Variation in Foliar ẟ 15 N Reflects Anthropogenic Nitrogen Absorption Potential of Mangrove Forests

: Research Highlights: Mangrove forests are absorbing anthropogenically produced excess nitrogen under moderate to intensive human interaction in the study sites, further indicating the degree of deviation from the natural ecosystem condition. Background and Objectives: Mangrove species, when directly connected to anthropogenic activities such as sewage disposal, agricultural inputs, and receiving of animal manure, absorb excess nutrients from the systems and act as ecological indicators of long ‐ term natural changes. However, there is a paucity of examples of how the mangroves respond to a land ‐ use gradient comparing to the non ‐ mangrove plants under indirect anthropogenic impacts. Materials and Methods: In this investigation, foliar total nitrogen (N), carbon to nitrogen (C/N) ratio, and δ 15 N of mangrove and non ‐ mangrove species collected from 15 watersheds on three islands in Okinawa, Japan, have been compared. The land ‐ use areas in the study watersheds were delineated by ArcGIS software, and the correlation between the foliar traits and the human ‐ affected area ratios were examined. Results: Foliar δ 15 N of the mangroves, which was significantly different from those of the non ‐ mangroves on each island, showed significantly higher values (5‰ to 14‰) in human ‐ affected forests, whereas the values were up to 3‰ in pristine forests. Furthermore, the significant positive relationship between foliar nitrogenous traits and the human ‐ affected area ratios suggested that the anthropogenic N might be regulating foliar N content and δ 15 N signature on the sites. Conclusion: Different degrees of foliar isotopic fractionation with the land ‐ use gradient have clarified that mangroves can be a powerful tool for monitoring ecosystem conditions under anthropogenic disturbances.


Introduction
Mangrove forests are ecologically important in trapping and storage sediments, nutrients, and heavy metals from inland runoff [1][2][3][4] and protect the surrounding ocean ecosystems from eutrophication [5,6]. The ongoing increase of anthropogenically produced reactive nitrogen (N) in the environment is one of the major causes of impairment to the natural condition and the valuable ecosystem services of mangrove forests [7][8][9][10][11]. However, inconsistent findings are available, which unveil mangroves as unpredictable and site-specific ecosystems throughout the tropical and subtropical countries. Though Valiela et al. [12] showed no relationship between the population density and the loss of mangrove forests, Alongi [13] claimed a reduction of them for the population load. In some places, mangroves have declined from the direct interaction of sewage inputs, maricultural practices, urbanization, hydrological changes, and land-use variation regardless of population density [14,15]. The mangroves, connected to the nutrient enrichment, were reported as long-term indicators of monitoring eutrophication trends in the ecosystems [16][17][18]. However, there is a gap in the comparative study of foliar traits of how mangroves differ from co-existing non-mangroves. Therefore, investigation of the site-specific responses of mangrove forests to anthropogenic disturbances is important for understanding the changes of the natural condition and how the mangroves are effective as ecological indicators of the changes of the ecosystems.
The δ 15 N composition of plant tissues appears systematically in identical environments [19] and provides reliable information about the origin of N [5,16,[20][21][22][23]. The higher N supply from anthropogenic activities [24][25][26][27][28][29] enhances the stable isotope fractionation processes [30,31] like ammonia volatilization, nitrification, and denitrification leading to an elevated δ 15 N signature of ecosystem components [16,25,28,32]. In contrast, the systems with lower N availability restrict the gaseous N loss resulting in a low value of δ 15 N [32]. Previous studies reported that the δ 15 N values ranging from 10‰ to 22‰ are characteristic to the human and animal origin, whereas the naturally produced N through fixation or from the atmospheric deposition attributes the values from −2‰ to 0‰ [33,34], even as negative as −5.6‰ by cyanobacterial fixation [35]. Therefore, the recent studies have followed the typical way of comparing foliar traits of mangroves in human-affected sites with those in the unaffected or less affected sites [5,[36][37][38][39]. However, the comparative study of foliar nitrogenous traits viz. foliar total N content, C/N ratio, and ẟ 15 N between mangroves and nonmangroves is yet absent.
Mangroves cover a small area of about 800 ha in Okinawa Prefecture, the south-western part of Japan in the Ryukyu Archipelago [40,41]. Okinawa Prefecture is associated with enormous environmental issues such as intense agriculture, infrastructure [42], deformation of natural landforms [43], red soil erosion and runoff, which have accounted for a large extent of destruction of marine and coral lives [44][45][46]. Therefore, the authors believe it is important to determine the foliar δ 15 N signature of co-existing mangroves and non-mangroves in watersheds under moderate to intensive human activities for reducing the gaps in understanding (1) whether the mangroves differ from the adjacent non-mangrove species in functioning anthropogenically originated reactive N; (2) whether any relationship exists between the foliar traits and respective land-use areas in the sampling sites; and (3) whether mangroves can be used as ecological indicators of monitoring anthropogenic impacts on Iriomote (natural environment), Ishigaki (moderately human-affected) and Okinawa (intensively human-affected) islands in Okinawa Prefecture, Japan.

Study Sites
A total of 15 watersheds named Urauchi, Kura, Shiira, Maera, Mare, Hinai, and Nakama on Iriomote Is., Nagura, Fukido, Miyara, Hirakubo, and Todoroki on Ishigaki Is. and Manko, Kesaji, and Okukubi on Okinawa Is. were selected as the study sites. The coordinates of the sampling points and collected leaf types are available in Table 1. Each watershed of the three islands commonly comprises short channels, creeks, and swamps, which support the growth of mangroves along the shorelines with distinct zonation from the river mouth to the upstream area [47]. The non-mangrove communities have grown in parallel at the backside of the mangroves. The nearly entire area of Iriomote Is. is covered by the National Reserved Forest Park [48], which is commonly used for scientific research and small-scale tourism. In contrast, Okinawa and Ishigaki islands are accompanied by intensive agricultural activities, animal husbandry, local construction, and tourism, which are frequently found close to the mangrove forests [44,49]. The islands are generally under the humid-subtropical climate [50]. The average yearly temperature, humidity, and rainfall of the last five years on the three islands ranged from 23 to 26 °C, about 78% to 80% and approximately 1900 to 2400 mm, respectively (Japan Meteorological Agency, data collected in 2019). The inland flow to the downstream creeks, channels, and swamps of these areas largely depends on short-term weather conditions [44] and may carry a risk of nutrient enrichment to the oceans. Particularly, livestock farms and agricultural fields are highly considered to influence local-scale N cycling in the mangrove forests on Ishigaki and Okinawa islands.

Leaf Sample Collection
The dominant mangrove species on the sites such as Bruguiera gymnorrhiza, Rhizophora stylosa, and Kandelia obovata and the randomly selected non-mangrove species closely grown at the backside of the mangroves were chosen for leaf sample collection during summer (May to September) in the year of 2017 and 2018. Three points respective to the down, middle, and upstream areas in each watershed were selected for leaf sampling, which was possible only in five watersheds of Urauchi, Nagura, Kesaji, Okukubi, and Manko. In the rest of the watersheds, one or two sampling points were chosen because the mangrove areas along the creek were too small to focus on three streams. Besides, the non-mangroves closely grown within an area of a maximum of 20 m 2 to the mangroves were sampled at each point except in the Mare watershed, where no mangroves were found. In this watershed, the non-mangrove trees, directly connected to the creek water, and closely grown to the adjacent mangroves in Hinai watershed, were sampled. Variation in foliar δ 15 N was examined in those watersheds where two plant groups were available within the designed sampling plots.
Three or more trees of each mangrove species were used for the replicated sampling from the water vicinity to the inland. On the other hand, five trees of collective non-mangrove species growing backside of the selected mangroves were chosen for sampling based on their availability at each point. The list of the non-mangrove species collected in each watershed is shown in Supplementary Table S1. A total of five sun-facing mature leaves per individual were collected from the middle-stage branch of the tree.

Processing of Leaf Samples and Analysis of Foliar N Content, C/N Ratio, and ẟ 15 N
The leaf samples were washed with pure water and wiped on sites, then taken to the laboratory, and dried in an oven at 60 °C (DNE 910, Yamato) for 48 hrs. The samples were ground into fine powder by a wonder blender and stored in a desiccator until analysis. Before every analysis, they were dried at 105 °C for at least 4 h to remove moisture.
The total content of foliar N (% of dry weight) and C (% of dry weight) of powdered leaf samples was measured using an NCH (Nitrogen, Carbon, Hydrogen) analyzer (SUMIGRAPH; SCAS), and C/N ratio was calculated. Foliar δ 15 N was measured using an on-line isotope ratio mass spectrometer coupled to continuous flow interface (Temperature Conversion/Elemental Analyzer-IRMS), a combined set up of the ConFlo IV interface, Flash 2000 and Flash 2000 HT Plus, CNSOH and Delta V Advantage Isotope Ratio Mass Spectrometer (IRMS, Thermo Fisher Scientific, Waltham, MA, USA).

Processing the Land-Use Map
The land-use maps of the study watersheds (available in Supplementary Figure S1) were delineated by using ArcMap Geographic Information Services (GIS) software (ArC Map 10.4.1). The source data for the watershed borders and land-use were obtained from the National Land Policy Bureau, Japan Ministry of Land, Infrastructure, Transport, and Tourism. The baseline of the creek, swamp, and coastlines was obtained from the land conservation maps (1:200,000) from the Geospatial Information Authority of Japan (GSI). The mangrove areas in each watershed were adjusted by field survey and their visual observation on the Google map. Therefore, little inaccuracy in mangrove area delineation is considered. The land-use was divided into four groups viz. forest (mangroves plus nonmangroves), agricultural, residential, and creek and/or swamp areas, as shown in Table 2.
Generally, the forest areas in watersheds on Iriomote, Ishigaki, and Okinawa islands ranged from 92% to nearly 100%, 35% to 72%, and 14% to 78%, respectively. The largest forest area (about 100%) was found in Hinai and Shiira watersheds on Iriomote Is., whereas the largest human-affected area (agriculture plus residents; 84.5%) was found in Manko swamp watershed on Okinawa Is, followed by that (65.5%) in Todoroki watershed on Ishigaki Is. Eventually, the mangroves on Iriomote, Ishigaki, and Okinawa islands were categorized as pristine, moderately human-affected, and intensively human-affected forests, respectively.

Data Processing and Statistical Analysis
The measurement of δ 15 N is expressed in per mil unit (‰) calculated by the following equation: where R is the ratio of heavier to lighter N isotope. Compressed N2 gas was served as the reference gas for measurement of 15 N abundance in the sample relative to that in atmospheric N2. Isotopic values were calibrated using three working standards such as L-Alanine M9R2064 (AZ100: δ 15 N-Air = 1.79 ± 0.2‰); L-Alanine SS13 (AZ101; δ 15 N-Air = 13.7 ± 0.2‰) and L-Histidine M6M9675 (AZ120; δ 15 N-Air = 7.58 ± 0.2‰) supplied by the International Atomic Energy Agency. Foliar N, C, C/N ratio, and δ 15 N values were reported as mean ± standard error (SE). The analysis of variance (ANOVA) was performed to test the differences of foliar N, C/N ratio, and δ 15 N between the two plant groups (mangroves and non-mangroves) to compare within and among and the islands. The foliar data sets were examined for normality, and the mean values were investigated for the significance level of differences by Welch's t-tests. Linear regression analyses were performed to detect the relationship between foliar data and land-use area. Data processing, statistical analysis, and presentation of graphs were done by using Excel 2010 (Microsoft, Redmond, WA, USA) and R software (version 3.5.1, R Core Team, Vienna, Austria).

Variation in Foliar N Content, C/N Ratio, and δ 15 N on Islands
The distribution of foliar data sets of all the collected mangroves and non-mangroves has been shown in the boxplots representing their respective similarities and dissimilarities on the three islands ( Figure 1). The highest values of foliar N (1.7 ± 0.7% N) and foliar δ 15 N (8.0 ± 1.6‰) of mangroves were found in Kandelia obovata on Okinawa Is., whereas the lowest values of those (0.92% ± 0.1% N and 1.4‰ ± 2.7‰) were determined in Rhizophora stylosa on Iriomote Is. Likewise, the highest foliar N and ẟ 15 N (1.9% ± 0.7% and 0.7‰ ± 1.4‰) and the lowest foliar N and ẟ 15 N (1.4% ± 0.7% and −2.6‰ ± 2.5‰) of non-mangroves were recorded on Okinawa Is. and Iriomote Is., respectively. However, the values of foliar N and δ 15 N for the mangroves and the non-mangroves collected from Ishigaki Is. placed between the values found on Iriomote Is. and Okinawa Is. The mean ± SE of foliar N, C/N ratio, and ẟ 15 N for the mangrove and the non-mangrove leaf samples used in this study are provided in Supplementary Table S2.

Species Variation in Foliar δ 15 N in Watersheds
Foliar δ 15 N values of the mangroves and the non-mangroves were compared in each watershed. The comparison showed that the values of mangrove were significantly different from the values of non-mangroves (p <0.05) in all study watersheds (Figure 2). The foliar δ 15 N of the mangroves ranged from nearly 1‰ to 14‰ across the eight watersheds on the human-affected islands (Ishigaki and Okinawa). However, those ranged from 1 to 2‰ in four watersheds on Iriomote Is., which was assumed as a pristine environment. The lowest value of the mangroves foliar δ 15 N was found in B. gymnorrhiza (−0.02‰) in Kura watershed, followed by that in the same species (0.02‰) in Hinai watershed on Iriomote Is. Conversely, the highest value of the mangroves foliar δ 15 N was determined in R. stylosa (ca 14‰) in Todoroki watershed on Ishigaki Is., followed by those measured in R. stylosa and K. obovata (ca 10‰) in Manko swamp watershed on Okinawa Is. Moreover, all the values of foliar δ 15 N of the mangroves, collected from Ishigaki and Okinawa islands, were above 5‰ in all the watersheds except Fukido and Kesaji (Figure 2).

The Relationship among the Foliar N Content, C/N Ratio, ẟ 15 N and the Land-Use
For both the mangroves and the non-mangroves, the linear regression analyses between foliar δ 15 N and foliar N, as well as between foliar C/N ratio and foliar N, showed a significant correlation between them (Figure 3a). In addition, when foliar N and δ 15 N of both the mangroves and the nonmangroves were plotted against the human-affected areas, a significant positive correlation (p <0.01) was indicated (Figure 3b). The results suggested that the human activities particularly through the agricultural and residential areas might be responsible for higher N availability resulting in the N isotope ratio in foliar tissues of mangroves and non-mangroves on the study sites.

Variation in Foliar Traits of Mangroves and Non-Mangroves
Many factors, including soil water availability, variable anthropogenic inputs, and disturbance history, can regulate the N availability and contribute to the foliar traits on sites [28]. The significant differences in foliar N, thereby in C/N ratio, and foliar ẟ 15 N between co-existing mangroves and non-mangroves reflected different levels of N availability for the nutrition of two plant groups on the sampling sites. Foliar C content of the two plant groups did not vary significantly (data not shown). It is known to be dependent on photosynthesis of the plants, regulated by the climatic variation, especially sunlight and air temperature [50]. Since the climatic parameters in the study sites do not vary remarkably, as mentioned in Section 2.1, the lifting of foliar N content and the consistent decrease of C/N ratio of both the mangroves and non-mangroves on Okinawa Is. (Table S2) compared to those on Iriomote Is. are probably due to relatively higher N availability in the intensively humanaffected sampling sites on Okinawa Is. [51]. The results are also supported by the previous observation of Thimdee et al. [5] and Clough et al. [52] who have reported a higher value of foliar total N (about 2%) of mangroves receiving long-term sewage effluents than that (about 1%) of undisturbed mangroves.
In general, plants use a large portion of energy to root respiration for uptake and assimilation of N [53] to withstand in low-nutrient environments [54]. Moreover, it was reported that the mangroves are ones of the wood species that have the highest photosynthetic nitrogen-use efficiency (PNUE: the ratio of photosynthetic ability to N content in leaf) [55,56], which is one of the adaptive mechanisms of the mangroves [57] to grow under nutrient-limited conditions [58]. Therefore, it seems that mangroves are greatly able to absorb excess N from the external sources on the human-affected sites. The significantly lower values of foliar N and higher C/N ratio of the mangroves than those of the non-mangroves on each study island agree with the assumption.

Foliar ẟ 15 N of Mangroves and Non-Mangroves under Anthropogenic Impacts
Anthropogenic interaction regulates the local-scale N cycling [28] and sets the distinguishable signature of δ 15 N in plant tissues [19]. According to the inventory reported by Peterson and Fry [59] and Nadelhoffer et al. [60], the wood plants growing in the pristine environments throughout the world have a range of foliar δ 15 N values from −10‰ to 3‰. Although the foliar δ 15 N of all the nonmangroves collected from three islands placed within the natural range (Iriomote: −2.8‰ ± 2.4‰; Ishigaki: −0.6‰ ± 1.9‰; and Okinawa: 0.7‰ ± 1.9‰), only the Iriomote mangrove's foliar δ 15 N (1.7‰ ± 2.1‰) followed them. The results are supported by the previous observation of Costanzo et al. [16] and Cole et al. [26] who reported foliar δ 15 N from 2‰ to 4‰ in the natural mangroves. In contrast, the foliar δ 15 N of mangroves on Ishigaki and Okinawa islands were 5.5‰ ± 3.4‰ and 7.2‰ ± 3.0‰, respectively, which were congruent with the previous studies of the disturbed mangroves [18]. In Florida, the foliar δ 15 N of disturbed mangroves directly connected to agricultural draining channels ranged between 11 and 16‰, whereas, the values ranged from −5‰ to 2‰ in a comparatively pristine mangrove forest [25]. In addition, the foliar δ 15 N of a Rhizophora apiculata community under the influence of higher N availability by anthropogenic inputs was recorded as 6‰ [15].
Okinawa and Ishigaki islands are associated with large extent disturbances of intensive agriculture, upland soil erosion, agricultural practices, urbanization, and tourism [42][43][44][45][46]49]. It seems that the mangroves on Okinawa Is. are receiving intensively elevated inputs of anthropogenic N with a mean foliar δ 15 N of 7.2‰, whereas there are slightly elevated inputs of anthropogenic N with a mean foliar δ 15 N of 5.5‰ on Ishigaki Is. [18]. Both were over the natural range of foliar δ 15 N (up to 3‰) of pristine mangroves, as found on Iriomote Is. The elevated foliar δ 15 N can be linked, firstly, to the isotopic composition of N pools and, secondly, the fractionation of isotopes during N uptake and assimilation by roots and leaves. Many studies have proved that no isotope fractionation occurs during root uptake [61] of ammonium [62], and nitrate [63] or the fractionation is very negligible (<0.3‰) in many plants [64]. Manko watershed on Okinawa Is. has the largest population density of 5460 km −2 , and Todoroki and Miyara watersheds on Ishigaki Is. have the top two largest agriculture areas of 65% and 46%, respectively, associated with a relatively lower population density of 7 and 24 km −2 , respectively ( Table 2). It seems that those watersheds are responsible for the different degrees of isotope fractionation of external N on that respective islands.
Furthermore, the foliar ẟ 15 N values of mangroves were significantly higher (p <0.05) than those of the non-mangroves in every watershed on three islands (Figure 2). Different degrees of isotope fractionation and corresponding variation in foliar traits of mangroves in different watersheds are probably based on variable distances of waterways from N source to sinks [16,51], inter-species variation in N physiology, leaf life cycle of plants [65,66], and local-scale variability of microclimatic factors [51]. Costanzo et al. [16] have reported that δ 15 N of marine plant tissues declined gradually from 10‰ to 3‰ at the downstream river mouth with a relative decrease in the distance from the proximity of sewage outfalls. . The highest foliar δ 15 N of mangrove (around 14‰) found in Todoroki watershed ( Figure 2) might be due to the kinetic fractionation of N isotope via ammonia volatilization by urination from the large number of domestic animals (6198 of beef cattle, 389 pigs, 26 horses, and 8800 layers in an area of 49 km 2 ; data collected from local municipality office; City Hall, Ishigaki Is.) closely grown to the sampling sites [21]. The second highest foliar δ 15 N of mangrove (around 9‰) was found in Manko swamp watershed, which was connected to large-scale urban development and agricultural fields ( Table 2). In the previous studies, δ 15 N signatures were recorded as high as between 10‰ and 30‰ in the sources viz. livestock wastes, domestic wastewater [67] and wastewater treatment plant effluents [68,69], whereas those ranged from 11‰ to 16‰ in a mangrove forest directly connected to agricultural draining channels [25]. Contrastly, the depleted values of foliar δ 15 N of the mangroves on Iriomote Is. and all the non-mangroves on three islands firmly indicated that the plants were probably not absorbing anthropogenically produced N and were dependent on naturally produced N sources through fixation [70], atmospheric deposition [71], and mycorrhizal association [72].
The results of regression analysis of foliar data of the mangroves and non-mangroves clarified that the relative increase of N availability increases the foliar ẟ 15 N values significantly (p < 0.05) regardless of plant types (Figure 3a). The slopes between foliar ẟ 15 N and foliar N, and between foliar C/N ratio and foliar N of the mangroves (y = −3.11 + 6.79x and y = 62.5 − 18.6x, respectively) were slightly steeper than those of the non-mangroves (y = −9.74 + 5.43x and y = 65.5 − 20.5x, respectively). This suggested that the foliar traits of mangroves were more sensitive to relative N availability on the sites. In addition, foliar ẟ 15 N and foliar N content of both mangroves and non-mangroves showed a significant positive correlation (p <0.01) with the ratios of total human-affected area (Figure 3b). Although foliar N of the mangroves and the non-mangroves were found equally sensitive to increasing human-affected area (mangroves: y = 0.88 + 0.01x; non-mangroves: y = 1.36 + 0.01x), the change in foliar ẟ 15 N of the mangroves with the increase of human-affected areas was higher than that of the non-mangroves (mangroves: y = 0.42 + 0.11x; non-mangroves: y = −2.63 + 0.05x). Typically, all plants have the adaptabilities to withstand low-nutrient environments through the reduced requirement of nutrients and low growth rates over time [54]. Consequently, excess N supply may enhance relative uptake of N by plants if demanded, which is why non-mangroves also exhibit sensitivity to N availability from the anthropogenic sources. However, foliar ẟ 15 N signature of the non-mangroves showed that they relied mostly on naturally originated N sources. Even though both mangroves and non-mangroves can use different sources of N from precipitation along with soil solution depending on micrometeorological conditions on sites [73][74][75][76][77], mangroves are capable of using alternate nutrient sources of groundwater, river water, and tidal seawater [78,79] accordingly due to their unique root systems, distribution along the water vicinity of the rivers, and salt-tolerant plant physiology. Despite being tactful to relatively higher N availability on human-affected sites, the non-mangroves showed negative to nearly zero values of foliar ẟ 15 N values. The non-mangroves might be impotent to use anthropogenic N sources from the deeper groundwater, river water, and tidal saline water because of their relatively shorter root systems, growth backside of the mangroves, and lack of salt-tolerant plant physiology, respectively.
The relationship between foliar ẟ 15 N and foliar N content in watersheds has been widely debated in the study of anthropogenic impact assessment by macrophytes, particulate organic matter, zooplankton, and mangroves [18,30,51,80]. However, the discrepancy of such a relationship was also reported while monitoring the effectiveness of tropical macroalgae as a bioindicator of nutrient enrichment from shrimp farms [81]. Such an observation may result because of the inability of longterm N storage [82], related to species-specific nutrient physiology [65,66]. Though the mangrove plants, unlike algae and crabs, did not respond by δ 15 N to any change in a sewage outfall within a short period of 2 years in the Moreton Bay Catchment, Australia [65], they successfully indicated the magnitude of anthropogenic impacts throughout the sampling sites of this study. Mangrove plants have resorbed N-use efficiency of even more than 70% [83], with an average leaf life cycle of 16 months [55,84]. Therefore, not only Avicennia marina [18], mangrove species of R. stylosa, B. gymnorrhiza, and K. obovata (in this study) are likely absorbing, storing and recycling anthropogenic N in the environment and could be used as indicators to monitor the deviation of the natural condition in the mangrove forests.

Conclusions
Mangrove forests significantly vary from the adjacent non-mangrove communities for their distinctive plant physiology and habitat structure. Anthropogenically originated N may largely affect the growth of mangroves by providing excess nutrients to the N-limited natural condition of the mangrove forests. Thus, the external inputs of N are, thereby, regulating foliar N and ẟ 15 N of mangroves, which can be an effective index to evaluate ecosystem responses to environmental disturbances. As we observed that the mangroves could absorb anthropogenic N under any extent of human interaction, it is expected that the findings of this study will increase the use of mangroves in monitoring ecosystem status globally. However, it is important to understand the complete pathway of anthropogenically produced reactive N contributing to the enrichment of foliar ẟ 15 N in the watersheds in order to take any management and conservation steps for mangrove forests. The source identification of reactive N of surface and groundwater in the study watersheds is under investigation.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1: Table S1: The list of non-mangrove species collected from the study watersheds on three islands, Table S2: Foliar total N, total C, C/N ratio and δ 15 N of mangrove and non-mangrove species in the study watersheds on three islands. Figure S1: Location of the study watersheds (left above) and the land use at sampling points on Iriomote, Ishigaki, and Okinawa islands. Funding: This research received no external funding except the annual laboratory budget from Tokyo University of Agriculture, Japan.