Human and climate impact on ¹⁵N natural abundance of plants and soils in high-mountain ecosystems: a short review and two examples from the Eastern Pamirs and Mt. Kilimanjaro

Population pressure increasingly endangers high-mountain ecosystems such as the pastures in the Eastern Pamirs and the mountain forests on Mt. Kilimanjaro. At the same time, these ecosystems constitute the economic basis for millions of people living there. In our study, we, therefore, aimed at characterising the land-use effects on soil degradation and N-cycling by determining the natural abundance of (15)N. A short review displays that δ(15)N of plant-soil systems may often serve as an integrated indicator of N-cycles with more positive δ(15)N values pointing towards N-losses. Results for the high-mountain pastures in the Eastern Pamirs show that intensively grazed pastures are significantly enriched in (15)N compared to the less-exploited pastures by 3.5 ‰, on average. This can be attributed to soil organic matter degradation, volatile nitrogen losses, nitrogen leaching and a general opening of the N-cycle. Similarly, the intensively degraded savanna soils, the cultivated soils and the soils under disturbed forests on the foothill of Mt. Kilimanjaro reveal very positive δ(15)N values around 6.5 ‰. In contrast, the undisturbed forest soils in the montane zone are more depleted in (15)N, indicating that here the N-cycle is relatively closed. However, significantly higher δ(15)N values characterise the upper montane forest zone at the transition to the subalpine zone. We suggest that this reflects N-losses by the recently monitored and climate change and antropogenically induced increasing fire frequency pushing the upper montane rainforest boundary rapidly downhill. Overall, we conclude that the analysis of the (15)N natural abundance in high-mountain ecosystems is a purposeful tool for detecting land-use- or climate change-induced soil degradation and N-cycle opening.     

13C and 15N Abundances in the Sediment Core (VER 92/1-St-10-GC2) from Northern Lake Baikal

Abstract

Carbon and nitrogen stable isotope compositions of organic matter, TOC/TN ratio, and manganese concentration in a sediment core that was collected in northern part of Lake Baikal (VER92ST10-GC2, water depth at 922 m, about 3 m long) were investigated to elucidate the origin of the sedimentary organic matter and its associated environmental factors.

The sediment core was composed of mainly two parts: turbidite sections and other sections. Constant δ¹³C and δ¹⁵N values of the turbidite sections were observed (- 26.8 ±0.2 ‰ for δ¹³C and 3.2 ± 0.1 ‰ for δ¹⁵N) throughout the core. The higher δ¹³C in turbidite sections (about - 27 ‰) than that of the other sections (- 31 to - 29 ‰) was clearly observed, and δ¹⁵N was different between turbidite sections (about 3‰) and other sections (3 to 5 ‰). δ¹³C of other sections was close to that of pelagic phytoplankton, indicating that sediment other than turbidite sections is composed of autochthonous components. The variation of stable isotopes in other sections may be possibly caused by the changes in either phytoplankton growth rate or contribution ratios of terrestrial to aquatic plants for δ¹³C. Either denitrification or fluctuation of δ¹⁵N in pelagic phytoplankton can be the cause of variable δ¹⁵N in other sections.     

Evaluation of 15N Methods for the measurement of Denitrification in Soils

The precision of the ¹⁵N-emission and that of the ¹⁵N-balance methods was evaluated and both methods were compared in a denitrification experiment. ¹⁵N-analysis was performed with an isotope ratio mass spectrometer which was coupled to an elemental analyzer. The measuring sensitivity in soil and gas analysis was tested by analyzing ¹⁵N-standards. The detection limit for gas samples with two different procedures of ¹⁵N-gas analysis was δ¹⁵N = (4.5 ± 1.0)‰ and (0.5 ± 0.05)‰, respectively. The error in measurement was 19% and 12% respectively. ¹⁵N-analysis of a ¹⁵N-labelled soil (4.15 ppm ¹⁵N) resulted in a CV of 1.32%. The measurements had to be calibrated with soil standards because the ¹⁵N-values showed a continuous downward fluctuation in a range of 10–20% within several days, when only acetanilid was used for calibration. Mean ¹⁵N-losses which were determined with both methods during the denitrification experiment were in good agreement. The precision of the ¹⁵N-emission method was adequate in all variants of the experiment. The precision of the ¹⁵N-balance method however was unsatisfactory at low ¹⁵N-losses (0.2–2% of added ¹⁵N).     

Seasonal variation in natural abundance of δ13C and 15N in Salicornia brachiata Roxb. populations from a coastal area of India

High and fluctuating salinity is characteristic for coastal salt marshes, which strongly affect the physiology of halophytes consequently resulting in changes in stable isotope distribution. The natural abundance of stable isotopes (δ¹³C and δ¹⁵N) of the halophyte plant Salicornia brachiata and physico-chemical characteristics of soils were analysed in order to investigate the relationship of stable isotope distribution in different populations in a growing period in the coastal area of Gujarat, India. Aboveground and belowground biomass of S. brachiata was collected from six different populations at five times (September 2014, November 2014, January 2015, March 2015 and May 2015). The δ¹³C values in aboveground (?30.8 to ?23.6?‰, average: ?26.6?±?0.4?‰) and belowground biomass (?30.0 to ?23.1?‰, average: ?26.3?±?0.4?‰) were similar. The δ¹³C values were positively correlated with soil salinity and Na concentration, and negatively correlated with soil mineral nitrogen. The δ¹⁵N values of aboveground (6.7–16.1?‰, average: 9.6?±?0.4?‰) were comparatively higher than belowground biomass (5.4–13.2?‰, average: 7.8?±?0.3?‰). The δ¹⁵N values were negatively correlated with soil available P. We conclude that the variation in δ¹³C values of S. brachiata was possibly caused by soil salinity (associated Na content) and N limitation which demonstrates the potential of δ¹³C as an indicator of stress in plants.     

Nitrogen Isotope Ratios in Different Compartments of a Mixed Stand of Spruce,Larch and Beech Trees and of Understorey Vegetation Including Fungi

Abstract

Natural nitrogen isotope ratios were measured in different compartments (needles or leaves and twigs of different age classes and crown positions, roots and soil of different horizons) of spruce (Picea abies), larch (Larix decidua) and beech (Fagus sylvatica) trees in an 11-year-old mixed stand in the Fichtelgebirge, NE Bavaria, Germany. In addition, samples of understorey vegetation (mainly ericaceous shrubs and grass) and of ectomycorrhizal and saprophytic fungi were analyzed. The δ¹⁵N values found for all samples ranged between ?7.5 and + 4.5‰. No significant differences were found for the nitrogen isotope ratios of the three tree species despite of their evergreen versus deciduous foliage and despite of their different rooting depth. Ericaceous shrubs had the most negative and fungi and soil from the mineral horizon the most positive δ¹⁵N values. Positive δ¹⁵N values of the fungi indicate their ability to utilize organic soil nitrogen, but the data do not unequivocally show that plants forming mycorrhizas profit from this organic nitrogen source.     

Spatial variations in larch needle and soil δ15N at a forest–grassland boundary in northern Mongolia

The spatial patterns of plant and soil δ¹⁵N and associated processes in the N cycle were investigated at a forest–grassland boundary in northern Mongolia. Needles of Larix sibirica Ledeb. and soils collected from two study areas were analysed to calculate the differences in δ¹⁵N between needle and soil (Δδ¹⁵N). Δδ¹⁵N showed a clear variation, ranging from ?8?‰ in the forest to ?2?‰ in the grassland boundary, and corresponded to the accumulation of organic layer. In the forest, the separation of available N produced in the soil with ¹⁵N-depleted N uptake by larch and ¹⁵N-enriched N immobilization by microorganisms was proposed to cause large Δδ¹⁵N, whereas in the grassland boundary, small Δδ¹⁵N was explained by the transport of the most available N into larch. The divergence of available N between larch and microorganisms in the soil, and the accumulation of diverged N in the organic layer control the variation in Δδ¹⁵N.     

Natural abundances of 15N and 13C in leaves of some N2-fixing and non-N2-fixing trees and shrubs in Syria

A survey study was conducted on man-made plantations located at two different areas in the arid region of Syria to determine the variations in natural abundances of the ¹⁵N and ¹³C isotopes in leaves of several woody legume and non-legume species, and to better understand the consequence of such variations on nitrogen fixation and carbon assimilation. In the first study area (non-saline soil), the δ¹⁵N values in four legume species (Acacia cyanophylla,?1.73 ‰ Acacia farnesiana,?0.55 ‰ Prosopis juliflora,?1.64 ‰; and Medicago arborea,+1.6 \textperthousand) and one actinorhizal plant (Elaeagnus angustifolia,?0.46 to?2.1 ‰) were found to be close to that of the atmospheric value pointing to a major contribution of N₂ fixing in these species; whereas, δ¹⁵N values of the non-fixing plant species were highly positive. δ¹³C ‰; in leaves of the C3 plants were found to be affected by plant species, ranging from a minimum of?28.67 ‰; to a maximum of?23 ‰. However, they were relatively similar within each plant species although they were grown at different sites. In the second study area (salt affected soil), a higher carbon discrimination value (Δ¹³C ‰) was exhibited by P. juliflora, indicating that the latter is a salt tolerant species; however, its δ¹⁵N was highly positive (+7.03 ‰) suggesting a negligible contribution of the fixed N₂. Hence, it was concluded that the enhancement of N₂ fixation might be achieved by selection of salt-tolerant Rhizobium strains.     

Characteristics of suspended matter in the River Sava watershed,Slovenia

A combination of C/N ratios, δ¹³C and δ¹⁵N values in suspended matter was used to examine the seasonal (late summer 2004 and spring 2005) relationship with hydrological characteristics of the River Sava watershed in Slovenia. The values of C/N ratios range from 1.2 to 19.1, δ¹³C values range from?29.2 to?23.0 ‰ and δ¹⁵N values from 0.5 to 16.7 ‰ and indicate that the samples are a mixture of two end members: modern soils and plant litter. A simple mixing model was used to indicate that soil organic carbon prevails over plant litter and contributes more than 50% of suspended material. The calculated annual particulate organic carbon flux is estimated as 5.2×10¹⁰ g C/year, the annual particulate nitrogen flux 8.5×10⁹ g N/year and the total suspended solid flux is estimated to be 1.3×10¹² g/year. Anthropogenic impact was detected only in a tributary stream of the River Sava, which is located in an agriculture–industrial area and is reflected in higher δ¹⁵N values in suspended matter and high nitrate concentrations in the late summer season.     

Food webs in Mongolian grasslands: The analysis of 13C and 15N natural abundances

Overgrazing often lowers species richness and productivity of grassland communities. For Mongolian grassland ecosystems, a lack of detailed information about food-web structures makes it difficult to predict the effects of overgrazing on species diversity and community composition. We analysed the δ¹³C and δ¹⁵N signatures of herbaceous plants, arthropods (grouped by feeding habit), wild and domestic mammals, and humans in central Mongolia to understand the predominant food-web pathways in this grassland ecosystem. The δ¹³C and δ¹⁵N values of mammals showed little variation within species, but varied considerably with slope position for arthropods. The apparent isotopic discrimination between body tissue and hair of mammals was estimated as 2.0 ‰ for δ¹³C and 2.1 ‰ for δ¹⁵N, which was large enough to cause overestimation of the trophic level of mammals if not taken into account when using hair samples to measure isotopic enrichment.     

Amino Acid 15N/14N Analysis at Natural Abundances: A New Tool for Soil Organic Matter Studies in Agricultural Systems

Abstract

The effects of landuse, fertilizer history and soil type on the quantity and isotopic quality of hydrolysable soil amino acids were examined in 3 grassland and 2 arable soils. Results showed, (i) that overall concentrations of individual amino acids were highest in the grassland soils, (ii) that ‰δ¹⁵N values of the individual amino acids differed considerably between the five soils, and (iii) that the combination of amino acid ‰δ¹⁵N values and concentrations could be used to distinguish between landuse, crop type and fertilizer history. This preliminary study indicates that the pathways of transformation of soil amino acid N are influenced by long term N inputs and that associated biological processes are reflected in differences in concentrations and ‰δ¹⁵N values of individual soil amino acids.     

Using natural 15N abundances to trace the fate of waste-derived nitrogen in forest ecosystems: New Zealand case studies

Treatment of wastewater generally results in elevated natural ¹⁵N abundance (δ¹⁵N) in the effluent and sludges. For example, high δ¹⁵N values are found in treated sewage effluent, biosolids, and other wastes that are commonly applied to land. In contrast, N deficient coniferous forest soils usually have a low δ¹⁵N. When wastes with high δ¹⁵N values are applied to land, their distinctive δ¹⁵N signature can potentially be used to trace the fate of waste-derived N in the ecosystem. In this paper, we provide an overview of the use of δ¹⁵N in land application of wastes, including New Zealand case studies on tracing nitrogen in forest ecosystems.     

Using the natural 15N abundance to assess the main nitrogen inputs into the sand dune area of the North-Western Negev desert (ISRAEL)

The variation of the natural ¹⁵N abundance is often used to evaluate the origin of nitrogen or the pathways of N input into ecosystems. We tried to use this approach to assess the main input pathways of nitrogen into the sand dune area of the north-western Negev Desert (Israel). The following two pathways are the main sources for nitrogen input into the system:

1. Biological fixation of atmospheric nitrogen by cyanobacteria present in biological crusts and by N₂-fixing vascular plants (e.g. the shrub Retama raetam);

2. Atmospheric input of nitrogen by wet deposition with rainfall, dry deposition of dust containing N compounds, and gaseous deposition.

Samples were taken from selected environmental compartments such as biological crusts, sand underneath these crusts (down to a depth of 90?cm), N₂-fixing and non-N₂-fixing plants, atmospheric bulk deposition as well as soil from arable land north of the sandy area in three field campaigns in March 1998, 1999 and 2000. The δ¹⁵N values measured were in the following ranges: grass ?2.5‰ to +1.5‰; R. reatam: +0.5‰ to +4.5‰; non-N₂-fixing shrubs +1‰ to +7‰; sand beneath the biological crusts +4‰ to +20‰ (soil depth 2–90?cm); and arable land to the north up to 10‰. Thus, the natural ¹⁵N abundance of the different N pools varies significantly. Accordingly, it should be feasible to assess different input pathways from the various ¹⁵N abundances of nitrogen. For example, the biological N fixation rates of the Fabaceae shrub R. reatam from the ¹⁵N abundances measured were calculated to be 46–86% of biomass N derived from the atmosphere. The biological crusts themselves generally show slight negative ¹⁵N values (?3‰ to ?0.5‰), which can be explained by biological N fixation. However, areas with a high share of lichens, which are unable to fix atmospheric nitrogen, show very negative values down to ?10‰. The atmospheric N bulk deposition, which amounts to 1.9–3.8?kg?N/ha?yr, has a ¹⁵N abundance between 4.4‰ and 11.6‰ and is likely to be caused by dust from the arable land to the north. Thus, it cannot be responsible for the very negative values of lichens measured either. There must be an additional N input from the atmosphere with negative δ¹⁵N values, e.g. gaseous N forms (NO _x , NH₃). To explain these conflicting findings, detailed information is still needed on the wet, particulate and gaseous atmospheric deposition of nitrogen.     

Comparison of pore water samplers and cryogenic distillation under laboratory and field conditions for soil water stable isotope analysis*

We used pore water samplers (PWS) to sample for isotope analysis (1) only water, (2) soil under laboratory conditions, and (3) soil in the field comparing the results with cryogenic extraction (CE). In (1) and (2), no significant differences between source and water extracted with PWS were detected with a mean absolute difference (MAD) always lower than 2?‰ for δ²H and 1?‰ for δ¹⁸O. In (2), CE water was more enriched than PWS-extracted water, with a MAD respect to source water of roughly 8?‰ for δ²H and 4?‰ for δ¹⁸O. In (3), PWS water was enriched relative to CE water by 3?‰ for δ²H and 0.9?‰ for δ¹⁸O. The latter result may be due to the distinct water portions sampled by the two methods. Large pores, easily sampled by PWS, likely retain recent, and enriched, summer precipitation while small pores, only sampled by CE, possibly retain isotopically depleted water from previous winter precipitation or irrigation inputs. Accuracy and precision were greater for PWS relative to CE. PWS is therefore suggested as viable tool to extract soil water for stable isotope analysis, particularly for soils used in this study (sandy and silty loams).     

Soil δ15N patterns in old-growth forests of southern Chile as integrator for N-cycling

Old-growth forests of southern Chile represent an important reserve of temperate (rain) forests in the world. Wetter and colder forest ecosystems appear to be more efficient in conserving and recycling N such that mostly non-plant available N species are lost, which could be indicated by more depleted δ¹⁵N values of the soil and plants. Hydrological N loss from the old-growth forests in southern Chile occurs mainly via dissolved organic nitrogen and not via dissolved inorganic N. Forest disturbances (e.g. fire, clear-cutting or enhanced N deposition) cause (abrupt) changes in ecosystem N-cycling processes. In this study, we hypothesized that δ¹⁵N signatures of soil profiles under old-growth forests could be used as an integrator for ecosystem N-cycling, and changes of these δ¹⁵N profiles could be valuable to assess ecosystem resilience towards disturbances. Six old-growth forests were selected in the phytogeographical region of the Valdivian rain forest in southern Chile. One of the sites has been partly burned in February 2002. First, we observed that ecosystems with higher mean annual precipitation and lower mean annual temperature were relatively more depleted in ¹⁵N. Secondly, we found that a forest fire caused a 100-fold increase of the nitrate export and induced an enrichment of the soil δ¹⁵N signal in the upper 20 cm.     

15N natural abundance of non-fixing woody species in the Brazilian dry forest (caatinga)

Foliar δ¹⁵N values are useful to calculate N₂ fixation and N losses from ecosystems. However, a definite pattern among vegetation types is not recognised and few data are available for semi-arid areas. We sampled four sites in the Brazilian caatinga, along a water availability gradient. Sites with lower annual rainfall (700 mm) but more uniform distribution (six months) had δ¹⁵N values of 9.4 and 10.1 ‰, among the highest already reported, and significantly greater than those (6.5 and 6.3 ‰) of sites with higher rainfall (800 mm) but less uniform distribution (three months). There were no significant differences at each site among species or between non-fixing legume and non-legume species, in spite of the higher N content of the first group. Therefore, they constitute ideal reference plants in estimations of legume N₂ fixation. The higher values could result from higher losses of ¹⁵N depleted gases or lower losses of enriched ¹⁵N material.     

Isotopic Assessment of Sources and Processes Affecting Sulfate and Nitrate in Surface Water and Groundwater of Luxembourg

Abstract

Surface water and deep and shallow groundwater samples were taken from selected parts of the Grand-Duchy of Luxembourg to determine the isotopic composition of nitrate and sulfate, in order to identify sources and/or processes affecting these solutes. Deep groundwater had sulfate concentrations between 20 and 40mg/L, δ³⁴S_sulfate values between ?3.0 and ?20.0‰, and δ¹⁸O_sulfate values between +1.5 and +5.0‰ nitrate was characterized by concentrations varying between <0.5 and 10mg/L, δ¹⁵N_nitrate values of ～?0.5‰, and δ¹⁸O_nitrate values ～+3.0‰. In the shallow groundwater, sulfate concentrations ranged from 25 to 30mg/L, δ³⁴S_sulfate values from ?20.0 to +4.5‰, and δ¹⁸O_sulfate values from ～+0.5 to +4.5‰ nitrate concentrations varied between ～10 and 75mg/L, δ¹⁵N_nitrate values between +2.5 and +10.0‰, and δ¹⁸O_nitrate values between +1.0 and +3.0‰. In surface water, sulfate concentrations ranged from 10 to 210mg/L, δ³⁴S_sulfate values varied between ?9.3 and +10.9‰, and δ¹⁸O_sulfate values between +3.0 and +10.7‰ were observed. Nitrate concentrations ranged from 10 to 40mg/L, δ¹⁵N_nitrate values from +6.5 to +12.0‰, and δ¹⁸O_nitrate values from ?0.4 to +4.0‰. Based on these data, three sulfate sources were identified controlling the riverine sulfate load. These are soil sulfate, dissolution of evaporites, and oxidation of reduced S minerals in the bedrock. Both groundwater types were predominantly influenced by sulfate from the two latter lithogenic S sources. The deep groundwater and a couple shallow groundwater samples had nitrate derived mainly from soil nitrification. All other sampling sites were influenced by nitrate originating from sewage and/or manure. A decrease in nitrate concentration observed along one of the rivers was attributed to denitrification. It appears that sulfate within Luxembourg's aquatic ecosystem is mainly of lithogenic origin, whereas nitrate is often derived from anthropogenic activities.     

Unusual patterns in 15N blood values after a diet switch in red knot shorebirds

When a diet switch results in a change in dietary isotopic values, isotope ratios of the consumer's tissues will change until a new equilibrium is reached. This change is generally best described by an exponential decay curve. Indeed, after a diet switch in captive red knot shorebirds (Calidris canutus islandica), the depletion of ¹³C in both blood cells and plasma followed an exponential decay curve. Surprisingly, the diet switch with a dietary ¹⁵N/¹⁴N ratio (δ¹⁵N) change from 11.4 to 8.8 ‰ had little effect on δ¹⁵N in the same tissues. The diet-plasma and diet-cellular discrimination factors of ¹⁵N with the initial diet were very low (0.5 and 0.2 ‰, respectively). δ¹⁵N in blood cells and plasma decreased linearly with increasing body mass, explaining about 40 % of the variation in δ¹⁵N. δ¹⁵N in plasma also decreased with increasing body-mass change (r ²=.07). This suggests that the unusual variation in δ¹⁵N with time after the diet switch was due to interferences with simultaneous changes in body-protein turnover.     

Spatial and temporal variability of stable isotopes and biological parameters for the Danube River in Serbia

This paper presents the results of hydrological, physicochemical, biological, and isotopic investigations of the Danube River along the stretch through Serbian territory conducted during four campaigns in September and November 2007, September 2008 and April 2009. The stable isotope values exhibited significant changes both in the Danube (?10.7 to?9.5‰ for δ¹⁸O and?73.7 to?67.1 ‰ for δ²H) and in its tributaries (?9.1 to?8.5‰ for δ¹⁸O and?69.4 to?59.4‰ for δ²H) depending on the time of survey, which could be partly attributed to the influences of seasonal effects. Results emphasise the dominant role of tributaries inflows from aquifers along the Danube. The very narrow range of δ¹³C_POC (from?28.9 to?27.4 ‰) was associated with relatively high C/N ratios (C/N>9), and together with δ¹⁵N_TPN values, the date suggested that, in early spring, a major fraction of particulate organic matter was derived from allochthonous matter. An orthogonal varimax rotation of the principal components analysis identified four latent factors (‘mineral related’, ‘biological’, ‘hardness’, and ‘soil inlets’) which are responsible for the data structure covering 79% of the observed variations among the variables studied. A reliable grouping of samples with respect to the season was found.     

Pleistocene age paleo-groundwater inferred from water-stable isotope values in the central part of the Baltic Artesian Basin

A new data set of δ²H and δ¹⁸O in the groundwater from the central part of the Baltic Artesian Basin is presented. The hydrogeological section is subdivided into stagnation, slow exchange and active exchange zones. Na–Ca–Cl brine found at the deepest part – the stagnation zone – is characterized by δ¹⁸O values above ?5?‰ and δ²H values approaching ?40?‰ with respect to Vienna Standard Mean Ocean Water. The slow exchange zone where waters of mostly intermediate salinity reside is characterized by δ¹⁸O values around ?11.7?‰ and δ²H values around ?85.3?‰. Mean δ¹⁸O and δ²H values of the fresh groundwater in the active water exchange zone are ?11.1 and ?79.9?‰, respectively. Characteristically, the groundwater in the active and slow exchange zone is isotopically more depleted compared with the precipitation values observed, and the depletion increases with depth down to the level where strongly enriched brines are encountered.     

Latitudinal distribution of nitrogen isotopic composition in suspended particulate organic matter in tropical/subtropical seas

Natural nitrogen isotopic composition (δ¹⁵N) of suspended particulate organic matter (POM) and nitrogen fixation rates via ¹⁵N₂ assay were measured in surface waters along 120° E from 30° N to 30° S in the Asian marginal seas (the East/South China Seas and the Sulu/Celebes/Java Seas) and the northeastern Indian Ocean in November–December 2005 and March 2006. The POM δ¹⁵N values ranged from?1.8 to 12.2‰ with an average of 3.6‰ and showed a decreasing trend towards the equator in both hemispheres. In parallel, the measured N₂ fixation rates showed an increase from the subtropical to the tropical seas. This implies that a higher contribution of ¹⁵N-depleted POM was derived from enhanced N₂ fixation. Water temperature and the stability of water column were partly responsible for the observed variations in nitrogen fixation. The large-scale spatial variations in suspended POM δ¹⁵N and N₂ fixation rates suggest that the suspended POM δ¹⁵N may be a potential indicator of nitrogen fixation in surface waters in tropical/subtropical seas.