SEARCH

SEARCH BY CITATION

Keywords:

  • nitrogen isotopes;
  • soil nitrogen;
  • plant nitrogen

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Background of N Isotope Research
  5. 3. Methods
  6. 4. Results and Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[1] We compiled new and published data on the natural abundance N isotope composition (δ15N values) of soil and plant organic matter from around the world. Across a broad range of climate and ecosystem types, we found that soil and plant δ15N values systematically decreased with increasing mean annual precipitation (MAP) and decreasing mean annual temperature (MAT). Because most undisturbed soils are near N steady state, the observations suggest that an increasing fraction of ecosystem N losses are 15N-depleted forms (NO3, N2O, etc.) with decreasing MAP and increasing MAT. Wetter and colder ecosystems appear to be more efficient in conserving and recycling mineral N. Globally, plant δ15N values are more negative than soils, but the difference (δ15Nplant15Nsoil) increases with decreasing MAT (and secondarily increasing MAP), suggesting a systematic change in the source of plant-available N (organic/NH4+ versus NO3) with climate. Nitrogen isotopes reflect time integrated measures of the controls on N storage that are critical for predictions of how these ecosystems will respond to human-mediated disturbances of the global N cycle.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Background of N Isotope Research
  5. 3. Methods
  6. 4. Results and Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[2] The growing rate of anthropogenically-derived N deposition worldwide [Galloway, 1998] has focused attention on the ability of terrestrial ecosystems to withstand or adapt to this nutrient influx. Ecosystem responses are variable, ranging from enhanced net primary productivity [Vitousek and Howarth, 1991] to, at the other extreme, N saturation with associated declines in productivity [Aber et al., 1989]. There is a growing body of knowledge of how specific ecosystems respond in the short term to these novel N inputs [Vitousek et al., 1997] and yet, in the absence of many long-term N cycling studies, it is difficult to extend these studies spatially across broad gradients of climate and geography.

[3] Decades of research have shown that the global pattern of soil organic nitrogen (SON) storage in undisturbed ecosystems is a function of a suite of ecosystem variables [Jenny, 1941; Post et al., 1985], among the most important being mean annual precipitation (MAP) and mean annual temperature (MAT) [Jenny, 1928]. Climate affects SON storage by moderating N input and output rates [Olson, 1958]. Input rates are controlled by combined rates of atmospheric deposition [Holland et al., 1999] and N fixation [Cleveland et al., 1999]. Output rates are mediated by microbial N processing and are proportionally related to the total SON pool size and to MAT and MAP. However, patterns of total SON storage versus climate do not by themselves provide insights into the mechanisms that transfer N through the soil (Figure 1).

image

Figure 1. Schematic diagram of soil and plant N and 15N “black box” mass balance model. Terms in parentheses are the flux terms for N and 15N, respectively, and are defined in text.

Download figure to PowerPoint

[4] Here we show, following on previous work [Austin and Vitousek, 1998; Handley et al., 1999; Martinelli et al., 1999; Schuur and Matson, 2001], that the natural abundance of N isotopes in soils and plants is also correlated with environmental variables, most importantly climate, at both local and global scales. In addition, we focus on the isotopic difference between plants and soils with climate. These patterns of N isotopes, when interpreted via mass balance models [Amundson and Baisden, 2000; Brenner et al., 2001], suggest systematic spatial variations in N cycling processes that may in turn be indicative of the response of ecosystems to increased N deposition or other forms of disturbance.

2. Background of N Isotope Research

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Background of N Isotope Research
  5. 3. Methods
  6. 4. Results and Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[5] Hoering [1955] first reported the natural abundance of N isotopes in natural compounds using present-day mass spectroscopic technology and standards (δ15N(‰) = [(Rsample/Rair)−1] × 1000, where R = the 15N/14N of a sample or standard, and where the standard is atmospheric N2). Atmospheric N2 = 0.3663 atom% 15N [Junk and Svec, 1958] and is invariable over a wide geographical area [Mariotti, 1983]. The use of N isotopes in ecology and agriculture arose primarily through the use of 15N to trace the fate of agricultural N in the environment [Kohl et al., 1971], later subjected to a variety of criticisms [e.g., Bremner and Tabatabai, 1973], and the use of plant δ15N values to quantify the amount of biologically fixed N utilized by that plant [e.g., Shearer and Kohl, 1986]. In the δ15N method of measuring biological N fixation, the δ15N value of biologically fixed N is well known (0 ± 2 ‰ [Shearer and Kohl, 1986]) while the δ15N value of the other plant N source (soil N) is highly variable.

[6] In order to better constrain the soil N isotopic signature for N fixation studies, Shearer et al. [1978] examined the δ15N values of soil surface samples (0–15 cm) from North America. They found that soil δ15N values decreased with increasing precipitation, but were unable to offer a hypothesis to account for this trend. Studies of δ15N values of whole soils (below surface horizons) across geographical gradients were not initially as commonly measured. Mariotti et al. [1980] examined the plant and soil δ15N value variation along an elevation gradient (climosequence) in France, and found that both plant and soil δ15N values declined with increasing elevation (and declining temperatures and increasing precipitation), though they did not specifically address the mechanisms responsible for the isotopic variations.

[7] More recent studies of soils and plants have been made along controlled environmental gradients. Vitousek et al. [1989] examined soil and plant δ15N values along a Holocene to Pliocene time gradient in Hawaii, while Hobbie et al. [1998] characterized plant and litter δ15N values along a Holocene time gradient in Alaska. Austin and Vitousek [1998] measured soil and plant δ15N values along a rainfall gradient on young (∼3000 years) soils in Hawaii, and Schuur and Matson [2001] conducted detailed δ15N studies on ecosystems along a rainfall gradient on much older landscapes. Two recent data sets have been compiled of soil and plant isotope patterns [Handley et al., 1999; Martinelli et al., 1999] that add to these climatic and geographical comparisons.

[8] The role of climate on soil and plant δ15N values has gained increasing attention. Austin and Vitousek [1998] showed that ecosystem δ15N values decreased with increasing precipitation, and discussed the likelihood that the forms (and isotopic composition) of N that is lost change systematically as precipitation increases. Handley et al. [1999], using a data set gleaned from the literature and unpublished data, generally expanded the importance of water availability on ecosystem δ15N values, although their discussion of the strength of the relationship to precipitation was somewhat equivocal. Handley et al. [1999] also developed a conceptual and schematic model of soil N cycling and noted that the cause of the geographical patterns of soil δ15N values likely resides in differences in the isotopic composition of N leaving the system. Martinelli et al. [1999] assembled a large data set in order to compare the δ15N values of soils and plants in tropical versus temperate forests. They noted that tropical forest ecosystems are commonly enriched in 15N relative to their temperate counterparts, and attributed these differences to the differential importance of N as a limiting plant element in the tropics versus temperate regions, and to the increased “openness” (rates of N input/output versus internal N cycling) of the N cycle in tropical forests.

[9] Amundson and Baisden [2000] and Brenner et al. [2001] presented a simplified ecosystem N and 15N mass balance model that contains elements of previous models [e.g., Handley et al., 1999; Herman and Rundel, 1989; Shearer et al., 1974], but which differs due to its focus only on system inputs/outputs, the only processes which can ultimately alter the isotopic composition of an ecosystem (Figure 1). The mass balance expressions for total N in the soil/plant system are

  • equation image
  • equation image

where Ns, Np = soil and plant N pools (kg m−2), respectively; Iex, Ifix = atmospheric N inputs (kg m−2 yr−1) and biological N fixation, respectively; kex = soil N loss decay constant (yr−1), kp = plant available N decay constant, and ks = constant describing annual N return to the soil from plants.

[10] The expressions that define 15N are

  • equation image
  • equation image

where R = 15N/14N of subscripted flux or pool and α = isotopic fractionation factor (unitless) accounting for isotopic discrimination during subscripted process.

[11] The steady state solution of this model for soil is (see Brenner et al. [2001] for more details)

  • equation image

where equation imagetotal is the weighted mean isotopic ratio of inputs, and for plants:

  • equation image

which, if no N fixation occurs, equals

  • equation image

[12] One of the key implications of this model is that the isotopic composition of the total soil N pool at steady state is determined only by the isotopic composition of the inputs and losses. This straightforward mathematical result has been recognized by some [Handley et al., 1999], but the concept is obscured with the rates of N cycling and loss in other papers. Virtually all soil systems, with the exception of those in hyperarid climates that have no leaching [Böhlke et al., 1997], are open and have N losses that eventually approach input rates. The key factor controlling the steady state isotopic composition of soils is the value of the fractionation factor αex. Along these lines, Austin and Vitousek [1998] defined ecosystem openness in terms of the form (and isotopic composition) of N lost, a definition that is consistent with isotopic mass balance.

[13] The mass balance model considers the integrated soil N pool and does not address the 15N variations that may occur with depth. Observations in a variety of climates suggest that the δ15N value of soil N can vary in a number of ways with depth: (1) Along a montane elevation gradient in Hawaii, there was little or no depth variation [Uebersax, 1996], (2) in gravelly desert soils with irregular root distributions, δ15N varied randomly (in concert with roots) with depth [Brenner, 1999], and (3) most commonly, in grassland soils, δ15N values display a remarkably consistent exponential increase with depth [Mariotti et al., 1980; Brenner et al., 2001]. Modeling these patterns requires the inclusion in mass balance models of (1) transport processes, (2) depth-dependent plant N inputs, and (3) multiple N pools. To our knowledge, only models describing the exponential increase in 15N with depth have been presented [e.g., Amundson and Baisden, 2000; Baisden, 2000; Baisden et al., 2002]. While these models offer mechanistic insights into soil N processing, data to constrain the models (particularly soil N transport rates) is difficult to obtain. In summary, the black box model is particularly appropriate for comparative analyses of commonly available data collected across broad geographical gradients, the objective of this paper.

3. Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Background of N Isotope Research
  5. 3. Methods
  6. 4. Results and Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[14] We combined new data from Hawaii and Tanzania [Uebersax, 1996], French Polynesia [Matzek,1999], California and Nevada [Brenner, 1999], and recent compilations of global soil and plant δ15N values [Handley et al., 1999; Martinelli et al., 1999] and other data from the literature (see supplementary data1). For the new data, soils were sampled by horizon. Samples (depending on remoteness of the site) were either air or freeze dried, and ground. Samples from Hawaii and Tanzania were reacted with 1 N HCl to remove carbonates, while those of French Polynesia were acidified with 0.5 N HCl. The C and N content was measured on a Carlo-Erba elemental analyzer. Samples from Hawaii and Tanzania were combusted in sealed tubes [Kendall and Grim, 1990], and the purified N2 isotopically analyzed on a VG Prism isotope ratio mass spectrometer. Samples from California, Nevada, and French Polynesia were isotopically analyzed via elemental analyzer/continuous flow mass spectrometry (VG Optima and Europa 20/20). Plant tissue samples were analyzed by the same methods. Soil bulk density, where possible, was measured on soil cores of known volumes. The weighted mean (by depth, N content, bulk density) δ15N value in all soils was calculated to depths of both 10 and 50 cm. Many reports from the literature lack bulk density (and commonly total N), and thus the weighted means in these data were based only on horizon thickness.

[15] Previous soil data analyses have reported the upper 10 cm of soils [Handley et al., 1999]. We used studies reporting soil δ15N values to 50 cm because that quantity gives a more integrated view of total soil N storage and cycling. Nonetheless, soil δ15N values to 10 cm are well correlated to those of 50 cm (r2 = 0.85, n = 50), but are about 1.4 ‰ more negative (y intercept) due to the commonly observed increase in δ15Nsoil with depth. If multiple soils were sampled within the same climatic zone, the values were averaged to eliminate overrepresentation artifacts. Statistical analyses of the data were performed using JMP® statistical software. The resulting regression equations were illustrated spatially using global mean annual temperature and precipitation (0.5 × 0.5 degree grids) data compiled by Willmott and Matsuura [2000].

4. Results and Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Background of N Isotope Research
  5. 3. Methods
  6. 4. Results and Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

4.1. Soil and Plant δ15N Values Versus Climate

[16] We began by further testing the hypothesis that climate, especially precipitation, exerts a first-order control on ecosystem δ15N values as suggested in previous papers [e.g., Mariotti et al., 1980; Austin and Vitousek, 1998; Handley et al., 1999; Amundson and Baisden, 2000; Schuur and Matson, 2001]. In the regression analyses, MAT and MAP were chosen as the independent variables representing climate because of their well-known controls on N cycling processes as discussed above. While the inclusion of latitude improved model explanatory capabilities in some cases (as noted by Handley et al. [1999]), the general relationship between climate and latitude caused us to drop latitude as an independent variable.

[17] Our analysis shows that δ15Nsoil decreases with decreasing MAT and (less significantly) with increasing MAP (Table 1). The r2 of the climate to soil δ15N relationships increased when soil to 50 cm was used (versus 10 cm depth) and when studies designed explicitly to examine climate effects on soil N (climosequences, Table 2) were used. The regression model that described the “climosequence” data was then used to portray spatial trends in soil δ15N values (Figure 2a). The global trends resulted in strong latitudinal banding, with high northern latitude ecosystems having the most depleted soil δ15N values and arid and tropical zones having the most positive soil δ15N values. We note that while this global perspective is admittedly generated from a small data set, the general trends with climate would also occur with the larger data set for soils collected to 10 cm (Table 1). Therefore, pending the availability of more soil 15N analyses, the present Figure 2 represents our best estimate of trends (but possibly not absolute values) in global soil δ15N values.

image

Figure 2. (a) Estimated geographical distribution of soil δ15N values to 50 cm and (b) estimated geographical trends in Δδ15Nplant-soil. Global mean annual temperature and precipitation (0.5 × 0.5 degree grids) data are obtained from Willmott and Matsuura [2000].

Download figure to PowerPoint

Table 1. Summary of Regression Models, Where Parameters are Reported Only in Cases Where p < 0.10a
Ecosystem PropertyInterceptMAT, CMAP, mmModel r2n
  • a

    Data for models are discussed in section 3.

Soil δ15N (10)3.19850.1340−0.00050.1185
Soil δ15N (50)3.88640.1680−0.0070.1947
Soil δ15N (50) climosequences4.32660.2048−0.00120.3929
Plant δ15N0.06970.1548−0.00160.34106
Δsoil-plant(10)−4.69260.0911−0.00070.1349
Δsoil-plant(50)−7.28130.17180.2330
Δsoil-plant(50) climosequences−8.40120.18520.3721
Table 2. N Isotope Composition of Soil and Plant N for Climosequences Analyzed in Table 1
SiteCountryVegetationElevation, mMAT, CMAP, mmSoil δ15N to 10 cm, ‰Soil δ15N to 50 cm, ‰δ 15N PlantsΔplant−soil to 10 cm, ‰Δplant−soil to 50 cm, ‰Referencea
Kyle Canyon, NVUSAcreosote84018160      
joshua tree1400143265.96.82.9−3−3.91
pinyon-juniper1750124363.74.9−2.2−5.9−7.1 
ponderosa215095493.84−2.8−6.6−6.8 
Sierra Nevada, CAUSAoak grassland470183301.52.8−1.9−3.4−4.71
oak woodland730155702.54.3    
mixed conifer1240129100.82.6−4.5−5.3−7.1 
mixed conifer1950810552.73.4    
mixed conifer2890312702.54.7−2.4−4.9−7.1 
Mt. KilimanjaroTanzaniamontane forest18291919702.23.6−1.2−3.4−4.82
montane forest24541416702.34.5−1.1−3.4−5.6 
grassland25451415702.14.9−1.5−3.6−6.4 
grassland29901213203.65.4−2.1−5.7−7.5 
heather3505910404.86−3.1−7.9−9.1 
 390168204.95.9    
Kohala MountainsHawaiipiligrass, keawe1222318012.0812.5   2
buffelgrass, lantana, keawe6742057013.2513.1    
pasture with kikuyu grass9921810609.188.4    
pasture with kikuyu grass12001725005.434.7    
ohia, fern-trees12541730003.071.7    
Maui, HawaiiUSAmetrosideros forest13701622095.795.172−3.79−3.173
metrosideros forest13701624356.124.471−5.12−3.47 
metrosideros forest13701627594.875.61−1−5.87−6.61 
metrosideros forest13201633387.356.270.8−6.55−5.47 
metrosideros forest13001640303.23.46−2−5.2−5.46 
metrosideros forest12701650662.950.65−5−7.95−5.65 
FranceS Genevanot reported110075141054.145.1−1.9−6.04−74
1440516834.274.8−1.8−6.07−6.6 
1600418112.663.8−5−7.66−8.8 
1800319721.553−2.8−4.35−5.8 

[18] In interpreting these apparent climatic trends, we begin with the assumption [Handley et al., 1999] that the δ15N value of N inputs to the soils varies within a restricted range geographically. Clearly this hypothesis requires long-term data collected at widely dispersed sites to fully support it. It is known that the δ15N value of atmospherically derived N exhibits large temporal and spatial variations [Kendall, 1998; Heaton, 1987], and (especially downwind of human impacted areas) that nitrate is commonly enriched in 15N relative to that of ammonium [Kendall, 1998]. Nonetheless, when means of annual inputs are calculated, it has been suggested that inputs range between −3 and + 3‰ [Handley et al., 1999]. Whether this mean applies to the entire globe remains uncertain, but the hypothesis of a restricted range of inputs for individual climosequences used in our regression analysis is supportable. All these sequences were established along elevation gradients, where sites are within 50 km or less of each other.

[19] If the range of δ15N values of N inputs is restricted globally, then the variation in soil δ15N values entirely reflects variations in αex (equation (5)). The data (Table 1) suggest that as MAP increases and MAT decreases (1) αex apparently approaches 1, (2) the steady state δ15N value of soil approaches that of estimated atmospheric N inputs (as required by equation (5)), and (3) the isotopic composition of losses equal that of the soil N. Nonfractionated losses (relative to the soil N pool) may occur in the form of dissolved organic N (DON) and/or particulate N in eroded soil material, though there have been few published measurements to date of the isotopic composition of DON [Feuerstein et al., 1997], and none have been directly applied to soil processes. In the Great Lakes, Feuerstein et al. [1997] noted that DON was a few per mil more negative than coexisting particulate organic matter in the water column. Recently, Perakis and Hedin [2002] have shown that in middle- to high-latitude South American temperate forests, the primary form of N loss from the soils is DON (up to about 3 kg N ha yr−1). This follows the work of Hedin et al. [1995], who showed that along an elevation gradient on the southern Chilean coast, the ratio of NO3/DON and NO3/NH4+ decreased systematically with increasing elevation (and decreasing temperature). These findings are consistent with the soil isotopic trends derived here if DON losses are relatively nonfractionating. In contrast, the apparent propensity of ecosystems to lose 15N-depleted forms of N (relative to the soil N pool) with decreasing MAP and increasing MAT is striking. These 15N-depleted forms of N loss might be expected to include NO3, N2O [Pérez et al., 2000], and other mineralized forms of N, all of which are thought to be 15N-depleted relative to total soil N [Shearer and Kohl, 1986]. However, complications to this interpretive scenario occur when kinetic limitations or competing N pathways cause the observed fractionation to be different than optimal values (see Shearer and Kohl [1986] for expanded discussion). Regardless of the form of N lost from soils, there is a consistent trend in the apparent isotopic composition of the loss (relative to the soil N pool) with climate.

[20] Plant δ15N values decreased with increasing MAP and with decreasing MAT (Table 1), like the soils, but the magnitude of the difference between plant and soil Δδ15N values (Δδ15Nplant-soil) decreased with increasing MAT (and less strongly with decreasing MAP) (Table 1, Figure 2b). We note that plant δ15N values can vary up to several per mil with season, plant part, and between different types of plants at the same site [Handley et al., 1999]. Most analyses in the data set used here are foliar tissue, and minimize the plant tissue effect. While recognizing the potential variability in plant isotope values, the δ15N value of non-N fixing plants should, as we hypothesize in equation (4), reflect the isotopic composition of mineral forms of N derived from the soil N pool (e.g., αp). Therefore, the simple measure (Δδ15Nplant-soil, which is equivalent to 103lnαp, might be interpreted as the isotopic composition of plant-available N provided that isotopic discrimination does not occur during plant uptake and assimilation. There is presently disagreement over the important issue of assimilation-based N isotope discrimination. Evans et al. [1996] reported little isotopic discrimination during the uptake of either NO3 or NH4+ by tomatoes at concentrations comparable to field conditions. In contrast, Yoneyama et al. [2001] and others [e.g., see Högberg, 1997] report small discrimination during the uptake of NO3 by rice at field concentrations, but relatively large discriminations during NH4+ uptake (with discrimination increasing with concentration). Clearly, this is an area deserving of additional work. Even if assimilation-based discrimination does not occur, the isotopic composition of plant available N deduced from (Δδ15Nplant-soil is only suggestive of N forms. For example, an increase in (Δδ15Nplant-soil may be interpreted as a shift in the N source for plants from organic N or NH4+ (which may be only a few ‰ more negative than soil N) to NO3, which may be about 10‰ or more negative than the organic N from which it is derived (see Shearer and Kohl [1986] for summary of observed fractionation factors for these processes).

[21] The apparent Δδ15Nplant-soil trend with changes in MAT can be reconciled with hypothesized shifts in soil N economies from organic/NH4+ -dominated in cool temperate forests [e.g. Hedin et al., 1995] to NO3 -dominated in tropical forests [Aber, 1992; Matson et al., 1999]. In cold ecosystems with large differences in Δδ15Nplant-soil, two mechanisms may be responsible: (1) The organic N/NH4+ plant-available sources in northern latitudes are 15N-depleted and/or (2) isotopic discrimination during NH4+ uptake is occurring. A mechanism consistent with (1) that has been recently proposed [Hobbie et al., 1999, 2000] is that plant N sources in northern latitudes may be depleted in 15N through mycorrhizal fungi/plant interactions (reviewed by Högberg [1997]). The possibility that the second mechanism occurs was discussed above. In contrast, the small Δδ15Nplant-soil at high temperatures may be explained by increased microbial competition with plants for N with increasing temperature, resulting in a relatively low concentration of mineral N available as NH4+ and with subsequent isotopic enrichment of any NO3 that is plant available. Alternatively, denitrification or other competing processes may enrich remaining soil NH4+ or NO3 with increasing temperature, resulting in low plant-soil values.

[22] What is particularly striking is the apparent independent behavior of both αp (as estimated by Δδ15Nplant-soil and αex with climate [see also Schuur and Matson, 2001]. The available data suggest that as MAP increases and MAT decreases, the isotopic composition of N leaving the soil approaches that of the soil N pool (e.g., reflecting little isotopic fractionation) while the form of N taken up by plants becomes increasingly depleted relative to the soil N pool. Assuming that the trends reflect true values of plant available N, the most likely explanation of this apparently incompatible relationship is that plants, and all ecosystem N consumers, become increasingly efficient at mineral N cycling with increasing moisture [Austin and Vitousek, 1998] such that only non plant-available forms of N are subject to leaching or erosional losses.

4.2. Other Controls on Ecosystem δ15N Values

[23] The modest r2 of the statistical relations between plant/soil δ15N values and climate (Table 1) indicates a possible suite of other controls on ecosystem δ15N values. One obvious source of variability, already discussed, are spatial variations in the isotopic composition of atmospheric N inputs. Yet, beyond this explanation are other controls (ecosystem state factors [Jenny, 1941]) on ecosystem N cycling that have been shown or are suspected of exerting secondary (or even primary) controls on total soil N and δ15N values. The known ranges of the effects of each of these factors on soil δ15N values (discussed below) are illustrated in Figure 3.

image

Figure 3. Estimated range in the effect of individual state factors [Jenny, 1941] on the δ15N value of soil N. Sources of values illustrated are discussed in the text.

Download figure to PowerPoint

[24] Soil age is an important control on soil and plant δ15N values for two reasons. First, for sites not yet at steady state, the δ15N value of the soil is dependent on soil age and the kinetics of N cycling (equations (1)(4)). There are, to our knowledge, no detailed studies of soil and plant N isotope variation over timespans pertinent to the approach to steady state (102 to 103 years based on soil N chronosequence data [Syers et al., 1970]). Hobbie et al. [1998, 1999] examined plant and soil organic horizons (as opposed to mineral soil horizons emphasized here) in Alaska and found pronounced trends in plants over a 250-year timespan, with the plant trend dependent on whether the plant derived N from soil N, N fixation, or symbiotic relations with mycorrhizal fungi.

[25] Soil age also exerts an influence on ecosystem δ15N values over longer timescales. It has been demonstrated convincingly in Hawaii [Vitousek et al., 1989; Martinelli et al., 1999], and also California [Brenner et al., 2001], that the δ15N values of soils and plants vary as sites age from 103 to 106 years. All these sites may be viewed as having obtained their own steady state, but the N cycling processes differ between sites due to changes in the relative importance of N versus P (or other elements) as limiting nutrients versus time. It has been hypothesized for the chronosequence in California [Brenner et al., 2001] that as P becomes plant limiting (due to weathering losses of this element), N becomes an “excess” element (which is reflected in elevated NO3 concentrations in soil lysimeters) [White and Brantley, 1995], and sites experience larger amounts of mineral N (and 15N depleted) forms of loss. In summary, because the Earth's surface is a complex suite of deposits and outcrops of greatly varying ages, it is likely that age plays a critical role in ecosystem N isotope variability. The largest reported range in soil δ15N values caused by soil age (Hawaii) is about 5.5‰ [Martinelli et al., 1999].

[26] Topographic position plays a key role in N cycling [Schimel et al., 1985], and preliminary work shows it affects soil δ15N values [Karamanos and Rennie, 1980a; Sutherland et al., 1993]. First, topographic positions subject to water collection and anaerobic conditions may exhibit denitrification and the residual accumulation of 15N [Karamanos and Rennie, 1980b]. However, more generally, convex portions of the landscape are subject to pervasive erosive removal of soil material and undergo constant rejuvenation [McKean et al., 1993; Heimsath et al., 1997]. The net rates of soil movement on slopes are proportional to slope [Heimsath et al., 1997]:

  • equation image

where Qs = sediment flux (cm2 yr−1) and D = soil diffusivity (cm2 yr−1). Briefly, the slope of a upland soil is proportional to the soil's residence time (τ) as follows:

  • equation image

[27] These relationships suggest that steeply sloping soils, which have low residence times, should be far from the steady state N (equation (1)) and δ15N values (equation (3)) for level soils at the same location, and should have δ15N values that approach atmospheric inputs. There are presently no published data to test this hypothesis. Here, we present part of a data set for a small zero order watershed in the central California coast range (Figure 4). At this site, previous research has shown that soil transport is proportional to slope [Heimsath et al., 1997], indicating that equation (9) is appropriate to interpret the data. The soils sampled follow a trend of decreasing δ15N values with increasing slope (decreasing residence time) (Figure 4). The δ15N value of the surface horizon, as residence time approaches 0, is 2.3‰ and similar to the presumed range in atmospheric N inputs. The y-intercept for the second soil horizon is enriched in 15N (δ15N = 3.9‰), suggesting (as expected) that a portion of the N inputs to this horizon are derived from microbially altered N transported downward from the overlying horizon. In summary, these preliminary data suggest that hillslope position can impart up to a 2‰ variation in soil δ15N values at a given location. In a topographic study in Saskatchewan that included both erosional landscape segments and the wetter depositional areas, the total range in soil δ15N values was reported to be about 4‰ [Sutherland et al., 1993].

image

Figure 4. Observed relationship between the δ15N value of soil N and 1/slope on the convex portions of a zero-order watershed in Marin County, California. Slopes were calculated based on elevation data and a kriging algorithm.

Download figure to PowerPoint

[28] The nature of the soil parent material has been shown to exert a control on soil and plant δ15N values. So-called “white sand” soils of the tropics, which are N limited, support plants lower in N, and depleted in 15N, relative to other soils in the region [Martinelli et al., 1999]. Matzek [1999] showed that in the island ecosystems of French Polynesia, soils and plants on coral substrate had more negative δ15N values than nearby ecosystems on comparably aged volcanic alluvium due to what was suspected to be greater DON losses. While there are not, to our knowledge, other explicit parent material comparisons on soil N isotopes, the wide variety of rock and sediment types globally suggest that this variable will exert considerable influence on N isotopes in soils and plants. Based on the work of Matzek [1999], the parent material effect on soil δ15N values at a given location is at least 3‰.

[29] Finally, human activity has caused a wholesale disruption of the global N cycle [Vitousek et al., 1997] and aspects of its N isotope composition. Because cultivation greatly accelerates loss of soil N through enhanced decomposition rates, there can be a pronounced increase in the δ15N value of the remaining soil N in certain soil N pools, although isotopic changes in the total N pools are not necessarily noted [Tiessen et al., 1984; Shearer and Kohl, 1986]. Riga et al. [1970] report up to a 3‰ increase in soil δ15N values to 60 cm due to cultivation. The addition of N fertilizers derived from atmospheric N2, or manures (which are commonly greatly enriched in 15N because of ammonia volatilization) also have an impact on soil and crop isotopic values. Riga et al. [1970] showed that fertilization with manure can increase soil δ15N values by about 1‰ relative to treatments without manures.

[30] While our analysis here focused deliberately on “undisturbed” ecosystems, the impact of humans on the N isotope biogeochemistry of these remaining parts of the planet is deserving of greater scrutiny. Most “undisturbed” ecosystems have experienced increased rates of atmospheric N deposition [Holland et al., 1999], likely driving some systems out of steady state [Vitousek et al., 1997; Perakis and Hedin, 2002]. It has been proposed that in temperate forests, the increasing N inputs result in an increased NO3/DON ratio of N losses from the soils [Perakis and Hedin, 2002; van Breeman, 2002], a process that likely affects the isotopic composition of the total soil and plant N pool. Additionally, the changes that atmospheric pollution have imparted on the long-term δ15N value of N deposition is not well known [Kendall, 1998].

4.3. Significance of Observations

[31] The climatic and spatial trends of soil and plant δ15N values we developed here are consistent with interpretations from previous work and data analyses, although it is clear that additional ecosystem measurements along well-constrained environmental gradients will greatly enhance the details of the climatic trends. However, even at this stage of our knowledge, it is important to address how these apparent ecosystem δ15N patterns might prove useful in environmental and ecological research.

[32] We hypothesize that as an integrator of long term N cycling [Robinson, 2001], N isotopes are indicators of systematic variations in ecosystem N cycling processes and efficiencies, and to the resilience of an ecosystem to enhanced N deposition through anthropogenic activities. Aber [1992] reviewed attributes of N-limited versus N-saturated ecosystems in the context of their response to atmospheric N deposition. In terms of characteristics relevant to N isotopes [Aber, 1992, Table 1], N-limited ecosystems are characterized by (1) NH4+ (versus NO3) uptake by plants, (2) high rates of NO3 and NH4+ immobilization, (3) high fraction of soil fungi that are mycorrhizal, and (4) very low rates of N2O production. Characteristics 2 and 4, which are considered characteristic of many temperate forests [Martinelli et al., 1999], are consistent with the spatial patterns of low soil δ15N values in cool, moist environments (and the opposite trends in tropical latitudes) (Figure 2a). Characteristics 1 and 3, considered characteristic of temperate to boreal ecosystems [Martinelli et al., 1999; Höberg, 1997], are consistent with the trends in (Δδ15Nplant-soil with climate (Figure 2b). Additionally, Aber [1992] proposed that high soil C/N ratios are indicative of N-limited ecosystems. There is a strong relationship between soil C/N ratios derived from work by Post et al. [1985] and climate (r2 = 0.47, p < 0.01, n = 26), and there is an inverse relationship (r2 = 0.38, p = < 0.01, n = 30) between total soil C/N ratios and mean soil δ15N values to 50 cm (using our climosequence data).

[33] Therefore, we argue that the δ15Nsoil and Δδ15Nplant-soil patterns (Figures 2a and 2b) accurately portray the sensitivity of ecosystems to increased rates of anthropogenically derived N, with those ecosystems with the most negative δ15N soil values and the largest Δδ15Nplant-soil being the least susceptible to immediate negative impacts of increased N deposition. While the locations and spatial patterns of sensitive ecosystems illustrated by N isotopes are not novel, they do illustrate that N isotopes, which integrate an array of properties and processes of N-sensitive ecosystems [Aber, 1992], are a useful tool in broad scale, or possibly site-specific, assessments of ecosystem N status.

[34] Beyond the possibly direct link to N sensitivity, the N isotope patterns and relationships outlined here suggest an array of fruitful experiments and observations to test or determine the processes behind the patterns. Some of the obvious avenues of research include (1) spatially and temporally broad measures of the δ15N value of total N deposition, as well as the δ18O and δ17O values of NO3; (2) well-designed temporal observations of the isotopic composition of NH4+, NO3, N2O and other species in soil waters of undisturbed terrestrial ecosystems to better constrain the relation of N fluxes to integrated soil δ15N values; (3) experiments on a broader range of naturally occurring plants (as opposed to select agricultural cultivars) to determine the degree of isotopic discrimination during plant N assimilation; and (4) further studies on the role and extent of mycorrhizal fungi on N uptake and isotope discrimination. Study of these topics will not only illuminate N isotope patterns, but will provide a more mechanistic understanding of the terrestrial N cycle.

5. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Background of N Isotope Research
  5. 3. Methods
  6. 4. Results and Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[35] Our data analysis indicates that climate plays a significant role in soil N and N isotope processing and retention. The presence, and explainable nature, of global δ15N patterns in soils and plants provides insights into the cycling of terrestrial N and ecosystem response to climatic change and to increasing fluxes of anthropogenically derived N. While the data indicate that there is a global pattern of ecosystem δ15N values that appears to be driven primarily by climate, our understanding of the mechanisms behind the processes is in its infancy. Multiple opportunities for research exist to test hypotheses about terrestrial N cycling that will greatly refine our understanding of these N isotope patterns.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Background of N Isotope Research
  5. 3. Methods
  6. 4. Results and Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[36] A.T.A. acknowledges support from the Inter-American Institute for Global Change Research (CRN-012) and the Fundación Antorchas of Argentina, and R.A. acknowledges the support of the California Agricultural Experiment Station and the Kearney Foundation of Soil Science. We thank two anonymous reviewers for helpful comments on the manuscript, and we thank P. Vitousek and J. Ehleringer for insights provided in early discussions. We thank Y. Guo for assistance in the preparation of Figure 2.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Background of N Isotope Research
  5. 3. Methods
  6. 4. Results and Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information
  • Aber, J. D., Nitrogen cycling and nitrogen saturation in temperate forest ecosystems, Trends Ecol. Evol., 7, 220224, 1992.
  • Aber, J. D., K. J. Nadelhoffer, P. Steudler, and J. M. Melillo, Nitrogen saturation in northern forest ecosystems, Bioscience, 39, 378385, 1989.
  • Amundson, R., and W. T. Baisden, Stable isotope tracers and mathematical models in soil organic matter studies, in Methods in Ecosystem Science, edited by O. E. Sala et al., pp. 117137, Springer-Verlag, New York, 2000.
  • Austin, A., and P. M. Vitousek, Nutrient dynamics on a precipitation gradient, Oecologia, 113, 519529, 1998.
  • Baisden, W. T., Soil organic matter turnover and storage in a California annual grassland chronosequence, Ph.D. dissertation, Univ. of Calif., Berkeley, Berkeley, Calif., 2000.
  • Baisden, W. T., R. Amundson, D. L. Brenner, A. C. Cook, C. Kendall, and J. Harden, A multi-isotope C and N modeling analysis of soil organic matter turnover and transport as a function of soil depth in a California annual grassland soil chronosequence, Global Biogeochem. Cycles, 16, 1135, doi:10.1029/2001GB001823, 2002.
  • Böhlke, J. K., G. E. Ericksen, and K. Revesz, Stable isotope evidence for an atmospheric origin of desert nitrate deposits in northern Chile and southern California, U.S.A. Chem. Geol., 136, 135152, 1997.
  • Bremner, J. M., and M. A. Tabatabai, 15N enrichment of soils and soil derived nitrate, J. Environ. Qual., 2, 363365, 1973.
  • Brenner, D. L., Soil nitrogen isotopes along natural gradients: Models and measurements, M.S. thesis, Univ. of Calif., Berkeley, Berkeley, Calif., 1999.
  • Brenner, D. L., R. Amundson, W. T. Baisden, C. Kendall, and J. Harden, Soil N and 15N variation with time in a California annual grassland ecosystem, Geochim. Cosmochim. Acta, 65, 41714186, 2001.
  • Cleveland, C. C., A. R. Townsend, D. S. Schimel, H. Fisher, R. W. Howarth, L. O. Hedin, S. S. Perakis, E. F. Latty, J. C. VonFischer, A. Elseroad, and M. F. Wasson, Global patterns of terrestrial biological nitrogen (N-2) fixation in natural ecosystems, Global Biogeochem. Cycles, 13, 623645, 1999.
  • Evans, R. D., A. J. Bloom, S. S. Sukrapanna, and J. R. Ehleringer, Nitrogen isotope composition of tomato (Lycopersicon esculentum MILL.cv.T-5) grown under ammonium or nitrate nutrition, Plant Cell Environ., 19, 13171323, 1996.
  • Feuerstein, T. P., P. H. Ostrom, and N. E. Ostrom, Isotopic biogeochemistry of dissolved organic nitrogen: A new technique and application, Org. Geochem., 27, 363370, 1997.
  • Galloway, J. N., The global nitrogen cycle: changes and consequences, Environ. Pollut., 102, 1524, 1998.
  • Handley, L., A. Austin, D. Robinson, C. Scrimgeour, J. Raven, T. Heaton, S. Schmidt, and G. Stewart, The 15-N natural abundance (δ15N) of ecosystem samples reflects measures of water availability, Aust. J. Plant Phys., 26, 185199, 1999.
  • Heaton, T. H. E., 15-N/14-N ratios of nitrate and ammonium in rain at Pretoria, South Africa, Atmos. Environ., 21, 843852, 1987.
  • Hedin, L. O., J. J. Armesto, and A. H. Johnson, Patterns of nutrient loss from unpolluted, old-growth temperate forests: Evaluation of biogeochemical theory, Ecology, 76, 493509, 1995.
  • Heimsath, A. M., W. E. Dietrich, K. Nishiizumi, and R. C. Finkel, The soil production function and landscape equilibrium, Nature, 388, 358361, 1997.
  • Herman, D. J., and P. W. Rundel, Nitrogen isotopic fractionation in burned and unburned chaparral soils, Soil Sci. Soc Am. J., 53, 12291236, 1989.
  • Hobbie, E. A., S. A. Macko, and H. H. Shugart, Patterns in N dynamics and N isotopes during primary succession in Glacier Bay, Alaska, Chem. Geol., 152, 311, 1998.
  • Hobbie, E. A., S. A. Macko, and H. H. Shugart, Insights into nitrogen and carbon dynamics of ectomycorrhizal and saprotrophic fungi from isotopic evidence, Oecologia, 118, 353360, 1999.
  • Hobbie, E. A., S. A. Macko, and M. Williams, Correlations between foliar δ15N and nitrogen concentrations may indicate plant-mycorrhizal interactions, Oecologia, 122, 273, 2000.
  • Hoering, T., Variation of nitrogen-15 abundance in naturally occurring substances, Science, 122, 12331234, 1955.
  • Högberg, P., 15-N natural abundance in soil-plant systems, New Phytol., 137, 179203, 1997.
  • Holland, E. A., F. J. Dentener, B. H. Braswell, and J. M. Sulzman, Contemporary and pre-industrial global reactive nitrogen budgets, Biogeochemistry, 46, 743, 1999.
  • Jenny, H., Relation of climatic factors to the amount of nitrogen in soils, J. Am. Soc. Agron., 20, 900912, 1928.
  • Jenny, H., Factors of Soil Formation: A System of Quantitative Pedology, McGraw-Hill, New York, 1941.
  • Junk, G., and H. V. Svec, The absolute abundance of the nitrogen isotopes in the atmosphere and compressed gas from various sources, Geochim. Cosmochim. Acta, 14, 234243, 1958.
  • Karamanos, R. E., and D. A. Rennie, Changes in natural 15N abundance associated with pedogenic processes in soil, I, Changes associated with saline seeps, Can. J. Soil Sci., 60, 337344, 1980a.
  • Karamanos, R. E., and D. A. Rennie, Changes in natural 15N abundance associated with pedogenic processes in soil, II, Changes on different slope positions, Can. J. Soil Sci., 60, 365372, 1980b.
  • Kendall, C., Tracing nitrogen sources and cycling in catchments, in Isotope Tracers in Catchment Hydrology, edited by C. Kendall, and J. McDonnell, Elsevier Sci., New York, 1998.
  • Kendall, C., and E. Grim, Combustion tube method for measurement of nitrogen isotope ratios using calcium oxide for total removal of carbon and water, Anal. Chem., 62, 526529, 1990.
  • Kohl, D. H., G. B. Shearer, and B. Commoner, Fertilizer nitrogen: contribution to nitrate in surface water in a corn belt watershed, Science, 174, 13311334, 1971.
  • Mariotti, A., Atmospheric nitrogen is a reliable standard for 15-N abundance measurements, Nature, 303, 685687, 1983.
  • Mariotti, A., D. Pierre, J. C. Vedy, S. Bruckert, and J. Guillemot, The abundance of natural nitrogen 15 in the organic matter of soils along an altitudinal gradient, Catena, 7, 293300, 1980.
  • Martinelli, L. A., M. C. Piccolo, A. R. Townsend, P. M. Vitousek, E. Cuevas, W. McDowell, G .P. Robertson, O. C. Santos, and K. Treseder, Nitrogen isotopic composition of leaves and soil: Tropical versus temperate forests, Biogeochemistry, 46, 4565, 1999.
  • Matson, P. A., W. H. McDowell, A. R. Townsend, and P. M. Vitousek, The globalization of N deposiiton: Ecosystem consequences in tropical environments, Biogeochemistry, 46, 676683, 1999.
  • Matzek, V. A., Differences in nutrient cycling and water availability in soils forming on different substrates: Evidence from coral and volcanic islands in tropical Pacific, M.S. thesis, Univ. of Calif., Berkeley, Berkeley, 1999.
  • McKean, J. A., W. E. Dietrich, R. C. Finkel, J. R. Southon, and M. W. Caffee, Quantification of soil production and downslope creep rates from cosmogenic 10Be accumulations on a hillslope profile, Geology, 21, 343346, 1993.
  • Olson, J. S., Rates of succession and soil changes on southern Lake Michigan sand dunes, Bot. Gaz., 119, 125170, 1958.
  • Perakis, S. S., and L. O. Hedin, Nitrogen loss from unpolluted South American forests mainly via dissolved organic compounds, Nature, 415, 416419, 2002.
  • Peréz, T., S. E. Trumbore, S. C. Tyler, E. A. Davidson, M. Keller, and P. B. De Camargo, Isotopic variability of N2O emissions in tropical forest soils, Global Biogeochem. Cycles, 14, 525535, 2000.
  • Post, W. M., J. Pastor, P. J. Zinke, and A. G. Stagenberger, Global patterns of soil nitrogen storage, Nature, 317, 613616, 1985.
  • Riga, A., H. J. VanPraag, and N. Brigode, Rapport isotopique naturel de l'azote dans quelques sols forestiers et agricoles de belgique soumis a divers traitements culturaux, Geoderma, 6, 213222, 1970.
  • Robinson, D., δ15N as an integrator of the nitrogen cycle, Trends Ecol. Evol., 16, 153162, 2001.
  • Schimel, D., M. A. Stillwell, and R. G. Woodsmansee, Biogeochemistry of C, N, and P in a soil catena of the shortgrass steppe, Ecology, 66, 276282, 1985.
  • Schuur, E. A. G., and P. A. Matson, Net primary productivity and nutrient cycling across a mesic to wet precipitation gradient in Hawaiian montane forest, Oecologia, 128, 431442, 2001.
  • Shearer, G., and D. Kohl, N2−fixation in field settings: Estimations based on natural 15-N abundance, Aust. J. Plant Phys., 13, 699756, 1986.
  • Shearer, G., J. Duffy, D. H. Kohl, and B. Commoner, A steady-state model of isotopic fractionation accompanying nitrogen transformations in soil, Soil Sci. Soc. Am. Proc., 38, 315322, 1974.
  • Shearer, G., D. H. Kohl, and S. H. Chien, The nitrogen-15 abundance in a wide variety of soils, Soil Sci. Soc. Am. J., 42, 899902, 1978.
  • Sutherland, R. A., C. vanKessel, R. E. Farrell, and D. J. Pennock, Landscape-scale variations in plant and soil nitrogen-15 natural abundance, Soil Sci. Soc. Am. J., 57, 169178, 1993.
  • Syers, J. K., J. A. Adams, and P. R. Walker, Accumulation of organic matter in a chronosequence of soils developed in windblown sand in New Zealand, J. Soil Sci., 21, 146153, 1970.
  • Tiessen, H., R. E. Karamanos, J. W. B. Stewart, and F. Selles, Natural nitrogen-15 abundance as an indicator of soil organic matter transformations in native and cultivated soils, Soil Sci. Soc. Am. J., 48, 312315, 1984.
  • Uebersax, A., The content and stable isotope systematics of carbon and nitrogen in soil organic matter from elevation transects in Hawaii (USA) and Mt. Kilimanjaro (Tanzania), M.S. thesis, Univ. of Calif., Berkeley, Berkeley, 1996.
  • van Breeman, N., Nitrogen cycle: Natural organic tendency, Nature, 425, 381382, 2002.
  • Vitousek, P. M., and R. W. Howarth, Nitrogen limitation on land and sea: How can it occur? Biogeochemistry, 13, 87115, 1991.
  • Vitousek, P. M., G. Shearer, and D. H. Kohl, Foliar 15-N natural abundance in Hawaiian rainforest: Patterns and possible mechanisms, Oecologia, 78, 383388, 1989.
  • Vitousek, P. M., J. D. Aber, R. W. Howarth, G. E. Likens, P. A. Matson, D. W. Schindler, W. H. Schlesinger, and D. G. Tilman, Human alteration of the global nitrogen cycle: Sources and consequences, Ecol. Appl., 7, 737750, 1997.
  • White, A. F., and S. L. Brantley, Chemical weathering rates of silicate minerals: An overview, in Chemical Weathering Rates of Silicate Minerals, Rev. Mineral., vol. 31, edited by A. F. White, and S. L. Brantley, pp. 122, Mineral. Soc. of Am., Washington, D. C., 1995.
  • Willmott, C. J., and K. Matsuura, Terrestrial air temperature and precipitation: Monthly and annual time series (1950–1996), version 1.01, report, Cent. for Clim. Res., Dep. of Geogr., Univ. of Del., Newark, 2000.
  • Yoneyama, T., T. Matsumaru, K. Usui, and W. M. H. G. Engelaar, Discrimination of nitrogen isotopes during absorption of ammonium and nitrate at different nitrogen concentrations by rice (Oryza sativa L.) plants, Plant Cell Environ., 24, 133139, 2001.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Background of N Isotope Research
  5. 3. Methods
  6. 4. Results and Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

Supporting material is available via Web browser or via Anonymous FTP from ftp://ftp.agu.org, directory “append” (Username = “anonymous”, Password = “guest”); subdirectories in the ftp site are arranged by paper number. Information on searching and submitting electronic supplements is found at http://www.agu.org/pubs/esupp_about.html.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.