INTRODUCTION

Stable and radioactive isotopes provide useful tracers in aquatic ecology and biogeochemistry to elucidate flow paths, transformations, and sources of dissolved and particulate materials, as well as to measure rates of biological and chemical processes (Newbold and others 1983; Bower and others 1987; Michener and Schell 1994; Hall 1995; Peterson and others 1997). Whole ecosystem tracer-level additions are a relatively new approach employing isotopes to examine elemental cycles and food webs in lakes, streams, and estuaries (for example, Newbold and others 1983; Hall 1995; Peterson and others 1997; Hughes and others 2000; Tank and others 2000; Cole and others 2002). This approach has focused primarily on nitrogen (15NH +4 , 15NO −3 ) and phosphorus (32PO −34 ) dynamics in streams (for example, Newbold and others 1983; Peterson and others 1997; Tank and others 2000; Peterson and others 2001) and estuaries (Holmes and others 2000; Hughes and others 2000); C cycling has been examined to a lesser degree (Hall 1995; Hall and Meyer 1998; Cole and others 2002). Complex mixtures of 14C-labeled compounds from algal and wetland plant leachates have been used to examine C dynamics in laboratory experiments (for example, Wetzel and others 1995; Tranvik and Kokalj 1998), however, radiolabeled mixtures like these are not typically released into the environment. To date, the flow of organic C through lotic and lacustrine food webs has been tracked with stable isotopes using either an inorganic substrate for photosynthesis (H13CO −3 ) or a monomeric labile organic molecule (13C-acetate) (Hall 1995; Hall and Meyer 1998; Cole and others 2002). The intricacies of the C cycle, however, cannot be completely elucidated with either bicarbonate or acetate as tracers because stream dissolved organic carbon (DOC) comprises a mixture of compounds primarily of terrestrial origin that vary in their biological availability (Thurman 1985; Kaplan and Newbold 1993; Meyer and others 1998; Palmer and others 2001; Kaplan and Newbold 2003). Thus, to examine stream organic matter dynamics, we labeled deciduous trees with 13CO2 to create a stable isotope tracer more representative of stream DOC. Deciduous leaf litter leachate like stream DOC is a complex mixture of organic compounds with substantial polymeric character and various biodegradable fractions (Cummins and others 1972; McClaugherty 1983; Thurman 1985; Qualls and others 1991; Hongve 1999). We recognize that leaf litter leachate is not a perfect model for stream DOC; it has not been extensively processed in the soil like most DOC entering streams. Nevertheless, we feel that it is a good tool to begin measuring in situ stream dynamics given that DOC leached from leaf litter can account for 30% of the daily DOC export from small forested streams on an annual scale and can comprise 42% of the DOC inputs in the fall (McDowell and Fisher 1976; Meyer and others 1998).

Numerous studies have isotopically labeled tree C, focusing on the physiology of photosynthate allocation (Schier 1970; Webb 1977; Isebrands and Nelson 1983; Smith and Paul 1988; Horwath and others 1994). To our knowledge, trees have not been isotopically labeled to generate a product for an environmental tracer experiment. Most tree C labeling studies have used 14C (that is, Schier 1970; Webb 1977; Isebrands and Nelson 1983; Smith and Paul 1988; Horwath and others 1994) rather than 13C (Maillard and others 1994; Vivin and others 1996; Simard and others 1997) because of its lower cost and widely available sensitive detection methods; 14C has a six-orders of magnitude lower detection limit than 13C. For health, safety, and environmental reasons, radiolabeled C is generally not used for field releases (Hesslein and others 1980; Bower and others 1987). However, with new methodology and instrumentation advances, the environmentally safe 13C is now being used as a tracer.

The overall goal of our study was to generate a mixture of terrestrially derived, 13C-labeled DOC molecules that could be used to examine in situ stream DOC dynamics associated with fresh tree litter inputs. Here, we present a method to label tree C with 13CO2 and test the efficacy of using labeled leachate from the trees in a whole-stream DOC tracer addition.

METHODS

Trees were labeled with 13C in two different years (2000, 2001) at the National Phytotron located at Duke University, Durham, North Carolina. During 2000 (Year 1), tree seedlings were grown with 0.82 moles of 13CO2 to assess the distribution and level of 13C enrichment. During 2001 (Year 2), tree seedlings were grown with 25-fold more 13CO2 to generate tree tissues with a strong enough 13C signal for a 13C-DOC stream tracer addition. In 2002, we used leachate generated from Year 2 13C-labeled tree tissues in a whole-stream isotope tracer addition.

Tree Cultivation

Thirty-two one-year-old bare root tulip poplar (Liriodendron tulipifera L.) seedlings were planted prior to leaf out in 10 L (16-cm diameter, 51-cm depth) PVC containers filled with a 3:1 (v/v) mixture of sterilized gravel and vermiculite. The seedlings were placed in a 1.2 m × 2.4 m × 2.1 m walk-in plant growth chamber. The trees were watered to saturation twice daily, using one-half strength Hoagland’s solution in the morning and distilled water in the afternoon (Downs and Hellmers 1975). The photoperiod, irradiance, temperature, relative humidity, and atmospheric CO2 concentration in the chamber were the same during Years 1 and 2 (Table 1). As soon as the seedlings were planted, CO2 was injected into the growth chamber to maintain the CO2 concentration set point (Tables 1 and 2). This CO2 provided the background CO2 in the growth chamber (Table 2; Bkg. CO2); 13CO2 was later added to the chamber air to label the trees with 13C.

Table 1 Environmental Conditions in Growth Chambers for Years 1 and 2 Tree C Labeling Experiments
Full size table
Table 2 13CO2 Injection Rates for Years 1 and 2 Tree C Labeling Experiments
Full size table

Year 1

At 52 days after planting (DAP), the locations of the youngest fully expanded leaf on each plant’s main stem and branches were marked to delineate aboveground plant tissues that formed prior to the 13CO2 injection (old tissues) from those that formed during the injection (new tissues). Immediately before the injection, an automatic watering system was installed and the chamber was sealed to minimize 13CO2 loss. 13CO2 was injected into the chamber from a lecture bottle of 99 atom % 13C-CO2 (Cambridge Scientific Laboratories) for 8 h during daylight starting on 53 DAP. A mass flow valve (0–10 sccm) was used to meter the 13CO2 injection rate into the chamber and help mix the 13CO2 with the chamber air (Table 2). Eighty-four DAP, the seedlings were harvested and their leaf, stem, and root tissues were separated into old and new tissues. Old and new leaf and stem tissues were separated using the marks placed on the seedlings prior to the 13CO2 injection. Root tissue was separated into two portions following washing; the upper four-fifths and the lower fifth of the root ball were considered old and new tissues, respectively. All tissues were dried to a constant weight at 60°C.

Year 2

At 24 DAP, seven seedlings were performing poorly and replaced with pot-grown seedlings of the same age. At 42 and 50 DAP, the trees were sprayed with Safer’s® sulfur spray to control powdery mildew. Fifty DAP, the trees were marked to distinguish old and new tissues using the same protocol as in Year 1. The 13CO2 injection began 51 DAP following the protocol from Year 1 with minor modifications to the daily injection period and rate (Table 2). One hundred and seven DAP, the seedlings were harvested and their leaf, stem, and root tissues were separated into old and new tissues. Old and new leaf, stem, and root tissues were separated and dried according to the protocol from Year 1.

Tree Tissue Preparation

Dried tissues were ground in a Wiley Mill through a 250-μm mesh sieve prior to the cold water extraction and isotope analysis. Tree tissues were ground from least to most enriched (old to new) and the mill was cleaned between samples. Composite samples (30 to 100 mg) of each tissue type were prepared for isotope analysis from equal aliquots of ground tissues from all 32 seedlings.

Cold Water Extraction of New Tree Tissues

In Year 1, approximately 3 g of both ground new leaves and stems were added to 1 L of 0.2-μm filtered (polyethersulfone membrane filter, Gelman Supor®-200) de-ionized water (DIW). The water and plant material were gently stirred with a Teflon-coated stir bar in a 2-L glass container in the dark at 4°C for 24 h. After 24 h, the water and plant material were centrifuged at 17,931 × g for 0.5 h. The supernatant liquid was sequentially filtered through DIW-rinsed paper (Whatman 4) and GF/F (Whatman) filters. Sub-samples of the leachate were taken for DOC concentration (total DOC = 12C-DOC + 13C-DOC), dissolved total saccharides (DTS) and free monosaccharides (DFMS) (0.2-μm filtered; HT Tuffryn® membrane Acrodisc®), humic substances, and δ13C-DOC isotope analyses. DOC and humic substance samples were analyzed immediately following collection and isotope and DFMS samples were stored frozen until analysis. DTS samples were acidified to pH 1 with 6 N HCl and stored at 4°C until analysis. All glassware and GF/F filters were baked at 500°C for 6 h prior to use.

In Year 2, approximately 1 g of ground new tree tissue (60.9% leaves, 24.6% stems, 14.5% roots; % weight of new tissues) was added to 1 L of 0.2-μm filtered DIW. The procedure for cold water extraction was similar to that used in Year 1 with the exception that the leachate was only filtered through a GF/F filter following centrifugation. Samples of the tree tissue before leaching, as well as the leachate were taken for C, DOC, and 13C isotope analyses. Samples of the leachate were also taken for humic substance and carbohydrate analyses. DOC, humic substance, carbohydrate, and isotope samples were stored following Year 1 procedures.

13C-DOC Whole-Stream Tracer Addition

In October 2002, 120 L of 13C-labeled tree tissue leachate (120 mg C L−1; δ13C 6160) were released into White Clay Creek, a third-order forested piedmont stream located in southeastern Pennsylvania (39°53′N, 75°47′W), and its concentration was measured along a 1265-m reach at eight stations. Tree tissue leachate was prepared using Year 2 new tree tissues; approximately 4 g of tissue (60.9% leaves, 24.6% stems, 14.5% roots; % weight of new tissues) were added to 4 L of 0.2-μm (polyethersulfone membrane, Gelman Supor®-200) filtered DIW. The procedure for cold water extraction was similar to that used in Year 2 with the exception that the leachate was Tyndallized for 0.5 h in a 70°C water bath following GF/F filtration. Twenty-four hours after Tyndallization, the process was repeated to ensure the biological stability of the leachate. The leachate was stored in 2-L sterile plastic containers in the dark at 4°C until the stream injection. The 13C-DOC was injected into the stream over the course of 115 min using a peristaltic pump at the rate of 1062 ml min−1. After the 13C-DOC was thoroughly mixed in the stream water (∼60 min after the pump started) and its concentration stabilized (plateau period), water samples for DOC and 13C-DOC were collected at the eight stations and filtered through GF/F filters (Stream Solute Workshop 1990). DOC samples were immediately analyzed following collection and 13C-DOC samples were stored frozen until analysis. A further description of data analysis to estimate 12C-DOC uptake from the 13C-DOC release is beyond the scope of this manuscript (Wiegner and others unpublished data).

Analytical Measurements

DOC was measured by Pt-catalyzed persulfate oxidation using either an OI 700 or an OI 1010 (O.I. Corp.) carbon analyzer (Kaplan 1992). Humic substances were measured using XAD-8 resin chromatography (Stanley and Kaplan 1998). Humic carbon was defined as the difference between DOC concentrations of the sample and XAD-8 resin effluent. Dissolved carbohydrates were assayed using high-performance liquid chromatography and pulsed amperometric detection (Cheng and Kaplan 2001, 2003). DFMS were measured on samples without pretreatment and DTS were measured following acid hydrolysis. Dissolved polysaccharide concentrations were calculated as the difference between the DTS and DFMS concentrations.

Tree tissue samples were prepared for isotope analysis and subsequently combusted by a modified Dumas technique (Macko and others 1987). Preparation of DOC for isotope analysis followed the method of Feuerstein and others (1997), except that the samples were not dialyzed. The gases resulting from the Dumas combustion of tree tissues and DOC were purified by cryogenic gas separation and subsequent isotope determinations were performed using either a dual inlet on a PRISM (Micromass) stable isotope ratio mass spectrometer or an Eurovector elemental analyzer interfaced to a ISOPRIME (Micromass) stable isotope mass spectrometer (Wong and others 1992). The elemental analyzer method yielded results (average ±S.D.) that were within 0.14‰ (±0.15) of the modified dual inlet method (Feuerstein and others 1997) at natural abundance levels and 0.024 atom % (±0.037) at enriched levels (Gandhi and others 2004). Stable carbon isotope ratios were expressed as:

$$ {\delta^{13}{\rm{C}}} = [(R_{{\rm{sample}}} / R_{{\rm{standard}}}) -1] \times 1000$$
(1)

where R is 13C/12C in the samples and standard. The standard used was V-PDB and the reproducibility of these measurements at natural abundance levels was not more than 0.2‰ and less than 0.03 atom % for enriched values (Table 3). The concentration of organic carbon in tree tissue samples was estimated using either a calibrated baratron capacitance manometer (MKS Instruments) during gas separation or the Eurovector elemental analyzer.

Table 3 Dry Weight and Average (±S.D.) Carbon Content, δ13C, Atom %, and Moles of 13C from the 13CO2 Tank (n13C) in the Old and New Tree Tissues from Years 1 (n = 2) and 2 (n = 3) Tree C Labeling Experiments
Full size table

Calculations

13C enrichment of the tree tissues (fractional abundance, F) was calculated from measured δ13C values and is reported as atom % (atom % = F × 100), where:

$$ {\rm{F}} = {{\rm{R}}/({\rm R}+1)} = ^{13}{\rm{C}} / (^{12}{\rm{C}} +^{13}{\rm{C}}) $$
(2)

These carbon isotope data were used in a two-source mixing model to provide estimates of the fraction of 13C from the enriched CO2 tank incorporated by the trees. The equation derived is similar to that used by Johnston and others (1997):

$$ {\rm{TDC}} = {\rm[({F_T}-{F_{CT}})/({F_{CO2}} -\Delta - {\rm{F}}_{{\rm{CT}}})]} $$
(3)

where TDC (tank derived carbon) is the fraction of carbon from the 13C-enriched CO2 tank in a given sample of the enriched tree tissue, F T is the measured fractional abundance of 13C in the enriched tree tissue (Table 3), F CT is the fractional abundance of 13C in the control tree tissue end member (0.010815, the average F for deciduous trees grown with atmospheric CO2; Broadmeadow and Griffiths 1993), F CO2 is the fractional abundance of 13C in the 13CO2 tank (0.99 in Year 1; 0.9994 in Year 2), and Δ is a photosynthesis fraction factor expressed in 13C fractional abundance terms (0.011332; Fogel and Cifuentes 1993). The moles of C in the tree tissues derived from the 13CO2 tank (n13C) were calculated as:

$$ {n^{13}{\rm C}} = TDC \times M_T $$
(4)

where M T is the total moles of C in the enriched tree tissues (Table 3). M T was calculated for each tree tissue type using their respective δ13C value, carbon content, and dry weight (Table 3). To calculate the 13C labeling efficacy of the tulip poplar C, we first assumed that all the C in the tree tissues derived from the 13CO2 tank was 13C and then applied the following equation:

$$ {\%^{13} C - {\rm {Labeling \ Efficacy}}} = ({{\rm{n}}^{13} {\rm{C}}_T} / {{\rm{n}}^{13} {\rm{CO}}_{2}}) \times 100 $$
(5)

where n13C T is the sum of all the moles of C in the trees derived from the 13CO2 tank (old and new) and n13CO2 is the number of 13C moles injected from 13CO2 tank into the growth chamber (Table 2).

Statistics

Atom % and % 13C labeling efficacy data for Years 1 and 2 tree tissues were analyzed by analysis of variance (ANOVA) using SAS (version 8.1; GLM). Post hoc analyses were performed using Tukey’s Studentized Range test (α = 0.05).

RESULTS

13C enrichment in the tulip poplar trees varied in each year (P < 0.0001) and by tissue age (P < 0.0001) and type (P ≤ 0.0006). Year 2 tree tissues were five times more enriched with 13C than Year 1 tissues; this enrichment corresponded to a 20-fold increase in the number of moles of 13C in the tree tissue derived from the 13CO2 tank and a 25-fold increase in the number of 13CO2 moles injected into the growth chamber during Year 2 (Table 2 and 3). New tissues were more enriched in 13C than old tissues (P ≤ 0.003) and stems were generally the most enriched tissue type (Table 3). In Year 1, 13C comprised 1.26 atom % (±0.07) and 1.47 atom % (±0.13) of the C within the old and new tree tissues, respectively. In Year 2, 13C comprised 5.89 atom % (±0.72) and 7.06 atom % (±0.57) of the C in the old and new tissues, respectively.

Overall, a higher percentage of the injected 13CO2 was incorporated into the trees in Year 1 (59% ±1) than in Year 2 (43% ±0) (P < 0.0001; Figure 1A, B). At the whole tree scale (old + new tissues) each year, a similar percentage of 13CO2 was found in the leaves and stems, which differed from that found in the roots (P ≤ 0.0008). The roots had a smaller percentage of the 13CO2 than the leaves or stems in Year 1 (Figure 1A) and a higher percentage of the 13CO2 in Year 2 (Figure 1B). We observed distinct patterns of preferential plant tissue labeling in both years when considering old and new tissues separately (P < 0.0001). Old root tissues incorporated the greatest percentage of 13C followed by stems and leaves (α = 0.05) whereas new leaf tissues incorporated the most 13C followed by stems and roots (α = 0.05). Percent 13C incorporation parallels the biomass distribution among the different tree tissues (Table 3; Figure 1A, B).

Figure 1
figure 1

Average (±S.D.) percent 13C incorporation into old and new tree tissues for (A) Year 1 and (B) Year 2. Percent 13C incorporation is calculated as the percent of the 13CO2 released into the growth chambers that was incorporated into the tree tissues.

Full size image

Tree C and 13C were equally water-soluble in Years 1 and 2 (Table 4). Overall, a higher percentage of tree C was water-soluble in Year 2 than in Year 1 (Table 4). The percent of 13C that was water-soluble in Year 1 and 2 represents 12% and 13% of the 13CO2 injected into the growth chamber, respectively. Humic substances comprised a similar percent of the water-soluble tree C in Years 1 (24% ±2) and 2 (26% ±2). Likewise, a similar percent of the water-soluble tree C was carbohydrate-C in Years 1 (14% ±0.5) and 2 (13% ±0.2), with monosaccharides and polysaccharides representing 5% (±2) and 8% (±1) of this C (average of both years), respectively (Table 4). The dominant carbohydrates in the tree tissue leachates, expressed as % of DTS C-concentration (average of both years), were glucose (63% ±1), rhamnose (14% ±0), galactose (11% ±1), fructose (9% ±1), and mannose (2% ±1). All of the rhamnose, galactose, and mannose were polysaccharide-derived whereas all of the fructose was monosaccharide. Glucose was found in both the mono-(23% ±8) and polysaccharide fractions (40% ±8).

Table 4 Average (±S.D.) Percent Total C (CT) and 13C Recovery for the Cold Water Extractions of the New Tree Tissues from Years 1 and 2 Tree C Labeling Experiments
Full size table

The 13C-DOC whole-stream tracer addition sufficiently enriched the DOC pool in White Clay Creek. The average background δ13C-DOC value for the stream prior to the injection was −25.8 (±0.4); the tracer addition elevated it up to 337 (±35; average of stations 1 and 2) and only increased the stream DOC concentration by less than 5% (±4), from a background of 1.48 (±0.06) mg C L-1 to a maximum of 1.55 (±0.03) mg C L−1 following the addition of the 13C-DOC. The 13C enrichment of the tree tissue leachate was strong enough for its signal to be easily detected 1265-m downstream from the injection site when the average stream discharge and width were 15.1 L s−1 and 3.3 (±1.4) m, respectively (Figure 2). The correction of these data for background 13C-DOC concentration and stream water dilution, and the ultimate calculation of uptake length, mass transfer coefficient, and total DOC supply to the benthos are topics for a different manuscript.

Figure 2
figure 2

Decline of the δ13C-DOC along a 1265-m reach in White Clay Creek, PA, during a 115 min whole-stream 13C-DOC tracer addition. Average (±S.D.) δ13C-DOC values before and during the whole-stream isotope addition are shown. Note δ13C-DOC values during the stream injection are not corrected for stream water dilution and background 13C-DOC concentration.

Full size image

DISCUSSION

13C Enrichment in Tree Tissues

The 13C enrichment patterns we observed in the tulip poplar tissues were consistent with previous observations of photosynthate allocation and natural abundance signals. Photosynthate generated during plant growth is allocated for growth and development of new tissues (Quinlan 1965), and this should explain the preferential enrichment of newly formed tissues compared to older tissues. Natural abundance studies have shown that stem/trunk tissues are more enriched in 13C than leaf tissues (Leavitt and Long 1982, 1986; Schleser 1992; Terwilliger and Huang 1996; Arndt and Wanek 2002). For example, the trunks of juniper and beech trees were 1.2 to 5‰ heavier than the leaves (Leavitt and Long 1982; Schleser 1992). Placing our data in similar terms, old stems were 136‰ and 1131‰ heavier than old leaves and new stems were 154‰ and 157‰ heavier than new leaves in Years 1 and 2, respectively. Diffusion processes cannot explain this fractionation pattern because diffusion away from the leaves would enrich them in 13C (Leavitt and Long 1982). The variant δ13C signals among major biochemical components of plant tissues, which are a function of biosynthetic fractionation factors, may explain differences in the 13C enrichment among the tree tissues examined here (Abelson and Hoering 1961; Benner and others 1987; Ghashghaie and others 2001; Fernandez and others 2003). Additionally, the role of newly developed phloem cells in the translocation of 13C-enriched photosynthate from the canopy to leaves lower on the tree and roots may cause the trunk tissue to become more enriched in 13C relative to the leaves adjacent to the trunk (Schleser 1992). In our study, stem tissues were comprised primarily of developing trunk.

Allocation of 13C in Trees

During the growing season, leaves incorporate most of the new photosynthate while allocation to the stems and roots varies with tree type (Schier 1970; Webb 1977; Isebrands and Nelson 1983; Smith and Paul 1988; Horwath and others 1994). In our study, the new leaves had the highest percent of 13C followed by the stems and roots, respectively (Figures 1A, B); this allocation pattern parallels the biomass distribution among the different tree tissues (Table 3). We expected this 13C distribution pattern within the tulip poplars given that active regions of growth act as sinks for photosynthate (Hansen 1967).

C turnover in the new tissues resulted in a high percentage of 13C in the old tree tissues. Overall, 20% and 26% of the 13CO2 taken up by the tulip poplars was found in the old tissues for Years 1 and 2, respectively. In addition, the percent 13C distribution in the old tissues was different from the pattern observed for the new tissues (Figures 1A, B); the roots had the greatest percentage of 13C in them followed by the stems and then leaves. This percent 13C incorporation pattern parallels the biomass distribution among the different tree tissues (Table 3) and is similar to C allocation patterns reported for deciduous trees during the fall using 14C (Horwath and others 1994).

13C Recovery

Controlled environment and pulse-chase experiments are commonly used to isotopically label tree C. In controlled environment experiments like ours, an entire tree is exposed to isotopically labeled CO2 while its concentration is monitored and maintained (Webb 1977; Smith and Paul 1988; Horwath and others 1994). In contrast, during a pulse-chase experiment, a segment of a tree (that is, leaf, branch) is exposed to the C label until approximately 99% of it has been taken up (for example see, Balatinecz and others 1966; Dickson and Nelson 1982; Kuhns and Gjerstad 1991; Simard and others 1997; Mikan and others 2000). In studies employing both labeling techniques, young (seedlings to 2-years-old) trees are exposed to CO2 for 30 min to 1 d, and harvested 8 h to 2 weeks after exposure (Balatinecz and others 1966; Dickson and Nelson 1982; Kuhns and Gjerstad 1991; Horwath and others 1994; Simard and others 1997; Mikan and others 2000). High (77%) C recoveries have been reported for controlled environment labeling experiments and are thought to result from complete C fixation under regulated labeling conditions and re-fixation of the labeled aboveground night respiration (Horwath and others 1994). C recoveries for pulse-chase experiments (24 to 75%) are more variable than controlled environment experiments, which may result from incomplete C fixation and high respiratory losses of the labeled C (Horwath and others 1994).

The 13C recoveries observed for our experiments (43% and 59%) are within the range, albeit on the low side, reported for tree C labeling experiments employing both controlled environment and pulse-chase designs (Balatinecz and others 1966; Dickson and Nelson 1982; Kuhns and Gjerstad 1991; Horwath and others 1994; Simard and others 1997; Mikan and others 2000). Discrepancies in the C isotope incorporation between our study and other C labeling experiments may be due to differences in the C isotopes used, experimental designs (as discussed above), tree types, CO2 exposure times, and tree tissue harvest times or may have resulted from leaky chambers and/or unused 13CO2 in the chamber atmosphere at the end of labeling. Photosynthetic and metabolic discrimination against 13C is much less than that against 14C (Van Norman and Brown 1952); therefore, it is unlikely that discrimination differences between 13C and 14C were responsible for our low label recovery. Loss of the 13C to aboveground respiration, the soil atmosphere, and/or root exudation may have contributed to our low recoveries; however, the 13C lost to the aforementioned processes should have been re-fixed by the trees given our controlled environment design. Similarly, the tulip poplars were harvested immediately following labeling to maximize the 13C recovery and minimize the amount of 13C that could have been lost to respiration. To rule out the remaining factors that may have contributed to our low 13C recoveries, we calculated the remaining amount of injected 13CO2 in the chamber after labeling using a two-source mixing model, similar to the one in the calculations section. Parameters used in this calculation were the fractional abundance of 13C in the chamber air at the end of labeling (0.0196758), background CO2 tank (0.011164), and 13CO2 tank (0.99), as well as the number of CO2 moles in the chamber given the CO2 concentration point (Table 2) and chamber volume (5728 L). This calculation could only be done for Year 1 where the δ13C of the chamber air had been measured during labeling. Our calculation suggests that approximately 49% of the 13CO2 injected into the chamber was still there at the end of labeling and that the chambers were relatively leak-free. Assuming the conditions were similar for Year 2, our low 13C recoveries were the result of unused injected 13CO2 in the growth chambers. A more even 13CO2 injection rate over the course of labeling or a more intense injection rate at the beginning could have resulted in higher 13C recoveries in the trees.

In addition, a higher percentage of the injected 13CO2 was incorporated into the trees in Year 1 than in Year 2 (Figures 1A, B). Given that Year 2 trees were exposed to 25 times more 13CO2 for a longer time (23-d more) under the same environmental conditions as Year 1 trees, they should have incorporated at least as much, if not more, 13CO2. The only difference in the growth conditions between Years 1 and 2 is that Year 2 trees were sprayed with Safer’s ® sulfur spray to control powdery mildew which may have lowered the trees’ carbon fixation rate, resulted in a lower 13CO2 recovery, and a higher percentage of 13CO2 in the chamber air at the end of the experiment. However, we cannot verify that the mildew and/or the anti-fungal agent affected the trees’ carbon fixation rate because, as mentioned above, no δ13C measurements for the chamber air were made in Year 2.

13C-DOC Tracer

Tree tissue leaching studies have shown that 6 to 47% percent of leaf C goes into solution during cold-water extractions (for example, Nykvist 1962, 1963; Cummins and others 1972; Petersen and Cummins 1974; Gessner and Schwoerbel 1989; Magill and Aber 2000). In our study, a similar percentage of tree C and 13C went into solution during the cold-water extractions, suggesting that the 13C label was evenly distributed among the water-soluble components of the new tree tissues (Table 4). In addition, our results indicate that most (∼70 and 80%) of the tree C, as well as 13C was allocated into non-soluble, structural components. The water-soluble tree C had both humic and polysaccharide components, confirming that our 13C-DOC tracer, like stream water, possesses polymeric constituents. Though clearly, the structure of the constituents isolated from our 13C-DOC tracer may differ from those found in stream water.

Enriching plant material in 13C is a costly endeavor. CO2 highly enriched in 13C is expensive and large quantities are needed to sufficiently label material for an isotope tracer experiment, placing a premium on maximizing recovery. For a stream injection experiment, the δ13C value of the DOC injectate needs to be sufficiently high to overwhelm the background fluctuations in the δ13C of the stream DOC. In addition, the DOC injectate needs to be adequately labeled so that its addition to the stream does not increase the DOC concentration and can be followed for a considerable distance downstream. It is important that the tracer addition does not increase the stream DOC concentration so that rates of in situ processes can be measured. Preliminary results from the whole-stream 13C-DOC tracer addition demonstrate that the tulip poplar tissues were sufficiently labeled with 13C to meet the criteria for a whole-ecosystem isotope tracer addition (Figure 2).

Results from our study illustrate that large quantities of CO2 highly enriched with 13C are needed to generate a stream DOC tracer for a whole-stream isotope addition. Calculations suggest that the contribution of the 13C-enriched CO2 (99 atom % 13C) to the 13C enrichment of the trees was not affected by the δ13C value of the background CO2; it was the volume (amount) of 13CO2 added relative to the growth chamber volume that influenced the 13C tree tissue enrichment. Thus, the greater the amount of 13CO2 used to label the trees, the greater the 13C enrichment in the tree tissues (Table 2 and 3). Additionally, if we had used CO2 with a δ13C value similar to the background CO2 used in Year 1 (Table 2), a CO2 source more enriched with 13C than the atmosphere and several orders of magnitude less expensive than the 99% atom 13C-CO2 used to label the trees, the 13C signal of our tracer would not have been strong enough for a whole-stream isotope addition.

CONCLUSIONS

Tree C can be successfully labeled with 13CO2 and used to create a DOC tracer with humic and polysaccharide constituents that is more representative of stream/riverine DOC than previously used carbon isotope tracers (Hall 1995; Hall and Meyer 1998). Although our tracer is not a perfect model for stream DOC, we feel that it is a good tool for starting to measure in situ stream DOC dynamics given the fact that DOC leached from leaf litter accounts for 30% of daily DOC export from small forested streams on an annual scale and can comprise 42% of the DOC inputs in the fall (McDowell and Fisher 1976; Meyer and others 1998). Specifically, our tracer will allow for the direct measurement of stream DOC uptake lengths and rates, transformations, turnover rates, consumer-food linkages, interactions with mineral surfaces, and metabolism, and may even provide some insight into why there is so little terrestrial C in the oceans (Bader 1956; Meyers-Schulte and Hedges 1986; Hedges and others 1997; Opsahl and Benner 1997). The utility of our 13C-DOC tracer could potentially be enhanced by aging it in the soil to generate a tracer that is chemically and biologically more similar to stream DOC. Additionally, a tracer like ours can be used to examine soil processes like organic matter degradation, DOC production, humic substance formation, and DOC sorption (that is, Bottner and others 2000; Coûteaux and others 2001). Ultimately, having a carbon stable isotope tracer more representative of natural organic matter will allow ecosystem scientists to examine carbon dynamics across the terrestrial-aquatic gradient.