Contrasting Considerations among Agricultural Stakeholders in Japan on Sustainable Nitrogen Management

: Sustainable nitrogen (N) management in agriculture is one of the most important issues affecting the environmental performance of modern agriculture. It is actually well perceived that coordinated efforts and holistic approaches are required to regulate N use by farmers. The purpose of this study was to provide an initial examination of stakeholders’ views in Japan regarding N use in agriculture and challenges to increase its sustainability. The analysis was based on a questionnaire study of ﬁve types of stakeholders (farmers, advisors, researchers, suppliers, policy makers). By means of multivariate analysis techniques it was revealed that consensus was lacking either in the acknowledgment of the causes and effects of unsustainable N management or in the challenges that need to be addressed. N losses from farms and the effects of N use were perceived but not conceived equally by all stakeholders. Organic farming and mandatory measures were the most controversial challenges, while those involving awareness, training and advisory were the most popular. This study cannot provide safe conclusions that can be generalized in the Japanese context, but it indicates domains where further research is required and orientations for future policy design towards more sustainable N use.


Introduction
A common characteristic of modern production systems is the over-exploitation of natural resources such as water and soil in order to meet the growing demand for food worldwide. At the same time, the need for environmental protection, biodiversity conservation and sustainable production is a global requirement. In the challenge of increasing crop yields while securing sustainable use of natural resources, the role of nitrogen (N) is considered very important. The overuse of N in agricultural activities and its excess in the environment serve as primary examples of the type of urgent problem of agriculture, but N is also a limiting factor in production systems, considering that gains in production of important crops are a result of increased inputs of N [1]. To face this challenge, the adoption of sustainable N management is necessary with measures at local and global scales. Indeed, agriculture is the biggest anthropogenic source of nitrate, a major pollutant of surface and groundwater, and of nitrous oxide (N 2 O), a potent greenhouse gas accounting for 52% of the world anthropogenic N 2 O emissions on average in Japan has examined the attitudes and preferences of specific stakeholder groups (e.g., farmers) with regards to environmentally friendly agriculture and financial incentives to support it [18][19][20]. So far, no case study of N management and considerations of more stakeholders has been undertaken, although N pollution is of importance in policy. Farmers' and stakeholders' attitudes with regards to N use and pollution have been examined in existing studies as part of their broader preferences for other environment-related issues. The examination of the attitudes of various stakeholders regarding the use of N in Japanese agriculture could yield important considerations for future research and policy-making.
The purpose of this study was to provide an initial quantitative examination of the opinions and attitudes of stakeholders on the use of N inputs in the Japanese farming sector. In particular, the study built on data from a questionnaire survey of a sample of key stakeholders engaged in the farming sector in Japan. By means of established methods of multivariate analysis, the paper revealed significant divergences in their opinions and attitudes regarding the actual use of N inputs, which also extended to their evaluation of challenges towards more sustainable N fertilization in Japanese farming. Since this is one of the first studies to examine this issue in a quantitative way, the results can be used as a primary step towards recognizing research and policy priorities for more sustainable N use in agriculture in the country. For this reason, the results are discussed in conjunction with previous research and findings.

Survey Design and Administration
Within the Towards International Nitrogen Management System (INMS) project (inms.international), a questionnaire was developed, aiming to record and analyze the attitudes, opinions, interests and aspirations of agricultural stakeholders worldwide regarding the sustainable management of N inputs in agriculture. The questionnaire was designed to accommodate the diversity of stakeholders involved in N management in agriculture (farmers; researchers/teachers; policy makers; input suppliers; practitioners/farm advisors), therefore it comprised three parts. The first part recorded the profile of the respondent (gender, age, education) as well the stakeholder group where he/she felt that he/she mostly belonged. The second part was specific to each group and varied across stakeholders, while the third was common to all survey participants. In the latter, respondents were presented with 5-point Likert-scale questions aiming to evaluate their attitudes and opinions regarding N loss from farms, the effects of N fertilizers, farming practices and sources of N pollution in the country. A separate Likert-scale question presented 17 challenges to respondents regarding the use of N fertilizers.
The survey was oriented towards the key stakeholder groups involved in sustainable N management, as identified within the INMS project. An online anonymous survey was conducted in Japan using the questionnaire translated into Japanese from 25 May 2020 to 26 August 2020, calling for expert networks (researchers and consultants) and personal networks and social media (other stakeholders). For this reason, the sample was not random and there was no control on the number of respondents from each stakeholder category. Responses were automatically recorded in a dedicated online spreadsheet, where Likert-scale variables were coded in a scale from 1 (=Not at all/Totally disagree) to 5 (=Very much/Totally agree).

Methodological Approach and Data Analysis
The methodological framework for the analysis of the categorical (ordinal) data in this paper combined (i) A descriptive analysis of responses to acquire a general view of respondents' attitudes and perceptions. The significance of differences across stakeholder groups was tested with the Kruskal-Wallis non-parametric test and with post hoc tests for pairwise comparisons, as the normality assumption for parametric tests did not stand, also due to differences in sample sizes across groups. (ii) A Categorical Principal Component Analysis (CatPCA) on categorical variables (items) describing the challenges that were presented and evaluated by respondents, in order to reduce the number of variables and derive fewer (in number) and meaningful variables (dimensions) explaining the broader framework. (iii) Using the dimensions yielded by the CatPCA, a Multinomial Logit Model was used (Categorical Principal Component Logistic Regression-CPCRL) to examine potential differences in how each group of stakeholders regarding different dimensions of challenges (i.e., policy options).

Categorical Principal Component Analysis
Respondents' attitudes regarding challenges for future N use sustainability were analyzed with CatPCA in order to investigate the internal validity of specific survey variables. The method is very useful in multivariate analysis, as it reduces an original set of variables into a smaller set of uncorrelated components (dimensions) that represent most of the information (variance) found in the original variables [21]. For each variable in the original data set a correlation coefficient (factor loading) is estimated with each dimension, which is the criterion by means of which the dimension can be identified. As a rule of thumb, loadings higher than 0.7 indicate that the variable is correlated to the dimension, however in social sciences the subjective opinion of the researcher is also important for the identification task. The uncorrelated components (dimensions) are new variables that can be used in further analysis, so the researcher is able to deal with fewer variables without losing valuable information. In the application in this paper, the statistical package SPSS 24 was used for the analysis.

Categorical Principal Component Logistic Regression
A Multinomial Logit (MNL) model, widely applied in the social sciences literature [22,23], was employed in order to examine how different stakeholder groups evaluated challenges regarding sustainable N use. The basic Multinomial Logit Model is specified as follows: where P is the probability that outcome j occurs, X i are independent variables and β k the corresponding coefficients. This formulation of the logit model is estimated by the maximum likelihood estimation method. The most common goodness-of-fit measure is the McFadden pseudo-R2, which is a counterpart to R-square for linear regression. The Wald-statistic is a counterpart to the well-known t-statistic. The MNL model used in this paper is an application of CPCRL. In particular, the set of challenges presented to respondents was reduced to a smaller set of dimensions using CatPCA and then the new variables (dimensions) were used as the regressors (independent variables) in the multinomial logistic regression. The dependent variable accounted for the five stakeholder categories (farmers, consultants, researchers, policy makers, suppliers). The model was estimated using the econometric software package NLOGIT 6.0. Table 1 presents the characteristics of the sample. Almost half of the respondents (49%) identified themselves as researchers, followed by advisors/consultants (20%). Policy makers and representatives from the supply industry were equally represented (10% each), while the remaining 16 respondents (11%) were farmers. The sample showed an overrepresentation of women, which exceeded the gender balance of the Japanese population. Also, almost all respondents held a Bachelor's or post-graduate degree, where, as expected, 96% of respondents categorized as "Researchers" had post-graduate studies. On the contrary, the age balance of respondents was more indicative of the population, as the majority of respondents belonged to the 31-45 (38%) and 46-60 (37%) age groups. Apart from education, there were no important deviations from the average characteristics across stakeholder groups.

Sample Profile and Characteristics
In Japan farmers are categorized as follows in terms of their involvement in farming: (i) Full-time farmers have more than 50% of their income as agricultural income, are under 65 years old and have more than 60 days of agricultural work; (ii) Semi-full-time farmers have less than 50% of their total income as agricultural income and household members under the age of 65 who are engaged in self-employed farming for 60 days or more a year; (iii) Sideline farmers have no household members under the age of 65 who are engaged in self-employed farming for 60 days or more a year. Farmers in the sample were mostly full-time farmers (56%), followed by semi-full-time (25%) and sideline farmers (19%). That was largely different from the mean composition in Japan in 2020, that is, 22%, 14%, and 64%, respectively [24], which implies that farmers who participated in the study were more active and practical than the national average. Also, the sample composition of this survey differed greatly from the mean occupational population ratios in Japan, meaning that the number of stakeholders related to agriculture exhibited a small proportion in Japan. The ratio of agriculture and forestry (as industry) in the number of employed persons aged 15 or over was low and decreasing from 4.7% in 2000 to 3.5% in 2015 [25]. That ratio of scientific research, professional and technical services (including all fields not limited to agriculture) was 3.1-3.3% [25].  Table 2 presents responses regarding four domains related to N use in agriculture. With regards to attitudes about N losses, most respondents were strongly in agreement with the fact that the reduction of N losses would make a difference (4.25 average score) and they were not aware of having N losses from their farm (3.61). When it came to the views of specific groups, there was general consensus that their reduction would lead to environmental benefits. Stakeholders also agreed-however, to a lesser degree-that farmers were not fully aware of such N losses in their farms and that they did not know of specific ways to reduce them, although they affected farm productivity. Contrasting views, however, were revealed regarding the extent of such N losses (between farmers, consultants and researchers) as well as the potential costs of interventions to reduce them (between consultants and researchers). In addition, suppliers supported the necessity to modernize farms in order to reduce N loss, contrary to farmers and researchers. Interestingly, diverging opinions were recorded regarding the extent to which farmers actually expected that their behavior would make a difference, as farmers, consultants and researchers expressed neutral positions, contrary to suppliers, who confirmed disbeliefs in the actual interest that farmers had on the matter.

Attitudes and Opinions Regarding the Use and Effects of N Fertilizers and N Pollution
The environmental and economic effects of N fertilizers were generally highly acknowledged by respondents, especially with regards to adverse effects on water resources (4.23) and to the necessity to be used carefully (4.28). There was consensus with regards to consequences on productivity, water resources and soil. However, concerning air quality and climate change, farmers expressed more mediocre opinions compared to other stakeholders-especially suppliers and researchers-while significant divergence was also revealed between farmers and policy makers about effects on biodiversity. This dichotomy was also confirmed regarding the careful use of N fertilizers, as farmers were significantly less preoccupied with the issue compared to almost all other stakeholder groups.
Concerning respondents' opinions about N fertilization practices, respondents were slightly opposed to items concerning soil analysis to inform the use of N fertilizers by farmers (2.83); the availability of fertilizers as a barrier reducing their use (2.80); and the fact that their practices were oriented to the reduction of the environmental effects of N fertilizers (2.45). However, different levels of acceptability and contrasting views were expressed. There was consensus about the fact that farmers used fertilizers compatible to their machinery, which also achieved the highest mean score across all respondents, and also about the fact that farmers' choices of N quantities were not always aimed at higher economic performance. Stakeholders' opinions also converged regarding the fact that lack of accessibility was not a serious issue reducing the use of N fertilizers. Respondents also expressed agreement with the statement that "It's not possible to be a crop farmer without using N fertilizers", but farmers and policy makers expressed much lower levels of agreement compared to other groups-especially consultants and suppliers. Two additional significant points of divergence included environmental pressures and soil analyses. Regarding the former, farmers supported that their use of N fertilizers was low with a view to decrease environmental pressures, while all other stakeholders disagreed on that, especially researchers. Similarly, farmers reported that they performed soil analyses and adjusted their practices accordingly, but researchers and policy makers stated opposite opinions. Regarding the use of cheap fertilizers, significantly contrasting opinions were detected, but post hoc tests did not reveal specific disagreements between groups.
Chemical N fertilizers (4.06) and livestock production (3.98) were acknowledged as the most important sources of N pollution in the country, as for this item the highest mean was calculated. However, a dichotomy was detected between the views of consultants and researchers. Contrasting views were also pointed out regarding organic fertilizers (e.g., manure/compost, slurry, crop by-products), although no significant differences across groups could be reported. On the contrary, farmers did not acknowledge the transport sector as a source of N pollution, which diverged significantly from the point of view of researchers and consultants. The importance of other polluters, such as livestock production, industry, urban activities, and energy production, was also pointed out, but without significant differences across groups.   Table 3 presents the results of the CatPCA on the set of 17 variables (items) describing challenges for sustainable N use in agriculture. The analysis yielded five principal components (Dimensions) with eigenvalues higher than 1, which accounted for 67.7% of total variance. These dimensions were kept for the analysis. Each one of them was named/characterized by the variables that had the higher loadings and were thus highly correlated with the specific dimension, as follows.

•
Dimension 1 "Targeted measures, research and training/advisory". The highest loadings were found for variables describing environmentally friendly techniques, targeted measures and research needs, as well as education, training and awareness requirements for farmers (including simple and understandable information and proposal of voluntary measures) and the general public (also communication). • Dimension 2 "Mandatory measures" included two items that described compulsory measures and critical thresholds to decrease the use of N. • Dimension 3 "Organic production" described challenges relating to organic practices, especially regarding the use of organic fertilizers and the further adoption of organic production in general. • Dimension 4 "Measures to control fertilization costs" involved challenges to reduce costs and increase efficiency of N fertilizers, including collective provision, price regulation and more effective practices. • Dimension 5 "Top-down approaches" encompassed policy-driven measures to regulate the use of N fertilizers, which involved subsidies to increase N use and a "top-down" shift to crops requiring less N (crop restructuring).  Table 4 presents the results of the CPCRL. According to the partial effects, there were not many dimensions that interpreted attitudes and pursuits of stakeholder groups significantly. For farmers, Dimensions 2 and 3 seemed to have a significant effect, showing that higher interest in challenges relating to mandatory measures and organic production affected the probability that one respondent belonged to this group. On the contrary, researchers expressed significant interest in mandatory measures and so did policy makers. Suppliers were mostly interested in issues categorized in Dimension 5, relating to topdown approaches to reduce N losses and increase efficiency. It should be noted that, for consultants, none of the five dimensions was found to be significant in explaining probability of categorization, while challenges in Dimension 1 and 4 were not significant explanatory variables for any of the five stakeholder categories. * denotes significance at p < 0.1; *** denotes significance at p < 0.01. Table 5 presents the specific challenges that were categorized in the five dimensions, along with the average score of responses for each one of them per stakeholder category. It is obvious that measures grouped in Dimension 1 achieved the highest averages, which showed that challenges relating to innovation and training/advisory were the most relevant to sustainable N management. Combined with CPCRL analysis, respondents' opinions did not diverge significantly and were relatively high for all items in this dimension. Challenges including organic production and cost reduction-grouped in Dimensions 3 and 4-also received high levels of acceptance overall-especially among farmers-except for the importance of the role of cooperatives. On the other hand, the least preferred challenges were those relating to top-down approaches and mandatory measures. In particular, mandatory measures to regulate N use-either to increase or decrease it-were the least favorite options and this also applied to crop restructuring, if it was "imposed" top-down. The establishment of critical thresholds by law was more acceptable than other challenges grouped in these two dimensions-however, it was the least preferred by consultants and farmers.

Discussion
The findings of this study pinpointed that advisory and training for farmers constitute key challenges that should be considered toward sustainable N use. The Ministry of Agriculture, Forestry and Fisheries (MAFF), Japan, promotes organic farming to save the natural functions in the agricultural biosphere [26], providing supports to farmers, especially to increase new farmers [27]. MAFF created a web portal to support agricultural corporations and new agricultural applicants to increase new farmers. Apart from training and advisory for farmers, this study also revealed that public awareness was deemed an important challenge for increasing the sustainability of N use in Japanese agriculture. Sasaki [19] showed that there were positive relations of cause and effect between agrienvironmental knowledge and consciousness of the environment based on a questionnaire survey in Shiga Prefecture, Japan.
The analysis in this paper revealed that voluntary policy measures were among the most preferred options to increase the sustainability of N use in agriculture. This was in line with the research presented by Kuriyama et al. [18], who deployed an experimental economic analysis for farmers and found that a subsidy for conservation has continuous effects on conservation behavior, while the conservation effects depend on the amount of subsidy. This system has been found effective in other areas-such as the European Union-in the form of Payments for Ecosystem Services, where farmers receive policy support contingent upon the provision of environmental services [28]. On the contrary, a non-monetary support system has a significant effect on conservation behavior only in its early stage, thereafter, financial incentives are needed to sustain the effect [18].
Awareness raising for the general public also received high score in this study, in line with Uetake and Sasaki [20]. In their study, farmers' profits would increase by means of an environmental payment to reduce chemical fertilizer use if the selling price of fertilizerreduced crops was higher than that of conventional crops-which relates to consumer awareness. The same study also reported that direct environmental payments were effective to increase rice production and to improve environmental quality by enhancing crop N uptake and carbon storage. A 100% reduction of chemical fertilizers was not necessarily good compared to a 50% reduction since ammonia loss from manure might increase. Therefore, the finding of this study that N management measures would achieve environmental benefits coincided with the study by Uetake and Sasaki [20], which also applied to the fact that this study revealed a very high support for environmentally friendly techniques for efficient fertilization. This also applied to the finding of this study that input suppliers were in favor of modernizing the sector in order to increase sustainability of N use.
The results of the analysis demonstrated that two issues were the most contradictory. The first was the imposition of mandatory issues and top-down solutions. The low acceptability of a shift to crops with less N requirements showed that crop restructuring as an induced policy option is not acceptable. Mandatory measures were highly appreciated by policy makers, but received considerably lower scores from other stakeholders, which showed the disparity of preferences and their conflicting attitudes. Farmers and consultants, actual implementers of measures, particularly expressed much lower acceptability for such the measures. On the other hand, researchers supported a shift toward less N-demanding crops relatively more than other stakeholders. A survey conducted in Denmark by Case et al. [29] showed the importance of policies to a higher adoption of better N management measures. In their study, most farmers agreed that the restrictions imposed by national and supranational policies (e.g., Nitrate Directives, The Clean Water Act) was a key factor in enhancing N efficiency from manure application/use.
In addition, as expected, suppliers welcomed-more than others-measures to subsidize the use of N fertilizers, as in the example of developing countries in East Africa. However, the survey conducted by Rakshit [30] from 2015-2016-to evaluate the effects of the 2010 fertilizer subsidy measure-showed that this was not the best solution to promote agriculture productivity or the reduction of fertilizer losses. This finding was interesting, when considered alongside the low acceptability of measures relating to collective pro-vision. Japan has a long history of agricultural cooperatives, that is, Japan Agricultural Cooperatives (JA), which was enacted in 1947 with the Agricultural Cooperative Law but can be traced back to the traditional cooperatives in the Edo era (1603-1868) [31,32]. The current JA movement is comprehensive, covering almost all needs of both the farming and non-farming communities including credit, mutual insurance, purchasing, processing, marketing, and welfare [31]. In fact, 74% of the fertilizer sales to farmers can be traced back to JA while the National Federation of Agricultural Cooperative Associations (ZEN-NOH) contributed to 50% of the procurement of raw fertilizer materials and 55% of the distribution of fertilizers in 2013 [33]. Although no significant discrepancies were detected across stakeholders in this study, the low score of this measure could reflect the fact that stakeholders did not consider that encouragement was required to enforce the role of JA in fertilizer procurement. In addition, the integration of cooperatives in the country, which led to a decrease from 3591 in 1990 to 719 in 2010, largely aimed to strengthen the purchasing and sales business of JA [34]. In addition, this situation partially explained why suppliers were strongly in favor of a further reinforcement of the procurement and distribution system already established by JA and their associations, as they obviously preferred to sell to cooperatives and thus to reduce uncertainty and transaction costs.
The second contradictory issue was organic production. In order to adopt organic practices, the positive attitude of farmers toward new measures and innovations to support organic production was proposed by this study and respondents were found to be in favor. There is significant evidence that certain practices relating to organic production, such as the use of composts, can be advantageous to promote residues treatment and nutrients recycling, besides the advantage of organic matter addition to soils. Benefits include reduction of mass, volume, and water content compared with fresh manure, which in turn reduces transportation requirements. Concomitant benefits also involve elimination of pathogens, parasites, weed seeds and odor emissions on land application via the fermentation process. Compost also enhances soil physical and biological properties and has a disease suppression effect [35,36]. However, the lack of synchrony between crop needs and nutrient release from composts may, in turn, increase ammonia and nitrate loss and, depending on the source of the material composted, there is an associated risk of undesired element build-up in soils. This leads to the necessity of taking close care about organic practices to avoid trade-offs and pollution swapping.
MAFF promotes organic farming in Japan based on the Act on the Promotion of Organic Agriculture enacted in 2006. However, the area for organic farming is still very small, only less than 0.6% of the total cropland area of Japan [33]. In addition, as explained in Miyake and Kohsaka [37], organic production in Japan is in need of better contextualization and definition. Therefore, the findings of this study could set a good starting point to discuss and design policies targeted to the needs of farmers. However, positive attitudes about organic farming by farmers in this study could be partially due to the over-representation of full-time farmers in the sample, who are generally more active and expected to be more inclined toward organic production and environmental protection through practices that include minimal use of N fertilizers and performance of soil analyses to adjust their practices accordingly. Thus, the policies to support farmers' eagerness to adopt new measures and innovations for organic production, instead of mandatory requirements, could be more feasible to start the discussion. In this domain, the role of agricultural cooperatives should be considered because of their current orientation toward bulk provision of chemical fertilizers.
Among the public goods treated by Japanese agri-environmental policy, that is, landscapes, biodiversity, water quality, soil quality and protection, climate change, air quality, and resilience to natural disaster [10], water quality and climate change are closely related to the agricultural N use. With respect to N management, Japanese agri-environmental policy treats water quality as the most important public good [10]. Livestock wastewater and runoff and leaching of N from croplands as non-point sources have caused water pollution [38]. That was reflected well in the survey, in the high agreement with "Bad for water resources" (Table 2). Water quality is regulated by relevant laws, which focus on achievement of the environmental standards (i.e., nitrate and total N concentrations in water) and regulate discharge from livestock farming. Water quality is linked to both human health as drinking water and ecosystem health via eutrophication. Concentrations of nitrate and total N in terrestrial water have been improved since the beginning of the 2000s, whereas the environmental standards of water pollution have often been violated in groundwater for nitrate and in lake water for total N [39,40]. The mean ratio of agricultural areas that exceeded drinking water limits for nitrate in groundwater was 5% in 2000 and the level of exceedance was 17th out of 23 countries [41]. However, the groundwater nitrate pollution tends to be persistent as shown in that exceedance of the nitrate standard was 17.3% in the survey of the areas adjacent to polluted wells and 42.6% in the monitoring survey of polluted wells in 2018 [40]. A general improvement in water quality regarding nitrate concentrations is usually accompanied by the existence of contamination/pollution hotspots due to agricultural activities, as demonstrated by [42,43]. This means that a closer look must be given to local and regional aspects of N losses, and all actors must be involved in the development of sustainable N management.
The highest agreement of policy makers with "Bad for biodiversity" was also interesting compared to the existing situation in the country, because the Japanese agrienvironmental policy does not directly relate the N issue to biodiversity conservation (rather relating agrochemicals and invading species intendedly or unintendedly introduced with agricultural practices to biodiversity conservation). Note that the biodiversity aspect of the Japanese agri-environmental policy is to conserve farm areas including surrounding areas as agro-ecosystems. In the same context, opinions regarding N use and soil quality were not contrasting and the average was lower than for other natural resources, although soil quality is also subject to agricultural N use since a long-term or heavy use of N fertilizer can acidify cropland soil without soil amendments to neutralize, like lime [44].

Conclusions
The analysis of survey data in this study provided some initial support to the argument that agricultural stakeholders in Japan recognize the importance of agricultural N effects and of related N loss from farms. In addition, average responses regarding the adoption of sustainable N management practices showed that significant room for improvement is deemed. However, stakeholders' views diverged in several statements in all domains under examination in this study, including the challenges that need to be addressed toward sustainable N use.
The main conclusion that can be derived from this study is that the different views of stakeholders in key issues could be one of the causes of N surplus in the country, as consensus was lacking either in the acknowledgment of the causes and effects of unsustainable N management or in the challenges that need to be addressed. Thus, the present study demonstrated that it is worth taking a step forward to elucidate the awareness in Japan of sustainable agriculture from the viewpoint of N management, including information on the potential co-benefits and drawbacks of N losses. Among the most important findings that could be considered in policy design is that there was a significant positive attitude of farmers toward new measures and innovations to support organic production. This could be a good starting point to discuss policies targeted to the needs of farmers, also providing a hopeful future for sustainable agriculture in Japan.
The basic limitation of this study is that the sample is too small and not representative of key agricultural stakeholders in Japan. For this reason, by no means could the results reported here be used directly in policy-making. However, this is one of the first studies tackling attitudes and opinions about sustainable N management in Japan and the first one revealing contrasting considerations among agricultural stakeholders. The study results converging with previous research show that these findings could provide the basis to initiate policy debates and more widespread research in the field, with the examination of larger representative samples of respondents nationwide or in particular areas with high levels of N use in agriculture. Such future studies could be supported and understood better through dedicated stakeholder analyses to grasp how each group perceives the role of other groups, but also their particular interests and perceptions regarding other environmental issues, cultural and societal values. Indeed, stakeholders' communication was found to be necessary and beneficial to better N management behaviors. Also, more research regarding the role of consumer-driven change in N management is required, as in the example described by Parker [45], who discussed an alternative form to enhance consumer-citizenship to have a say in what and how produce is grown and how the land is managed. Assessments of the roles and opinions of the general public are required toward this end, as in the study by Sasaki [19], according to which there were negative relations of cause and effect between food safety consciousness and consumers' evaluation of agri-environmental policy.