Assessing nitrogen flow and nitrogen footprint in the food system of a subtropical island with a scenario to mitigate nitrogen load impacted by trade-dependent agriculture

Recent hikes in fertilizer, feed, and food prices threaten the food security of island-dwelling people who rely heavily on imports to sustain food supply and production. The influx of reactive nitrogen (Nr) through imports increases nitrogen load and degrades the environment. To overcome these problems, a robust and sustainable food system must be developed. In this study, we aimed to evaluate the present nitrogen flow in the food system of Ishigaki Island, located in the subtropical zone of Japan, and propose a measure to improve it based on the nitrogen footprint concept. Results showed that the major Nr-loss pathways for agricultural activity on the island were ‘crop-unused’ (37%) and ‘manure’ (43%). In food production, most of the Nr loss to the environment was related to export products, and less than 30% was related to island consumers. To meet the demand of food supply on the island, 5.1 times greater amount of food Nr than that of produced for island consumers was imported from overseas regions, placing the burden of Nr loss on such regions. We found that agricultural activities on the island mainly used chemical fertilizer; less than 13% of cattle manure was reused. To reduce the influx of Nr, we created a scenario in which 30% of chemical fertilizer was replaced by cattle manure. Results indicated 70% of the cattle manure produced on the island was necessary to achieve this scenario. This system could reduce Nr imports and Nr loss on the island by 16% and 17%, respectively. The proposed food system can be extended to other islands to overcome the recent price hikes and conserve the environment. This study is the first to present a detailed nitrogen flow in the food system of a tropical/subtropical island by using the nitrogen footprint concept.


Introduction
The recent price hikes in fertilizers, feeds, and food due to the COVID-19 pandemic and the Russia-Ukraine War (FAO 2022) threaten the global food system and people's livelihoods and food security. This unstable global situation seriously affects small islands because they heavily rely on the import of goods such as food and chemical fertilizers, to sustain food supply and production (e.g. Maesato et al 2016). Moreover, agriculture also has a negative impact on the environment. Various human activities generate reactive nitrogen (Nr; all forms of nitrogen except di-nitrogen [N 2 ]), but the Nr amount per year already exceeds natural levels (Fowler et al 2013). More than 75% of this Nr is utilized in food production processes, such as chemical fertilizer nitrogen application and biological nitrogen fixation (BNF) in farmlands. Consequently, the nitrogen cycle is disturbed to a degree considerably beyond the 'planetary boundary' (Rockström et al 2009). Inefficient use of fertilizers causes considerable loss of Nr to the environment through processes such as volatilization, leaching, and run-off. This phenomenon pollutes the atmosphere, groundwater, and surface water bodies, causing health hazards to inhabitants and damaging the environment (UNEP 2019). To reduce the release of Nr into the environment, we should minimize the use of 'new' Nr, such as chemical fertilizers, and promote the efficient reuse of 'old' Nr, such as livestock manure (Shibata et al 2017, Eguchi et al 2018. Recently, differences in nitrogen balance have been identified among regions (Schulte-Uebbing et al 2022). The nitrogen cycle in a specific area must be assessed first to reduce nitrogen load (Shibata et al 2017, McCourt andMacDonald 2021). Among various regions, islands have been the focus of concern regarding the increase in nitrogen load due to the influx of Nr through imports of fertilizers, feeds, and food. Marine areas of tropical/subtropical islands, such as the Coral Triangle in the western Pacific Ocean, play a major role in maintaining biodiversity; however, they are severely degraded by climate change and nitrogen loading (Hoegh-Guldberg 1999, Hoegh-Guldberg et al 2009. On many tropical/subtropical islands, tourism is the main industry and source of income. Overdependence on imported Nr and its release into the environment could collapse the food systems and economies of islands. Moreover, Eguchi et al (2018) noted that food imports provide extra Nr release to overseas regions with agricultural production. Thus, a robust and sustainable food system that does not rely heavily on imports and conserves the natural environment is essential not only for tropical/subtropical islands but also for overseas regions. To achieve this goal, the nitrogen flows in the food systems of islands should be evaluated. However, few studies have conducted such an evaluation, and others (e.g. Goto et al 2017) have only focused on nitrogen flow related to a specific crop, such as sugarcane.
In this study, we selected Ishigaki Island as the study area. The island has developed export-oriented agriculture and livestock industries, such as sugarcane production and beef calf breeding, which depend on imported chemical fertilizers and concentrate feeds. Similar to the Coral Triangle, Ishigaki Island is also rich in tourism resources, such as coral reefs, but faces environmental problems, such as nutrient runoff from chemical fertilizers and livestock manure (Miyamoto et al 2018). Therefore, agricultural promotion and environmental conservation on Ishigaki Island deserve attention.
Farmers alone cannot improve the nitrogen flow in the food system. The cooperation of consumers, who are the main driving force of nitrogen flow in the food system, is also important (Eguchi and Hirano 2019). The nitrogen footprint (Leach et al 2012) is a simple quantitative indicator of the reality and problems of the nitrogen cycle and is useful for sharing results among stakeholders, including consumers. A major advantage of the nitrogen footprint concept is that it can express 'new' Nr input and Nr loss (virtual nitrogen) during food production to before the consumption process for each food item.
In this study, we aimed to evaluate the present nitrogen flow in the food system of Ishigaki Island, located in the subtropical zone of Japan, and propose a measure to improve it based on the nitrogen footprint concept. We considered nitrogen flow not only inside the island, but also in overseas regions where imported feed and food is produced, which is important to consider global scale nitrogen issue (Eguchi et al 2018). We created and examined a scenario to mitigate Nr loss on the island considering the Japanese agri-environmental policy to reduce chemical fertilizer use by 30%, which is a key performance indicator of MeaDRI, the Strategy for Sustainable Food Systems in Japan (MAFF 2021).
This study is the first to present a detailed nitrogen flow in the food system of a tropical/subtropical island by using the nitrogen footprint concept, which has been applied on the national (e.g. Galloway et al 2014, Shibata et al 2014) and provincial (McCourt and MacDonald 2021 scales. This study focused on the food system of Ishigaki Island because the majority of the nitrogen footprint is related to food (Galloway et al 2014, Shibata et al 2017; in addition, the island requires a robust food system.

Study site
Ishigaki City (24 • 20 ′ 26" N; 124 • 09 ′ 20" E) is located 280 km east of Taiwan and in the subtropical climate zone. The city is composed of one main island (Ishigaki Island, 222.24 km 2 ) and 13 uninhabited islands. The population is 49 392 (Ishigaki 2018). Agriculture and tourism are the two main industries on this island. Approximately 1 million tourists spent an average of 3.5 d on the island before the COVID-19 pandemic (Ishigaki 2018), indicating that tourists contributed to an increase in the island population by 10 000 people. The agricultural land on the island spans 5 330 ha. In crop production, sugarcane and pineapple are the main products, and most of them are for export. Sugarcane residue (bagasse) is mainly used as fuel in a sugar factory on the island. In livestock production, approximately 23 000 heads of cattle (mostly calves) are bred mainly for export, with a relatively small number of dairy cattle, pigs, broilers, egg-laying hens, goats, and horses. Live animals other than beef calves, goats, and horses are sent outside the island for slaughter. The self-sufficiency rate of coarse feed on the island is close to 100% (according to interviews), whereas that of concentrate feed is 0%. Natural tuna and non-fed seaweed are the main seafood products of the island. Garbage, such as food processing waste and food loss is treated at a garbage animal, and (c) seafood production in the food system of Ishigaki Island, associated with the import/export of reactive nitrogen (Nr). New Nr includes chemical fertilizer Nr and BNF Nr, which are newly generated from N2. Old Nr refers to cattle manure in this study. We considered the recycling of livestock excreta Nr, crop residue Nr, and food processing waste Nr. Black boxes with white letters indicate Nr loss to the environment. We calculated Nr removal from human excreta by denitrification during sewage treatment. Seafood (seaweed and fish) production was calculated from captured 'live animal Nr' because on Ishigaki Island seafood is produced in natural or non-fed aquaculture.
incineration center on the island. All the chemical fertilizers necessary for crop and coarse feed production are imported. As described above, large amounts of resources are transferred between the island and overseas regions. Figure 1 illustrates the nitrogen flow in the food system of Ishigaki Island. Nitrogen flow in the year 2018 was calculated using official statistical data (Ishigaki 2018). Major crops and livestock listed in the statistics were used as agricultural products in the calculation. For unavailable data on the city scale, statistical data on food/feed production on the prefectural and national scales, as well as data from previous studies, were applied. See supplementary tables 1-4 for details.

Calculation assumption
For sugarcane, we considered BNF based on Ando et al (2001), who reported that 1/3 of nitrogen in sugarcane bodies is supplied by BNF. Based on the interview, we assumed that the bagasse ash left over after the combustion process in the sugar factory is returned to the sugarcane fields as fertilizer. We took pumpkin as a representative of vegetables because it is the most widely cultivated vegetable on Ishigaki Island (Okinawa 2012). With the exception of sugarcane, we assumed food processing waste and food loss on the island were incinerated, i.e. their Nr was released into the environment. All livestock excreta were assumed to be converted into manure. The nitrogen volatilization from livestock barns and during the conversion was calculated in accordance with NIES (2022) (supplementary table 5). For simplification, all manure Nr used on the island was assumed to be cattle manure Nr because the amount of other livestock manure Nr was relatively small (see section 3.1.1.). Based on the interviews, we assumed that no manure was applied to the sugarcane or pineapple fields or grasslands. In beef cattle production, we considered grazing and assumed that excreta was returned to the grassland during the grazing period. In this study, we did not consider denitrification and ammonia volatilization from agricultural land. We assumed that no feed was applied in seafood production.
All concentrate feed and chemical fertilizer were assumed to be imported. Import of food was calculated to meet the demand of food supply of the island (see section 2.2.5.). Exports were considered for the major export-oriented products of beef calves (live animals), sugarcane (coarse sugar), pineapples, vegetables, and seafood (seaweed and natural tuna). We also regarded the livestock shipped out for slaughter (dairy cattle, pigs, broilers, and egg-laying hens) as export. We could not find data regarding milk and egg exports. Thus, for this study, we assumed no export of milk or eggs occurs on Ishigaki Island. The removal of human excreta Nr by denitrification during sewage treatment was calculated as previously described by Shibata et al (2014).

Nr loss to the environment for each production
For each of the major food items produced on the island, Nr losses from different pathways (cropunused in the field, crop processing waste, livestock manure, slaughter waste, food processing waste, food loss, and human excreta; figure 1) were calculated to determine large pathways of nitrogen release on the island. We also estimated the Nr loss in overseas regions for concentrate feed production to evaluate the influence of livestock production on environmental load. Assuming that concentrate feed was primarily formulated from corn, we calculated the amount of Nr loss and Nr input during grain corn production in overseas regions using parameters for the U.S. reported by Leach et al (2012) because Japan imports corn mainly from the U.S. (Ministry of Agriculture, Forestry and Fisheries (MAFF 2021)).

Nitrogen footprint for food for island consumers
The nitrogen footprint is defined as the sum of Nr losses to the environment due to various human activities. It is generally determined in relation to consumer behavior (Leach et al 2012). In this study, we focused on products that are produced and consumed within the island, because these products affect the nitrogen load on the island directly through the consumers' diets. We estimated the nitrogen footprint excluding Nr loss to the environment for exported production and imported food to identify the island products for the consumers' diets that affect the nitrogen load the most. Per-capita nitrogen footprint for food for island consumers was calculated based on Nr loss to the environment (both on Ishigaki Island and in overseas regions) and the population (=59 392; number of tourists included).

Imported coarse food Nr and Nr loss in overseas regions
We estimated the amount of imported coarse food Nr by compensating for the deficit of coarse food on the island, that is, by subtracting coarse food produced for consumption on the island from the average values of coarse food demand for Japan (Eguchi andHirano 2019, Hirano et al 2023). The amounts of net food Nr and consumed Nr of imported coarse food were calculated using data from previous studies (Eguchi andHirano 2019, Hirano et al 2023).
To estimate Nr loss in overseas regions for imported coarse food production, we applied the following equation: where Loss indicates Nr loss in overseas regions (kgN); Consumed indicates consumed Nr of imported coarse food on Ishigaki Island (kgN); VNF indicates virtual nitrogen factor and is defined as total Nr loss to the environment excluding consumed Nr divided by consumed Nr; Coarse indicates the imported coarse food Nr (kgN). The value of VNF for each production was adopted from Eguchi and Hirano (2019). The values used are listed in supplementary table 6.

Nitrogen flow of whole food system and scenario analysis
The nitrogen flows for all food items were summed to comprehensively assess the present nitrogen flow in the island's food system. We estimated the manure utilization ratio (used cattle manure/total produced livestock manure) to understand the present manure use situation and to judge whether it is possible to replace 30% of chemical fertilizer. A scenario to mitigate Nr loss on the island was created following the agri-environmental policy of the Japanese government strategy MeaDRI (MAFF 2021). In the analysis, we assumed that manure was applied to all crops and that crop yields were kept constant even after chemical fertilizer Nr was replaced with manure Nr based on the previous studies (e.g. Anai et al 2010, de Ponti et al 2012, Liu et al 2021. We chose the use of cattle manure in the analysis because beef cattle production is a key and indispensable industry on the island, and increasing the manure utilization ratio can be controlled by society. In contrast, other measures, such as increasing fertilizer use efficiency, require more scientific knowledge.

Nr loss to the environment regarding agricultural activity
For agricultural activity, Nr loss on the island was 1 523 967 kgN, whereas that in overseas regions was 184 133 kgN for concentrate feed production (figure 2). The major Nr loss pathways in Ishigaki Island were 'crop-unused' and 'manure' occupying 37% and 43% of Nr loss on the island, respectively. Sugarcane and beef had the highest Nr losses among the crops/animals because their cultivated areas/numbers of animal heads were also the largest. According to Nakanishi (2008) and Miyamoto et al (2018), chemical fertilizer and livestock manure are the main sources of nitrogen loads on Japanese subtropical islands including Ishigaki Island. This is consistent with our results (figure 2). Kaji and Nagatomo (2014) reported that approximately 60% of the nitrogen applied to sugarcane is unused (i.e. leached, volatilized, or remained in the soil). The calculation results showed that 45% of Nr applied by the chemical fertilizer Nr and BNF was lost as crop-unused Nr (supplementary figure (a)) indicating higher nitrogen use efficiency than in the report. Note that the nitrogen flow of sugarcane production had a large amount of food processing waste, occupying 41% of total Nr loss, which is mainly induced by bagasse combustion, although emitted NO x concentration is lower than the regulation level (based on interview). For pineapple and vegetables, 90% of Nr loss on the island was estimated to be crop-unused (figure 2). This indicated that the nitrogen use efficiency was very low and that the improvement of fertilizer application could significantly reduce the Nr loss. Unused Nr of paddy rice production showed the lowest ratio (71%) because of lower fertilizer application rate (supplementary table 4).
In animal production, Nr loss also occurs overseas because of imported concentrate feeds. To cattle, goats, and horses, coarse feed produced on the island was applied; therefore, crop-unused Nr was involved. The largest manure producer was cattle (beef and dairy cattle), which accounted for 88% of the total manure Nr (figure 2).

Nitrogen footprint of productions on Ishigaki Island
The calculation revealed that, in total, 28% of Nr loss to the environment is related to food production for island consumers (table 1). Per-capita nitrogen footprint was large for vegetables and beef cattle. We considered that this high footprint is due to low nitrogen use efficiency and import rate. Regarding nitrogen use efficiency, the prefectural standard for fertilizer application for pumpkin is higher than the Japanese average (supplementary table 4(d)). On Ishigaki Island, the total fertilizer application was 700 kgN ha -1 (=chemical fertilizer 280 kgN ha -1 + manure 420 kgN ha -1 ), whereas that of the Japanese average is 540 kgN ha -1 (=chemical fertilizer 400 kgN ha -1 + manure 140 kgN ha -1 ) (Mishima 2002, Mishima et al 2012. Our result indicated that unused Nr accounted for 90% of Nr loss (figure 2). For beef cattle, 86% of feed Nr (=coarse feed Nr + concentrate feed Nr) was released as manure Nr. In contrast, the 'other animals' released about 70% of feed Nr as manure Nr (except goat and hoarse = 92% and 91%, respectively) (supplementary figures (e)-(k)). Our estimation of food import (section 3.1.3.) revealed that the import rate of vegetables and beef were 51.5% and 0%, respectively, while those of most of the other products were higher than 65% (except egg and other meat = 56% and 50%, respectively) (table 2). This low import rate with low nitrogen use efficiency resulted in the large per-capita nitrogen footprint of vegetables and beef. Note that for egg, the per-capita nitrogen footprint was smaller than that of vegetables because egg had higher consumed Nr than vegetables (7 490 and 4 524 kgN, respectively) (supplementary figures (d) and (i)) and Nr loss to the environment for island consumers' diets per consumed Nr of egg was 37% of that of vegetables (4.2 and 11.3, respectively). Other meat (goat and horse) has low per-capita nitrogen footprint because their coarse food production was small (supplementary table 7).
Our results showed that Nr loss to the environment for food production for the consumers was small, and most of the Nr loss was related to export production indicating that Ishigaki Island shouldered a burden of Nr loss for food consumption in overseas regions. Whereas food production on the island was not enough to satisfy consumer demand. We, therefore, estimated imported food Nr on the island.

Imported coarse food
Except for beef, coarse food production on Ishigaki Island was not sufficient to meet the demand of the consumers (table 2). The produced coarse food was 55 977 kgN (supplementary table 7), whereas imported food was estimated to be 5.1 times larger than produced food on the island, reflecting the island's characteristic of heavy dependence on food imports (Maesato et al 2016). The amount of Nr loss in overseas regions for imported food (=470 281 kgN; table 2) was higher than that on Ishigaki Island (=408 665 kgN; table 1). This indicated that Nr loss on the island is made smaller by placing part of the burden of Nr loss on overseas regions to support the island's needs.  a Milk and dairy products were applied to the calculation because slaughter was sent to overseas regions. b Egg was applied because slaughter was sent to overseas regions. We estimated the import of each product to compensate for differences between production on Ishigaki Island and coarse food demand, which is the Japanese average value per capita obtained from Eguchi and Hirano (2019) and Hirano et al (2023). a This shows Japanese average values in 2015. b This displays coarse food to be imported to Ishigaki Island. The surplus of beef production was applied to the deficit of the other products equally (=0.003 kgN capita -1 ). Then import for individual products was determined. c This displays Nr loss in overseas regions for the imported coarse food production.

Comprehensive nitrogen flow on Ishigaki Island
On the island, half of the imported Nr originated from chemical fertilizers for agriculture, followed by concentrate feeds for animal production ( figure 3(a)). In overseas regions, the amount of Nr input was relatively similar to that on the island (concentrate feed, BNF, chemical fertilizer, and manure = 1 747 665 kgN). Moreover, Nr loss in overseas regions was close to the amount of Nr in the imported products (food and concentrate feeds = 773 962 kgN). A breakdown of chemical fertilizer and Nr flow in overseas regions is shown in supplementary table 8. As previously described in section 3.1.2., most of the Nr loss on the island was a result of export production. These results show that Ishigaki's food system is characterized as tradedependent, that is, the island shoulders a burden of Nr loss for the food consumed in overseas regions; it simultaneously places a burden of Nr loss for imported food and feed on overseas regions. Agriculture on Ishigaki Island uses mainly chemical fertilizer; however, less than 13% of cattle manure was reused for agricultural production. This indicated the necessity of resource recycling, and the amount of remaining manure was enough to replace 30% of chemical fertilizer.

Scenario analysis
To replace 30% of chemical fertilizer, approximately 70% of the cattle manure produced on the island was necessary ( figure 3(b)). This scenario would reduce the total Nr from outside and Nr loss on the island to 84% and 83% of that at present state, respectively. The Nr loss on the island from beef and dairy cattle production would be 59% (727 246 to 428 869 kgN) and 67% (24 493 to 16 475 kgN) of the current level, respectively. The analysis results indicated that increasing cattle manure utilization would be an effective measure to reduce nitrogen load.
Our results presented the maximum value of manure use. To achieve the Japanese strategy, however, we should combine manure use and other measures for the following reasons. On Ishigaki Island, sugarcane and grassland occupy most of the agricultural area (3 155 ha; supplementary table 4). Muck spreaders will be necessary to apply manure to an area this large. Transportation of manure to the field is one of the problems because of the scarcity of available labor. Moreover, pineapple prefers acidic soil; however, manure application will raise soil pH.

Future plan
Owing to these reasons, other strategies aside from manure use should be developed to reduce the nitrogen load. This study showed that the greatest Nr loss in crop production was due to crop-unused Nr (figure 2). Previous studies have proposed improving fertilizer use efficiency (Oita et al 2020). On Ishigaki Island, we believe that improving fertilizer application . BNF is biological nitrogen fixation; 'usage' is the manure usage rate calculated by dividing 'used cattle manure' by 'all cattle manure'; 'exported Nr' is the amount of Nr in exported products; 'denitrification' is the removed Nr by denitrification from human excreta during sewage treatment; and 'imported food' is all assumed to be coarse food. In accordance with the agri-environmental policy of the Japanese government strategy MeaDRI, we created a scenario in which 30% of all chemical fertilizers were replaced with cattle manure. The underlined values indicate differences between the present and scenario conditions. The calculation also considered the input of Nr and Nr loss for imported food and feed production in overseas regions. Arrows A-C represent: food processing waste and food loss of imported coarse food (A); all Nr losses excluding human excreta in the food system (figure 1) and volatilization during manure production (B); and human excreta after food consumption (C). could further reduce the nitrogen load. For example, for sugarcane, the most cultivated crop on Ishigaki Island, fertilizer is applied at the early cropping stage. This application results in a nitrogen use efficiency of less than 4% . Okamoto et al (2021) showed that halving the first fertilizer application increases nitrogen use efficiency while maintaining yield. In addition, the combination of chemical fertilizers and manure increases sugarcane yield by 9%-12% (Higa et al 2011).
This study had some limitations. For instance, some data, such as those on goat and horse production, were lacking. In the future, studies should not only collect further data to obtain accurate results but also analyze different scenarios to reduce cropunused Nr while maintaining or improving yields. Moreover, further measures to reduce nitrogen load may be developed by comparing the nitrogen footprint between the island and Japan (e.g. Eguchi and Hirano 2019). For comparison, we will estimate the nitrogen footprint by including Nr loss to the environment for imported food production in future research.

Conclusion
This study evaluated the present nitrogen flow in the food system of Ishigaki Island by using the nitrogen footprint concept and then examined a scenario where the utilization of cattle manure from the island is increased to reduce nitrogen load. The results of this study are summarized as follows: (1) The food system on Ishigaki Island is characterized by heavy dependence on imports of chemical fertilizer, concentrate feed, and food (e.g. imported food Nr is 5.1 times larger than that of food produced for consumption on the island). (2) The major Nr loss pathways on the island are unused Nr by crops and livestock manure, occupying 37% and 43% of all Nr release to the environment for agricultural activity, respectively. (3) Less than 30% of Nr loss to the environment for coarse food production is related to food for the consumers' diets on the island.
(4) By replacing 30% of the current chemical fertilizer use with 70% of the cattle manure produced in Ishigaki, the island's imported Nr and Nr loss on the island would be reduced by 16% and 17%, respectively.
The proposed robust and sustainable food system can be extended to other islands to overcome the unstable global situation, including the recent price hikes, and conserve the environment. Further studies are needed to estimate the maximum possible reduction in chemical fertilizer Nr and Nr loss to the environment through effective fertilizer application.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).