occurrence els) simultaneously, while analyzing correlation between matrix uses observed adoption data and calculates the pro- errors in the different models. Failure to account for these portion of all farmers who jointly adopt any two practices, correlated error terms can result in inefficient coefficient evaluating all possible dyads of practices, a method that estimates and biased error terms (Cappellari and Jenkins has been applied widely in ecology to evaluate species co- 2003). This approach has been applied in other studies look- occurrence (Hines and Keil 2020). Both relatedness matri- ing at simultaneous adoption in developing agricultural set- ces are visualized as undirected weighted networks with the tings (Koppmair et al. 2017; Kassie et al. 2015; Kara et al. edge weights between every pair of practices reflecting the 2008; Teklewold et al. 2013; Jara-Rojas et al. 2013). relatedness of those two practices. We use Quadratic Assign- Considering all N management practices, each equation ment Procedure (QAP) matrix correlation to assess which in the system can be written as: practices frequently occur together and which practices have highly correlated errors, indicating a potential underlying Y = B X + B X … B X + e, i 1i a 2i b ni n dimension influencing their adoption. (i = LT, ST, CC, IN, MP, SA, PB, ET), To test H2, we draw on descriptive statistical analyses including Pearson’s chi-squared tests to investigate differ - where Y indicates the i different practices (LT = Leaf Test- ences in individual practice adoption rates between farm ing, ST = Soil Testing, CC = Cover Crops, IN = Ir r igation types and evaluate our MVP coefficient estimates to under - N Testing, MP = Moisture Probe, SA = Split Application, stand the predictive power of key farm operation charac- PB = Pressure Bomb, ET = Evapotranspiration-based sched- teristics of interest (crop type, farm size, irrigation system uling) of interest and X are the predictor variables of inter- and water source), while controlling for all other farmer est. Our unit of analysis is an individual farmer. For farmers behavior and demographic variables, as well as interde- who operate across multiple fields, we evaluate their practice pendency across practices. We then qualitatively evaluate adoption only on the largest field of their most important differences in practice portfolios across different operation crop, thus including only one observation per farmer. This types by looking at differences in the co-occurrence practice yields a matrix of estimated model coefficients, with the networks. We highlight results for practice portfolio differ - coefficient for each covariate (B … B ) estimated for each 1i ni ences across farm types with different irrigation systems, of the eight practices. The MVP assumes that the error terms but additional side-by-side practice portfolio comparisons for each practice (e , e , e …) jointly follow a multivari- LT ST CC between other operation types are included in the Online ate normal distribution with a mean of 0 and variance of 1. Appendix Figures A3–A4. The model also generates a variance–covariance matrix that Descriptive statistics and data visualization were car- provides the correlation coefficients (rho) between the error ried out in R Statistical Software Version 3.5.3; multivari- terms of all pairs of equations. These correlations can offer ate modelling was conducted in Stata16. All model code is insight on the complementary (i.e. positive correlations) or linked in the Online Appendix. substitutable (i.e. negative correlations) nature of pairs of practices. Simulated maximum likelihood techniques are used to estimate the model, and following Cappellari and 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 791 The MVP variance–covariance matrix indicates appropri- Results ate use of the MVP model, as the likelihood ratio test for the overall correlation of error terms [ƿ = ƿ = ƿ = ƿ = ƿ Our results are structured around our hypotheses. First, to LT ST SA CC IN 2 2 = ƿ = ƿ = ƿ = 0; χ (28) = 208.781, Prob > χ = 0.000] evaluate how adoption decisions are interrelated across ET PB MP rejects the null hypothesis that the error terms across adop- practices (H1), we present the MVP model error correla- tion equations were not correlated. This result is also sup- tion matrix and practice co-occurrence matrix, as well as ported by the many significant correlation coefficients the QAP correlation between the two measures. Then, we between the pairwise model correlation terms (see Table 2). evaluate how farm operation characteristics influence indi- Nearly all practices are positively correlated, and thus can vidual practice and portfolio adoption (H2) by reviewing be considered complementary. The fertilizer-specific prac- the MVP coefficient estimates and descriptively comparing tices however, emerge as the core practices that are the most co-occurrence networks across different farm types. frequently adopted and have the highest correlations in their error terms with each other, indicating the greatest interde- Interdependence in practice adoption (H1): MVP pendency: leaf testing and soil testing (ρ = 0.63), leaf testing model error correlation and co‑occurrence matrices and irrigation well N testing (ρ = 0.39), leaf testing and split application (ρ = 0.38). On average, farmers report adopting 3.5 out of eight N man- The co-occurrence matrix further indicates that leaf test- agement practices. Adoption rates for individual practices ing, soil testing and split application make up the core of ranged from 75% (split application) to 14% (pressure bomb), the portfolios for most farmers, with co-occurrence rates but the majority of all farmers adopt split application, leaf all at 50% or greater. Each of these three practices also co- testing and soil testing (See Fig. 2). occurred over 30% of the time with moisture probes and The MVP variance–covariance correlation matrix and roughly 25–30% of the time with both irrigation well N test- the practice co-occurrence matrix support H1, in that both ing and ET-based irrigation scheduling. Cover crops and measures show strong relatedness between practices and pressure bombs were most peripheral and had the lowest co- demonstrate that practices cluster most strongly within the occurrence rates with any other practice (see Online Appen- fertilizer farm management area. dix Table A8 for all co-occurrence rates). The co-occurrence Fig. 2 Individual practice adoption rates across all farms; colors indicate management area of the farm. (Color figure online) 1 3 792 J. Rudnick et al. 1 3 Table 2 Correlation coefficients (rho) and standard errors (in parenthesis) of the eight adoption equations’ error terms from the MVP model Farm management Ƿ Ƿ Ƿ Ƿ Ƿ Ƿ Ƿ LT ST SA IN CC ET PB area Fertilizer  Ƿ 1 LT  Ƿ 0.629 (0.06)*** 1 ST  Ƿ 0.382 (0.07)*** 0.209 (0.08)*** 1 SA  Ƿ 0.388 (0.08)*** 0.315 (0.07)*** 0.249 (0.09)*** 1 IN Soil  Ƿ 0.362 (0.08)*** 0.140 (0.07)** 0.203 (0.08)** 0.252 (0.07)*** 1 CC  Ƿ 0.268 (0.08)*** 0.056 (0.07) 0.138 (0.08)* 0.244 (0.07)*** 0.105 (0.07) 1 ET Irrigation  Ƿ 0.220 (0.10) − 0.065 (0.09) − 0.09 (0.11) 0.087 (0.09) 0.122 (0.09) 0.219 (0.09)*** 1 PB  Ƿ 0.334 (0.11)*** 0.288 (0.07)*** 0.110 (0.08) 0.211 (0.07)*** 0.036 (0.07) 0.346 (0.07)*** 0.180 (0.09)** MP Matrix is symmetrical. Positive correlations are interpreted as complementary practices, negative correlations are interpreted as substitutable Bolded values highlight correlation coefficients that are statistically significant at p < 0.1 Practices abbreviated as: LT leaf testing, ST soil testing, SA split applications, CC cover crops, IN irrigation well N testing, ET evapotranspiration-based irrigation scheduling, PB pressure bomb, MP moisture probe 2 2 Likelihood ration test of ƿLT = ƿST = ƿSA = ƿCC = ƿIN = ƿET = ƿPB = ƿMP = 0; χ (28) = 208.781, Prob > χ = 0.000 ***p < 0.01; **p < 0.05; *p < 0.1 A farm systems approach to the adoption of sustainable nitrogen management practices in… 793 Fig. 3 Practice interdependency networks shown. Nodes represent sure bomb, MP moisture probe. (Left) The co-occurrence network has management practices, colored based on farm management areas and edge weights representing the frequency at which two practices are scaled based on adoption rates, with larger nodes representing higher jointly adopted on the same parcel. (Right) The MVP variance–covar- adoption. Practice names are abbreviated as: LT leaf testing, ST soil iance network has edge weights showing the correlation in errors testing, SA split applications, CC cover crops, IN irrigation well N between individual practice equations estimated through the MVP testing, ET evapotranspiration-based irrigation scheduling, PB pres- model and MVP error correlations are visualized as networks with Farm type influences individual practice edge weights representing the strength of the relationship and portfolio adoption (H2): adoption rates between two practices based on the two relatedness meas- and MVP coefficient estimates ures (See Fig. 3). The Quadratic Assignment Procedure demonstrates the In addition to evaluating the interdependency between prac- MVP variance–covariance and co-occurrence matrices are tices, we compared how adoption of individual practices positively correlated and thus consistent with one another and practice portfolios differed across farm types. We find (r = 0.62; p < 0.001; see Figure A1 in Online Appendix). The significantly higher rates of adoption for all eight practices in largest residuals were observed with split application when perennial crop systems as compared to annual crop systems paired with other practices such as soil testing, leaf testing (p < 0.001), very large farms (> 1000 acres) as compared to and moisture probe; split application has higher co-occur- smaller farms (p < 0.05), and operations with pressurized rence rates with each of these practices than error correlation irrigation infrastructure as compared to gravity-fed irrigation rates in the MVP model. (p < 0.05). Farmers with access to both surface and ground- water adopted six of the eight practices at significantly 1 3 794 J. Rudnick et al. Fig. 4 MVP model coefficient estimates and 95% confidence intervals plotted for each practice (denoted by different shape markers), with prac- tices colored by farm management areas higher rates (p < 0.01), with the exceptions of cover crops, The effect of farm type is also evident when we evaluate where adoption was higher with access to groundwater only differences in practice portfolios. Results indicate partial (p < 0.05), and soil testing, where there were no significant support for H2, with irrigation system being a driving fac- differences across water sources (Online Appendix Figure tor in determining if and how farmers integrate the irrigation A2). management practices into their N management portfolios. While adoption rates vary significantly when we look at On pressurized-irrigation farms, we see greater overall co- each farm operation characteristic in isolation, our MVP occurrence of most practices, but more noticeably, we see results support H2, indicating crop type, farm size and irri- an expansion of the core practices adopted by most farmers gation system are the most important adoption predictors, to include moisture probes, ET-based irrigation scheduling, when controlling for multiple farm operation and socio- and irrigation well N testing (see Fig. 5). behavioral factors. Perennial crop type is a positive and sig- nificant predictor for all practices except soil testing. Farm size is a positive and significant predictor for all practices Discussion except cover crops. Pressurized irrigation system is the farm operation characteristic that most distinguishes adoption Improved N management to maximize crop use and mini- drivers between different practices, as a positive and sig- mize losses to the environment is complex, requiring farm- nificant predictor for irrigation-specific practices (moisture ers regularly monitor, evaluate and adapt. The relevant suite probe, pressure bomb and ET-based irrigation scheduling), of management practices needed to address these multiple as well as irrigation well N testing. Water source variables dimensions stretch across different management areas of the however, were seldom significant predictors for any practice. farm and vary with other operational characteristics. Our Among the information, socio-behavioral and farmer results highlight the complexity and interdependency of demographic control variables, information source tally was farmers’ adoption of these practices. the only variable that had a positive and significant effect on On average, California farmers adopt between three and all practices. The other socio-behavioral and demographic four of the eight N management practices we measured. variables were less consistently significant across all prac- Regarding H1, we find fertilizer practices form the core of tices, but generally in the direction predicted by existing most farmers’ practice portfolios, with greater practice adop- literature (see Fig. 4; all model coefficients are presented in tion and co-occurrence rates and greater covariance in our Online Appendix Table A9). multivariate modelling. Regarding H2, we find that indi - vidual practice and portfolio adoption varies across farm 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 795 Fig. 5 Practice co-occurrence networks compared between pressur- leaf testing, ST soil testing, SA split applications, CC cover crops, ized-irrigation farms (left) and gravity-fed irrigation farms (right). IN irrigation well N testing, ET evapotranspiration-based irrigation Nodes represent management practices, colored based on farm man- scheduling, PB pressure bomb, MP moisture probe. Edge weights agement areas and scaled based on adoption rates, with larger nodes represent the frequency at which two practices are jointly adopted on representing higher adoption. Practice names are abbreviated as: LT the same parcel, for the farm operation type types, with larger farms and perennial crop farms adopting interdisciplinary research team’s agronomic expertise and more practices overall and farms with pressurized irrigation our qualitative field work, prompt the following postula- systems integrating irrigation-related practices more readily tions about why fertilizer-focused practices are more read- into their portfolios. Applying a farm systems conceptual ily adopted and why practice portfolios vary across farm and analytical approach motivates our explicit measurement types. of the interdependencies across multiple practices and facili- First, Cooperative Extension and on-farm technical tates our interpretation of how and why farmers make deci- consultants have historically emphasized fertilizer appli- sions to adopt multiple practices to fit their farm operation. cation and monitoring practices. Cooperative Extension has heavily focused on providing fertilizer application Potential mechanisms driving adoption rate recommendations for different crop types (Geisseler and interdependence patterns 2016) and has promoted the 4R’s framework as a decision- support tool for adjusting fertilizer applications. On-farm The survey and modelling results show strong evidence of technical consultants, such as Certified Crop Advisers who practice interdependence and complementarities, which can approve the N Management Plans required by farm- is consistent with multiple mechanisms that might sup- ers under the ILRP (if the farmer is not Self-Certified), port systems thinking. These results, combined with our also play a large role in disseminating practice knowledge. 1 3 796 J. Rudnick et al. These consultants provide fertilizer product and applica- larger farms, perennial crop farms, and farmers accessing tion rate recommendations, and oftentimes carry out field more information sources adopting all practices at higher and crop monitoring practices (i.e. leaf and soil testing) on rates. These findings are consistent with much past adop - behalf of the farmer. However, some of these consultants tion research that highlights the importance of economies of work for fertilizer companies as product representatives, scale, capacity to accept risk, and social and informational which can create misaligned incentives for promoting networks as key factors that help to lower the barriers to practices that reduce fertilizer use. adoption (Houser et al. 2019; Lubell and Fulton 2008; Marra Irrigation well N testing is noteworthy, as it has recently et al. 2003; Prokopy et al. 2019). become part of standard fertilizer practices. In response to Differences in adoption of irrigation management prac- continuing concerns about nitrate contamination in ground- tices specifically, were even more apparent across farms with water, the State Water Board updated the ILRP mandatory different irrigation infrastructure, with pressurized irriga- reporting requirements in 2018 to require farmers to measure tion system farms being significantly more likely to adopt nitrate levels in their irrigation wells and include it as part the irrigation-focused practices and integrate them into of their N budget. The idea to “pump and apply” has thus their overall N management portfolios. Many of the irri- only relatively recently become integrated into Coopera- gation management practices require increased investment tive Extension outreach efforts and CCA recommendations in equipment and supplies, technical learning, and time to (Nitrogen Management Training Materials 2019). This may implement. Pressurized irrigation infrastructure may be a help to explain our observation of increasing adoption and proxy indicating an additional level of capacity for capital co-occurrence rates with the traditional set of fertilizer prac- investment and operational sophistication. tices. The heavy emphasis of extension and technical con- Moreover, the ability for farms with pressurized irrigation sulting on fertilizer-related practices overall is likely to be a systems to more precisely control water delivery may serve key factor driving our observations of these practices being to make some of the irrigation-specific practices seem more featured at the core of most farmers’ practice portfolios. advantageous in these systems. For example, a farmer work- Furthermore, it illuminates opportunities for extension and ing with a groundwater-fed, drip irrigation system has more outreach to diversify practice recommendations to increase ability to adjust their irrigation timing and total water deliv- adoption of the more peripheral irrigation and soil-focused ered in response to evapotranspiration or moisture probe practices. data; thus adopting these practices improves their water use In concordance with the statistical results, our qualita- efficiency. In contrast, a farmer with a flood irrigation system tive field work suggests connections between fertilizer, soil dependent on surface water deliveries is not able to irrigate and irrigation management are still not widely understood any time they want and instead is beholden to the water dis- across technical consultant or farmer communities. Irriga- trict delivery schedule. Their reduced flexibility for irriga- tion practices specifically have only been integrated into tion may make the irrigation-specific practices seem obso- N management in relatively recent research and extension lete. These examples are used to illustrate why the irrigation efforts that have expanded to consider the full farm system. practices may be perceived to be better ‘fit’ in the portfolios Yet, in functionally differentiated farming operations (Kling of farms with pressurized irrigation systems. We do want and Mackie 2019), technical consultants and in-house farm to make clear however, the irrigation practices in and of managers may only specialize in one component of the farm themselves are not reliant on pressurized irrigation systems system (e.g. fertility management) and lack expertise to inte- and could be put to use in gravity-feed irrigation systems. grate across fertility and irrigation. Thus, holistic practice recommendations that integrate multiple areas of the farm Limitations and future work are still rather limited to university researchers and Coop- erative Extension and a small subset of innovative farmers As with much survey-based research, our results may be and technical consultants. The limited understanding of this limited by low response rates, particularly in SJDWQC connectivity between multiple areas of the farm and their and ESJWQC. Further, we acknowledge the possibility impact on N management helps to explain the limited inclu- of a response bias from farmers who are more engaged sion of irrigation practices and cover cropping in farmers N in extension, outreach and ILRP activities. However, our management practice portfolios. sample is a relatively good representation of the diver- Second, the peripheral status of irrigation management sity of types and size of agricultural operations in Cali- practices, exacerbated on farms with gravity fed irrigation, fornia’s Central Valley. Our analyses are also constrained annual crops and smaller farm size, exemplifies how the ben- to evaluating adoption only on farmers’ “most important” efits and costs of practices vary across farming contexts. parcels. This may lead to overrepresentation of perennial The impact of access to capital- including financial, techni- crop types and larger parcels in our analyses and under- cal and social capital- is apparent across all practices, with predict how these same farmers manage their smaller or 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 797 annual parcels differently. Overcoming this data limita- measure the mechanisms driving interdependencies across tion, while not dramatically increasing the burden placed multiple areas of the farm system may provide for greater on survey respondents, could be achieved by pairing our insight to understand and predict farmer decision-making survey data with reporting data collected by programs like across a variety of behaviors of interest. the ILRP. However, concerns about confidentiality often Finally, encouraging more holistic approaches to farm limit the capacity to connect data collected by researchers management challenges will require that the farm systems to data collected by other organizations, and particularly framework move beyond research in the academy and those involved in regulatory programs. become adopted as a way of practice by extension and on- We contend that measuring and studying adoption of the farm consultants, who provide critical technical assistance interdependencies between practices is essential to moving and trusted practice recommendations. This emphasizes the forward on understanding drivers and barriers to adoption, need for an ‘Extension 3.0 model’ (Lubell et al. 2014), in and designing policy and outreach that adequately account which university extension focus their efforts on ‘training for how interdependencies vary with practice-specific attrib- the trainer’ with a farm systems approach to addressing the utes (c.f. Reimer et al. 2012a). It will be helpful to char- management challenge at hand. acterize how the costs and benefits of practices vary with farm operation characteristics, and how they accrue when practices are used in tangent with one another, as part of the Conclusions same portfolio. It is also worth noting that relevant man- agement practices will vary across contexts. The practices This study contributes the most comprehensive analysis of considered as best practices for improving N management farmer behavior on N management in California to date, in California differ from the frequently studied practices for illuminating adoption trends for a suite of practices across a water quality protection in the Midwest, such as filter or wide range of farm and farmer characteristics. We analyze buffer strips, side dressing, nitrification inhibitors, or tillage adoption of eight N management practices that stretch across (Denny et al. 2019; Stuart et al. 2014). fertilizer, soil and irrigation components of the farm. We It is additionally important to analyze the relationship argue an integrated farm systems approach improves our between different portfolios of practices and agronomic understanding of farmers’ decision-making across different and environmental outcomes. It is possible that two prac- farm operation contexts and management practices, allow- tices that are frequently co-adopted on a specific type of ing us to better capture the complexity and interdependency farm are actually a mal-adaptive combination when con- of adoption decision-making on portfolios of practices. sidering the goal of improved N use efficiency. A criti- Our findings highlight how adoption patterns and adoption cal next step for agronomic research should be to develop drivers differ across practices, and how the combination or better understandings of the agronomic, economic and portfolio of practices adopted may differ quite dramatically environmental outcomes associated with different practice across heterogeneous farms, based on operational character- portfolios on different types of farms. This can serve as istics that influence fit. the basis for developing practice portfolio recommenda- Our interdisciplinary and engaged approach further tions tailored to specific farm types, where the interactions informs the outreach and policy recommendations that fol- between multiple practices and the constraints and oppor- low from our results. We have observed that the large hetero- tunities of the full farm system are taken into account. geneity in California’s farms result in neighboring operations Future research should seek to further extend the farm facing dramatically different barriers or achieving greatly systems approach by considering not only how N manage- different benefits with the adoption of an individual practice ment practices relate to one another, but also how N man- or specific portfolio of practices. The “disproportionality” of agement practices might relate to other farm management impact (Nowak et al. 2006) on N pollution means that moti- decisions, like pest management, farm investment, labor vating and focusing extension efforts to increase adoption constraints, and on-farm data collection. We must consider on large, input-intensive operations increases the likelihood how farmers organize their operations and management of achieving sustainability goals. Yet, we see that large size staff, the timing of management decisions on different and greater capital are important enabling factors increasing parts of the farm, and how management practices that opti- likelihood of adoption already, and these operations may mize for one outcome (e.g. N use efficiency) may influence be able to achieve higher private returns due to increased other agronomic, economic or ecological outcomes. The efficiency or other co-benefits. In contrast, small farms with farm systems approach is broadly applicable beyond the smaller individual N footprints, may face extreme financial, study of N management and should be incorporated into technical or information barriers to implement even the easi- the study of other conservation practice and on-farm tech- est practices. These factors must be considered in justifying nology adoption. Expanding the research lens to explicitly and tailoring financial incentive and technical assistance 1 3 798 J. Rudnick et al. many members of our Survey Advisory Committee who gave ample programs that are aimed to increase and maintain practice feedback throughout the project and continuously work to support adoption. Our results suggest that N management extension progress on nitrogen management in the field; this work would not and technical assistance programs in California should focus be possible without all of their support. We also acknowledge and on developing practice portfolio recommendations that are appreciate that this work was funded by a competitive grant from the California Department of Food and Agriculture Fertilizer Research adapted to fit small farms, annual cropping systems and and Education Program. operations that don’t currently have pressurized irrigation infrastructure in place. Funding Funding was provided by California Department of Food Furthermore, a novel opportunity exists in California and Agriculture Fertilizer Research and Education Program (Grant to coordinate policy and extension to promote a holistic No. 16-0620). approach to N management through the implementation of Open Access This article is licensed under a Creative Commons Attri- the ILRP. As our survey results and field work have illumi- bution 4.0 International License, which permits use, sharing, adapta- nated, N management policy and extension efforts have his- tion, distribution and reproduction in any medium or format, as long torically focused on fertilizer practices, resulting in greater as you give appropriate credit to the original author(s) and the source, adoption and integration of these practices, at the expense of provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are holistic portfolios that integrate soil and irrigation practices included in the article’s Creative Commons licence, unless indicated as well. Developing practice portfolio recommendations that otherwise in a credit line to the material. If material is not included in integrate irrigation, soil and fertilizer practices together, and the article’s Creative Commons licence and your intended use is not are tailored to different types of farms should be a priority permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a for N management research and extension going forward. copy of this licence, visit http://creativ ecommons .or g/licenses/b y/4.0/. The effort to develop tailored portfolio recommendations could also clarify how to measure improvement around N management, since using all, or more management practices References may not always be applicable or optimal for all operations. Evaluating whether farmers are adopting the best-suited Ajzen, I. 1989. The Theory of Planned Behavior. Organizational portfolio of practices for their specific farm context may Behavior and Human Decision Processes 50: 179–211. provide a more tangible goal for farmers, while also provid- American Association for Public Opinion Research. 2016. Response Rate Calculator V4.0. 2016. http://www.aapor .org/Educa tion- ing more clarity to the public and policymakers as to when Resou rces/For-Resea rcher s.aspx. and how reductions in N pollution may be reached. Arbuckle, J. G., and Hanna Rosman. 2014. Iowa Farmers’ Nitro- We suggest a coordinated approach engaging all of the gen Management Practices and Perspectives. http://lib.dr.iasta key actors across the wide information and extension net- te.edu/exten sion_commu nitie s_pubs/24. Aubert, B.A., S. Andreas, and J. Grimaudo. 2012. IT as Enabler work working on this issue: California Department of Food of Sustainable Farming: An Empirical Analysis of Farmers’ and Agriculture, State and Regional Water Boards, Water Adoption Decision of Precision Agriculture Technology. Deci- Quality Coalitions, USDA Natural Resources Conserva- sion Support Systems 54 (1): 510–520. https://doi.or g/10.1016/j. tion Service, Resource Conservation Districts, commodity dss.2012.07.002. Baumgart-Getz, A., L.S. Prokopy, and K. Floress. 2012. Why Farm- groups, University of California Cooperative Extension, ers Adopt Best Management Practice in the United States: A on-farm PCAs and CCAs, NGOs and private sector com- Meta-Analysis of the Adoption Literature. Journal of Environ- panies. Combined, this network of actors has the resources, mental Management 96 (1): 17–25. https ://doi.org/10.1016/j. reach, and earned trust of farmers to develop comprehen- jenvm an.2011.10.006. Bawden, R. 1995. On the Systems Dimension in Farm Systems sive strategies that seek to (1) diversify and incorporate Research. Journal of Farm Systems Resource Extension 5 (2): practices across the entire farm that holistically address N 1–18. management, (2) shift social norms away from excess ferti- Blank, S. 2001. Producers Get Squeezed up the Farming Food Chain: lizer application as ‘cheap yield insurance’, and (3) provide A Theory of Crop Portfolio Composition and Land Use. Applied Economic Perspectives and Policy 23 (2): 404–422. creative incentives through shared risk approaches, techni- Bosch, N.S., J.D. Allan, J.P. Selegean, and D. Scavia. 2013. Scenario- cal assistance in implementation and maintenance, market Testing of Agricultural Best Management Practices in Lake Erie recognition and reduced regulatory burdens. Together these Watersheds. Journal of Great Lakes Research 39 (3): 429–436. efforts can promote and sustain a farm systems approach to California Department of Food and Agriculture. 2018. California Agricultural Statistics Review 2017–2018. Sacramento, CA. N management. https://www .cdfa.ca.gov/statis tics/PDFs/2017-18A gRepor t.pdf . Cappellari, L., and S.P. Jenkins. 2003. Multivariate Probit Regres- Supplementary Information The online version contains supplemen- sion Using Simulated Maximum Likelihood. The Stata Jour- tary material available at https:// doi.org/10.1007/s10460-02 1-10190-5 . nal: Promoting Communications on Statistics and Stata 3 (3): 278–294. https ://doi.org/10.1177/15368 67x03 00300 305. Acknowledgements We would like to acknowledge all farmers who Central Valley Regional Water Quality Control Board. 2020. Irri- voluntarily participated in our survey, the Water Quality Coalitions gated Lands Regulatory Program (ILRP) Frequently Asked who facilitated our study and ensured relevancy of this work, and the 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 799 Questions. https: //www.waterb oards .ca.gov/centra lvall ey/water Salinas Valley Groundwater. Report for the California State Water _issue s/irrig ated_lands /ilrp_faq.pdf. Resources Control Board Report to the Legislature. Church, S.P., J. Lu, P. Ranjan, A.P. Reimer, and L.S. Prokopy. 2020. Hillis, V., M. Lubell, and M. Hoffman. 2018. Sustainability Partner - The Role of Systems Thinking in Cover Crop Adoption: Impli- ship and Viticulture Management in California. Journal of Envi- cations for Conservation Communication. Land Use Policy 94: ronmental Management 217: 214–225. https ://doi.org/10.1016/j. 104508.jenvm an.2018.03.033. Coates, R.W., M.J. Delwiche, and P.H. Brown. 2005. Precision Irri- Hines, J., and P. Keil. 2020. Common Competitors and Rare Friends. gation and Fertilization in Orchards. In ASAE Annual Meeting. Nature Ecology and Evolution 4 (1): 8–9. https://doi.or g/10.1038/ American Society of Agricultural and Biological Engineers. s4155 9-019-1071-3. https ://doi.org/10.13031 /2013.19815 Hoffman, M., M. Lubell, and V. Hillis. 2014. Linking Knowledge and Conley, T., and C. Udry. 2010. Learning about a New Technol- Action Through Mental Models of Sustainable Agriculture. Pro- ogy: Pineapple in Ghana. American Economic Review 100 (1): ceedings of the National Academies of Science 111: 13016–13021. 35–69. Houser, M., S.T. Marquart-Pyatt, R.C.H. Denny, A. Reimer, and D. Costello, A.B., and J.W. Osborne. 2005. Best Practices in Explora- Stuart. 2019. Farmers, Information, and Nutrient Management in tory Factor Analysis: Four Recommendations for Getting the the US Midwest. Journal of Soil and Water Conservation 74 (3): Most from Your Analysis. Practical Assessment, Research and 269–280. https ://doi.org/10.2489/jswc.74.3.269. Evaluation 10: 7. Howitt, R., D. MacEwan, and J. Medellin-Azuara. 2008. Calculat- Denny, R.C.H., S.T. Marquart-Pyatt, and M. Houser. 2019. Under- ing California Cropping Patterns in 2050. Davis. https ://water standing the Past and Present and Predicting the Future: Farm-.ca.gov/Legac yFile s/water plan/docs/cwpu2 009/0310 final/v4c03 ers’ Use of Multiple Nutrient Best Management Practices in a02_cwp20 09.pdf. the Upper Midwest. Society and Natural Resources 32 (7): Jara-Rojas, R., B.E. Bravo-Ureta, A. Engler, and J. Diaz. 2013. An 807–826. https ://doi.org/10.1080/08941 920.2019.15740 45. Analysis of the Joint Adoption of Water Conservation and Soil DeVellis, R.F. 2003. Scale Development: Theory and Applications, Conservation in Central Chile. Land Use Policy 32: 292–301. 2nd ed. Thousand Oaks: Sage Publications. Kanter, D.R., W. Winiwarter, B.L. Bodirsky, L. Bouwman, E. Boyer, S. Dillman, D.A., G. Phelps, R. Tortora, K. Swift, J. Kohrell, J. Berck, Buckle, J.E. Compton, T. Dalgaard, W. de Vries, D. Leclère, and and B.L. Messer. 2008. Response Rate and Measurement Differ - A. Leip. 2020. A Framework for Nitrogen Futures in the Shared ences in Mixed Mode Surveys Using Mail, Telephone, Interac- Socioeconomic Pathways. Global Environmental Change 61: tive Voice Response and the Internet. Social Science Research 102029. https ://doi.org/10.1016/j.gloen vcha.2019.10202 9. 38 (1): 1–18. Kara, E., M. Ribaudo, and R.C. Johansson. 2008. On How Environ- Feder, G. 1982. Adoption of Interrelated Agricultural Innovations: mental Stringency Influences Adoption of Best Management Prac- Complementarity and the Impacts of Risk, Scale, and Credit. tices in Agriculture. Journal of Environmental Management 88 American Journal of Agricultural Economics 64 (1): 94–101. (4): 1530–1537. https ://doi.org/10.1016/j.jenvm an.2007.07.029. https ://doi.org/10.2307/12411 77. Kassie, M., H. Teklewold, M. Jaleta, P. Marenya, and O. Erenstein. Feder, G., and D. Umali. 1993. The Adoption of Agricultural Innova- 2015. Understanding the Adoption of a Portfolio of Sustainable tions: A Review. Technological Forecasting and Social Change Intensification Practices in Eastern and Southern Africa. Land 43 (3–4): 215–239. Use Policy 42: 400–411. https ://doi.org/10.1016/j.landu sepol Fernández, V., and P.H. Brown. 2013. From Plant Surface to Plant .2014.08.016. Metabolism: The Uncertain Fate of Foliar-Applied Nutrients. Khalsa, S.D.S., and P.H. Brown. 2017. Grower Analysis of Organic Frontiers in Plant Science 4: 289. Matter Amendments in California Orchards. Journal of Envi- Fresco, L.O., and E. Westphal. 1988. A hierarchical Classification of ronmental Quality 46 (3): 649–658. https ://doi.or g/10.2134/ Farm Systems. Experimental Agriculture 24 (4): 399–419.jeq20 16.11.0456. Geisseler, D. 2016. Nitrogen Concentrations in Harvested Plant Parts— Khalsa, S.D.S., and P.H. Brown. 2019. Understanding Nitrogen A literature overview. University of California Davis. http://groun Cycling in an Irrigated Deciduous Permanent Crop. Acta Hor- dwate rnitr ate.ucdav is.edu/files /26913 8.pdf. ticultare 1253: 207–212. Ghadim, A.K., D.J. Pannell, and M.P. Burton. 2005. Risk, Uncer- Khalsa, S.D.S., T. Hartz, and P.H. Brown. 2018. Principles of Nitro- tainty, and Learning in Adoption of a Crop Innovation. gen Cycling and Management. University of California Agri- Agricultural Economics 33 (1): 1–9. https ://doi.or g/10.111 culture and Natural Resources. https://ucanr .edu/sites /S trategicI 1/j.1574-0862.2005.00433 .x.nitia tives /files /30116 8.pdf. Giller, K. 2013. Can We Define the Term ‘Farming Systems’? A Ques- Kipling, R.P., H.E. Taft, D.R. Chadwick, D. Styles, and J. Moorby. tion of Scale. SAGE Journals: Outlook on Agriculture 42 (3): 2019. Challenges to Implementing Greenhouse Gas Mitigation 149–153. https ://doi.org/10.5367/oa.2013.0139. Measures in Livestock Agriculture: A Conceptual Framework Good, A.G., and P.H. Beatty. 2011. Fertilizing Nature: A Tragedy of for Policymakers. Environmental Science and Policy 92: 107– Excess in the Commons. PLoS Biology 9 (8): 1–9. https ://doi. 115. https ://doi.org/10.1016/j.envsc i.2018.11.013. org/10.1371/journ al.pbio.10011 24. Kling, C., and C. Mackie. 2019. Improving Data Collection and Halbrendt, J., S.A. Gray, S. Crow, T. Radovich, A. Kimura, and B.B. Measurement of Complex Farms. Washington: National Acad- Tamang. 2014. Differences in Farmer and Expert Beliefs and the emies of Sciences, Engineering, and Medicine. National Acad- Perceived Impacts of Conservation Agriculture. Global Environ- emies Press. mental Change 28: 50–62. Knowler, D., and B. Bradshaw. 2007. Farmers’ Adoption of Conserva- Hanson, B., D. May, J. Simnek, J. Hopmans, and R. Hutmacher. 2009. tion Agriculture: A Review and Synthesis of Recent Research. Drip Irrigation Provides the Salinity Control Needed for Profit- Food Policy 32 (1): 25–48. https ://doi.or g/10.1016/j.foodp able Irrigation of Tomatoes in the San Joaquin Valley. California ol.2006.01.003. Agriculture 63 (3): 131–136. Koppmair, S., M. Kassie, and M. Qaim. 2017. The Influence of Farm Harter, T., J.R. Lund, J. Darby, G. Fogg, R. Howitt, K. Jessoe, S. Pet- Input Subsidies on the Adoption of Natural Resource Management tygrove, J. Quinn, and J. Viers. 2012. Addressing Nitrate in Cali- Technologies. Australian Journal of Agricultural and Resource fornia’s Drinking Water: With a Focus on Tulare Lake Basin and Economics 61 (4): 539–556. https ://doi.or g/10.1111/1467- 8489.12220. 1 3 800 J. Rudnick et al. Letey, J., and P. Vaughan. 2013. Soil Type, Crop and Irrigation Tech- Prokopy, L.S., K. Floress, D. Klotthor-Weinkauf, and A. Baumgart- nique Affect Nitrogen Leaching to Groundwater. California Agri- Getz. 2008. Determinants of Agricultural Best Management culture 67 (4): 231–241. Practice Adoption: Evidence from the Literature. Journal of Soil Levy, M.A., M.N. Lubell, and N. McRoberts. 2018. The Structure of and Water Conservation 63 (5): 300–311. https://doi.or g/10.2489/ Mental Models of Sustainable Agriculture. Nature Sustainability jswc.63.5.300. 1 (8): 413–420. Prokopy, L.S., K. Floress, J.G. Arbuckle, S.P. Church, F.R. Eanes, Y. Liu, T., R.J.F. Bruins, and M.T. Heberling. 2018. Factors Influencing Gao, B.M. Gramig, P. Ranjan, and A.S. Singh. 2019. Adoption Farmers’ Adoption of Best Management Practices: A Review and of Agricultural Conservation Practices in the United States: Evi- Synthesis. Sustainability (Switzerland) 10 (2): 1–26. https ://doi. dence from 35 Years of Quantitative Literature. Journal of Soil org/10.3390/su100 20432 . and Water Conservation 74 (5): 520–534. https://doi.or g/10.2489/ Lopus, S.E., M.P. Santibáñez, R.H. Beede, R.A. Duncan, J. Edstrom, jswc.74.5.520. F.J.A. Niederholzer, C.J. Trexler, and P.H. Brown. 2010. Survey Ranjan, P., S.P. Church, K. Floress, and L.S. Prokopy. 2019. Synthe- Examines the Adoption of Perceived Best Management Practices sizing Conservation Motivations and Barriers: What Have We for Almond Nutrition. California Agriculture 64 (3): 149–154. Learned from Qualitative Studies of Farmers’ Behaviors in the https ://doi.org/10.3733/ca.v064n 03p14 9. United States? Society and Natural Resources 32 (11): 1171– Lubell, M., and A. Fulton. 2008. Local Diffusion Networks Act as 1199. https ://doi.org/10.1080/08941 920.2019.16487 10. Pathways to Sustainable Agriculture in the Sacramento River Val- Reimer, A.P., D.K. Weinkauf, and L.S. Prokopy. 2012a. The Influence ley. California Agriculture 61: 131–137. https ://doi.org/10.3733/ of Perceptions of Practice Characteristics: An Examination of ca.v061n 03p13 1. Agricultural Best Management Practice Adoption in Two Indiana Lubell, M., M. Niles, and M. Hoffman. 2014. Extension 3.0: Man- Watersheds. Journal of Rural Studies 28 (1): 118–128. aging Agricultural Knowledge Systems in the Network Age. Reimer, A., A.W. Thompson, and L.S. Prokopy. 2012b. The Multi- Society and Natural Resources 27 (10): 1089–1103. https ://doi. Dimensional Nature of Environmental Attitudes among Farmers org/10.1080/08941 920.2014.93349 6. in Indiana: Implications for Conservation Adoption. Agriculture Marra, M., D.J. Pannell, and A.A. Ghadim. 2003. The Economics of and Human Values 29 (1): 29–40. https ://doi.org/10.1007/s1046 Risk, Uncertainty and Learning in the Adoption of New Agri- 0-011-9308-z. cultural Technologies: Where Are We on the Learning Curve? Reimer, A., J.E. Doll, B. Basso, S.T. Marquart-Pyatt, G.P. Robertson, Agricultural Systems 75 (2–3): 215–234. https://doi.or g/10.1016/ D. Stuart, and J. Zhao. 2017. Moving Toward Sustainable Farming S0308 -521X(02)00066 -5. Systems: Insights from Private and Public Sector Dialogues on McAllister, R.R.J., D. Smith, C.J. Stokes, and F.J. Walsh. 2009. Pat- Nitrogen Management. Journal of Soil and Water Conservation terns of Accessing Variable Resources Across Time and Space: 72 (1): 5A–9A. https ://doi.org/10.2489/jswc.72.1.5A. Desert Plants, Animals and People. Journal of Arid Environments Reimer, A., R. Denny, and D. Stuart. 2018. The Impact of Federal 73 (3): 338–346. https ://doi.org/10.1016/j.jarid env.2008.10.007. and State Conservation Programs on Farmer Nitrogen Manage- McIver, J., and E.G. Carmines. 1981. Unidimensional Scaling, vol. 24. ment. Environmental Management 62 (4): 694–708. https ://doi. Newbury Park, CA: Sage University Paper Series on Quantiative org/10.1007/s0026 7-018-1083-9. Applications in Social Science. Rogers, E.M. 2003. Diffusion of Innovations, 5th ed. New York: Simon McLellan, E.L., K.G. Cassman, A.J. Eagle, P.B. Woodbury, S. Sela, C. and Schuster. Tonitto, R.D. Marjerison, and H.M. Van Es. 2018. The Nitrogen Santos, J.R.A. 1999. Cronbach’s Alpha: A Tool for Assessing the Reli- Balancing Act: Tracking the Environmental Performance of Food ability of Scales. Journal of Extension 37 (2): 1–5. Production. BioScience 68 (3): 194–203. https://doi.or g/10.1093/ Schellenberg, D.L., R.D. Morse, and G.E. Welbaum. 2009. Organic biosc i/bix16 4. Broccoli Production on Transition Soils: Comparing Cover Crops, Mikkelsen, R. 2011. The ‘4R’ Nutrient Stewardship Framework for Tillage and Sidedress. Renewable Agriculture and Food Systems Horticulture. HortTechnology 21 (6): 658–662. 24 (2): 85–91. Murendo, C., S. Gwara, N. Mpofu, T. Pedzisa, and P. Chivenge. 2016. Shaffer, S. 2013. Encouraging California Specialty Crop Growers to The Adoption of a Portfolio of Sustainable Agricultural Practices Adopt Environmentally Beneficial Management Practices for Effi- by Smallholder Farmers in Zimbabwe. In 5th International Con- cient Irrigation and Nutrient Management. ference of the African Association of Agricultural Economists. Shaner, W., P. Philipp, and W.R. Schmehl. 1982. Farming Systems Addis Ababa, Ethiopia. Research and Development. Boulder, CO: Westview Press. Nitrogen Management Training Materials. 2019. California Institute Shaner, W., P. Philipp, and W.R. Schmehl. 2019. Farming Systems for Water Resources, University of California Division of Agri- Research and Development: Guidelines for Developing Countries. culture and Natural Resources. http://ciwr.ucanr.edu/Nitr og enMa New York: Routledge. ISBN: 12:978-0-367-01930-3. nagem ent/. Snapp, S., and B. Pound. 2008. Agricultural Systems: Agroecology and Norman, D.W. 1980. The Farming Systems Approach: Relevancy Rural Innovation for Development. Burlington, MA: Academic for the Small Farmer. East Lansing, Michigan: Michigan State Press. University. Snyder, C.S. 2017. Enhanced Nitrogen Fertiliser Technologies Sup- Nowak, P., S. Bowen, and P.E. Cabot. 2006. Disproportionality as a port the ‘4R’ Concept to Optimise Crop Production and Minimise Framework for Linking Social and Biophysical Systems. Society Environmental Losses. Soil Research 55 (6): 463–472. and Natural Resources 19 (2): 153–173. Stuart, D., R.L. Schewe, and M. McDermott. 2014. Reducing Nitrogen Osmond, D.L., D.L.K. Hoag, A.E. Luloff, D.W. Meals, and K. Neas. Fertilizer Application as a Climate Change Mitigation Strategy: 2015. Farmers’ Use of Nutrient Management: Lessons from Understanding Farmer Decision-Making and Potential Barriers Watershed Case Studies. Journal of Environment Quality 44 (2): to Change in the US. Land Use Policy. https ://doi.org/10.1016/j. 382. https ://doi.org/10.2134/jeq20 14.02.0091.landu sepol .2013.08.011. Pannell, D.J., G.R. Marshall, N. Barr, and A. Curtis. 2006. Adoption Taylor, R., and D. Zilberman. 2017. Diffusion of Drip Irrigation: The of Conservation Practices by Rural Landholders. Australian Jour- Case of California. Applied Economic Perspectives and Policy 39 nal of Experimental Agriculture, 2001: 1407–1424. https ://doi. (1): 16–40. https ://doi.org/10.1093/aepp/ppw02 6. org/10.1071/EA050 37. Teklewold, H., M. Kassie, and B. Shiferaw. 2013. Adoption of Mul- tiple Sustainable Agricultural Practices in Rural Ethiopia. 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 801 Journal of Agricultural Economics 64 (3): 597–623. https ://doi. Publisher’s Note Springer Nature remains neutral with regard to org/10.1111/1477-9552.12011. jurisdictional claims in published maps and institutional affiliations. Teshager, A.D., P.W. Gassman, S. Secchi, and J.T. Schoof. 2017. Sim- ulation of Targeted Pollutant-Mitigation-Strategies to Reduce Nitrate and Sediment Hotspots in Agricultural Watershed. Sci- ence of the Total Environment 607: 1188–1200. Jessica Rudnick is an Extension Specialist with California Sea Grant, Tomich, T., S. Bordt, R. Dahlgren, and K. Scow. 2016. The Califor- focused on human decision-making and governance in the Sacramento- nia Nitrogen Assessment: Challenges and Solutions for People, San Joaquin Delta. Dr. Rudnick completed her PhD in 2020 at UC Agriculture and the Environment. Oakland, CA: University of Davis in the Center for Environmental Policy and Behavior where California Press. her research focused on farmer decision-making and agricultural Tucker, M., and T.L. Napier. 2002. Preferred Sources and Channels of governance. Soil and Water Conservation Information among Farmers in Three Midwestern US Watersheds. Agriculture, Ecosystems & Environ- Mark Lubell is a Professor of Environmental Science and Policy at UC ment 92 (2–3): 297–313. Davis. He is the director of the Center for Environmental Policy and Ulrich-Schad, J.D., S.G. de Jalon, N. Babin, A. Pape, and L.S. Prokopy. Behavior, where his research focuses on cooperation and decision- 2017. Measuring and Understanding Agricultural Producers’ making in the context of agricultural and environmental policy. Adoption of Nutrient Best Management Practices. Journal of Soil and Water Conservation 72(5): 506–518. Sat Darshan S. Khalsa is an Assistant Project Scientist in the Depart- United States Department of Agriculture. 2018. “USDA Organic ment of Plant Sciences at UC Davis. His research focuses on nitrogen INTEGRITY Database.” https://or ganic.ams.usda.go v/integr ity/. cycling and management in tree crops integrating plant, soil and social United States Environmental Protection Agency. 2017. Nonpoint sciences. Dr. Khalsa has authored various papers on agricultural nitro- Source: Agriculture. https ://www .epa.gov/n ps/non po int-sour c gen use. e-agric ultur e. USDA Midwest Climate Hub. 2017. Agriculture in the Midwest. U.S. Stephanie Tatge completed her M.S. in International Agricultural Department of Agriculture. 2017. https://www .climatehub s.usda. Development at UC Davis in 2018, where her thesis leveraged spatial gov/hubs/midwe st/topic /agric ultur e-midwe st. analysis to investigate social networks and diffusion of information van Rooyen, J. 1984. Agricultural Economic Research in Less Devel- among farmers. She is now a Research and Outreach Specialist at the oped Countries: A Farm Systems Approach. Development South- Freshwater Trust in Sacramento, CA. ern Africa 1 (1): 56–64. https ://doi.or g/10.1080/03768 35840 84390 69. Liza Wood is a PhD student in Ecology at UC Davis, with a Designated Wauters, E., and E. Mathijs. 2014. The Adoption of Farm Level Soil Emphasis in Computational Social Science. Her dissertation focuses Conservation Practices in Developed Countries: A Meta-Analytic on governance of seed systems for impacts on biodiversity and social- Review. International Journal of Agricultural Resources, Govern- ecological resilience, in the context of climate change. Liza also works ance and Ecology 10 (1): 78–102. https://doi.or g/10.1504/IJARG on agricultural policy and farmer behavior in California, and the use of E.2014.06105 8. science in environmental policy. Wilson, R.S., G. Howard, and E.A. Burnett. 2014. Improving Nutrient Management Practices in Agriculture: The Role of Risk-Based Molly Sears is a PhD Candidate in the Department of Agricultural and Beliefs in Understanding Farmers’ Attitudes toward Taking Addi- Resource Economics at the University of California, Berkeley. Her tional Action. Water Resources Research 50: 6735–6746. https:// research is largely centered at the intersection of agriculture, envi- doi.org/10.1002/2013W R0152 00.Recei ved. ronmental quality, and policy. Molly’s recent papers include work on Zheng, C., Y. Jiang, C. Chen, Y. Sun, J. Feng, A. Deng, Z. Song, and W. recycling policy, groundwater salinity, and nitrogen management. Zhang. 2014. The Impacts of Conservation Agriculture on Crop Yield in China Depend on Specific Practices, Crops and Cropping Patrick H. Brown is a professor of Plant Sciences in the Department Regions. Crop Journal 2 (5): 289–296. https ://doi.org/10.1016/j. of Plant Sciences at UC Davis. His research focuses on the function cj.2014.06.006. and transport of nutrients in plants and the management of nutrients Zulfiqar, F., R. Ullah, M. Abid, and A. Hussain. 2016. Cotton Produc- in agricultural ecosystems. Dr. Brown has contributed significantly tion under Risk: A Simultaneous Adoption of Risk Coping Tools. to advancing research in his field and is dedicated to communicating Natural Hazards 84 (2): 959–974. https ://doi.org/10.1007/s1106 agronomic research through a robust and applied extension program. 9-016-2468-9. 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Agriculture and Human Values Springer Journals

A farm systems approach to the adoption of sustainable nitrogen management practices in California

Loading next page...
 
/lp/springer-journals/a-farm-systems-approach-to-the-adoption-of-sustainable-nitrogen-j2S2av80CK

References (100)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
ISSN
0889-048X
eISSN
1572-8366
DOI
10.1007/s10460-021-10190-5
Publisher site
See Article on Publisher Site

Abstract

Improving nitrogen (N) fertilizer management in agricultural systems is critical to meeting environmental goals while maintaining economically viable and productive food systems. This paper applies a farm systems framework to analyze how adoption of N management practices is related to different farming operation characteristics and the extent to which ferti- lizer, soil and irrigation practices are related to each other. We develop a multivariate probit regression model to analyze the interdependency of these adoption behaviors from 966 farmers across three watersheds and diverse cropping systems in the Central Valley of California. Our analysis demonstrates that farmers adopt varying combinations or portfolios of practices, with the most common portfolio featuring a core set of fertilizer-focused practices. Irrigation infrastructure is an especially important farm operation characteristic for encouraging adoption of innovative practice portfolios that integrate water and fertilizer management. These findings highlight the ability for a farm systems approach to improve our understanding of farmer decision-making across diverse agricultural landscapes. Keywords Agricultural decision-making · Nitrogen management · Farmer adoption · Farm systems · Multivariate probit regression Abbreviations NUE Nitrogen use efficiency BMP Best management practice QAP Quadratic assignment procedure CGSP Colusa Glenn Subwatershed Program SJDWQC San Joaquin County and Delta Water Quality ESJWQC East San Joaquin Water Quality Coalition Coalition ET Evapotranspiration ILRP Irrigated Lands Regulatory Program MVP Multivariate probit regression N Nitrogen * Jessica Rudnick California Sea Grant, University of California San Diego, jrudnick@ucsd.edu 9500 Gilman Dr., #0232, La Jolla, CA 92093-0232, USA Mark Lubell Department of Environmental Science and Policy, University mnlubell@ucdavis.edu of California Davis, One Shields Drive, Davis, CA 995616, USA Sat Darshan S. Khalsa sdskhalsa@ucdavis.edu Department of Plant Sciences, University of California Davis, One Shields Drive, Davis, CA 995616, USA Stephanie Tatge statge@ucdavis.edu The Freshwater Trust, 1717 I St., Suite A, Sacramento, CA 95811, USA Liza Wood belwood@ucdavis.edu Department of Agricultural and Resource Economics, University of California Berkeley, 100 Academic Hall, Molly Sears Berkeley, CA 94720, USA molly_vandop@berkeley.edu Patrick H. Brown phbrown@ucdavis.edu Vol.:(0123456789) 1 3 784 J. Rudnick et al. To address these problems, we apply a ‘farm systems’ Introduction framework that envisions farming operations as complex systems in which multiple parts or subsystems of the farm Nitrogen (N) fertilizers are a dominant input in industri- are interrelated and interdependent, such that change in alized agricultural systems, significantly enhancing crop one part of the system is likely to induce change in another growth and yields, while also generating one of the most (Giller 2013; Shaner et al. 2019). The concept of farm sys- challenging sources of current environmental pollution tems has emerged over time in agricultural development (Good and Beatty 2011; Osmond et al. 2015; Kanter et al. and agroecology literatures, usually with an emphasis on 2020). Excess fertilizer not taken up by the crop is suscep- the importance of interdisciplinary research and recogni- tible to leaching, runoff and volatilization, leading to con- tion of the multiple biophysical, ecological, social and eco- tamination of drinking water resources, ecosystem dam- nomic factors at play in agricultural systems (Norman 1980; ages and release of nitrous oxide, a potent greenhouse gas van Rooyen 1984; Bawden 1995; Snapp and Pound 2008; (U.S. EPA 2017; Harter et al. 2012; Tomich et al. 2016). Giller 2013; Shaner et al. 2019). However, this integrated While agricultural research and extension has devoted sig- approach remains rather nascent in applied farmer behavior nificant attention to developing farm management strate- and adoption research (Church et al. 2020). This paper aims gies for improving N use efficiency (NUE), understanding to reconcile this gap by applying a farm systems framework the factors influencing farmers’ adoption of these prac- to evaluate farmers’ adoption of a suite of N management tices remains an active area of agricultural social science practices, paying careful attention to both the interdepend- research (Reimer et al. 2017). Moreover, a growing body ency between individual management practices and how of ecological modelling research suggests that it will be practice portfolios vary across farm types. necessary in most circumstances to simultaneously imple- This study is grounded in the empirical context of the ment multiple best management practices, in order to see Central Valley of California, where N management has the desired improvements in ecological and social out- become a key focus of the state given extensive nitrate pol- comes (Bosch et al. 2013; Teshager et al. 2017; McLellan lution in groundwater resources and associated threats to et al. 2018). Yet, every practice will not necessarily be effi- drinking water (Harter et al. 2012). The diversity in agro- cient or effective under the ecological or operational con- nomic, economic and ecological factors across the region ditions of every farm, thus farmers must ultimately be able allow us to test two core hypotheses. First, farmers adopt to determine the practices that best fit their unique context portfolios of practices that reflect interdependencies, with and tailor a portfolio of practices across their farm that practices in the same farm management area (e.g. ferti- work together synergistically to improve N management. lizer, soil, irrigation) more likely to be co-adopted. Second, This paper develops a conceptual and analytical frame- the benefits and costs of practices, and thus the portfolio work that encompasses two problems related to the com- adopted, vary across heterogeneous agronomic, economic plexity of N management and vexed agricultural policies and ecological conditions that shape different farm opera- that incentivize or mandate practice adoption. First, man- tions. While these hypotheses do not exhaust the possible agement practices across a farm are interdependent and implications of a farm systems framework, they are impor- therefore it is important to analyze the mix or portfolio tant initial ideas applying the idea of farm systems to farmer of practices farmers use to meet their management goals. adoption on N management. Thus, research must move beyond analyzing only a single The remainder of this article is organized as follows: we practice at a time or counting multiple practices in ways situate our farm systems study on N management within that do not account for interdependencies. the context of broader adoption literature and develop our Second, not every practice or portfolio of practices will two core hypotheses related to N management in Califor- be effective under the ecological or operational conditions nia. We then provide more details on our study context and of every farm. Thus policies that encourage widespread research design, which relies on survey data from 966 farm- adoption of a specific practice or as many practices as ers in three watersheds of California’s Central Valley. We possible, do not adequately account for different farming describe a statistical analysis method, called multivariate contexts. The heterogeneity of practice benefits and costs probit regression, which allows us to estimate the probability across different agro-ecological contexts contributes to a farmer adopts different individual practices, accounting the inconsistent empirical results in terms of what farm for interdependence among those practices. This statistical operation and operator factors predict practice adoption framework allows us to better capture these interdepend- (Knowler and Bradshaw 2007; Prokopy et al. 2008, 2019; ences and go beyond analyses that look at practices in isola- Baumgart-Getz et al. 2012; Wauters and Mathijs 2014; tion or as a simple sum or index. The results focus on which Ulrich-Schad et al. 2017; Liu et al. 2018). portfolios of practices are likely to be co-adopted, and how farm operation characteristics, especially related to irrigation 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 785 systems, impact the portfolio adoption decisions of differ - the balance between crop N demand and N losses (Denny ent farms. The conclusion discusses theoretical and policy et al. 2019; Snyder 2017). Thus, effective N management implications of applying the farm systems perspective. will require a portfolio of management practices that influ- ence different management areas of the farm, tailored to the combination of ecological, agronomic and economic dimen- Theory and hypotheses: linking farm sions of a specific operation, all of which can influence how systems to nitrogen management practices a management practice works. We draw on this literature to analyze two hypotheses: The concept of farm systems research has played an impor- adoption decisions on multiple practices are interdependent tant role across agricultural development and agroecology with one another (H1) and adoption of portfolios of practices fields (Fresco and Westphal 1988; Snapp and Pound 2008; vary across farming operation characteristics (H2). These Giller 2013; Shaner et al. 2019), with the farm system being hypotheses are tested in the empirical context of California, defined as “a complex interaction of soils, water sources, which provides a diverse agricultural landscape with produc- crops, livestock, labour, and other resources and character- tion of both annual and perennial commodities, farms rang- istics within an environmental setting which the farm family ing in scale and structure from small acreage, family-owned manages in accordance with its preferences, capabilities and to very large acreage, corporate operations, and a complex available technologies” (Shaner et al. 1982). Farm systems hydrologic system where farmers irrigate using both surface approaches emphasize interdisciplinarity to understand and groundwater and a variety of irrigation technologies. interdependence between different parts of a farm and dif- We evaluate farmer’s adoption decisions of eight different N ferent characteristics of the farmer, and integrate farmers management practices across three farm management areas: throughout the research process to translate results to exten- fertilizer, soil, and irrigation (See Table 1 for descriptions sion, outreach, policy and management goals (van Rooyen of practices). 1984). While some recent work on sustainable agriculture has Hypothesis 1: interdependence of practice adoption explored integrated mental models and systems thinking (Levy et  al. 2018; Halbrendt et  al. 2014; Hoffman et  al. Interdependence across system components is one of the 2014), the farms system perspective has not been widely most fundamental premises of any systems approach. Within embraced in the applied social science research on farmer a farm system, interdependence is present across different decision-making and conservation practice adoption aspects of the farm and the management practices employed. (Church et al. 2020). Instead, much of the research on best In the context of N management specifically, N availability management practice (BMP) adoption emphasizes theories and N losses depend on applications of N fertilizer to crops, of decision making on a single action, like Diffusion of Inno- irrigation management which may introduce additional N vation (Rogers 2003) or Theory of Planned Behavior (Ajzen from nitrate-contaminated groundwater or push N through 1989). Often these studies set the dependent variable as the the soil profile, and soil management which may introduce adoption of one practice or an index composed of multiple organic N sources (e.g. compost, cover crops) or influence practices, and independent variables are some mix of farmer soil properties that relate to nutrient retention (Khalsa et al. demographics, attitudes, perceptions and farm operation 2018). Furthermore, ecosystem modelling suggests portfo- characteristics (Prokopy et al. 2019; Ranjan et al. 2019). We lios of practices will be necessary in order to reduce N losses aim to apply the farm systems framework to understand how enough to have a positive impact on water quality (Bosch farmers consider relationships between practices and how et al. 2013; Teshager et al. 2017; Denny et al. 2019). the specific characteristics of a practice influence its adopt- Portfolio approaches have been applied to measure the ability on different types of farms (Reimer et al. 2012a). relationships between adoption decisions to estimate which N management is a clear case demanding the application practices are complementary (i.e. work in congruence with of a farm systems approach. A substantial body of agro- one another to enhance efficacy), conflictual (i.e. result in nomic research has developed tools to approximate crop N worse outcomes when used together) or substitutable (i.e. demand and recommend best management practices that redundant; one can replace the use of the other) with oth- reduce N losses (Snyder 2017; Khalsa and Brown 2019). ers in contributing to a management goal (McAllister et al. For example, the “4Rs” conceptual framework (“Right 2009; Teklewold et al. 2013; Kassie et al. 2015; Zulfiqar rate”, “Right time”, “Right place”, and “Right source”) et  al. 2016; Murendo et al. 2016; Koppmair et al. 2017). offers farmers and farm advisors a decision support tool From a decision-making perspective, co-adoption of multi- to adjust fertilizer applications (Mikkelsen 2011). More ple practices may be linked through multiple mechanisms, recently, research has highlighted how other aspects of the such as information sources that jointly recommend related farm, especially soil and irrigation management, influence practices (Tucker and Napier 2002), farmer experimentation 1 3 786 J. Rudnick et al. Table 1 Descriptions of N management practices evaluated in study, including the management area of the farm with which the practice is typi- cally associated Practice name Description Farm management area Leaf testing Test crop leaf for crop nutrient status to determine if plant is up-taking enough nutrients Fertilizer Split application Divide fertilizer applications into smaller doses and apply in different applications at needed Fertilizer times in season Soil testing Test soil for residual nitrate at beginning of season and adjust fertilizer application rate as Fertilizer appropriate Irrigation well N testing Test irrigation water in wells for nitrate content and adjust fertilizer application rates as needed Fertilizer Cover crops Plant cover crops to help hold moisture and nutrients in the soil; provides an organic source of Soil nitrogen that breaks down more slowly over time Moisture probe Test soil water content to determine depth of soil saturation and more precisely control irriga- Irrigation tion to give crop just enough water, which still retaining fertilizer in root zone Pressure bomb Determine plant-water stress and adjust irrigation scheduling as appropriate, including when Irrigation fertilizer is applied so that fertilizer stays in root zone ET-based irrigation scheduling Use evapotranspiration (ET) data to determine plant water losses, and calculate how much Irrigation water needs to be replaced with irrigation. Appropriately place fertilizer in the irrigation set so that fertilizer stays in root zone with multiple practices (Pannell et al. 2006), the adoption of potential disconnect between fertilizer application and water one practice lowering the perceived risk or cost of another management is exacerbated further by modern, large scale practice (Feder 1982), technical knowledge ‘spillovers’ that farm operations where often there are specialized teams of contribute to a better understanding of an additional prac- employees and consultants that oversee different areas of the tice (Conley and Udry 2010), co-dependence or use of the operation, with little interaction (Kling and Mackie 2019). same farm infrastructure (Hanson et al. 2009), or increased With these considerations, we hypothesize (H1) that fer- sophistication that drives the farmer to use information from tilizer management practices will be at the core of practice one practice to amend their use of another practice (Aubert adoption portfolios (i.e. higher adoption and co-adoption et al. 2012). with each other), with irrigation management practices at While our data does not allow us to test these mechanisms the periphery (i.e. lower adoption and co-adoption rates). explicitly, we draw on our interdisciplinary understand- ing of N management in California to hypothesize which Hypothesis 2: practice portfolios fit farming practices may be more frequently co-adopted. The relevant operation characteristics N management practices for which we measure adoption (see Table  1) were identified by University of California The complexity of the ecological and agronomic processes Cooperative Extension farm advisors and experts in nutri- that shape N management make it such that there is no pana- ent management. Direct N fertilizer application and moni- cea, or ‘one-size-fits-all’ approach to the selection of appro- toring practices have traditionally been a strong emphasis priate management practices. The portfolio of management of extension, including the 4R’s principles and monitoring practices responds to site-specific farm characteristics that nutrient availability in the plant-soil system. As a result, we influence operational compatibility and economic feasibil - predict farmers have adopted these practices at higher rates ity. This site-specificity is not unique to N management; for and have better knowledge of their interdependencies. In example, a study in different regions of China found that contrast, attention toward irrigation and soil management conservation management practices broadly have differ - practices has been more recent, as improved understandings ent impacts on crop yields and conservation goals under and technologies to monitor nutrient movement through- different climate conditions and cropping systems (Zheng out the agro-ecosystem have been developed (Coates et al. et  al. 2014). The results suggest that farmers must adapt 2005; Khalsa and Brown 2017; Schellenberg et al. 2009; their selection of management practices, and subsequently Fernández and Brown 2013). The extent to which irriga- practice portfolios, to ‘fit’ their specific operations to achieve tion practices influence N losses also depends heavily on the intended results. farm’s water source, soil type and irrigation infrastructure Unlike Midwest agricultural landscapes dominated by in place, complicating the ability for farm advisers to make staple field crops (USDA Midwest Climate Hub 2017), general practice recommendations around irrigation. This California features a diverse agricultural landscape which 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 787 provides an excellent research system for analyzing how Research design portfolios of practices may vary in their fit with different types of farms (Lopus et al. 2010; Shaffer 2013). Farm size Research context: nitrogen management is one of the most consistent predictors of practice adoption in California’s Central Valley (Prokopy et al. 2019), based on the argument that larger farms have more financial capital and economics of scale, California ranks as the most economically valuable agricul- which reduces barriers to practice adoption associated with tural state in the United States by annual crop cash sales. cost, time to return on investment, and risk (Feder and The state boasts more than 400 commodity crops grown Umali 1993; Ghadim et al. 2005; Kipling et al. 2019). Cali- across 77,000 farms and ranches on 25 million acres of land, fornia farm operations also feature crops across the “crop spread along a 500 mile longitudinal gradient (California hierarchy”, from low-value annuals to high-value perenni- Department of Food and Agriculture 2018). The Mediter- als (Blank 2001). High-value perennial crop systems like ranean climate is ideal for perennial and annual crops in fruit and nut orchards, are increasingly replacing low-value most areas of the state, yet creates reliance on irrigation annuals in California (Howitt et al. 2008), and receive a and a highly engineered water system. Top commodities higher return on investment for implementing practices that include dairy, grapes, almonds, berries, livestock, lettuce, increase efficiency or have yield gains. walnuts, tomatoes, pistachios and citrus. Farms also vary The California context also offers the opportunity to study widely in scale and structure—from small and mid-sized the effect of irrigation infrastructure on practice adoption. family-owned operations to very large, multi-commodity The state’s year-round growing season and Mediterranean international corporations (California Department of Food climate (i.e. dry summers, wet winters) creates reliance on and Agriculture 2018). irrigation, encouraging many combinations of irrigation sys- Importantly for our focus on N management, California tems and water sources, from gravity-fed flood irrigation to is one of the first states in the U.S. to implement an agri- systems with pressurized drip or sprinkler infrastructure, cultural non-point source pollution regulatory program, the fed by surface or groundwater. Irrigation systems can have Irrigated Lands Regulatory Program (ILRP). The ILRP is a large influence over N leaching (Letey and Vaughan 2013), implemented through local entities known as “Water Quality and different systems may be more or less compatible with Coalitions” and includes mandatory elements around report- different practices. Drip irrigation systems have been widely ing use of best management practices and an N budget, as adopted in California’s perennial nut crops and high value well as attendance at one educational meeting per year, held annual crops, particularly on farms with sandy soils that rely in each Water Quality Coalition (Central Valley Regional on groundwater (Taylor and Zilberman 2017). Some N man- Water Quality Control Board 2020; for more information, agement practices are implemented more easily through drip see Online Appendix). The N management practices we irrigation systems, like split application where the farmer study in this paper are consistent with those tracked as part can deliver fertilizer sets through the drip irrigation, provid- of the ILRP mandatory reporting. This policy landscape ing a low-labor way to distribute fertilizer throughout the offers a unique context to study the potential effects of gov - season. Other practices have co-evolved with the diffusion ernance on farmer decision-making, compared to well-docu- of drip irrigation, like adjusting irrigation rates according to mented studies evaluating practice adoption under voluntary evapotranspiration (ET) estimates of crop water needs (Han- policy settings (Reimer et al. 2018; Hillis et al. 2018). son et al. 2009; Taylor and Zilberman 2017). As a result, farmers working in operations with pressurized irrigation infrastructure may find their operations more compatible Survey and data collection with the aforementioned practices, whereas farmers working with gravity fed irrigation systems may show a propensity This paper employs data collected through a mail survey toward a different suite of practices. conducted in 2018 across the Central Valley of California. In considering how management practices may vary due The project integrated stakeholder feedback throughout to underlying differences in farm attributes, we hypoth- the research process and included multiple phases of inter- esize (H2) that individual practices and practice portfolios views, focus groups, and preliminary survey data collec- will be distinguishable across different farm types, based tion that both informed our survey design and dissemina- on crop type, farm size, irrigation system and water source tion strategy, and helped in interpreting results. An external characteristics. advisory committee, representing policymakers, farmers, directors of the Water Quality Coalitions, and nationwide researchers and extension specialists also provided survey 1 3 788 J. Rudnick et al. review. Institutional Review Board approval for the study Department of Agriculture 2018). We removed all obvious was obtained through the University of California Davis. non-agricultural entries (e.g. golf courses or public lands The survey was distributed to farmer members from using pesticides) from the mailing lists. This list contained three Water Quality Coalitions: the Colusa Glenn Sub- our best estimate of all eligible farmers who would report to watershed Program (CGSP), the San Joaquin County and SJDWQC or ESJWQC under the ILRP. The survey was sent Delta Water Quality Coalition (SJDWQC), and the East San to all farmers in SJDWQC region (n = 2322) and to a random Joaquin Water Quality Coalition (ESJWQC) (see Fig. 1). sample of 33% of farmers in ESJWQC region (n = 1243), Together, these Coalitions covered over 900,000 acres of due to the size of the Coalition. In aggregate, this totaled irrigated cropland and approximately 7500 individual farm- 4994 surveys mailed across all three regions. ing operations in 2017. These regions represent a longitu- We followed a four-wave mailing process using a modi- dinal transect of the Central Valley that captures a range of fied Tailored Design Method, which included a cover letter agricultural, ecological and socio-political dimensions. The and survey, followed by a reminder postcard, then second most important crop types in these regions include almonds, letter and survey, and final reminder postcard (Dillman et al. walnuts, grapes, tomatoes, sunflowers, pistachios, alfalfa, 2008). In CGSP, the Coalition permitted us access to join corn and wheat. Rice is also a top production crop in the our survey response data to their anonymized mandatory CGSP region, but was removed from our sampling frame reporting data on management practices adopted on each since our management practices focus on non-flooded crop- field. These data allowed us to compare survey-reported ping systems. practice adoption rates with adoption rates reported on man- Mailing addresses were provided by CGSP, to which datory paperwork by farmers who did not respond to the the survey was sent to all members (n = 1471). In SJD- survey, offering opportunity to evaluate the self-selection WQC and ESJWQC, mailing addresses were obtained from bias that is prevalent in survey-based research. We found county Agricultural Commissioner offices, who maintain that adoption rates did not differ between survey respondents publically-available databases of all commercial farming and non-respondents, thus indicating our survey respondents operations in compliance with Pesticide Use Reporting were representative of adoption behavior occurring across requirements. In these regions, organic farmer addresses the watershed (see Online Appendix Table A1). were also obtained through the U.S. Department of Agri- culture Organic INTEGRITY Database (United States Fig. 1 Map of Central Valley of California, highlighting three Water Quality Coalitions where survey was distributed 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 789 combined in a composite variable; all alpha scores were > Variable measurement 0.70, a widely-accepted cut-off to indicate internal validity (Santos and Reynaldo 1999). See Online Appendix Table A2 The survey questionnaire included 30 questions covering a range of topics related to farmers’ views on N management, for information on survey questions and composite variables. Finally, farmer demographic variables included a binary including their adoption of eight different N management practices on their largest parcel of their most important crop variable for college education, a categorical variable for income class, and a continuous variable for years in agri- in the 2017 crop year, measured as a binary variable. In this paper, we treat these eight N management practices as the culture. Binary variables were included to distinguish the three Water Quality Coalitions, with the baseline as farmers simultaneous adoption outcome variables in a multivariate probit regression model. who didn’t identify with the three Coalitions of focus or left the question unanswered. The binary measurement of these practices is a data limi- tation of this paper, especially for practices that are applied Survey respondents in a temporally (e.g. multiple times/year versus every other year) or spatially (e.g. full operation versus particular fields) We received a total of 966 partial and full survey responses heterogeneous fashion. Given the known heterogeneity of operations on our mailing lists (e.g. operations varied back (CGSP: n = 377, SJDWQC: n = 312, ESJWQC: n = 183), constituting an average response rate of 20% from a single crop up to 14 unique crops, and from < 1 acre to > 20,000 acres), we were constrained to developing a (CGSP: 30.7% SJDWQC: 14.4%, ESJWQC: 15.4%). We removed 101 responses from farmers reporting on irrigated survey tool that was general enough to fit every possible respondent. Furthermore, the ILRP collects practice adop- pasture, bringing our useable number of respondents to 865. Response rates were adjusted for the possible non-eligible tion data in a binary fashion as well. Thus, aligning our data structure with that of the regulatory program allows us the addresses included in our original mailing lists (Ameri- can Association for Public Opinion Research 2016). Our best opportunity for data comparisons and to draw policy- relevant conclusions. response rate is on par with recent surveys using similar designs and regarding similar topics (Denny et al. 2019; The survey measured farm operation characteristic vari- ables of interest: most important crop type (aggregated into Wilson et al. 2014; Arbuckle and Rosman 2014). All survey data was digitized for analysis. a binary variable: perennial/annual crop), farm size (log- transformed), primary irrigation type (aggregated into a Respondents are fairly representative of the full farm- ing populations in the surveyed regions, when compared to binary variable: pressurized systems- drip, micro-sprinkler, versus gravity-fed systems- flood, furrow, border strip) and USDA 2012 Census of Agriculture data (see Online Appen- dix Tables A3–A5). The average farm size of our respond- water source (aggregated into two binary variables: access to surface water versus groundwater, and access to both water ents is 355 acres (minimum < 1 acre, maximum ~ 12,000 acres). In aggregate, our survey respondents manage 329,800 sources versus single source). We also measured a number of behavioral variables typi- acres of land across the Central Valley, approximately 35% of the acreage of the study area. Seventy-nine percent of cally used in agricultural adoption research to include as controls (Prokopy et al. 2019). These included information- respondents own their land; 80% of respondents are male; 84% of respondents identify as White or Caucasian, 4% as related variables measuring access to information from three perspectives: a tally of the total number N manage- Hispanic or Latino and 3% as Asian or Asian American. Sixty-one percent of respondents have at least some college ment information sources, a binary variable for the use of Certified Crop Advisers to create N budgets (“consultants”), education. On average, respondents have 35 years of farm- ing experience, and the median gross farm income bracket and a binary variable for the completion of a Self-Certifi- cation course, which is a voluntary educational component is $100,000–$200,000. Respondents listed all crops and acreage they cultivate, of the ILRP which allows farmers to self-certify their own N budgets. Socio-behavioral concepts included problem though we only asked about practice adoption on their most important crop, as self-identified on the survey. Sixty-four awareness (“acceptance of agricultural N sources”), envi- ronmental values (“conservation motivation”), and perceived percent of farmers report only growing one crop, 27% two to four crops, and 4% have five or more crops. Eighty-five per - behavioral control (“self-efficacy”) (Reimer et al. 2012b). These latent variables were constructed using exploratory cent of respondents indicate a perennial crop as their most important. Seventy-four percent of respondents have pres- factor analysis to combine multiple survey question items measured on five-point Likert scales (Costello and Osborne surized irrigation systems on their most important parcels. Forty percent of respondents rely on groundwater only, 44% 2005), which improves reliability (McIver and Carmines 1981; DeVellis 2003; Santos 1999). Cronbach alpha scores have access to surface water (only) through riparian rights or irrigation district delivery water, and 16% have both surface were used to verify internal consistency between the items 1 3 790 J. Rudnick et al. and groundwater access (See Online Appendix Table A6 for Jenkins (2003), our MVP models are estimated using the all descriptive statistics). Geweke–Hajivassiliou–Keane (GHK) simulator in Stata 16. Multivariate normal probabilities are calculated at each Analysis approach: multivariate probit iteration of the simulation. Simulation bias is minimized by for estimating interdependencies increasing the number of random draws from the simulator, to at least as large as the square root of the sample size; we Much of the existing adoption literature uses a standard ran the model with 35 random draws (Cappellari and Jenkins quantitative approach of estimating some type of linear 2003). We also tested for any ordering effects in the depend- model with an individual practice, or count of practices, as ent variables by running the model with multiple different the dependent variable, and multiple predictor variables to orders for the practice dependent variables; results were con- test hypotheses about drivers of adoption (Prokopy et al. sistent across all runs. As an additional robustness check, we 2019). Here, we need an empirical model that simultane- fit individual univariate probit regression models for each of ously estimates farm and farmer variable influences on the the eight practices, which produces very similar coefficient adoption of multiple practices, and how those practices are estimates (See Online Appendix for additional discussion related to each other. To accomplish this goal, we employ a on robustness and Table A7 for univariate probit results). multivariate probit (MVP) model that allows estimation of To test H1, we evaluate the MVP variance–covariance multiple binary probit regression models (in our case 8 mod- matrix alongside a co-occurrence matrix. The co-occurrence els) simultaneously, while analyzing correlation between matrix uses observed adoption data and calculates the pro- errors in the different models. Failure to account for these portion of all farmers who jointly adopt any two practices, correlated error terms can result in inefficient coefficient evaluating all possible dyads of practices, a method that estimates and biased error terms (Cappellari and Jenkins has been applied widely in ecology to evaluate species co- 2003). This approach has been applied in other studies look- occurrence (Hines and Keil 2020). Both relatedness matri- ing at simultaneous adoption in developing agricultural set- ces are visualized as undirected weighted networks with the tings (Koppmair et al. 2017; Kassie et al. 2015; Kara et al. edge weights between every pair of practices reflecting the 2008; Teklewold et al. 2013; Jara-Rojas et al. 2013). relatedness of those two practices. We use Quadratic Assign- Considering all N management practices, each equation ment Procedure (QAP) matrix correlation to assess which in the system can be written as: practices frequently occur together and which practices have highly correlated errors, indicating a potential underlying Y = B X + B X … B X + e, i 1i a 2i b ni n dimension influencing their adoption. (i = LT, ST, CC, IN, MP, SA, PB, ET), To test H2, we draw on descriptive statistical analyses including Pearson’s chi-squared tests to investigate differ - where Y indicates the i different practices (LT = Leaf Test- ences in individual practice adoption rates between farm ing, ST = Soil Testing, CC = Cover Crops, IN = Ir r igation types and evaluate our MVP coefficient estimates to under - N Testing, MP = Moisture Probe, SA = Split Application, stand the predictive power of key farm operation charac- PB = Pressure Bomb, ET = Evapotranspiration-based sched- teristics of interest (crop type, farm size, irrigation system uling) of interest and X are the predictor variables of inter- and water source), while controlling for all other farmer est. Our unit of analysis is an individual farmer. For farmers behavior and demographic variables, as well as interde- who operate across multiple fields, we evaluate their practice pendency across practices. We then qualitatively evaluate adoption only on the largest field of their most important differences in practice portfolios across different operation crop, thus including only one observation per farmer. This types by looking at differences in the co-occurrence practice yields a matrix of estimated model coefficients, with the networks. We highlight results for practice portfolio differ - coefficient for each covariate (B … B ) estimated for each 1i ni ences across farm types with different irrigation systems, of the eight practices. The MVP assumes that the error terms but additional side-by-side practice portfolio comparisons for each practice (e , e , e …) jointly follow a multivari- LT ST CC between other operation types are included in the Online ate normal distribution with a mean of 0 and variance of 1. Appendix Figures A3–A4. The model also generates a variance–covariance matrix that Descriptive statistics and data visualization were car- provides the correlation coefficients (rho) between the error ried out in R Statistical Software Version 3.5.3; multivari- terms of all pairs of equations. These correlations can offer ate modelling was conducted in Stata16. All model code is insight on the complementary (i.e. positive correlations) or linked in the Online Appendix. substitutable (i.e. negative correlations) nature of pairs of practices. Simulated maximum likelihood techniques are used to estimate the model, and following Cappellari and 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 791 The MVP variance–covariance matrix indicates appropri- Results ate use of the MVP model, as the likelihood ratio test for the overall correlation of error terms [ƿ = ƿ = ƿ = ƿ = ƿ Our results are structured around our hypotheses. First, to LT ST SA CC IN 2 2 = ƿ = ƿ = ƿ = 0; χ (28) = 208.781, Prob > χ = 0.000] evaluate how adoption decisions are interrelated across ET PB MP rejects the null hypothesis that the error terms across adop- practices (H1), we present the MVP model error correla- tion equations were not correlated. This result is also sup- tion matrix and practice co-occurrence matrix, as well as ported by the many significant correlation coefficients the QAP correlation between the two measures. Then, we between the pairwise model correlation terms (see Table 2). evaluate how farm operation characteristics influence indi- Nearly all practices are positively correlated, and thus can vidual practice and portfolio adoption (H2) by reviewing be considered complementary. The fertilizer-specific prac- the MVP coefficient estimates and descriptively comparing tices however, emerge as the core practices that are the most co-occurrence networks across different farm types. frequently adopted and have the highest correlations in their error terms with each other, indicating the greatest interde- Interdependence in practice adoption (H1): MVP pendency: leaf testing and soil testing (ρ = 0.63), leaf testing model error correlation and co‑occurrence matrices and irrigation well N testing (ρ = 0.39), leaf testing and split application (ρ = 0.38). On average, farmers report adopting 3.5 out of eight N man- The co-occurrence matrix further indicates that leaf test- agement practices. Adoption rates for individual practices ing, soil testing and split application make up the core of ranged from 75% (split application) to 14% (pressure bomb), the portfolios for most farmers, with co-occurrence rates but the majority of all farmers adopt split application, leaf all at 50% or greater. Each of these three practices also co- testing and soil testing (See Fig. 2). occurred over 30% of the time with moisture probes and The MVP variance–covariance correlation matrix and roughly 25–30% of the time with both irrigation well N test- the practice co-occurrence matrix support H1, in that both ing and ET-based irrigation scheduling. Cover crops and measures show strong relatedness between practices and pressure bombs were most peripheral and had the lowest co- demonstrate that practices cluster most strongly within the occurrence rates with any other practice (see Online Appen- fertilizer farm management area. dix Table A8 for all co-occurrence rates). The co-occurrence Fig. 2 Individual practice adoption rates across all farms; colors indicate management area of the farm. (Color figure online) 1 3 792 J. Rudnick et al. 1 3 Table 2 Correlation coefficients (rho) and standard errors (in parenthesis) of the eight adoption equations’ error terms from the MVP model Farm management Ƿ Ƿ Ƿ Ƿ Ƿ Ƿ Ƿ LT ST SA IN CC ET PB area Fertilizer  Ƿ 1 LT  Ƿ 0.629 (0.06)*** 1 ST  Ƿ 0.382 (0.07)*** 0.209 (0.08)*** 1 SA  Ƿ 0.388 (0.08)*** 0.315 (0.07)*** 0.249 (0.09)*** 1 IN Soil  Ƿ 0.362 (0.08)*** 0.140 (0.07)** 0.203 (0.08)** 0.252 (0.07)*** 1 CC  Ƿ 0.268 (0.08)*** 0.056 (0.07) 0.138 (0.08)* 0.244 (0.07)*** 0.105 (0.07) 1 ET Irrigation  Ƿ 0.220 (0.10) − 0.065 (0.09) − 0.09 (0.11) 0.087 (0.09) 0.122 (0.09) 0.219 (0.09)*** 1 PB  Ƿ 0.334 (0.11)*** 0.288 (0.07)*** 0.110 (0.08) 0.211 (0.07)*** 0.036 (0.07) 0.346 (0.07)*** 0.180 (0.09)** MP Matrix is symmetrical. Positive correlations are interpreted as complementary practices, negative correlations are interpreted as substitutable Bolded values highlight correlation coefficients that are statistically significant at p < 0.1 Practices abbreviated as: LT leaf testing, ST soil testing, SA split applications, CC cover crops, IN irrigation well N testing, ET evapotranspiration-based irrigation scheduling, PB pressure bomb, MP moisture probe 2 2 Likelihood ration test of ƿLT = ƿST = ƿSA = ƿCC = ƿIN = ƿET = ƿPB = ƿMP = 0; χ (28) = 208.781, Prob > χ = 0.000 ***p < 0.01; **p < 0.05; *p < 0.1 A farm systems approach to the adoption of sustainable nitrogen management practices in… 793 Fig. 3 Practice interdependency networks shown. Nodes represent sure bomb, MP moisture probe. (Left) The co-occurrence network has management practices, colored based on farm management areas and edge weights representing the frequency at which two practices are scaled based on adoption rates, with larger nodes representing higher jointly adopted on the same parcel. (Right) The MVP variance–covar- adoption. Practice names are abbreviated as: LT leaf testing, ST soil iance network has edge weights showing the correlation in errors testing, SA split applications, CC cover crops, IN irrigation well N between individual practice equations estimated through the MVP testing, ET evapotranspiration-based irrigation scheduling, PB pres- model and MVP error correlations are visualized as networks with Farm type influences individual practice edge weights representing the strength of the relationship and portfolio adoption (H2): adoption rates between two practices based on the two relatedness meas- and MVP coefficient estimates ures (See Fig. 3). The Quadratic Assignment Procedure demonstrates the In addition to evaluating the interdependency between prac- MVP variance–covariance and co-occurrence matrices are tices, we compared how adoption of individual practices positively correlated and thus consistent with one another and practice portfolios differed across farm types. We find (r = 0.62; p < 0.001; see Figure A1 in Online Appendix). The significantly higher rates of adoption for all eight practices in largest residuals were observed with split application when perennial crop systems as compared to annual crop systems paired with other practices such as soil testing, leaf testing (p < 0.001), very large farms (> 1000 acres) as compared to and moisture probe; split application has higher co-occur- smaller farms (p < 0.05), and operations with pressurized rence rates with each of these practices than error correlation irrigation infrastructure as compared to gravity-fed irrigation rates in the MVP model. (p < 0.05). Farmers with access to both surface and ground- water adopted six of the eight practices at significantly 1 3 794 J. Rudnick et al. Fig. 4 MVP model coefficient estimates and 95% confidence intervals plotted for each practice (denoted by different shape markers), with prac- tices colored by farm management areas higher rates (p < 0.01), with the exceptions of cover crops, The effect of farm type is also evident when we evaluate where adoption was higher with access to groundwater only differences in practice portfolios. Results indicate partial (p < 0.05), and soil testing, where there were no significant support for H2, with irrigation system being a driving fac- differences across water sources (Online Appendix Figure tor in determining if and how farmers integrate the irrigation A2). management practices into their N management portfolios. While adoption rates vary significantly when we look at On pressurized-irrigation farms, we see greater overall co- each farm operation characteristic in isolation, our MVP occurrence of most practices, but more noticeably, we see results support H2, indicating crop type, farm size and irri- an expansion of the core practices adopted by most farmers gation system are the most important adoption predictors, to include moisture probes, ET-based irrigation scheduling, when controlling for multiple farm operation and socio- and irrigation well N testing (see Fig. 5). behavioral factors. Perennial crop type is a positive and sig- nificant predictor for all practices except soil testing. Farm size is a positive and significant predictor for all practices Discussion except cover crops. Pressurized irrigation system is the farm operation characteristic that most distinguishes adoption Improved N management to maximize crop use and mini- drivers between different practices, as a positive and sig- mize losses to the environment is complex, requiring farm- nificant predictor for irrigation-specific practices (moisture ers regularly monitor, evaluate and adapt. The relevant suite probe, pressure bomb and ET-based irrigation scheduling), of management practices needed to address these multiple as well as irrigation well N testing. Water source variables dimensions stretch across different management areas of the however, were seldom significant predictors for any practice. farm and vary with other operational characteristics. Our Among the information, socio-behavioral and farmer results highlight the complexity and interdependency of demographic control variables, information source tally was farmers’ adoption of these practices. the only variable that had a positive and significant effect on On average, California farmers adopt between three and all practices. The other socio-behavioral and demographic four of the eight N management practices we measured. variables were less consistently significant across all prac- Regarding H1, we find fertilizer practices form the core of tices, but generally in the direction predicted by existing most farmers’ practice portfolios, with greater practice adop- literature (see Fig. 4; all model coefficients are presented in tion and co-occurrence rates and greater covariance in our Online Appendix Table A9). multivariate modelling. Regarding H2, we find that indi - vidual practice and portfolio adoption varies across farm 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 795 Fig. 5 Practice co-occurrence networks compared between pressur- leaf testing, ST soil testing, SA split applications, CC cover crops, ized-irrigation farms (left) and gravity-fed irrigation farms (right). IN irrigation well N testing, ET evapotranspiration-based irrigation Nodes represent management practices, colored based on farm man- scheduling, PB pressure bomb, MP moisture probe. Edge weights agement areas and scaled based on adoption rates, with larger nodes represent the frequency at which two practices are jointly adopted on representing higher adoption. Practice names are abbreviated as: LT the same parcel, for the farm operation type types, with larger farms and perennial crop farms adopting interdisciplinary research team’s agronomic expertise and more practices overall and farms with pressurized irrigation our qualitative field work, prompt the following postula- systems integrating irrigation-related practices more readily tions about why fertilizer-focused practices are more read- into their portfolios. Applying a farm systems conceptual ily adopted and why practice portfolios vary across farm and analytical approach motivates our explicit measurement types. of the interdependencies across multiple practices and facili- First, Cooperative Extension and on-farm technical tates our interpretation of how and why farmers make deci- consultants have historically emphasized fertilizer appli- sions to adopt multiple practices to fit their farm operation. cation and monitoring practices. Cooperative Extension has heavily focused on providing fertilizer application Potential mechanisms driving adoption rate recommendations for different crop types (Geisseler and interdependence patterns 2016) and has promoted the 4R’s framework as a decision- support tool for adjusting fertilizer applications. On-farm The survey and modelling results show strong evidence of technical consultants, such as Certified Crop Advisers who practice interdependence and complementarities, which can approve the N Management Plans required by farm- is consistent with multiple mechanisms that might sup- ers under the ILRP (if the farmer is not Self-Certified), port systems thinking. These results, combined with our also play a large role in disseminating practice knowledge. 1 3 796 J. Rudnick et al. These consultants provide fertilizer product and applica- larger farms, perennial crop farms, and farmers accessing tion rate recommendations, and oftentimes carry out field more information sources adopting all practices at higher and crop monitoring practices (i.e. leaf and soil testing) on rates. These findings are consistent with much past adop - behalf of the farmer. However, some of these consultants tion research that highlights the importance of economies of work for fertilizer companies as product representatives, scale, capacity to accept risk, and social and informational which can create misaligned incentives for promoting networks as key factors that help to lower the barriers to practices that reduce fertilizer use. adoption (Houser et al. 2019; Lubell and Fulton 2008; Marra Irrigation well N testing is noteworthy, as it has recently et al. 2003; Prokopy et al. 2019). become part of standard fertilizer practices. In response to Differences in adoption of irrigation management prac- continuing concerns about nitrate contamination in ground- tices specifically, were even more apparent across farms with water, the State Water Board updated the ILRP mandatory different irrigation infrastructure, with pressurized irriga- reporting requirements in 2018 to require farmers to measure tion system farms being significantly more likely to adopt nitrate levels in their irrigation wells and include it as part the irrigation-focused practices and integrate them into of their N budget. The idea to “pump and apply” has thus their overall N management portfolios. Many of the irri- only relatively recently become integrated into Coopera- gation management practices require increased investment tive Extension outreach efforts and CCA recommendations in equipment and supplies, technical learning, and time to (Nitrogen Management Training Materials 2019). This may implement. Pressurized irrigation infrastructure may be a help to explain our observation of increasing adoption and proxy indicating an additional level of capacity for capital co-occurrence rates with the traditional set of fertilizer prac- investment and operational sophistication. tices. The heavy emphasis of extension and technical con- Moreover, the ability for farms with pressurized irrigation sulting on fertilizer-related practices overall is likely to be a systems to more precisely control water delivery may serve key factor driving our observations of these practices being to make some of the irrigation-specific practices seem more featured at the core of most farmers’ practice portfolios. advantageous in these systems. For example, a farmer work- Furthermore, it illuminates opportunities for extension and ing with a groundwater-fed, drip irrigation system has more outreach to diversify practice recommendations to increase ability to adjust their irrigation timing and total water deliv- adoption of the more peripheral irrigation and soil-focused ered in response to evapotranspiration or moisture probe practices. data; thus adopting these practices improves their water use In concordance with the statistical results, our qualita- efficiency. In contrast, a farmer with a flood irrigation system tive field work suggests connections between fertilizer, soil dependent on surface water deliveries is not able to irrigate and irrigation management are still not widely understood any time they want and instead is beholden to the water dis- across technical consultant or farmer communities. Irriga- trict delivery schedule. Their reduced flexibility for irriga- tion practices specifically have only been integrated into tion may make the irrigation-specific practices seem obso- N management in relatively recent research and extension lete. These examples are used to illustrate why the irrigation efforts that have expanded to consider the full farm system. practices may be perceived to be better ‘fit’ in the portfolios Yet, in functionally differentiated farming operations (Kling of farms with pressurized irrigation systems. We do want and Mackie 2019), technical consultants and in-house farm to make clear however, the irrigation practices in and of managers may only specialize in one component of the farm themselves are not reliant on pressurized irrigation systems system (e.g. fertility management) and lack expertise to inte- and could be put to use in gravity-feed irrigation systems. grate across fertility and irrigation. Thus, holistic practice recommendations that integrate multiple areas of the farm Limitations and future work are still rather limited to university researchers and Coop- erative Extension and a small subset of innovative farmers As with much survey-based research, our results may be and technical consultants. The limited understanding of this limited by low response rates, particularly in SJDWQC connectivity between multiple areas of the farm and their and ESJWQC. Further, we acknowledge the possibility impact on N management helps to explain the limited inclu- of a response bias from farmers who are more engaged sion of irrigation practices and cover cropping in farmers N in extension, outreach and ILRP activities. However, our management practice portfolios. sample is a relatively good representation of the diver- Second, the peripheral status of irrigation management sity of types and size of agricultural operations in Cali- practices, exacerbated on farms with gravity fed irrigation, fornia’s Central Valley. Our analyses are also constrained annual crops and smaller farm size, exemplifies how the ben- to evaluating adoption only on farmers’ “most important” efits and costs of practices vary across farming contexts. parcels. This may lead to overrepresentation of perennial The impact of access to capital- including financial, techni- crop types and larger parcels in our analyses and under- cal and social capital- is apparent across all practices, with predict how these same farmers manage their smaller or 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 797 annual parcels differently. Overcoming this data limita- measure the mechanisms driving interdependencies across tion, while not dramatically increasing the burden placed multiple areas of the farm system may provide for greater on survey respondents, could be achieved by pairing our insight to understand and predict farmer decision-making survey data with reporting data collected by programs like across a variety of behaviors of interest. the ILRP. However, concerns about confidentiality often Finally, encouraging more holistic approaches to farm limit the capacity to connect data collected by researchers management challenges will require that the farm systems to data collected by other organizations, and particularly framework move beyond research in the academy and those involved in regulatory programs. become adopted as a way of practice by extension and on- We contend that measuring and studying adoption of the farm consultants, who provide critical technical assistance interdependencies between practices is essential to moving and trusted practice recommendations. This emphasizes the forward on understanding drivers and barriers to adoption, need for an ‘Extension 3.0 model’ (Lubell et al. 2014), in and designing policy and outreach that adequately account which university extension focus their efforts on ‘training for how interdependencies vary with practice-specific attrib- the trainer’ with a farm systems approach to addressing the utes (c.f. Reimer et al. 2012a). It will be helpful to char- management challenge at hand. acterize how the costs and benefits of practices vary with farm operation characteristics, and how they accrue when practices are used in tangent with one another, as part of the Conclusions same portfolio. It is also worth noting that relevant man- agement practices will vary across contexts. The practices This study contributes the most comprehensive analysis of considered as best practices for improving N management farmer behavior on N management in California to date, in California differ from the frequently studied practices for illuminating adoption trends for a suite of practices across a water quality protection in the Midwest, such as filter or wide range of farm and farmer characteristics. We analyze buffer strips, side dressing, nitrification inhibitors, or tillage adoption of eight N management practices that stretch across (Denny et al. 2019; Stuart et al. 2014). fertilizer, soil and irrigation components of the farm. We It is additionally important to analyze the relationship argue an integrated farm systems approach improves our between different portfolios of practices and agronomic understanding of farmers’ decision-making across different and environmental outcomes. It is possible that two prac- farm operation contexts and management practices, allow- tices that are frequently co-adopted on a specific type of ing us to better capture the complexity and interdependency farm are actually a mal-adaptive combination when con- of adoption decision-making on portfolios of practices. sidering the goal of improved N use efficiency. A criti- Our findings highlight how adoption patterns and adoption cal next step for agronomic research should be to develop drivers differ across practices, and how the combination or better understandings of the agronomic, economic and portfolio of practices adopted may differ quite dramatically environmental outcomes associated with different practice across heterogeneous farms, based on operational character- portfolios on different types of farms. This can serve as istics that influence fit. the basis for developing practice portfolio recommenda- Our interdisciplinary and engaged approach further tions tailored to specific farm types, where the interactions informs the outreach and policy recommendations that fol- between multiple practices and the constraints and oppor- low from our results. We have observed that the large hetero- tunities of the full farm system are taken into account. geneity in California’s farms result in neighboring operations Future research should seek to further extend the farm facing dramatically different barriers or achieving greatly systems approach by considering not only how N manage- different benefits with the adoption of an individual practice ment practices relate to one another, but also how N man- or specific portfolio of practices. The “disproportionality” of agement practices might relate to other farm management impact (Nowak et al. 2006) on N pollution means that moti- decisions, like pest management, farm investment, labor vating and focusing extension efforts to increase adoption constraints, and on-farm data collection. We must consider on large, input-intensive operations increases the likelihood how farmers organize their operations and management of achieving sustainability goals. Yet, we see that large size staff, the timing of management decisions on different and greater capital are important enabling factors increasing parts of the farm, and how management practices that opti- likelihood of adoption already, and these operations may mize for one outcome (e.g. N use efficiency) may influence be able to achieve higher private returns due to increased other agronomic, economic or ecological outcomes. The efficiency or other co-benefits. In contrast, small farms with farm systems approach is broadly applicable beyond the smaller individual N footprints, may face extreme financial, study of N management and should be incorporated into technical or information barriers to implement even the easi- the study of other conservation practice and on-farm tech- est practices. These factors must be considered in justifying nology adoption. Expanding the research lens to explicitly and tailoring financial incentive and technical assistance 1 3 798 J. Rudnick et al. many members of our Survey Advisory Committee who gave ample programs that are aimed to increase and maintain practice feedback throughout the project and continuously work to support adoption. Our results suggest that N management extension progress on nitrogen management in the field; this work would not and technical assistance programs in California should focus be possible without all of their support. We also acknowledge and on developing practice portfolio recommendations that are appreciate that this work was funded by a competitive grant from the California Department of Food and Agriculture Fertilizer Research adapted to fit small farms, annual cropping systems and and Education Program. operations that don’t currently have pressurized irrigation infrastructure in place. Funding Funding was provided by California Department of Food Furthermore, a novel opportunity exists in California and Agriculture Fertilizer Research and Education Program (Grant to coordinate policy and extension to promote a holistic No. 16-0620). approach to N management through the implementation of Open Access This article is licensed under a Creative Commons Attri- the ILRP. As our survey results and field work have illumi- bution 4.0 International License, which permits use, sharing, adapta- nated, N management policy and extension efforts have his- tion, distribution and reproduction in any medium or format, as long torically focused on fertilizer practices, resulting in greater as you give appropriate credit to the original author(s) and the source, adoption and integration of these practices, at the expense of provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are holistic portfolios that integrate soil and irrigation practices included in the article’s Creative Commons licence, unless indicated as well. Developing practice portfolio recommendations that otherwise in a credit line to the material. If material is not included in integrate irrigation, soil and fertilizer practices together, and the article’s Creative Commons licence and your intended use is not are tailored to different types of farms should be a priority permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a for N management research and extension going forward. copy of this licence, visit http://creativ ecommons .or g/licenses/b y/4.0/. The effort to develop tailored portfolio recommendations could also clarify how to measure improvement around N management, since using all, or more management practices References may not always be applicable or optimal for all operations. Evaluating whether farmers are adopting the best-suited Ajzen, I. 1989. The Theory of Planned Behavior. Organizational portfolio of practices for their specific farm context may Behavior and Human Decision Processes 50: 179–211. provide a more tangible goal for farmers, while also provid- American Association for Public Opinion Research. 2016. Response Rate Calculator V4.0. 2016. http://www.aapor .org/Educa tion- ing more clarity to the public and policymakers as to when Resou rces/For-Resea rcher s.aspx. and how reductions in N pollution may be reached. Arbuckle, J. G., and Hanna Rosman. 2014. Iowa Farmers’ Nitro- We suggest a coordinated approach engaging all of the gen Management Practices and Perspectives. http://lib.dr.iasta key actors across the wide information and extension net- te.edu/exten sion_commu nitie s_pubs/24. Aubert, B.A., S. Andreas, and J. Grimaudo. 2012. IT as Enabler work working on this issue: California Department of Food of Sustainable Farming: An Empirical Analysis of Farmers’ and Agriculture, State and Regional Water Boards, Water Adoption Decision of Precision Agriculture Technology. Deci- Quality Coalitions, USDA Natural Resources Conserva- sion Support Systems 54 (1): 510–520. https://doi.or g/10.1016/j. tion Service, Resource Conservation Districts, commodity dss.2012.07.002. Baumgart-Getz, A., L.S. Prokopy, and K. Floress. 2012. Why Farm- groups, University of California Cooperative Extension, ers Adopt Best Management Practice in the United States: A on-farm PCAs and CCAs, NGOs and private sector com- Meta-Analysis of the Adoption Literature. Journal of Environ- panies. Combined, this network of actors has the resources, mental Management 96 (1): 17–25. https ://doi.org/10.1016/j. reach, and earned trust of farmers to develop comprehen- jenvm an.2011.10.006. Bawden, R. 1995. On the Systems Dimension in Farm Systems sive strategies that seek to (1) diversify and incorporate Research. Journal of Farm Systems Resource Extension 5 (2): practices across the entire farm that holistically address N 1–18. management, (2) shift social norms away from excess ferti- Blank, S. 2001. Producers Get Squeezed up the Farming Food Chain: lizer application as ‘cheap yield insurance’, and (3) provide A Theory of Crop Portfolio Composition and Land Use. Applied Economic Perspectives and Policy 23 (2): 404–422. creative incentives through shared risk approaches, techni- Bosch, N.S., J.D. Allan, J.P. Selegean, and D. Scavia. 2013. Scenario- cal assistance in implementation and maintenance, market Testing of Agricultural Best Management Practices in Lake Erie recognition and reduced regulatory burdens. Together these Watersheds. Journal of Great Lakes Research 39 (3): 429–436. efforts can promote and sustain a farm systems approach to California Department of Food and Agriculture. 2018. California Agricultural Statistics Review 2017–2018. Sacramento, CA. N management. https://www .cdfa.ca.gov/statis tics/PDFs/2017-18A gRepor t.pdf . Cappellari, L., and S.P. Jenkins. 2003. Multivariate Probit Regres- Supplementary Information The online version contains supplemen- sion Using Simulated Maximum Likelihood. The Stata Jour- tary material available at https:// doi.org/10.1007/s10460-02 1-10190-5 . nal: Promoting Communications on Statistics and Stata 3 (3): 278–294. https ://doi.org/10.1177/15368 67x03 00300 305. Acknowledgements We would like to acknowledge all farmers who Central Valley Regional Water Quality Control Board. 2020. Irri- voluntarily participated in our survey, the Water Quality Coalitions gated Lands Regulatory Program (ILRP) Frequently Asked who facilitated our study and ensured relevancy of this work, and the 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 799 Questions. https: //www.waterb oards .ca.gov/centra lvall ey/water Salinas Valley Groundwater. Report for the California State Water _issue s/irrig ated_lands /ilrp_faq.pdf. Resources Control Board Report to the Legislature. Church, S.P., J. Lu, P. Ranjan, A.P. Reimer, and L.S. Prokopy. 2020. Hillis, V., M. Lubell, and M. Hoffman. 2018. Sustainability Partner - The Role of Systems Thinking in Cover Crop Adoption: Impli- ship and Viticulture Management in California. Journal of Envi- cations for Conservation Communication. Land Use Policy 94: ronmental Management 217: 214–225. https ://doi.org/10.1016/j. 104508.jenvm an.2018.03.033. Coates, R.W., M.J. Delwiche, and P.H. Brown. 2005. Precision Irri- Hines, J., and P. Keil. 2020. Common Competitors and Rare Friends. gation and Fertilization in Orchards. In ASAE Annual Meeting. Nature Ecology and Evolution 4 (1): 8–9. https://doi.or g/10.1038/ American Society of Agricultural and Biological Engineers. s4155 9-019-1071-3. https ://doi.org/10.13031 /2013.19815 Hoffman, M., M. Lubell, and V. Hillis. 2014. Linking Knowledge and Conley, T., and C. Udry. 2010. Learning about a New Technol- Action Through Mental Models of Sustainable Agriculture. Pro- ogy: Pineapple in Ghana. American Economic Review 100 (1): ceedings of the National Academies of Science 111: 13016–13021. 35–69. Houser, M., S.T. Marquart-Pyatt, R.C.H. Denny, A. Reimer, and D. Costello, A.B., and J.W. Osborne. 2005. Best Practices in Explora- Stuart. 2019. Farmers, Information, and Nutrient Management in tory Factor Analysis: Four Recommendations for Getting the the US Midwest. Journal of Soil and Water Conservation 74 (3): Most from Your Analysis. Practical Assessment, Research and 269–280. https ://doi.org/10.2489/jswc.74.3.269. Evaluation 10: 7. Howitt, R., D. MacEwan, and J. Medellin-Azuara. 2008. Calculat- Denny, R.C.H., S.T. Marquart-Pyatt, and M. Houser. 2019. Under- ing California Cropping Patterns in 2050. Davis. https ://water standing the Past and Present and Predicting the Future: Farm-.ca.gov/Legac yFile s/water plan/docs/cwpu2 009/0310 final/v4c03 ers’ Use of Multiple Nutrient Best Management Practices in a02_cwp20 09.pdf. the Upper Midwest. Society and Natural Resources 32 (7): Jara-Rojas, R., B.E. Bravo-Ureta, A. Engler, and J. Diaz. 2013. An 807–826. https ://doi.org/10.1080/08941 920.2019.15740 45. Analysis of the Joint Adoption of Water Conservation and Soil DeVellis, R.F. 2003. Scale Development: Theory and Applications, Conservation in Central Chile. Land Use Policy 32: 292–301. 2nd ed. Thousand Oaks: Sage Publications. Kanter, D.R., W. Winiwarter, B.L. Bodirsky, L. Bouwman, E. Boyer, S. Dillman, D.A., G. Phelps, R. Tortora, K. Swift, J. Kohrell, J. Berck, Buckle, J.E. Compton, T. Dalgaard, W. de Vries, D. Leclère, and and B.L. Messer. 2008. Response Rate and Measurement Differ - A. Leip. 2020. A Framework for Nitrogen Futures in the Shared ences in Mixed Mode Surveys Using Mail, Telephone, Interac- Socioeconomic Pathways. Global Environmental Change 61: tive Voice Response and the Internet. Social Science Research 102029. https ://doi.org/10.1016/j.gloen vcha.2019.10202 9. 38 (1): 1–18. Kara, E., M. Ribaudo, and R.C. Johansson. 2008. On How Environ- Feder, G. 1982. Adoption of Interrelated Agricultural Innovations: mental Stringency Influences Adoption of Best Management Prac- Complementarity and the Impacts of Risk, Scale, and Credit. tices in Agriculture. Journal of Environmental Management 88 American Journal of Agricultural Economics 64 (1): 94–101. (4): 1530–1537. https ://doi.org/10.1016/j.jenvm an.2007.07.029. https ://doi.org/10.2307/12411 77. Kassie, M., H. Teklewold, M. Jaleta, P. Marenya, and O. Erenstein. Feder, G., and D. Umali. 1993. The Adoption of Agricultural Innova- 2015. Understanding the Adoption of a Portfolio of Sustainable tions: A Review. Technological Forecasting and Social Change Intensification Practices in Eastern and Southern Africa. Land 43 (3–4): 215–239. Use Policy 42: 400–411. https ://doi.org/10.1016/j.landu sepol Fernández, V., and P.H. Brown. 2013. From Plant Surface to Plant .2014.08.016. Metabolism: The Uncertain Fate of Foliar-Applied Nutrients. Khalsa, S.D.S., and P.H. Brown. 2017. Grower Analysis of Organic Frontiers in Plant Science 4: 289. Matter Amendments in California Orchards. Journal of Envi- Fresco, L.O., and E. Westphal. 1988. A hierarchical Classification of ronmental Quality 46 (3): 649–658. https ://doi.or g/10.2134/ Farm Systems. Experimental Agriculture 24 (4): 399–419.jeq20 16.11.0456. Geisseler, D. 2016. Nitrogen Concentrations in Harvested Plant Parts— Khalsa, S.D.S., and P.H. Brown. 2019. Understanding Nitrogen A literature overview. University of California Davis. http://groun Cycling in an Irrigated Deciduous Permanent Crop. Acta Hor- dwate rnitr ate.ucdav is.edu/files /26913 8.pdf. ticultare 1253: 207–212. Ghadim, A.K., D.J. Pannell, and M.P. Burton. 2005. Risk, Uncer- Khalsa, S.D.S., T. Hartz, and P.H. Brown. 2018. Principles of Nitro- tainty, and Learning in Adoption of a Crop Innovation. gen Cycling and Management. University of California Agri- Agricultural Economics 33 (1): 1–9. https ://doi.or g/10.111 culture and Natural Resources. https://ucanr .edu/sites /S trategicI 1/j.1574-0862.2005.00433 .x.nitia tives /files /30116 8.pdf. Giller, K. 2013. Can We Define the Term ‘Farming Systems’? A Ques- Kipling, R.P., H.E. Taft, D.R. Chadwick, D. Styles, and J. Moorby. tion of Scale. SAGE Journals: Outlook on Agriculture 42 (3): 2019. Challenges to Implementing Greenhouse Gas Mitigation 149–153. https ://doi.org/10.5367/oa.2013.0139. Measures in Livestock Agriculture: A Conceptual Framework Good, A.G., and P.H. Beatty. 2011. Fertilizing Nature: A Tragedy of for Policymakers. Environmental Science and Policy 92: 107– Excess in the Commons. PLoS Biology 9 (8): 1–9. https ://doi. 115. https ://doi.org/10.1016/j.envsc i.2018.11.013. org/10.1371/journ al.pbio.10011 24. Kling, C., and C. Mackie. 2019. Improving Data Collection and Halbrendt, J., S.A. Gray, S. Crow, T. Radovich, A. Kimura, and B.B. Measurement of Complex Farms. Washington: National Acad- Tamang. 2014. Differences in Farmer and Expert Beliefs and the emies of Sciences, Engineering, and Medicine. National Acad- Perceived Impacts of Conservation Agriculture. Global Environ- emies Press. mental Change 28: 50–62. Knowler, D., and B. Bradshaw. 2007. Farmers’ Adoption of Conserva- Hanson, B., D. May, J. Simnek, J. Hopmans, and R. Hutmacher. 2009. tion Agriculture: A Review and Synthesis of Recent Research. Drip Irrigation Provides the Salinity Control Needed for Profit- Food Policy 32 (1): 25–48. https ://doi.or g/10.1016/j.foodp able Irrigation of Tomatoes in the San Joaquin Valley. California ol.2006.01.003. Agriculture 63 (3): 131–136. Koppmair, S., M. Kassie, and M. Qaim. 2017. The Influence of Farm Harter, T., J.R. Lund, J. Darby, G. Fogg, R. Howitt, K. Jessoe, S. Pet- Input Subsidies on the Adoption of Natural Resource Management tygrove, J. Quinn, and J. Viers. 2012. Addressing Nitrate in Cali- Technologies. Australian Journal of Agricultural and Resource fornia’s Drinking Water: With a Focus on Tulare Lake Basin and Economics 61 (4): 539–556. https ://doi.or g/10.1111/1467- 8489.12220. 1 3 800 J. Rudnick et al. Letey, J., and P. Vaughan. 2013. Soil Type, Crop and Irrigation Tech- Prokopy, L.S., K. Floress, D. Klotthor-Weinkauf, and A. Baumgart- nique Affect Nitrogen Leaching to Groundwater. California Agri- Getz. 2008. Determinants of Agricultural Best Management culture 67 (4): 231–241. Practice Adoption: Evidence from the Literature. Journal of Soil Levy, M.A., M.N. Lubell, and N. McRoberts. 2018. The Structure of and Water Conservation 63 (5): 300–311. https://doi.or g/10.2489/ Mental Models of Sustainable Agriculture. Nature Sustainability jswc.63.5.300. 1 (8): 413–420. Prokopy, L.S., K. Floress, J.G. Arbuckle, S.P. Church, F.R. Eanes, Y. Liu, T., R.J.F. Bruins, and M.T. Heberling. 2018. Factors Influencing Gao, B.M. Gramig, P. Ranjan, and A.S. Singh. 2019. Adoption Farmers’ Adoption of Best Management Practices: A Review and of Agricultural Conservation Practices in the United States: Evi- Synthesis. Sustainability (Switzerland) 10 (2): 1–26. https ://doi. dence from 35 Years of Quantitative Literature. Journal of Soil org/10.3390/su100 20432 . and Water Conservation 74 (5): 520–534. https://doi.or g/10.2489/ Lopus, S.E., M.P. Santibáñez, R.H. Beede, R.A. Duncan, J. Edstrom, jswc.74.5.520. F.J.A. Niederholzer, C.J. Trexler, and P.H. Brown. 2010. Survey Ranjan, P., S.P. Church, K. Floress, and L.S. Prokopy. 2019. Synthe- Examines the Adoption of Perceived Best Management Practices sizing Conservation Motivations and Barriers: What Have We for Almond Nutrition. California Agriculture 64 (3): 149–154. Learned from Qualitative Studies of Farmers’ Behaviors in the https ://doi.org/10.3733/ca.v064n 03p14 9. United States? Society and Natural Resources 32 (11): 1171– Lubell, M., and A. Fulton. 2008. Local Diffusion Networks Act as 1199. https ://doi.org/10.1080/08941 920.2019.16487 10. Pathways to Sustainable Agriculture in the Sacramento River Val- Reimer, A.P., D.K. Weinkauf, and L.S. Prokopy. 2012a. The Influence ley. California Agriculture 61: 131–137. https ://doi.org/10.3733/ of Perceptions of Practice Characteristics: An Examination of ca.v061n 03p13 1. Agricultural Best Management Practice Adoption in Two Indiana Lubell, M., M. Niles, and M. Hoffman. 2014. Extension 3.0: Man- Watersheds. Journal of Rural Studies 28 (1): 118–128. aging Agricultural Knowledge Systems in the Network Age. Reimer, A., A.W. Thompson, and L.S. Prokopy. 2012b. The Multi- Society and Natural Resources 27 (10): 1089–1103. https ://doi. Dimensional Nature of Environmental Attitudes among Farmers org/10.1080/08941 920.2014.93349 6. in Indiana: Implications for Conservation Adoption. Agriculture Marra, M., D.J. Pannell, and A.A. Ghadim. 2003. The Economics of and Human Values 29 (1): 29–40. https ://doi.org/10.1007/s1046 Risk, Uncertainty and Learning in the Adoption of New Agri- 0-011-9308-z. cultural Technologies: Where Are We on the Learning Curve? Reimer, A., J.E. Doll, B. Basso, S.T. Marquart-Pyatt, G.P. Robertson, Agricultural Systems 75 (2–3): 215–234. https://doi.or g/10.1016/ D. Stuart, and J. Zhao. 2017. Moving Toward Sustainable Farming S0308 -521X(02)00066 -5. Systems: Insights from Private and Public Sector Dialogues on McAllister, R.R.J., D. Smith, C.J. Stokes, and F.J. Walsh. 2009. Pat- Nitrogen Management. Journal of Soil and Water Conservation terns of Accessing Variable Resources Across Time and Space: 72 (1): 5A–9A. https ://doi.org/10.2489/jswc.72.1.5A. Desert Plants, Animals and People. Journal of Arid Environments Reimer, A., R. Denny, and D. Stuart. 2018. The Impact of Federal 73 (3): 338–346. https ://doi.org/10.1016/j.jarid env.2008.10.007. and State Conservation Programs on Farmer Nitrogen Manage- McIver, J., and E.G. Carmines. 1981. Unidimensional Scaling, vol. 24. ment. Environmental Management 62 (4): 694–708. https ://doi. Newbury Park, CA: Sage University Paper Series on Quantiative org/10.1007/s0026 7-018-1083-9. Applications in Social Science. Rogers, E.M. 2003. Diffusion of Innovations, 5th ed. New York: Simon McLellan, E.L., K.G. Cassman, A.J. Eagle, P.B. Woodbury, S. Sela, C. and Schuster. Tonitto, R.D. Marjerison, and H.M. Van Es. 2018. The Nitrogen Santos, J.R.A. 1999. Cronbach’s Alpha: A Tool for Assessing the Reli- Balancing Act: Tracking the Environmental Performance of Food ability of Scales. Journal of Extension 37 (2): 1–5. Production. BioScience 68 (3): 194–203. https://doi.or g/10.1093/ Schellenberg, D.L., R.D. Morse, and G.E. Welbaum. 2009. Organic biosc i/bix16 4. Broccoli Production on Transition Soils: Comparing Cover Crops, Mikkelsen, R. 2011. The ‘4R’ Nutrient Stewardship Framework for Tillage and Sidedress. Renewable Agriculture and Food Systems Horticulture. HortTechnology 21 (6): 658–662. 24 (2): 85–91. Murendo, C., S. Gwara, N. Mpofu, T. Pedzisa, and P. Chivenge. 2016. Shaffer, S. 2013. Encouraging California Specialty Crop Growers to The Adoption of a Portfolio of Sustainable Agricultural Practices Adopt Environmentally Beneficial Management Practices for Effi- by Smallholder Farmers in Zimbabwe. In 5th International Con- cient Irrigation and Nutrient Management. ference of the African Association of Agricultural Economists. Shaner, W., P. Philipp, and W.R. Schmehl. 1982. Farming Systems Addis Ababa, Ethiopia. Research and Development. Boulder, CO: Westview Press. Nitrogen Management Training Materials. 2019. California Institute Shaner, W., P. Philipp, and W.R. Schmehl. 2019. Farming Systems for Water Resources, University of California Division of Agri- Research and Development: Guidelines for Developing Countries. culture and Natural Resources. http://ciwr.ucanr.edu/Nitr og enMa New York: Routledge. ISBN: 12:978-0-367-01930-3. nagem ent/. Snapp, S., and B. Pound. 2008. Agricultural Systems: Agroecology and Norman, D.W. 1980. The Farming Systems Approach: Relevancy Rural Innovation for Development. Burlington, MA: Academic for the Small Farmer. East Lansing, Michigan: Michigan State Press. University. Snyder, C.S. 2017. Enhanced Nitrogen Fertiliser Technologies Sup- Nowak, P., S. Bowen, and P.E. Cabot. 2006. Disproportionality as a port the ‘4R’ Concept to Optimise Crop Production and Minimise Framework for Linking Social and Biophysical Systems. Society Environmental Losses. Soil Research 55 (6): 463–472. and Natural Resources 19 (2): 153–173. Stuart, D., R.L. Schewe, and M. McDermott. 2014. Reducing Nitrogen Osmond, D.L., D.L.K. Hoag, A.E. Luloff, D.W. Meals, and K. Neas. Fertilizer Application as a Climate Change Mitigation Strategy: 2015. Farmers’ Use of Nutrient Management: Lessons from Understanding Farmer Decision-Making and Potential Barriers Watershed Case Studies. Journal of Environment Quality 44 (2): to Change in the US. Land Use Policy. https ://doi.org/10.1016/j. 382. https ://doi.org/10.2134/jeq20 14.02.0091.landu sepol .2013.08.011. Pannell, D.J., G.R. Marshall, N. Barr, and A. Curtis. 2006. Adoption Taylor, R., and D. Zilberman. 2017. Diffusion of Drip Irrigation: The of Conservation Practices by Rural Landholders. Australian Jour- Case of California. Applied Economic Perspectives and Policy 39 nal of Experimental Agriculture, 2001: 1407–1424. https ://doi. (1): 16–40. https ://doi.org/10.1093/aepp/ppw02 6. org/10.1071/EA050 37. Teklewold, H., M. Kassie, and B. Shiferaw. 2013. Adoption of Mul- tiple Sustainable Agricultural Practices in Rural Ethiopia. 1 3 A farm systems approach to the adoption of sustainable nitrogen management practices in… 801 Journal of Agricultural Economics 64 (3): 597–623. https ://doi. Publisher’s Note Springer Nature remains neutral with regard to org/10.1111/1477-9552.12011. jurisdictional claims in published maps and institutional affiliations. Teshager, A.D., P.W. Gassman, S. Secchi, and J.T. Schoof. 2017. Sim- ulation of Targeted Pollutant-Mitigation-Strategies to Reduce Nitrate and Sediment Hotspots in Agricultural Watershed. Sci- ence of the Total Environment 607: 1188–1200. Jessica Rudnick is an Extension Specialist with California Sea Grant, Tomich, T., S. Bordt, R. Dahlgren, and K. Scow. 2016. The Califor- focused on human decision-making and governance in the Sacramento- nia Nitrogen Assessment: Challenges and Solutions for People, San Joaquin Delta. Dr. Rudnick completed her PhD in 2020 at UC Agriculture and the Environment. Oakland, CA: University of Davis in the Center for Environmental Policy and Behavior where California Press. her research focused on farmer decision-making and agricultural Tucker, M., and T.L. Napier. 2002. Preferred Sources and Channels of governance. Soil and Water Conservation Information among Farmers in Three Midwestern US Watersheds. Agriculture, Ecosystems & Environ- Mark Lubell is a Professor of Environmental Science and Policy at UC ment 92 (2–3): 297–313. Davis. He is the director of the Center for Environmental Policy and Ulrich-Schad, J.D., S.G. de Jalon, N. Babin, A. Pape, and L.S. Prokopy. Behavior, where his research focuses on cooperation and decision- 2017. Measuring and Understanding Agricultural Producers’ making in the context of agricultural and environmental policy. Adoption of Nutrient Best Management Practices. Journal of Soil and Water Conservation 72(5): 506–518. Sat Darshan S. Khalsa is an Assistant Project Scientist in the Depart- United States Department of Agriculture. 2018. “USDA Organic ment of Plant Sciences at UC Davis. His research focuses on nitrogen INTEGRITY Database.” https://or ganic.ams.usda.go v/integr ity/. cycling and management in tree crops integrating plant, soil and social United States Environmental Protection Agency. 2017. Nonpoint sciences. Dr. Khalsa has authored various papers on agricultural nitro- Source: Agriculture. https ://www .epa.gov/n ps/non po int-sour c gen use. e-agric ultur e. USDA Midwest Climate Hub. 2017. Agriculture in the Midwest. U.S. Stephanie Tatge completed her M.S. in International Agricultural Department of Agriculture. 2017. https://www .climatehub s.usda. Development at UC Davis in 2018, where her thesis leveraged spatial gov/hubs/midwe st/topic /agric ultur e-midwe st. analysis to investigate social networks and diffusion of information van Rooyen, J. 1984. Agricultural Economic Research in Less Devel- among farmers. She is now a Research and Outreach Specialist at the oped Countries: A Farm Systems Approach. Development South- Freshwater Trust in Sacramento, CA. ern Africa 1 (1): 56–64. https ://doi.or g/10.1080/03768 35840 84390 69. Liza Wood is a PhD student in Ecology at UC Davis, with a Designated Wauters, E., and E. Mathijs. 2014. The Adoption of Farm Level Soil Emphasis in Computational Social Science. Her dissertation focuses Conservation Practices in Developed Countries: A Meta-Analytic on governance of seed systems for impacts on biodiversity and social- Review. International Journal of Agricultural Resources, Govern- ecological resilience, in the context of climate change. Liza also works ance and Ecology 10 (1): 78–102. https://doi.or g/10.1504/IJARG on agricultural policy and farmer behavior in California, and the use of E.2014.06105 8. science in environmental policy. Wilson, R.S., G. Howard, and E.A. Burnett. 2014. Improving Nutrient Management Practices in Agriculture: The Role of Risk-Based Molly Sears is a PhD Candidate in the Department of Agricultural and Beliefs in Understanding Farmers’ Attitudes toward Taking Addi- Resource Economics at the University of California, Berkeley. Her tional Action. Water Resources Research 50: 6735–6746. https:// research is largely centered at the intersection of agriculture, envi- doi.org/10.1002/2013W R0152 00.Recei ved. ronmental quality, and policy. Molly’s recent papers include work on Zheng, C., Y. Jiang, C. Chen, Y. Sun, J. Feng, A. Deng, Z. Song, and W. recycling policy, groundwater salinity, and nitrogen management. Zhang. 2014. The Impacts of Conservation Agriculture on Crop Yield in China Depend on Specific Practices, Crops and Cropping Patrick H. Brown is a professor of Plant Sciences in the Department Regions. Crop Journal 2 (5): 289–296. https ://doi.org/10.1016/j. of Plant Sciences at UC Davis. His research focuses on the function cj.2014.06.006. and transport of nutrients in plants and the management of nutrients Zulfiqar, F., R. Ullah, M. Abid, and A. Hussain. 2016. Cotton Produc- in agricultural ecosystems. Dr. Brown has contributed significantly tion under Risk: A Simultaneous Adoption of Risk Coping Tools. to advancing research in his field and is dedicated to communicating Natural Hazards 84 (2): 959–974. https ://doi.org/10.1007/s1106 agronomic research through a robust and applied extension program. 9-016-2468-9. 1 3

Journal

Agriculture and Human ValuesSpringer Journals

Published: Sep 1, 2021

Keywords: Agricultural decision-making; Nitrogen management; Farmer adoption; Farm systems; Multivariate probit regression

There are no references for this article.