Variations in Nitrogen-15 Natural Abundance of Plant and Soil Systems in Four Remote Tropical Rainforests, Southern China

Oecologia DOI 10.1007/s00442-013-2778-5

ECOSYSTEM ECOLOGY - ORIGINAL RESEARCH

Variations in nitrogen-15 natural abundance of plant and soil systems in four remote tropical rainforests, southern China

Ang Wang • Yun-Ting Fang • De-Xiang Chen • Keisuke Koba • Akiko Makabe • Yi-De Li • Tu-Shou Luo • Muneoki Yoh

Received: 11 January 2013 / Accepted: 5 September 2013 Ó Springer-Verlag Berlin Heidelberg 2013

Abstract The foliar stable N isotope ratio (d15N) can uptake and mycorrhizal N transfer, and by direct uptake of ? 15 provide integrated information on ecosystem N cycling. atmospheric NH3/NH4 . The variation in foliar d N Here we present the d15N of plant and soil in four remote among species (by about 6 %) was smaller than in many typical tropical rainforests (one primary and three sec- N-limited ecosystems, which is typically about or over ondary) of southern China. We aimed to examine if (1) 10 %. The primary forest had a larger N capital in plants foliar d15N in the study forests is negative, as observed in than the secondary forests. Foliar d15N and the enrichment other tropical and subtropical sites in eastern Asia; (2) factor (foliar d15N minus soil d15N) were higher in the variation in d15N among different species is smaller com- primary forest than in the secondary forests, albeit differ- pared to that in many N-limited temperate and boreal ences were small, while there was no consistent pattern in ecosystems; and (3) the primary forest is more N rich than soil d15N between primary and secondary forests. the younger secondary forests and therefore is more 15N enriched. Our results show that foliar d15N ranged from Keywords Foliar stable nitrogen isotope ratio -5.1 to 1.3 % for 39 collected plant species with different Nitrogen availability Nitrogen status Southern growth strategies and mycorrhizal types, and that for 35 China Tropical forests - 15 species it was negative. Soil NO3 had low d N(-11.4 to - -3.2 %) and plant NO3 uptake could not explain the 15 ? negative foliar d N values (NH4 was dominant in the soil Introduction inorganic-N fraction). We suggest that negative values ? 15 might be caused by isotope fractionation during soil NH4 The foliar stable N isotope ratio (d N) can provide an integrated index of the openness of ecosystem N cycling (i.e., the ratio of N input or output relative to the standing N pool) (Amundson et al. 2003) and has the potential to Communicated by Hormoz BassiriRad. reveal spatial and temporal patterns of N cycling as well as how N cycling is altered by disturbances (Robinson 2001; & A. Wang D.-X. Chen ( ) Y.-D. Li T.-S. Luo Craine et al. 2009; Pardo and Nadelhoffer 2009; Fang et al. Research Institute of Tropical Forestry, Chinese Academy 15 of Forestry, Guangzhou 510520, China 2012a). For instance, foliar d N has been demonstrated to e-mail: [email protected] increase with increasing N availability, as evidenced by positive correlations of foliar d15N with net N minerali- & Y.-T. Fang ( ) zation and/or nitrification (Garten and Miegroet 1994; State Key Laboratory of Forest and Soil Ecology, Institute of Applied Ecology, Chinese Academy of Sciences, Shenyang Kitayama and Iwamoto 2001; Pardo et al. 2006; Craine 110164, China et al. 2009; Cheng et al. 2010; Watmough 2010), with N e-mail: [email protected] deposition (Pardo et al. 2006), with increased N leaching following clear-cutting (Pardo et al. 2002) and N additions K. Koba A. Makabe M. Yoh Institute of Agriculture, Tokyo University of Agriculture and (Hogberg and Johannisson 1993), and with foliar N con- Technology, Saiwai-cho 3-5-8, Fuchu, Tokyo 183-8509, Japan centration (Pardo et al. 2006; Craine et al. 2009). With the 123 Oecologia globalization of N deposition, there is increasing concern variability in foliar d15N values within species in 21 tem- about the impact of excess N on ecosystems, and a number perate forests. However, an ever wider range of foliar d15N of studies have been conducted to exploit d15N (natural (over 10 %) is observed in some tropical forests (Viani abundance of 15N) as a ‘‘natural’’ means to evaluate the et al. 2011), while this is not observed in others (e.g., dynamics of ecosystem N cycling and to assess N satura- Houlton et al. 2007). tion at both regional and global scales (Robinson 2001; Nevertheless, previous studies on ecosystem d15N are Pardo et al. 2006; Craine et al. 2009; Fang et al. 2011a). based primarily on research from North America, Europe, Compared to temperate and boreal forests, tropical and South America, Australia, and Africa (e.g., Martinelli et al. subtropical forests in general are considered to be richer in 1999; Amundson et al. 2003; Pardo et al. 2006; Craine et al. N, so that plants are expected to be more enriched in 15N. 2009). Limited studies exist for a large area of eastern Asia The foliar d15N in most tropical forests was found to be (Fang et al. 2012a). With increasing concerns over the positive and significantly higher than that in temperate impacts of elevated N deposition in this region, it is impor- forests (3.7 ± 3.5 vs. -2.8 ± 2.0 %; Martinelli et al. tant to understand its current forest ecosystem N status. 1999), supporting this expectation. However, recent Foliar d15N ranged from -6.6 to 0.7 % in 14 subtropical research found that the d15N of some tropical forests is not forests along an urban–rural transect in southern China (Fang always close to the averages compiled by Martinelli et al. et al. 2011a). In eastern Asia, low foliar d15N values are also (1999) for tropical forests. A previous study by Fang et al. reported for the tropical forests in Borneo, Malaysia (-4.8 to (2011a) shows that both foliage and 0- to 10-cm-deep 0.02 %; Kitayama and Iwamoto 2001) and for a subtropical mineral soil d15N (on average -3.8 and 0.9 % for foliage forest in Guandaushi, Taiwan (-6to-2.2 %; Liu et al. and soil, respectively) are much lower in 14 subtropical 2006). Thus it is critically important to explore if negative forests in southern China than the averages of tropical foliar d15N is widespread in eastern Asia, and why foliar forests (3.7 and 9.0 % for foliage and soil, respectively) d15N values are often reported to be negative, in order to but close to those of temperate forests (-2.8 and 2.0 %) assess ecosystem N status at a regional scale. compiled by Martinelli et al. (1999). Furthermore, foliar In this study, we selected four typical tropical forests in d15N was found to be lower in the urban forests than in the a remote area of Hainan Island, southern China to inves- rural forests, and lower values in the urban areas were tigate d15N in plant and soil systems. These forests con- - proposed to be caused by increased plant NO3 uptake due sisted of one primary forest and three secondary forests, to elevated N deposition (Fang et al. 2011a). Negative and were expected to have different N capital. The objec- values for plants are also reported in other tropical regions, tives of the study were to evaluate whether: e.g., south-eastern Brazil (Scarano et al. 2001), but the 1. Foliar d15N in the study forests is negative, as observed reasons for this have not been elucidated. in other tropical and subtropical sites in eastern Asia. Foliar d15N varies broadly with different species. Some 2. The difference in foliar d15N among species is smaller studies have reported that conifers tend to have lower d15N compared to that in many N-limited temperate and than hardwoods (Pardo et al. 2007; Templer et al. 2007) boreal ecosystems. although others reported no difference (Gebauer and Die- 3. The primary forest is more N rich than the three trich 1993). Also, in several studies, strong and consistent younger secondary forests (indicated by the total N patterns of the relative foliar d15N value by species have pool and N availability) and thereby is more 15N been reported (Templer et al. 2007). Meanwhile, the enriched in plants and soil. magnitude of variation in foliar d15N among species at a 15 - given site seems to be dependent on ecosystem N status (N In addition, the d N of soil NO3 was determined for cycling rate, forest floor and foliar chemistry). Nadelhoffer one secondary forest and the primary forest to investigate if et al. (1996) hypothesized that foliar d15N should have a it is negative and, if so, to what extent. Proportionally - ? larger variation (wide range) in N-limited conditions such increased NO3 uptake as compared to NH4 and dis- as tundra ecosystems than in N-rich conditions such as in solved organic N (DON) is expected to lower foliar d15N the tropics, because strongly N-limited systems have a (Fang et al. 2011a). small N pool, so that different plants exploit different forms of available N with distinct isotopic compositions and are colonized by mycorrhizas to facilitate N uptake, Materials and methods resulting in niche differentiation in N acquisition. As a consequence of the niche differentiation, foliar d15N values Study site varied by about 10 % among species within each of two moist tussock tundra sites (-7.0 to 3.0 %; Nadelhoffer The research was conducted in the Jianfengling National et al. 1996). Pardo et al. (2006) found a similar level of Natural Reserve, which is located in the southwestern part 123 Oecologia of Hainan Island (18°230–18°500N, 108°360–109°050E) Sampling (Fang et al. 2004; Chen et al. 2010). The total area of the reserve is about 470 km2, of which about 150 km2 is Sampling of leaves and soils for the analysis of d15N was mountain rain forest. The reserve experiences a tropical performed in July 2010 (growing season). The canopy monsoon climate, characterized by high annual tempera- leaves of common species were collected from three of the tures (19.8 °C), humidity (88 %), and plentiful precipita- five plots for the vegetation survey and were mixed for the tion (mean annual precipitation, 2,449 mm). same species within the same forest. Meanwhile, the sur- Approximately 80–90 % of the annual precipitation falls face soil was sampled using a soil core (5-cm inner during the wet season (May–October). The wet season is diameter) and divided into four soil layers: Oi, Oa?e, characterized by subtropical high pressure systems, and 0–5 cm and 5–10 cm mineral soil. Fine roots (\2 mm) rainfall is mainly from typhoons or thunderstorms, or is were collected from soil cores. topographic rain (Chen et al. 2010). Three soil sampling campaigns were conducted in order In this study, four forests in the reserve were selected, to determine soil N availability and/or 15N abundance for - one primary forest and three secondary forests (Chen et al. soil NO3 . In September 2011 the 0- to 10-cm mineral soil 2010). The four forests are 2 to 7 km apart from each other layer was sampled using a soil core from the same plot to ? - and locate at 800–900 m a.s.l. Three secondary forests determine available NH4 and NO3 concentrations. In were naturally regenerated after cutting in the 1960s and April 2012, in situ incubation was performed to determine 1970s. No fertilization was recorded in all four forests. the rate of net N mineralization and nitrification using PVC Dominant species in the forests are Livistona saribus pipes. At the beginning of the study, five 15-cm-long PVC (Lour.) Merr. ex A. Chev (Palmaceae), Lasianthus calyci- pipes (7.5 cm in diameter) were inserted to a depth of nus Geddes (Rubiaceae), Symplocos anomala Brand 10 cm in each forest. The top ends were covered by a cap (Symplocaceae), Symplocos lancifolia Sieb. & Zucc to avoid the input of throughfall and three holes were made (Symplocaceae), Ardisia quinquegona Bl. (Myrsinaceae), in the upper part of the PVC pipes (above the forest floor) Calamus rhabdocladus Burret (Palmaceae), Psychotria to keep them aerated. Meanwhile soil around each PVC hainanensis Li (Rubiaceae), Dacrycarpus imbricatus (Bl.) pipe (within 10 cm) was collected and mixed to determine de Laub. var. patulus de Laub. (Podocarpaceae), Symplocos initial inorganic N concentration. After 30 days, PVC soil congesta Benth. (Symplocaceae), Syzygium araiocladum samples were retrieved and kept in cool conditions during Merr. & L. M. Perry (Myrtaceae), Pinanga discolor Burret the transportation to the laboratory. In June 2012, nine (Palmaceae), Alseodaphne hainanensis Merr. (Lauraceae). composite samples from the 0- to 10-cm mineral soil layer The soil is a lateritic yellow soil. in the primary forest and one secondary forest were collected to determine potential rates of net N mineralization 15 - Vegetation survey and nitrification, and N abundance for soil NO3 .

Five plots of 160 m by 20 m were set up in each forest to Laboratory treatment and chemical analysis measure height and diameter at breast height (DBH) of all trees with DBH above 5 cm, and species were recorded for In the laboratory, all plant and organic soil layer (Oi and all measured trees. There are as many as 250 tree species Oa?e) samples for 15N analysis were dried at 60 °Ctocon- with a DBH greater than 1 cm within 1 ha in our study stant weight and weighed. Mineral soil from each plot was forests, so we used mixed-species regression models to passed through a 2-mm-mesh sieve to remove fine roots and estimate tree biomass (Chen et al. 2010) instead of species- coarse fragments. Subsamples of oven-dried foliage, floor specific regression models which are used in temperate materials, and mineral soils were ball milled and analyzed for zones. The biomass (W) of each individual tree for the d15N and total N and C concentrations by elemental analysis– trunk (1), branches (br), leaves (l), bark (bk) and roots isotope ratio mass spectrometry (EA1112 coupled with Delta- (r) was estimated by the following equations (Zeng et al. XP; ThermoFisher Scientific and Yokohama, Japan). Cali- 1997). brated DL-alanine (d15N =-1.7 %), glycine (d15N = 10.0 15 2 0.992674 %), and histidine (d N =-8.0 %) were used as the internal Trunk: Wt = 0.022816(D H) 15 2 0.902418 standards. The d N of the sample relative to the standard Bark: Wbk = 0.006338(D H) 2 0.999046 (atmospheric N2) was expressed as the following: Branch: Wbr = 0.005915(D H) 2 0.804661 15 Leaf: Wl = 0.005997(D H) d N ¼½ðRsample=RstandardÞ11000; 2 1.11527 Root: Wr = 0.003612(D H) 15 14 where R represents the isotope ratio ( N/ N) and Rstandard 15 14 where D represents DBH (cm) and H represents height is the N/ N for atmospheric N2. The analytical precision (m). for d15N was, in general, better than 0.2 %. 123 Oecologia

The soils collected in 2011 and 2012 were passed Casciotti (2011). The N isotopic composition was mea- through a 2-mm-mesh sieve immediately after arrival at the sured for the produced N2O. We ran several standards - laboratory and the sieved fresh mineral soil was used to (USGS32, 34, 35 and IAEA NO3 ) to obtain the calibra- ? - determine the extractable inorganic N (NH4 and NO3 ) tion curve for drift and blank (Fang et al. 2012b). The pool. Brieﬂy, 10 g mineral soil was extracted with 50 ml of average SDs for replicate analysis of an individual sample ? 15 - 15 - 2 M KCl solution. The NH4 concentration in the extract were ±0.2 % for d NinNO3 . The d N value for NO3 was determined by the indophenol blue method followed newly produced from nitriﬁcation during the incubation - by colorimetry, and the NO3 concentration was deter- was calculated using the following equation.

15 15 d NNO3newly produced ¼½ðd NNO3after incubation NO3 concentrationafter incubationÞ 15 ðd NNO3before incubation NO3 concentrationbefore incubationÞ= ðNO3 concentrationafter incubation NO3 concentrationbefore incubationÞ

- mined after cadmium reduction to NO2 , followed by the Statistical analysis sulfanilamide–nicotinamide adenine dinucleotide reaction (Liu et al. 1996). Ten milligrams soil sample was mixed One-way ANOVA was performed to examine the differ- with 50 ml deionized water and the pH in the suspension ences in the investigated variables among forests, among was measured using a glass electrode after 30 min shaking ecosystem components, growth strategy and between (Liu et al. 1996). The net N mineralization rate was cal- mycorrhizal association types. All analyses were conducted culated as the difference between the initial and incubated using the PASW Statistics 19.0 for Windows. Statistically inorganic N concentrations, and nitrification was calculated significant differences were set at a P-value of 0.05 unless - as the difference between the initial and incubated NO3 otherwise stated. concentrations. Percent nitrification was defined as the percentage net nitrification rate of the net N mineralization rate. Results Soils sampled in June 2012 from the primary forest and one secondary forests were used for laboratory incubation General properties of vegetation and soil in order to determine potential rates of net N mineralization 15 - and nitrification, and N abundance for soil NO3 , using a The biomass of the primary forest was higher (by similar method as described in Fang et al. (2012b). Briefly, 43–65 %) than those of three secondary forests (Table 1). about 10 g of freshly sieved soil from each composite The primary forest also had a greater mean height and sample was put into a glass jar (120 ml) and incubated at a DBH than the secondary forests (by 9–17 %). The primary constant temperate of 25 °C for 1 week. During the labo- forest and secondary forests 2 and 3 generally contained ratory incubation, the jars were opened for about 30 min more C and N than the secondary forest 1 (33–77 % for C, every 1–2 days to prevent the occurrence of denitrification. and 40–60 % for N). Soil pH and C/N was similar among We did not add more water in order to keep the soil the four forests. moisture constant during the incubation because our previous study showed that water loss was minor during the Soil N availability 7-day incubation (about 2 %; Fang et al. 2012b). Incubated ? - ? soils were extracted and NH4 and NO3 in the soil extract The available NH4 concentration in the 0- to 10-cm were determined in the same way as described above. mineral soil layer was between 3.1 and 5.7 mg N kg-1 Soil extracted with KCl before and after laboratory during the sampling period (Fig. 1a, b), with no significant 15 - incubation was used to determine d N for NO3 by the difference between successional stages or between sam- - denitrifier method (Sigman et al. 2001). The N2O produced pling date. The soil available NO3 concentration was - -1 by denitrifiers from NO3 was analyzed by an isotope-ratio 0.2–0.7 mg N kg in September 2011, and was signifi- mass spectrometer (Sercon) with the same machine setting cantly lower than in April 2012 (0.7–2.2 mg N kg-1). ? - and protocols for the measurements as in McIlvin and Similar to NH4 , no difference in NO3 concentration was

123 Oecologia

Table 1 Properties of vegetation and the 0- to 10-cm mineral soil layer in the four tropical forests Forest Biomass (mg ha-1) Mean DBH (cm) Mean height (m) Soil pH Soil C conc. (%) Soil N conc. (%) Soil C/N

Secondary 1 240 (35) a 13.3 (0.4) b 9.10 (0.3) b 4.49 (0.07) a 1.37 (0.13) b 0.10 (0.01) b 13.8 Secondary 2 258 (44) a 13.7 (0.9) a 11.04 (0.5) a 4.44 (0.03) a 2.12 (0.08) a 0.15 (0.01) a 14.1 Secondary 3 223 (35) a 16.9 (0.6) a 10.2 (0.4) ab 4.33 (0.06) a 1.82 (0.12) ab 0.14 (0.01) a 12.5 Primary 368 (62) a 16.0 (1.0) ab 11.86 (0.5) a 4.84 (0.27) a 2.43 (0.27) a 0.16 (0.01) a 14.6 One SEM in parentheses; n = 5 for vegetation variables and n = 6 for soil variables DBH Diameter at breast height, conc. concentration Different letters indicate signiﬁcant differences among forests at P \ 0.05

a 8 Sep 2011 b 8 Apr 2012 c 25 Apr 2012 A 7 NH + 7 NH + Min.

4 4 ) - - Nitrifi. NO NO A -1 3 3 20 ) ) -1 6 -1 6 AB mon -1 5 5 BC A A AB C 15 (mg N kg - (mg N kg -

3 A 4 B B 3 4

a 10 or NO

3 or NO 3 + + 4 4 A NH 2 NH 2 5

Min. or Nitrifi. (mg N kg a b b 1 a 1 b b ab b b b b 0 0 0

Primary Primary Primary SecondarySecondary 1 Secondary 2 3 SecondarySecondary 1 Secondary 2 3 SecondarySecondary 1 Secondary 2 3 Forests

? - ? Fig. 1 Available NH4 and NO3 concentrations (mean ± SE) in letters indicate significant difference in NH4 concentration (a, b)or the 0- to 10-cm mineral soil layer sampled in September 2011 (a) and mineralization rate (c) among forests, and different lowercase letters - in April 2012 (b), and the rates of net N mineralization and indicate significant difference in NO3 concentration (a, b)or nitrification (mean ± SE) in April 2012 (c). Different uppercase nitrification rate (c) among forests observed between successional stages (Fig. 1a, b). The soil forest and the secondary forest 3 (-2.3 %) was not sig- net N mineralization rate ranged from 5.8 to nificant (Table 2). The mean foliar N concentration was 16.1 mg N kg-1 month-1, and the net nitrification rate higher in the primary forest (1.74 %) than in the secondary ranged from 0.1 to 7.8 mg N kg-1 month-1, with the forests (1.35–1.49 %; Table 2). There were three species highest values in secondary forest 3 (Fig. 2c). collected from all four forests, of which two displayed the same trend as the means among forests (Fig. 3). Plant d15N Mean foliar d15N was between -2.8 and -2.0 % in trees, between -3.2 and -1.2 % and between -2.8 and Samples of 39 species were collected from four forests 0.1 % in woody shrubs and non-woody shrubs (Fig. 4a). (Table 2). Foliar d15N of these species ranged from -5.1 to Across all the four forests, average d15N values were -2.4, 1.3 %, with an arithmetic mean of -2.3 % (Table 2; -2.2 and -2.2 % for trees, woody shrubs and non-woody Fig. 2). Of 39 species, 35 species (90 %) had negative shrubs, respectively. Two-way ANOVA (using forest type values (Fig. 2). Foliar %N ranged from 0.69 to 2.61 %, and and growth strategy as main factors) showed that there was the mean was 1.50 % (Table 2). There was no correlation no significant difference between forests and growth between d15N and %N for the species (Fig. 2). strategy in foliar d15N. The mean foliar d15N was -1.4 % in the primary forest, No N-fixing species were collected. Of 39 species, 36 and was significantly greater than in secondary forests 1 species were associated with arbuscular mycorrhizal (AM) and 2 (both -2.5 %). The difference between the primary fungi, one species with ectomycorrhizal (ECM) fungi

123 Oecologia

Fig. 2 Foliar stable N isotope a 2 ratio (d15N)(a) and N concentration (b) (mean ± SE) 0 for all plant species (n = 1–4). ‰) -2 Species abbreviations use the N ( 15 first three letters of the genus δ -4 name and the first three letters -6 of the specific epithet (see Table 2) b 3

1 N con. (%)

Lit.var Lic.spi Lit.bav Eur.cil Lit.elm Liv.sar Lit.lon Ole.tsoSyz.araBla.cocAdi.haiNep.topCyc.blaMac.chiAls.haiDas.rosCyc.patFic.vasCal.rhaLin.robCry.chi Bei.laeMan.haiArd.qui Pin.disPsy.haiGlo.coc Las.cal Acr.pedXan.hai Sym.ano Dac.imbNeo.cam Pan.aus Sym.sumPou.ann Sym.con

Sym.lancifolia Sym.lancilimba Species

Table 2 Foliar stable N isotope ratio (d15N), C and N concentrations and C/N ratios for each sampled species in four tropical forests Forests/growth strategies Species Mycorrhizal types d15N(%) N conc. (%) C conc. (%) C/N ratio

Primary forest Tree Alseodaphne hainanensis AM -3.1 1.30 54.2 41.6 Nephelium topengii AM -3.7 1.40 54.9 39.1 Pouteria annanmensis AM -0.7 2.61 52.6 20.2 Xanthophyllum hainanense AM 0.3 2.51 52.1 20.8 Litsea baviensis ECTO -2.5 2.40 49.5 20.7 Woody shrub Lithocarpus longipedicellatus AM 0.1 1.59 52.1 32.9 Non-woody shrub Calamus rhabdocladus – -2.4 1.45 46.8 32.4 Licuala spinosa AM -1.7 1.51 49.1 32.5 Pandanus austrosinensis AM -0.1 0.99 48.2 48.9 Pinanga discolor AM -1.7 1.54 46.6 30.3 Livistona saribus AM -0.4 1.81 49.3 27.3 Secondary forest 1 Tree Cryptocarya chinensis AM -2.4 1.65 55.3 33.5 Lindera robusta AM -2.6 1.28 54.8 42.6 Litsea variabilis AM -5.1 1.80 54.0 29.9 Manglietia hainanensis AM -2.0 1.32 48.5 36.7 Woody shrub Psychotria hainanensis – -1.5 1.56 48.6 31.1 Adinandra hainanensis AM -3.9 1.07 48.9 45.9 Blastus cochinchinensis AM -5.3 1.70 44.5 26.1 Glochidion coccineum AM -1.3 1.05 47.5 45.2 Lasianthus calycinus AM -1.4 1.38 43.9 31.7 Symplocos congesta AM 1.3 1.33 42.8 32.2 Symplocos lancifolia AM -2.0 1.66 41.6 25.0 Symplocos anomala AM -3.1 1.23 41.5 33.7 Non-woody shrub Calamus rhabdocladus – -2.2 1.91 48.9 25.7 Licuala spinosa AM -3.4 0.87 30.4 34.8 Pandanus austrosinensis AM -3.5 0.69 48.9 70.6 Livistona saribus AM -2.1 1.14 49.0 42.9 Secondary forest 2 Tree Acronychia pedunculata AM 0.1 2.39 48.5 20.3

123 Oecologia

Table 2 continued Forests/growth strategies Species Mycorrhizal types d15N(%) N conc. (%) C conc. (%) C/N ratio

Cyclobalanopsis patelliformis AM -3.0 1.26 50.5 39.9 Dacrycarpus imbricatus AM -2.5 1.36 51.9 38.3 Dasymaschalon rostratum AM -3.1 1.99 48.2 24.2 Eurya ciliata AM -1.9 1.12 44.6 39.8 Ficus vasculosa AM -2.8 1.81 44.2 24.5 Lithocarpus elmerrillii AM -1.3 1.54 51.8 33.6 Machilus chinensis AM -3.4 1.02 52.3 51.4 Symplocos lancilimba AM -1.3 1.52 40.6 26.7 Woody shrub Blastus cochinchinensis AM -4.1 1.56 45.6 29.3 Lasianthus calycinus AM -0.9 1.27 43.5 34.1 Psychotria hainanensis – -1.3 1.62 48.1 29.8 Syzygium araiocladum AM -5.0 1.18 50.0 42.4 Symplocos lancifolia AM -3.2 1.53 40.2 26.4 Non-woody shrub Calamus rhabdocladus – -4.1 2.00 47.6 23.8 Licuala spinosa AM -3.9 1.08 50.1 46.3 Livistona saribus AM -1.5 1.15 48.4 42.1 Secondary forest 3 Arbor Cyclobalanopsis blakei AM -3.6 1.80 51.3 28.5 Beilschmiedia laevis AM -2.4 1.52 51.2 33.7 Cryptocarya chinensis AM -2.6 1.58 54.9 34.7 Woody shrub Psychotria hainanensis – -2.5 1.41 45.8 32.6 Lasianthus calycinus AM -0.9 1.46 44.4 30.4 Symplocos sumuntia AM -0.8 2.02 47.2 23.4 Olea tsoongii AM -5.0 1.23 48.5 39.5 Symplocos lancifolia AM -2.0 1.17 38.7 33.0 Ardisia quinquegona AM -1.9 1.69 46.3 27.4 Neolitsea cambodiana Lec. AM -2.5 1.33 51.9 39.0 Non-woody shrub Calamus rhabdocladus – -2.0 1.74 47.9 27.4 Licuala spinosa AM -3.7 1.23 48.3 39.4 Livistona saribus AM -0.4 0.91 49.34 54.1 AM Arbuscular mycorrhizal, ECTO ectomycorrhizal, – unknown

(Table 2). Across all forests, we found no significant dif- was no significant difference in d15N between the O layer ference between AM (on average -2.3 %) and ECM and plants. The mean d15N was between 0.9 and 1.9 % in species (-2.5 %; Table 2). the 0- to 5-cm mineral soil layer and between 1.8 and AM and EM species were considered to be significantly 2.7 % in the 5- to 10-cm mineral soil layer (Fig. 4a). different if the difference in their foliar d15N exceeded 2 Mineral soils were significantly 15N enriched relative to SDs of the AM species’ foliar d15N. plants (Fig. 4a). The difference in soil d15N among forests The d15N in fine roots varied from -2.1 to -0.6 %, and was small compared to the difference in foliar d15N averaged -1.1 % (1.1 % higher than foliar d15N; Fig. 4a). (Fig. 4a). Similarly, the difference in the soil N concen- There was no significant difference between two succes- tration was small although the difference was sometimes sional stages. statistically significant among forests (Fig. 4b).

15 - 15 Soil d N Soil available NO3 –d N

The d15N ranged from -1.5 to -1.2 % in the Oa?e layer, In June 2012, the soil collected from the primary forest and and from -1.8 to -1.0 % in the Oi layer (Fig. 4a). There the secondary forest 3 was selected to determine potential

123 Oecologia

a 0 b -1 )

‰ -2 a secondary 1 N ( -3 secondary 2 15 a a a δ -4 secondary 3 primary -5 c 2.5 d 2.0 a a aa 1.5 1.0

N con. (%) 0.5 0.0

Primary L.saribus L.spinosa Secondary 1 Secondary 2 Secondary 3 C.rhabdocladus Forests Species

Fig. 3 Foliar d15N and N concentration (mean ± SE) among forests (a, c) for three species (Livistona saribus, Licuala spinosa and Calamus rhabdocladus) which were collected in each (b, d). Different letters indicate signiﬁcant difference among forests at P \ 0.05

Discussion δ15N (‰) N con. (%) a -4 -3 -2 -1 0 1 2 3 b 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Foliar d15N value in comparison with other regions

Arbor * In our study of tropical forests, we found that most of Woody shrub * species were 15N depleted in their leaves, ranging from Non-woody shrub -5.1 to 1.3 % with an average of -2.3 %. Negative foliar 15 Fine root d N data were consistent with the existing data available for tropical and subtropical sites in eastern Asia, e.g., Oi layer * eastern China (averaged -4.8 % across 22 oak forests; Oa+e layer Kang et al. 2011), Guangdong Province (averaged -3.8 % 0-5 mineral soil across 14 forests; Fang et al. 2011a), Taiwan (-1.3 %; Liu * * secondary 1 secondary 2 5-10 mineral soil et al. 2006), Malaysia (-3.5 %; Kitayama and Iwamoto * secondary 3 15 primary 2001). These results suggest that negative foliar d Nis widespread in eastern Asia. 15 Fig. 4 d N(a) and N concentration (b) for plant and soil systems. Across the global data set for forest plants with AM Asterisks indicate significant difference among forests at P \ 0.05 fungi, foliar d15N is significantly correlated with foliar N concentration, MAT and MAP (Fig. 5). Foliar d15N values 15 - N production rates and d N for available NO3 . Mean- in our study plants with an AM association were below the while, pH, C and N concentrations, and d15N, were deter- regression lines for the global dataset (Fig. 5). Foliar d15N mined for these soil samples (Table 3). Consistent with the values were predicted to be 0.2, 0.7 and 2.8 % for our field observations (Fig. 1c), the secondary forest had higher study forests, according to the relationships between foliar net mineralization and nitrification rates (Table 3). Avail- d15N and N concentration, MAT and MAP, respectively, - 15 able NO3 before incubation was highly depleted in N, for the global forest dataset (Fig. 5), which are consider- which ranged from -11.4 to -3.2 % (Table 3), with ably higher (by 2.4–5.0 %) than the average value of averages of -5.7 and -10.6 % in the secondary forest 3 -2.2 % observed for AM species in the study forests and the primary forest (Table 3). The mean d15N value for (Table 2). - 15 available NO3 was -8.5 and -13.4 % lower than that for Why is foliar d N negative and relatively low in eastern bulk mineral soil in these two forests (Table 3). The newly Asian forests? There are several potential mechanisms for - 15 - produced NO3 during the laboratory incubation was also lower foliar d N. One of them is increased NO3 utili- 15 ? more enriched in N in the secondary forest than in the zation relative to NH4 and DON. This has been proposed primary forest (Table 3). by Fang et al. (2011a) to explain lower foliar d15N in the

123 15 - Oecologia Table 3 Soil properties and d N of soil-available NO3 in the 0- to 10-cm mineral soil layer of the primary forest and one secondary forest 15 ? - Forest/ pH d N N C C/N Moisture NH4 conc. NO3 conc. Net N mineral. Net nitriﬁcation Before After Newly replicate (%) conc. conc. (water/ (mg N kg-1 ) (mg N kg-1 ) rate (mg N kg-1 rate (mg N kg-1 incubation incubation produced -1 -1 - 15 - 15 - 15 (%) (%) fresh soil) day ) day ) NO3 -d N NO3 -d N NO3 -d N (%) (%) (%)

Secondary forest 3 1 4.15 1.6 0.18 2.43 13.6 0.23 9.0 4.1 0.9 1.7 -7.6 -2.5 -0.6 2 4.27 2.6 0.15 1.93 12.7 0.22 2.9 4.0 1.6 1.4 -7.7 -2.9 -1.0 3 4.12 2.7 0.19 2.21 11.9 0.20 3.8 4.1 1.0 1.0 -5.2 -1.9 0.1 4 4.13 2.4 0.18 1.89 10.8 0.23 5.3 4.6 1.1 1.6 -6.4 -2.1 -0.2 5 4.24 2.3 0.12 1.51 12.4 0.22 6.2 4.4 1.1 1.8 -6.7 0.0 2.4 6 4.03 3.7 0.19 2.07 10.8 0.21 2.6 4.4 0.8 0.9 -4.7 -0.6 2.4 7 4.08 3.1 0.17 2.20 13.3 0.23 3.7 5.5 1.3 1.3 -4.7 -0.5 2.0 8 4.08 3.9 0.15 1.77 11.5 0.22 2.4 4.2 0.9 1.0 -3.2 -0.2 1.5 9 3.92 3.3 0.18 2.13 12.0 0.23 5.7 5.0 0.9 1.4 -4.9 -0.5 1.8 Mean 4.11 2.8 0.20 2.00 12.1 0.20 4.6 4.5 1.1 1.3 -5.7 -1.2 0.9 SE 0.04 0.24 0.01 0.09 0.33 0.00 0.71 0.16 0.08 0.11 0.50 0.37 0.45 Primary forest 1 4.41 2.4 0.17 2.30 13.5 0.23 5.2 2.3 0.8 0.8 -11.0 -11.5 -11.7 2 4.31 2.1 0.16 2.16 13.2 0.23 8.4 1.7 0.6 0.4 -10.7 -13.8 -15.6 3 4.40 3.1 0.14 1.79 13.0 0.24 5.2 2.7 0.9 0.8 -10.1 -9.7 -9.6 4 4.49 2.5 0.20 2.95 14.6 0.23 9.0 2.2 0.6 0.8 -10.5 -11.3 -11.6 5 4.76 2.0 0.15 2.07 13.5 0.20 8.9 1.1 0.8 0.5 -10.2 -10.5 -10.6 6 4.65 3.6 0.13 1.83 14.0 0.23 7.0 2.1 0.5 0.6 -11.4 -10.8 -10.5 7 4.59 4.4 0.13 1.73 13.7 0.21 7.2 1.9 0.4 0.6 -10.9 -11.5 -11.9 8 4.75 2.9 0.13 1.68 13.3 0.22 10.7 2.0 0.8 0.5 -10.8 -14.4 -16.5 9 4.76 2.4 0.12 1.50 12.2 0.21 9.3 2.3 0.7 0.6 -9.9 -12.3 -13.5 Mean 4.57 2.8 0.10 2.00 13.4 0.20 7.9 2.0 0.7 0.6 -10.6 -11.8 -12.4 SE 0.06 0.26 0.01 0.15 0.22 0.00 0.63 0.15 0.05 0.05 0.16 0.51 0.78 P values 0.000 0.95 0.101 0.93 0.004 0.84 0.003 \0.001 0.001 \0.001 \0.001 \0.001 \0.001 Soils were sampled in June 2012 123 Oecologia

a 20 b c Global dataset Jianfengling 15

5 N (‰) 15 δ 0 Foliar -5

-10

-15 0 204060 -20 0 20 0 2000 4000 6000 8000 -1 Foliar N concentration (mg g ) MAT (oC) MAP (mm)

Fig. 5 Relationships between foliar d15N and N concentration (a), species collected in this study were associated with AM. Across the mean annual temperature (MAT; b) and mean annual precipitation global dataset, the relationships were all signiﬁcant (P \ 0.0001) and (MAP; c) for the study forests in Jianfengling and the unadjusted are shown in gray. a y = 0.05478x - 0.42395, R2 = 0.012; b, global dataset. The global dataset is restricted to forest plants with an y = 0.01252x2 - 0.01706x - 3.70331, R2 = 0.564; AM association {extracted from Craine [global foliar 15N(http://knb. c y = 0.00000089x2 ? 0.00501x - 4.0723, R2 = 0.188 ecoinformatics.org/knb/metacat/jcraine.18.2/knb)]}, since most of the

- ? urban and suburban forests they studied, because NO3 ,as accounted for 87 and 83 % of total inorganic N (NH4 plus 15 - a product of nitrification, is more N depleted than sub- NO3 ) uptake, respectively, in the primary forest and the ? ? strate NH4 and DON (Ho¨gberg 1997; Koba et al. 2010) secondary forest no. 3. This result confirms that NH4 and is a more dominant form of available inorganic N might be the main N nutrient in the study forests. under N-rich conditions (Aber et al. 1998). In the present In the above two-end-member model, no isotope effect 15 - ? - study, we found that the d N of soil NO3 was -11.4 to is assumed during either NH4 or NO3 uptake. However, -3.2 % and, on average, 8.5 and 13.4 % lower than the some studies suggest that isotopic discrimination against 15 15 ? corresponding bulk soil d N, respectively (Table 3). Thus N is actually large, and that NH4 uptake exerts a larger ? - if plants use DON and NH4 as N nutrition, plants may be isotope discrimination than NO3 uptake (Ariz et al. 2011). more 15N enriched. So we speculate that the isotopic fractionation during - ? However, plant NO3 utilization could not explain the NH4 uptake may be the main reason for negative foliar negative foliar d15N values for our study forests; soil- d15N values in the study forests. ? available N was dominated by NH4 , which accounted for, In addition, we suggest that plant leaves may make use ? ? on average, 92 and 83 % of the total inorganic N (NH4 of NH3 and/or NH4 directly from the atmosphere with - 15 plus NO3 ) in September 2011 and April 2012, respec- low d N in the study region, which in turn contributes to tively (Fig. 1). We run a two-end-member model to low foliar d15N. N deposition through precipitation in 2012 ? -1 -1 ? examine if plants mainly use NH4 as their main N was 6.1 kg N ha year , of which about half was NH4 nutrition. At the present time, d15N has not been measured (Fang et al., unpublished data). We did not determine d15N ? ? for soil NH4 . We assume that the N isotopic composition for NH4 for the study site, but the results showed that rain ? ? 15 of soil NH4 is identical to that of soil organic matter (both NH4 was greatly depleted in N at two sites in southern 2.8 %; Table 3), since there is minor isotope fractionation China, with d15N ranging from -16.6 to -1.3 % at during mineralization (Ho¨gberg 1997). The dominant type Dinghushan, Guangdong province (Koba et al. 2012) and of mycorrhizae in the study forest soils was arbuscular ranging from -19.8 to -10.4% in Guiyang, Guizhou - mycorrhizae (Table 3), which may impart a slight frac- province (Xiao et al. 2012). NO3 in precipitation was tionation of 2 % (Pate et al. 1993) or less (Michelsen et al. observed at two sites in Guangdong province, southern 1998). We thus assume the fine root to be 2 % lower than China to be enriched in 15N (Fang et al. 2011b; Koba et al. 15 - the preferred N source. Using measured d N for fine roots 2012). Thus, direct leaf uptake of atmospheric NO3 would - 15 15 (Fig. 4) and soil NO3 (Table 3) and assumed d N for soil raise foliar d N. ? ? 15 NH4 , we calculated the proportion of NH4 uptake and We suspect that higher foliar d N in the lowland - ? NO3 uptake and found that the utilization of NH4 tropical forests in Brazil (e.g., Terra-firme forests;

123 Oecologia

Martinelli et al. 1999) may be caused by seasonal flooding, different growth strategies and mycorrhizal types which is favorable for denitrification process. During (Table 3). The range of foliar d15N was about 6 % among - denitrification NO3 is consumed. Denitrification has a species (from -5.1 to 1.3 %) across four forests with - strong isotope effect on residual NO3 , albeit with variable different successional stages, and was smaller than that isotopic fractionation, mainly ranging from 10 to 30 % reported in N-limited boreal and temperate forests and (Houlton et al. 2006). If plants took up the residual 15N- some tropical savanna ecosystems, where the ranges are - enriched NO3 , then plants were likely to have a higher about or over 10 % (Nadelhoffer et al. 1996; Viani et al. d15N. This may be the one of reasons for the difference in 2011). Our results are consistent with those of a previous foliar d15N between Terra-firme forests and the forests study conducted in a N-saturated subtropical forest in growing on white-sand soil, known as ‘‘Amazon caatinga’’ Dinghushan, southern China, which showed a small vari- and ‘‘wallaba forest.’’ The latter often show much lower ation in foliar d15N (about 3 %) among five dominant d15N in plant tissues (Martinelli et al. 1999; Mardegan canopy species, and for Guandaushi subtropical forest in et al. 2009). Taiwan, where it was about 4 % (Liu et al. 2006). In eight On the other hand, our study forests may have a low N forest sites along an elevation from 700 to 3,100 m on Mt status with respect to relatively low foliar d15N. This is Kinabalu in Borneo, a variation of less than 4 % among ? consistent with other measures: (1) dominant NH4 in species at the same site was observed (Kitayama and - inorganic N in soil (Fig. 1); (2) relatively low soil NO3 Iwamoto 2001). Thus, it appears that in large areas of - concentration and NO3 production, particular in the sec- subtropical and tropical eastern Asia, forests display a ondary forests (Fig. 1); (3) moderate atmospheric N small difference in foliar 15N abundance between different deposition, as mentioned above. These results indicate that species, irrespective of variable ecosystem N status. This the study forests may sequestrate more CO2 under finding is in accordance with observations in Hawaiian 15 ? increased N deposition. Collectively, low foliar d N in the forests, where plants rely on a common N source: NH4 or - present study may be caused by different mechanisms NO3 (Houlton et al. 2007). compared to the observation in Guangdong province, In the present study, we found that the fine roots had southern China, which was supposed to occur by increased slightly higher d15N than leaves in all four forests (by - soil NO3 uptake due to high N deposition in the urban and 0.5–1.8 %; Fig. 4). This result supports many previous suburban areas (Fang et al. 2011a). studies listed in Pardo et al. (2013), although a few studies report the opposite pattern (Emmett et al. 1998). However, Variation in foliar d15N among species the difference between fine root and leaf d15N was relatively small compared to some studies, in which it was Some researchers (e.g., Nadelhoffer et al. 1996) suggest reported to be up to 5 % (Pardo et al. 2013). The principal that plants in the ecosystems located at high latitudes have reason recognized by most researchers for the difference is a wider range of d15N in their tissues than in ecosystems fractionation during transfer within the plant leading to located at lower latitudes, either under temperate or tropi- assimilation of 15N-enriched N in roots and 15N-depleted N cal climates. The variation of foliar d15N, to some extent, in leaves (Pardo et al. 2013). Another possible explanation reflects source partitioning. This is consistent with the for differences in root and leaf d15N within an individual niche complementary hypothesis, which states that plant plant is different sources of N for the root compared to the species occupy functionally distinct fundamental niches leaf. If roots take up their N from soil, while leaves take up and use resources in a complementary way. Hence, varia- their N directly from atmospheric deposition, then these tion in resource acquisition by plants in space and time can tissues could have different d15N values without any frac- lead to resource partitioning, thereby allowing different tionation occurring within the plant (Pardo et al. 2013). species and growth forms to co-occur (e.g., McKane et al. 1990; Schulze et al. 1994). However, a recent study per- Difference in foliar d15N among forests formed in six Hawaiian tropical forests along a gradient of precipitation suggests that a group of tropical plant species Primary forests often have larger total N pools and N with diverse growth strategies (trees and ferns, canopy, and availability than secondary forests due to long-term N subcanopy) relied on a common pool of inorganic N, rather accumulation (Vitousek and Walker 1989). We found that than specializing on different N pools (Houlton et al. the primary forest in this study also had a large N pool in 2007). This reliance on a common pool was suggested to be the plant biomass (a large biomass and higher N concen- responsible for a narrow range of foliar d15N among tration in canopy tree leaves) than the secondary forests dominant species (1.6–3 %). (Table 1; Fig. 4). However, soil total N concentration In the present study, in total plant 39 species were col- showed a smaller difference between the primary forest lected from the four forests. These species covered and secondary forests, as compared to the plant N pool 123 Oecologia

(Table 1). In addition, soil-available N and N production mineralization–mycorrhizal–plant uptake pathway, rate (mineralization and nitrification) were not always whereas soil d15N increased, likely due to a redistribution higher in the primary forest than in the secondary forests of depleted N from the mineral soil to the developing O (Fig. 1). horizon (Compton et al. 2007). Our study is the sixth forest Although soils showed large variation in available N chronosequences, showing higher foliar d15N and enrich- concentration, mineralization and nitrification rates and ment factor in the primary forest than in the secondary 15 - 15 d N for soil NO3 among forests, the difference in d N forests, although the difference was small among forests, both for plants and for soils was relatively small among while there was no consistent pattern in soil d15N (Fig. 4). forests (Fig. 4). To normalize initial differences in soil and plant d15N values due to previous land management and Acknowledgments This study was jointly supported by the Special soil age, we calculated an enrichment factor, which, in this Research Program for Public-Welfare Forestry (no. 201104009-06), Grants-in-Aid for Scientific Research (21310008), the Program to case, was defined as the difference between leaf and soil Create an Independent Research Environment for Young Researchers d15N. The enrichment factor has been demonstrated to from the Ministry of Education, Culture, Sports, Science and Tech- positively correlate to nitrification (Garten 1993; Emmett nology, Japan, and the NEXT Program (GS008) from the Japan et al. 1998), litterfall N flux (Emmett et al. 1998) and Society for the Promotion of Science, the State Key Laboratory of Forest and Soil Ecology, Institute of Applied Ecology, CAS (no. ecosystem N status (Emmett et al. 1998). Enrichment Y1SRC111J6), the Special Research Program of the Research Insti- factor (mean foliar d15N minus mean d15N of the 0- to tute for Tropical Forestry, CAF (no. RITFYWZX2011-02), the 10-cm mineral soil layer in each forest; Fig. 4) was -3.9, National Natural Science Foundation of China (no. 41171040). The -4.4, -5.6 and -3.0 % in the three secondary forests and study was also supported by CFERN and GENE Award Funds for ecological papers. The authors wish to thank Mr. Xiaoming Fang, Mr. primary forest, respectively. Thus, with respect to foliar Jiong Li, Dr. Han Xu, Mr. Mingxian Lin, Mr. Zhang Zhou, Mr. Lai d15N and the enrichment factor, the primary forest was Jiang, Mr. Huai Yang and the staff of Jianfengling National Natural more N rich compared to the three secondary forests. Reserve for their kind help in sampling and laboratory analysis. Our results across forests seem to support the theory that as forests age they become N rich and the ecosystem should become enriched in 15N as N is lost (Vitousek and References Walker 1989; Martinelli et al. 1999). Actually, existing chronosequence studies suggest that long-term variation in Aber J, McDowell W, Nadelhoffer K, Magill A, Berntson G, 15 Kamakea M, McNulty S, Currie W, Rustad L, Fernandez I d N which differs among systems is likely driven by a (1998) Nitrogen saturation in temperate forest ecosystems. range of processes. To our knowledge, so far, there are five Bioscience 48:921–934 forest chronosequences in three different biomes designed Amundson R, Austin AT, Schuur EAG, Yoo K, Matzek V, Kendall C, to investigate the patterns of d15N in soils and plants: Uebersax A, Brenner D, Baisden WT (2003) Global patterns of the isotopic composition of soil and plant nitrogen. Glob boreal forests in Alaska USA aged 55 to 225 years (Hobbie Biogeochem Cycles 17:1031 et al. 1999), boreal forests in Sweden aged hundreds to Ariz I, Cruz C, Moran JF, Gonza´lez-Moro MB, Garcıá-Olaverri C, thousands of years (Hyodo and Wardle 2009), temperate Gonza´lez-Murua C, Martins-Louçaõ MA, Aparicio-Tejo PM (2011) Depletion of the heaviest stable N isotope is associated forests in the New England region, USA, aged ten to ? ? with NH4 /NH3 toxicity in NH4 -fed plants. BMC Plant Biol 114 years (Compton et al. 2007), tropical rainforests in 11:83 Hawaii aged 28–67,000 years (Vitousek and Walker 1989) Chen D, Li Y, Liu H, Xu H, Xiao W, Luo T, Zhou Z, Lin M (2010) and 300–4,100,000 years (Martinelli et al. 1999). Along Biomass and carbon dynamics of a tropical mountain rain forest both Hawaiian chronosequences, both foliar and soil d15N in China. Sci China Life Sci 53:798–810 Cheng SL, Fang HJ, Yu GR, Zhu TH, Zheng JJ (2010) Foliar and soil increased with soil age (Vitousek and Walker 1989; Mar- 15N natural abundances provide field evidence on nitrogen tinelli et al. 1999), likely as the result of an increase in N dynamics in temperate and boreal forest ecosystems. Plant Soil losses. In Sweden, foliar and litter d15N increased with 337:285–297 forest age, probably due to increased N fixation by free- Compton JE, Hooker TD, Perakis SS (2007) Ecosystem N distribution and d15N during a century of forest regrowth after agricultural living N fixers, increased reliance on DON, or both (Hyodo abandonment. Ecosystems 10:1197–1208 and Wardle 2009). In contrast, foliar d15N decreased Craine JM, Elmore AJ, Aidar MPM, Bustamante M, Dawson TE, Hobbie through early succession in Alaska (Hobbie et al. 1999), EA, Kahmen A, Mack MC, McLauchlan KK, Michelsen A (2009) likely because of a shift in the importance of fractionation Global patterns of foliar nitrogen isotopes and their relationships 15 with climate, mycorrhizal fungi, foliar nutrient concentrations, and during mycorrhizal transfer of N to plants, and soil d N nitrogen availability. New Phytol 183:980–992 showed no consistent pattern (Hobbie et al. 1999). In the Emmett BA, Kjønaas OJ, Gundersen P, Koopmans C, Tietema A, 15 temperate forest chronosequence, foliar and Oi horizon Sleep D (1998) Natural abundance of N in forests across a d15N also decreased as stands aged, as a result of a pro- nitrogen deposition gradient. For Ecol Manage 101:9–18 Fang J, Li Y, Zhu B, Liu G, Zhou G (2004) Community structures and gressive tightening of the N cycle during forest regrowth species richness is the montane rain forest of Jianfengling, on agricultural lands and increased fractionation along the Hainan Island, China. Biodivers Sci 12:29 123 Oecologia

Fang Y, Yoh M, Koba K, Zhu W, Takebayashi YU, Xiao Y, Lei C, Martinelli LA, Piccolo MC, Townsend AR, Vitousek PM, Cuevas E, Mo J, Zhang W, Lu X (2011a) Nitrogen deposition and forest McDowell W, Robertson GP, Santos OC, Treseder K (1999) nitrogen cycling along an urban–rural transect in southern China. Nitrogen stable isotopic composition of leaves and soil: tropical Glob Change Biol 17:872–885 versus temperate forests. Biogeochemistry 46:45–65 Fang YT, Koba K, Wang XM, Wen DZ, Li J, Takebayashi Y, Liu McIlvin MR, Casciotti KL (2011) Technical updates to the bacterial XY, Yoh M (2011b) Anthropogenic imprints on nitrogen and method for nitrate isotopic analyses. Anal Chem 83(5):1850–1856 oxygen isotopic composition of precipitation nitrate in a McKane RB, Grigal DF, Russelle MP (1990) Spatiotemporal nitrogen-polluted city in southern China. Atmos Chem Phys differences in 15N uptake and the organization of an old-field 11(3):1313–1325 plant community. Ecology 71:1126–1132 Fang Y, Koba K, Yoh M, Makabe A, Liu X (2012a) Patterns of foliar Michelsen A, Quarmby C, Sleep D, Jonasson S (1998) Vascular plant d15N and their control in Eastern Asian forests. Ecol Res. doi:10. 15N natural abundance in heath and forest tundra ecosystems is 1007/s11284-012-0934-8 closely correlated with presence and type of mycorrhizal fungi in Fang Y, Koba K, Makabe A, Zhu F, Fan S, Liu X, Yoh M (2012b) roots. Oecologia 115:406–418 Low d18O values of nitrate produced from nitrification in Nadelhoffer K, Shaver G, Fry B, Giblin A, Johnson L, McKane R temperate forest soils. Environ Sci Technol 46:8723–8730 (1996) 15N natural abundances and N use by tundra plants. Garten Jr CT (1993) Variation in foliar 15N abundance and the Oecologia 107:386–394 availability of soil nitrogen on Walker Branch watershed. Pardo LH, Nadelhoffer KJ (2009) Using nitrogen isotope ratios to Ecology 74:2098–2113 assess terrestrial ecosystems at regional and global scales. In: Garten CT Jr, Miegroet HV (1994) Relationships between soil West JB, Bowen G, Dawson T, Tu K (eds) Isoscapes: nitrogen dynamics and natural 15N abundance in plant foliage understanding movement, pattern, and process on earth through from Great Smoky Mountains National Park. Can J For Res isotope mapping. Springer, New York 24:1636–1645 Pardo LH, Hemond H, Montoya J, Fahey T, Siccama T (2002) Gebauer G, Dietrich P (1993) Nitrogen isotope ratios in different Response of the natural abundance of 15N in forest soils and compartments of a mixed stand of spruce, larch and beech trees foliage to high nitrate loss following clear-cutting. Can J For Res and of understorey vegetation including fungi. Isot Environ 32:1126–1136 Health Stud 29:35–44 Pardo LH, Templer PH, Goodale CL, Duke S, Groffman PM, Adams Hobbie EA, Macko SA, Shugart HH (1999) Interpretation of nitrogen MB, Boeckx P, Boggs J, Campbell J, Colman B (2006) Regional isotope signatures using the NIFTE model. Oecologia 120:405–415 assessment of N saturation using foliar and root d15N. Biogeo- Ho¨gberg P (1997) Tansley Review No. 95 15N natural abundance in chemistry 80:143–171 soil-plant systems. New Phytol 137:179–203 Pardo LH, Hemond H, Montoya J, Pett-Ridge J (2007) Natural Hogberg P, Johannisson C (1993) 15N abundance of forests is abundance 15N in soil and litter across a nitrate-output gradient correlated with losses of nitrogen. Plant Soil 157:147–150 in New Hampshire. For Ecol Manage 251:217–230 Houlton BZ, Sigman DM, Hedin LO (2006) Isotopic evidence for Pardo LH, Semaoune P, Schaberg P, Eagar C, Sebilo M (2013) large gaseous nitrogen losses from tropical rainforests. Proc Natl Patterns in d15N in roots, stems, and leaves of sugar maple and Acad Sci 103:8745–8750 American beech seedlings, saplings, and mature trees. Biogeo- Houlton BZ, Sigman DM, Schuur EA, Hedin LO (2007) A climate- chemistry 112:275–291 driven switch in plant nitrogen acquisition within tropical forest Pate J, Stewart G, Unkovich M (1993) 15N natural abundance of plant communities. Proc Natl Acad Sci 104:8902–8906 and soil components of a Banksia woodland ecosystem in Hyodo F, Wardle DA (2009) Effect of ecosystem retrogression on relation to nitrate utilization, life form, mycorrhizal status and stable nitrogen and carbon isotopes of plants, soils and consumer N2-fixing abilities of component species. Plant Cell Environ organisms in boreal forest islands. Rapid Commun Mass 16:365–373 Spectrom 23:1892–1898 Robinson D (2001) d15N as an integrator of the nitrogen cycle. Trends Kang HZ, Liu CJ, Yu WJ, Wu LL, Chen DM, Sun X, Ma XP, Hu HB, Ecol Evol 16:153–162 Zhu XL (2011) Variation in foliar d15N among oriental oak Scarano FR, Duarte HM, Ribeiro KT, Rodrigues PJFP, Barcellos (Quercus variabilis) stands over eastern China: patterns and EMB, Franco AC, Brulfert J, Deleéns E, Lu¨ttge U (2001) Four interactions. J Geochem Explor 110:8–14 sites with contrasting environmental stress in southeastern Brazil: Kitayama K, Iwamoto K (2001) Patterns of natural 15N abundance in relations of species, life form diversity, and geographic distribu- the leaf-to-soil continuum of tropical rain forests differing in N tion to ecophysiological parameters. Bot J Linn Soc 136:345–364 availability on Mount Kinabalu, Borneo. Plant Soil 229:203–212 Schulze ED, Chapin F III, Gebauer G (1994) Nitrogen nutrition and Koba K, Isobe K, Takebayashi Y, Fang YT, Sasaki Y, Saito W, Yoh M, isotope differences among life forms at the northern treeline of Mo JM, Liu L, Lu X, Zhang T, Zhang W, Senoo K (2010) d15Nof Alaska. Oecologia 100:406–412 soil N and plants in a N-saturated, subtropical forest of southern Sigman DM, Casciotti KL, Andreani M, Barford C, Galanter M, China. Rapid Commun Mass Spectrom 24(17):2499–2506 Bohlke JK (2001) A bacterial method for the nitrogen isotopic Koba K, Fang Y, Mo J, Zhang W, Lu X, Liu L, Zhang T, Takebayashi analysis of nitrate in seawater and freshwater. Anal Chem Y, Toyoda S, Yoshida N (2012) The 15N natural abundance of 73:4145–4153 the N lost from an N-saturated subtropical forest in southern Templer PH, Arthur MA, Lovett GM, Weathers KC (2007) Plant and China. J Geophys Res Biogeosci. doi:10.1029/2010JG001615 soil natural abundance d15N: indicators of relative rates of Liu G, Jiang N, Zhang L, Liu Z (1996) Soil physical and chemical nitrogen cycling in temperate forest ecosystems. Oecologia analysis and description of soil profiles, vol 24. China Standard 153:399–406 Methods Press, Beijing Viani RA, Rodrigues RR, Dawson TE, Oliveira RS (2011) Functional Liu CP, Yeh HW, Sheu BH (2006) N isotopes and N cycle in a differences between woodland savannas and seasonally dry 35-year-old plantation of the Guandaushi subtropical forest forests from south-eastern Brazil: evidence from 15N natural ecosystem, central Taiwan. For Ecol Manage 235:84–87 abundance studies. Austral Ecol 36:974–982 Mardegan SF, Nardoto GB, Higuchi N, Moreira MZ, Martinelli LA Vitousek PM, Walker LR (1989) Biological invasion by Myrica faya (2009) Nitrogen availability patterns in white-sand vegetations in Hawai’i: plant demography, nitrogen fixation, ecosystem of Central Brazilian Amazon. Trees 23:479–488 effects. Ecol Monogr 59:247–265

123 Oecologia

Watmough SA (2010) An assessment of the relationship between Zeng Q, Li Y, Chen B (1997) Management of tropical potential chemical indices of nitrogen saturation and nitrogen forest ecosystems (in Chinese), vol 202. China Forestry Press, deposition in hardwood forests in southern Ontario. Environ Beijing Monit Assess 164:9–20 Xiao HW, Xiao HY, Long AM, Wang YL (2012) Who controls the ? monthly variations of NH4 nitrogen isotope composition in precipitation? Atmos Environ 54:201–206

123