Environ. Sci. Technol. 2008, 42, 7837–7841
Use of Stable Nitrogen Isotope Signatures of Riparian Macrophytes As an Indicator of Anthropogenic N Inputs to River Ecosystems A Y A T O K O H Z U , * ,†,‡ T O S H I H I R O M I Y A J I M A , § I C H I R O T A Y A S U , ‡ C H I K A G E Y O S H I M I Z U , †,‡ FUJIO HYODO,| KIYOSHI MATSUI,⊥ TAKANORI NAKANO,| EITARO WADA,| NOBORU FUJITA,‡ AND TOSHI NAGATA‡ Japan Science and Technology Agency, Saitama 332-0012, Japan, Center for Ecological Research, Kyoto University, Shiga 520-2113, Japan, Ocean Research Institute, The University of Tokyo, Tokyo 164-8639, Japan, Research Institute for Humanity and Nature, Kyoto 603-8047, Japan, and School of Science Education, Nara University of Education, Nara 630-8528, Japan
Received April 24, 2008. Revised manuscript received August 07, 2008. Accepted August 19, 2008.
Deterioration of aquatic ecosystems resulting from enhanced anthropogenic N loading has become an issue of increasing concern worldwide, and methods are needed to trace sources of N in rivers. Because nitrate from sewage is enriched in 15N relative to nitrate from natural soils, δ15N values of stream nitrate (δ15Nnitrate) should be an appropriate index of anthropogenic N loading to rivers, as should the δ15N values of riparian plants (δ15Nplant) because they are consumers of nitrate. We determined the δ15N values of stream nitrate and six species of riparian macrophytes in 31 rivers in the Lake Biwa Basin in Japan. We then tested the correlation between these values and various land-use parameters, including the percentage of land used for residential and agricultural purposes as well as for natural areas. These δ15N values were significantly positively correlated with land use (%) that had a high N load (i.e., residential or agricultural use) and significantly negatively correlated with forest (%). These findings indicate that δ15N valuesofstreamnitrateandriparianplantsmightbegoodindicators of anthropogenic inputs of nitrogen.
Introduction Deterioration of aquatic ecosystems resulting from enhanced anthropogenic N loading has become an issue of increasing concern worldwide (1, 2), and methods are needed to trace the sources of N in rivers. Nitrate from sewage is enriched in 15N relative to nitrate from natural soils (3). Therefore, δ15N values of stream nitrate (δ15Nnitrate) are an appropriate index for indicating anthropogenic N loading in rivers (4). Because nitrate in river water is an important source of N for riparian plants, δ15N values of riparian plants (δ15Nplant) can also be used as a proxy for 15N abundance in stream nitrate * Corresponding author fax: +81-29-850-2576;
[email protected]. † Japan Science and Technology Agency. ‡ Kyoto University. § The University of Tokyo. | Research Institute for Humanity and Nature. ⊥ Nara University of Education. 10.1021/es801113k CCC: $40.75
Published on Web 10/30/2008
2008 American Chemical Society
e-mail:
and an indicator of anthropogenic inputs of N (5). However, δ15N values of macrophytes in a given area can vary greatly depending on species, and the mechanisms underlying this variability remain unclear (6). The N nutrition of plants near a river channel varies depending on the plant’s established edaphic conditions that determine the primary N species for plant’s nutrition and on the accessibility of dissolved N in river water (7). If we could discriminate the characteristics of plant species for which δ15Nplant correlates well with δ15Nnitrate, it would be possible to determine the timeintegrated δ15N values of nitrate in surface water. For woody plants, a combination of dendrochronology and δ15N values of tree rings revealed historical changes in N dynamics in a forest (8, 9). Thus, if δ15N values of riparian willow tree rings are an appropriate recorder of δ15Nnitrate values, we could reconstruct the historical record for δ15Nnitrate values; however, few studies concerning δ15N values of riparian tree rings have been conducted. The objectives of this study were to examine relationships between δ15N values of riparian herbaceous plants and willows (δ15Nplant) and those of nitrate (δ15Nnitrate) in river surface water, and to infer the mode of variability in the relationships between δ15Nnitrate and δ15Nplant for the different plant species and sampling times. Sampling sites were located at the mouth of 31 rivers flowing into Lake Biwa (Supporting Information 1) covering a broad range of nutrient regimes and land-use patterns, as well as a large inter-river range in δ15Nnitrate values. Herbaceous macrophytes of five generalist species were sampled five times (twice in summer and once in each of the other three seasons) to determine the temporal variation. After the sampling of the water and herbaceous plants was completed, riparian willow (Salix spp.) stem cores were obtained to determine the tree ring δ15N values of willow corresponding to the water sampling periods. From the results of the correlation analysis between δ15Nnitrate and δ15Nplant, the appropriate timing of sampling and the use of macrophyte species as a proxy for 15N abundance in stream nitrate were examined.
Experimental Section Site Description. The sampling sites were located 0.5-4.0 km upstream from the mouths of 31 rivers flowing into Lake Biwa (3174 km2 total basin area) in Shiga Prefecture, Japan (Supporting Information 1). These 31 rivers are quite diverse in their watershed area (from 2.3 to 400 km2), human population densities (HPD; from 44 to 3174 persons/km2), and land-use patterns (forest [%], 0-91.5; paddy field area [%], 3.1-76.3; residential area [%], 0.7-34.5; see Supporting Information 2 for details). The HPD and land-use patterns in each river basin were analyzed using ArcView software (ver. 8.3, ESRI, U.S.) by applying data from three different sources: (1) a watershed map of Shiga Prefecture from the Shiga prefectural government, (2) 1997 digital national land information from the Ministry of Land Infrastructure and Transport, Japan, and (3) data from the 2000 population census from the Statistic Bureau, Ministry of Internal Affairs and Communications, Japan. Land use was expressed as percentage of land in the following categories: Paddy field (%), residental area (%), forest (%). We also subtracted the values for relatively natural land uses (forest, rocky areas, and open water) from 100% and defined the difference as the “sum of land use (%) with a high N load”. Sewage in our study area is generally processed through sewage treatment facilities that may only discharge into certain rivers, although the level of treatment was variable depending on the area. Sample Collection and Measurement. Herbaceous plant VOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7837
sampling was conducted five times (August-September 2003; July and November 2004; and February and May 2005) in the 31 rivers. We collected the following five species of herbaceous plants near the river channel: Juncus effusus var. decipiens, Persicaria thunbergii var. thunbergii, Oenanthe javanica, Phragmitesjaponica,andPhragmitesaustralis.InAugust-September 2003, only the two species of Phragmites were collected. In other sampling, all the species were collected except Phragmites species that were not collected in winter (i.e., in February 2005) because of dieback of the above-ground plant parts (Supporting Information 2). During each sampling period and at each sampling site, 3-4 individuals of each species were selected and one foliar part was collected from each individual. For all species, the collected foliage samples, that were not covered with dirt and well developed, corresponded to the second-5th leaf from the apical end, except for J. effusus var. decipiens, for which only stem samples were collected due to the lack of leaf lamella. One and a half-years after the water and herbaceous plant sampling was completed, in December 2006, stem cores were obtained from riparian willows (Salix spp.) along 24 rivers to determine the δ15N values of tree rings corresponding to the water and herbaceous plant sampling periods. The main trunk of the willows (age 3-6 years) that were growing close to the river channel near the water sampling points were cut down and stocked in the freezer until preparation. Wood samples from tree rings corresponding to the period from 2003 to 2005 were collected with a wood chisel. Foliar and wood samples were dried at 45 °C for more than 2 days prior to preparation for analysis. To remove extractable N-compounds, the willow tree ring samples were extracted using a Soxhlet apparatus with a 1:1 mixture (by volume) of toluene and ethanol for 18 h, followed by extraction with ethanol for 18 h and then with distilled water for 18 h (8). The nitrogen isotope ratios (δ15N values) in the plant materials were measured with a continuous-flow isotope ratio mass spectrometer (CF/IRMS; EA1108, Fisons Co., Italy; delta S, Finnigan Co., Germany). Stable isotope ratios are expressed in δ notation as the difference in parts per thousand (‰) from the standard: δ15N ) (Rsample ⁄ Rstandard-1) × 1000 where R is 15N/14N, and atmospheric nitrogen was used as the appropriate standard. Analytical precision was better than ( 0.2‰ for δ15N values. Stream water sampling for nitrate analysis was conducted at the same time as foliar sampling by collecting river surface water and then filtrating by glass-fiber filters (nominal retention, 0.7 µm; GF/F, Whattman). The δ15N values for nitrate in water samples were measured according to the denitrification method of Sigman et al. (10). Briefly, nitrate in water samples was converted to N2O by specific cultures of denitrifying bacteria. The δ15N values of generated N2O were measured with isotope ratio mass spectrometry (Precon/GasBench/IRMS; Delta plus XP, Thermo Electron Co., Germany). Analytical precision was better than ( 0.2‰ for δ15N values of nitrate.
Results The mean δ15Nplant values of Salix spp., P. japonica, P. thunbergii, and O. javanica were strongly correlated with that of nitrate in nearby stream surface water, whereas those of P. australis and J. effusus were not (Figure 1). The significant orthogonal regression line of mean δ15N values of the riparian plants against those of nitrate was Y ) 0.95X - 0.68 in P. japonica, Y ) 1.11X - 1.71 in P. thunbergii, Y ) 0.78X + 0.42 in O. javanica, and Y ) 0.53X - 0.15 in Salix spp. (Figure 1). Although we could not statistically exclude the effect of sampling time, the inter-river difference in mean δ15Nplant 7838
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 21, 2008
values was significant for all plant species (p < 0.001, Kruskal-Wallis one-way analysis of variance on ranks). The individual δ15N variation (SD, n ) 3) of the riparian plants in a single river was 0.04-2.20‰ in Salix spp., 0.11-2.16‰ in P. japonica, 0.49-2.59‰ in P. thunbergii, 0.21-3.52‰ in O. javanica, 0.12-6.20‰ in J. effusus, and 0.65-3.28‰ in P. australis. The interriver difference in average δ15Nnitrate values was significant (p < 0.001, Kruskal-Wallis one-way analysis of variance on ranks), and the temporal variation (SD, n ) 5) of δ15Nnitrate values in a single river ranged from 0.6 to 4.0‰ (Supporting Information 2). The mean values of δ15Nnitrate, concentration of nitrate, and δ15Nplant (except J. effusus and P. australis) were positively correlated with residential area (%) and sum of land-use area (%) with a high N load, and negatively correlated with forest area (%) in the watershed (Table 1). Moreover, δ15Nnitrate and δ15N values of Salix spp. and P. japonica were positively correlated with HPD and paddy field area (%). However, there was no significant correlation between the δ15Nplant of any plant species or δ15Nnitrate and basin area. The intrariver δ15Nplant variation could be ascribed to either differences in δ15N values among individuals or seasonal differences in δ15N values. The δ15N difference among sampling times was significant in some rivers (Table 2), although the difference generally was within 4.0‰ in almost all rivers and species. However, there was no consistent seasonal variation within a species, and no significant seasonal δ15N variation in respective species could be found in about half of the rivers (Table 2).
Discussion The δ15N values of four species of riparian macrophytes (i.e., Salix spp., P. japonica, P. thunbergii, and O. javanica) showed significant positive correlations with the sum of land use (%) having a high N load and with residential area (%), and a significant negative correlation with forest area (%). These δ15N values, therefore, might be good indicators of anthropogenic inputs of nitrogen (Table 1). Moreover, these δ15N indicators were robust to sampling time and river size difference that was estimated by basin size parameters (Tables 1, 2). The δ15N values of these four riparian macrophytes also showed positive correlation with δ15Nnitrate, whereas those of the other two species did not (Figure 1). The δ15N values of nitrate in surface water, which showed a significant positive correlation with HPD and sum of land use (%) with a high N load (Table 1), might also be one of the best indicators of anthropogenic inputs of nitrogen. Because nitrate was the primary component of inorganic N in our river water samples (Supporting Information 2), these facts strongly suggest that the anthropogenic N first enriched δ15Nnitrate in river water and secondarily increased δ15Nplant of the riparian macrophytes that depend on nitrate for their primary N nutrition. The δ15Nplant values of riparian plants did not, however, always correlate with δ15Nnitrate values (e.g., for J. effusus and P. australis). Among the six riparian macrophytes species studied, the four species with highly correlated δ15Nplant and δ15Nnitrate values had depleted mean δ15Nplant values relative to δ15Nnitrate values in many rivers, but the number of rivers with depleted δ15Nplant values varied from 15 out of 24 for O. javanica to 24 out of 24 for Salix spp. (Figure 1). Moreover, the degree of depletion of δ15Nplant in Salix spp. relative to δ15Nnitrate was significant in rivers with high δ15Nnitrate values, whereas it was within about -3‰ irrespective of δ15Nnitrate for the three other species. These interspecies differences in the relationships between δ15Nplant and δ15Nnitrate could be attributable to one or more of the following three processes: (1) hydrological differences in habitats, (2) interspecies differences in N nutrition, and (3) interspecies differences in
FIGURE 1. Mean δ15N values of 6 riparian plants (δ15Nplant) were plotted as the function of mean δ15N values of nitrate in nearby river surface water (δ15Nnitrate). The orthogonal regression line and the y ) x line were drawn as solid and dotted lines, respectively, for the four species in which δ15N values were positively correlated with δ15Nnitrate values. The orthogonal regression was fitted by using the intrariver variance ratio (σy/σx) averaged over all rivers from which the samples were collected The mean intrariver variance ratio of in P. japonica, P. thunbergii, O. javanica, and Salix spp. were 1.12, 1.41, 1.69, and 1.81, respectively. The equations for the orthogonal regression lines and the correlation coefficient are also shown. The horizontal bars correspond to the temporal SD (n ) 5) of δ15Nnitrate values at each sampling river from 2003 to 2005, and the vertical bars indicated the SD (n ) 3) among individual riparian plants from 2003 to 2005.
TABLE 1. Correlation Coefficients (r) between δ15N Values for Nitrate and Riparian Plants and Land-Use Parameters for Each Watersheda δ15Nplant concentration Phragmites Persicaria Oenanthe Juncus Phragmites Salix spp. of nitrate japonica thunbergii javanica effusus australis (n ) 24) (n ) 31) (n ) 22) (n ) 28) (n ) 24) (n ) 10) (n ) 16)
Land-use parameters
δ15Nnitrate (n ) 31)
paddy field (%) residental (%) sum of land use (%) with high N loade forest (%) human population density (people/km2)
0.42b 0.64d 0.59d -0.61d 0.67d
0.36b 0.38b 0.42b -0.42b ns
0.55c 0.58c 0.63d -0.65d 0.53c
0.65c 0.66d 0.76d -0.78d 0.59c
0.44b 0.38b 0.49c -0.50c ns
ns 0.49b 0.47b -0.48b ns
ns ns ns ns ns
-0.54b ns -0.53b 0.53b ns
basin area (km2)
ns
ns
ns
ns
ns
ns
ns
ns
a
The correlation coefficient between nitrate concentration and the land-use parameters are also shown for reference. b p < 0.05. c p < 0.01. d p < 0.001; ns, not significant. e The sum of land use (%) with a high N load is the value (%) of relatively natural land uses (forest, rocky areas, and open water) subtracted from 100%.
N metabolism. However, we were not able to discriminate which of these processes were the primary cause of the interspecies differences in δ15Nplant values. Two possible explanations for the interspecies differences in the relationships between δ15Nplant and δ15Nnitrate involve water sources, plant growth characteristics and hydrology. The source of nitrate in these habitats is a mixture of surface river water, hyporheic or subsurface water, and water from neighboring catchments. The δ15Nnitrate values from the latter two water sources would be different from those of δ15Nnitrate
in the surface water because of different water retention times and denitrification activity (3). Interspecies differences in spatial root distribution could cause differential utilization of nitrate sources derived from different water sources, which would result in the lowered correlation between δ15Nplant and δ15Nnitrate values in some riparian species. For example, P. australis can extend its roots to the deeper hypoxic layer (11) and has the potential to use nitrate derived from hyporheic water sometimes enriched in 15N due to denitrification (12). This seems a likely possibility in the case of P. australis, VOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7839
TABLE 2. Seasonal Variation in Herbaceous Macrophyte δ 15N Values in Rivers, Excluding Those with a Single Samplinga P. japonica
2003
(n ) 20)
August
2004 July
2003 2004
August 0 July 1 November 1 3 2005 May 1 0 there was no significant seasonal variance in
P. thunbergii (n ) 25)
2005
November 0 0
0 3 1
0 11 rivers
2004 July
May
2005
November
February
May
2004
July 2 0 1 November 1 0 1 2005 February 0 0 0 May 1 1 0 there was no significant seasonal variance in 17 rivers
O. javanica (n ) 21)
2004 July
2005
November
February
May
2004
July 0 2 1 November 0 0 0 2005 February 1 3 0 May 2 3 0 there was no significant seasonal variance in nine rivers
J. effusus (n ) 6)
2004 July
2005
November
February
May
2004
July 0 0 0 November 0 0 0 2005 February 0 0 0 May 0 0 0 there was no significant seasonal variance in five rivers
P. australis
2003
(n ) 12)
August
2003 2004
2004 July
2005
November
August 1 July 0 November 0 0 2005 May 0 1 there was no significant seasonal variance in
1 0
May 2 1 2
1 seven rivers
a The numbers represent the number of rivers in which the mean δ15N value of individuals sampled at the time shown in the row was significantly higher than those sampled at the time shown in the column.
especially in three rivers where the mean δ15Nplant values of P. australis were higher than 9.0‰ despite an overall mean of less than 6.0‰ (Figure 1). The particle size distribution of habitat soils is also an important factor determining water permeability and hydrologic connectivity with river water. Riparian plants that prefer microhabitats with high hydrologic connectivity with the river channel are able to utilize inorganic N in the river water as their N nutrition, whereas species with poor hydrologic connectivity are not able to utilize it. Because P. japonica is a pioneer species on sandy or rougher sediments near the river channel, their habitat would be high in water permeability and hydrologic connectivity to the river channel. In contrast, P. australis prefers habitats with muddy sediments that decrease hydrologic connectivity to the river channel (13). The habitats of other species were intermediate between the above two species. 7840
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 21, 2008
Interspecies differences in N nutrition could also affect the coorelation between δ15Nplant and δ15Nnitrate values. At least three types of difference in N nutrition should be considered: (1) the association of N from atmospheric deposition with low δ15N values, (2) the fact that primary N sources for macrophytes can vary among three potential sources (i.e., NH4+, NO3-, and organic N), and (3) isotope fractionation in the assimilation of nitrate. In the first situation, δ15Nplant would be depleted, especially in those plant species that depend on atmospheric N, which is known to be generally depleted in 15N (3), as their source for N nutrition. If the contribution of atmospheric N was constant across the rivers, the range of depletion in δ15Nplant relative to δ15Nnitrate should be large in rivers with high δ15Nnitrate. Meanwhile, if the contribution of atmospheric N was low in the rivers with high δ15Nnitrate values because of the significant positive correlation between δ15Nnitrate and nitrate concentration (r ) 0.47, n ) 24, p ) 0.02), the range of depletion in δ15Nplant relative to δ15Nnitrate might be reduced. In the second possible explanation, δ15N values of ammonium, nitrate, and organic N would be different each other (3). In nitrification, δ15N values of ammonium would be enriched relative to those of nitrate; so the utilization of ammonium would reduce the correlation between δ15Nplant and δ15Nnitrate values. Furthermore, mycorrhizal type and infection rate may result in a depletion of δ15Nplant values via utilization of organic N because the δ15N value of benthic particulate organic matter (BPOM) was lower than that of nitrate (Supporting Information 2). This process could explain why some mycorrhizal species (i.e., Salix spp.) have depleted levels of δ15N relative to nitrate. In addition to the difference in source δ15N values, the degree of fractionation during assimilation is known to vary depending on the source of N (3), which could result in a decreased correlation between δ15Nplant and δ15Nnitrate values. Nitrogen isotope fractionation in ammonium assimilation is generally larger than that in nitrate assimilation (14). And, isotope fractionation during mycorrhizal transfer of N to host plants is also significant (15, 16). Isotope fractionation in assimilation of nitrate is also one of the more plausible mechanisms that would decrease δ15Nplant relative to δ15Nnitrate. The isotope fractionation in assimilation of nitrate is known to vary from about -6.5 to +3.2‰ (generally from -5 to 0‰) depending on the vascular plant species and nitrate concentration (14, 17, 18). Isotope fractionation in the assimilation of nitrate is expected to be large under high concentrations of nitrate (16). If the nitrate concentration around roots was high in rivers with high δ15Nnitrate values, δ15Nplant values would be much depleted relative to δ15Nnitrate via isotope fractionation. On the contrary, when the nitrate pool around roots was exhausted via plant uptake, δ15Nplant and δ15Nnitrate values would be similar around the roots (3). Finally, inherent species-level characteristics in N metabolism should be also considered as a group of processes that could alter the relationships between δ15Nplant and δ15Nnitrate values. Among the six analyzed species, only P. thunbergii is an annual plant, whereas the others are perennials. This difference could potentially cause interspecies differences in N integration time. Although the correlation between δ15Nplant and δ15Nnitrate values was not particularly low in P. thunbergii (Figure 1), riparian macrophytes with significantly short N integration time would lower the correlation between δ15Nplant and δ15Nnitrate values. We analyzed δ15N of the foliar part of this plant as a representative value for the entire plant. However, we should carefully consider the selection of plant parts for analysis, especially in a species with large interpart difference in N isotope fractionation and integration time. Among the four species of riparian plants whose δ15N values were correlated with land-use parameters and δ15Nnitrate, P. japonica was the best indicator of δ15Nnitrate
in surface river water irrespective of the sampling timing. If we select riparian species favoring habitats with good hydrologic connectivity between the river channel and roots (i.e., a sandy habitat and shallow root distribution), it would be possible to determine the time-integrated δ15N values of nitrate in surface water. These data suggest that some riparian macrophyte species are useful recorders of δ15N values of nitrate in river waters and provide an opportunity to develop a tool for assessing the inputs of sewage-derived nitrogen taking changes in time into consideration. Furthermore, by dendrochronologically analyzing the δ15N values of riparian willows that were well correlated with δ15Nnitrate, it should be possible to successfully reconstruct the historical changes in the inputs of sewage-derived nitrogen into river ecosystems.
Acknowledgments We thank S. Yachi from the Research Institute for Humanity and Nature for his useful comments. This study was supported by the Research Project 3-1 of the Research Institute for Humanity and Nature and by the CREST project of the Japan Science and Technology Agency.
Supporting Information Available A map of the Lake Biwa watershed and a table of land-use and water-quality parameters. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Turner, R. E.; Rabalais, N. N. Changes in Mississippi River water quality this century. BioScience 1991, 41, 140–147. (2) Kinzing, A. P; Socolow, R. H. Human impacts on the nitrogen cycle. Phys. Today 1994, 24–31. (3) Kendall, C.; McDonnell, J. J. Isotope Tracers in Catchment Hydrology; Elsevier: Tokyo, 1998. (4) Mayer, B.; Boyer, E.; Goodale, C.; Jaworski, N. A.; Breemen, N. V.; Howarth, R. W.; Seitzinger, S.; Billen, G.; Lajtha, K.; Naedelhoffer, K.; Dam, D. V.; Hetling, L. J.; Nosal, M.; Paustian, K. Sources of nitrate in rivers draining sixteen watersheds in the northeastern U.S.: Isotopic constraints. Biogeochemisty 2002, 57/58, 171–197.
(5) McClelland, J. W.; Valiela, I. Linking nitrogen in estuarine producers to land-derived sources. Limnol. Oceanogr. 1998, 43, 577–585. (6) Cloern, J. E.; Canuel, E. A.; Harris, D. Stable carbon and nitrogen isotope composition of aquatic and terrestrial plants of San Francisco Bay estuarine system. Limnol. Oceanogr 2002, 47, 713–729. (7) Ho¨gberg, P. 15N natural abundance in soil-plant systems. New Phytol. 1997, 137, 179–203. (8) Saurer, M.; Cherubini, P.; Ammann, M.; Cinti, B. D.; Siegwolf, R. First detection of nitrogen from NOx in tree rings: a 15N/14N study near a motorway. Atmos. Environ. 2004, 38, 2779–2787. (9) Choi, W. J.; Lee, S. M.; Chang, S. X.; Ro, H. M. Variation of δ13C and δ15N in Pinus densiflora tree-rings and their relationship to environmental changes in eastern Korea. Water, Air, Soil, Pollut. 2005, 164, 173–187. (10) Sigman, D. M.; Casciotti, K. L.; Andreani, M.; Barford, C.; Galanter, M.; Bohlke, J. K. A bacterial method for the nitrogen isotopic analysis of nitrate in seawater and freshwater. Anal. Chem. 2001, 73, 4145–4153. (11) Kamio, A. On the process of polder land drainage and structural changes of Phragmites communis community in the Hachirogata center polder. Jpn. J. Ecol 1982, 32, 357–364. (12) Ostrom, N. E.; Hedin, L. O.; Fischer, J. C.; Robertson, P. Nitrogen transformations and NO3- removal at a soil-stream interface: a stable isotope approach. Ecol. Appl. 2002, 12, 1027–1043. (13) Miyawaki, A.; Okuda S. Vegetation of Japan Illustrated; Shibundo Co.: Tokyo, 1990. (14) Yoneyama, T.; Matsumaru, T.; Usui, K.; Engelaar, W. M. H. G. Discrimination of nitrogen isotopes during absorption of ammonium and nitrate at different nitrogen concentrations by rice (Oryza sativa L.) plants. Plant Cell Environ. 2001, 24, 133– 139. (15) Kohzu, A.; Tateishi, T.; Yamada, A.; Koba, K.; Wada, E. Nitrogen isotope fractionation during nitrogen transport from ectomycorrhizal fungi, Suillus granulates, to the host plant Pinus densiflora. Soil Sci. Plant Nutr. 2000, 46, 733–739. (16) Hobbie, E. A.; Colpaert, J. V. Nitrogen availability and colonization by mycorrhizal fungi correlate with nitrogen isotope patterns in plants. New Phytol. 2003, 157, 115–126. (17) Kohl, D. H.; Shearer, G. Isotopic fractionation associated with symbiotic N2 fixation and uptake of NO3- by plants. Plant Physiol. 1980, 66, 51–56. (18) Mariotti, A.; Mariotti, F.; Champigny, M-L.; Amarger, N.; Moyse, A. Nitrogen isotope fractionation associated with nitrate reductase activity and uptake of NO3- by Pearl Millet. Plant Physiol. 1982, 69, 880–884.
ES801113K
VOL. 42, NO. 21, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7841
