Published online August 9, 2007 Contribution of Root vs. Leaf Litter to Dissolved Organic Carbon Leaching through Soil Very little is known about dissolved organic matter (DOM) originating from fine roots in forest Shauna M. Uselman* soils in comparison to DOM originating from leaf litter. To compare the fate of root- vs. leaf- Ecology, Evolution, and Conservation derived dissolved organic carbon (DOC), we added 14C-labeled root litter at depths of 10 cm, Biology Program, and Natural Resources and a “shallow root treatment,” and 40 cm, a “deep root treatment,” and leaf litter to 50-cm column Environmental Science Dep. soil microcosms. We analyzed the solution leached from the columns during a 47-d simulated MS 186, Univ. of Nevada snowmelt treatment, trapped respired 14CO2, and measured translocation of 14C within the Reno, NV 89557 columns. In general, total C losses, as a percentage of 14C added, were substantial and high- est for the leaf treatment (8.2% leached as DOC, 13.4% translocated, and 14.8% respired), Robert G. Qualls intermediate for the shallow root treatment (2.3, 5.2, and 3.9%, respectively) and lowest for Natural Resources and Environmental the deep root treatment (2.4, 1.9, and 2.9%, respectively). The C lost to DOC leaching was Science Dep. similar, however, for the deep and shallow root treatments. As a percentage of total losses (i.e., MS 370, Univ. of Nevada the sum of DOC leaching, respiration, and translocation), 14C lost as DOC leaching was Reno, NV 89557 significantly higher in the deep root treatment than other treatments. These observations sug- gest that leaf-derived DOC may contribute to the formation of an A horizon and even to Juliane Lilienfein accumulation of soil organic carbon (SOC) in the B horizon during soil development, either by Natural Resources and Environmental adsorption or microbial biomass incorporation. The 14C data further showed that root-derived Science Dep. DOC, especially from root litter at greater depths, may help explain both the presence of SOC MS 370, Univ. of Nevada at depth and a portion of the DOC draining from soil profiles. Reno, NV 89557 Current address: Abbreviations: DIC, dissolved inorganic carbon; DOC, dissolved organic carbon; DOM, dissolved Synergy Resource Solutions, Inc. organic matter; SOC, soil organic carbon; SUVA, specific ultraviolet absorbance. 5393 Hamm Rd. Belgrade, MT 59714 P roduction of dissolved organic matter (DOM) is important in forest ecosystem development because it affects soil for- mation and transport of metals, it is a source of C for microbes, originating from the forest floor may be lost from the ecosystem (i) in runoff, which could short-circuit soil and move directly into surface waters, (ii) in soil microbial respiration, or (iii) by transport and it contributes to soil organic matter (SOM) accumulation. through the soil profile and loss into groundwater. The DOM Dissolved organic matter may be particularly important to the moving through soil may also be adsorbed onto soil surfaces, after development of a deepening distribution of SOM during eco- which it may be subject to desorption or microbial utilization, or it system development (Sollins et al., 1983). Furthermore, eco- may remain adsorbed, contributing to SOM. If desorbed, DOM systems lose C, N, and P through leaching of DOM, which may be further translocated through the soil profile, where it then FOREST, RANGE & WILDLAND SOILS affects overall ecosystem C, N, and P budgets (Qualls, 2000). has the potential to contribute to deeper SOM. Although fine Despite the importance of DOM in ecosystem nutrient roots are concentrated in surface soils, their vertical distribution in cycling, very little is known about DOM that originates from fine soil typically shows an exponential decrease with depth (Jackson et roots and how this DOM differs from DOM that originates from al., 1996); thus they are present in lower density at greater depths leaf litter. In fact, no study has addressed the fate of DOM origi- in soil. As such, deeper fine roots are a potential source of deeper nating from root litter. In this study, we focused on determining DOM. The DOM produced at greater depths may be more prone the fate of DOM produced by root and leaf litter. The DOM to being lost from the ecosystem in leaching because it would have circumvented a substantial volume of soil. In this study, we compared the fate of DOC originating Soil Sci. Soc. Am. J. 71:1555–1563 from root litter and leaf litter. Our specific objectives were (i) doi:10.2136/sssaj2006.0386 to demonstrate that DOC from root litter can be translocated Received 10 Nov. 2006. through soil, and be potentially lost from the system, and (ii) *Corresponding author (uselman@unr.nevada.edu). to compare the percentages of the C originating from root lit- © Soil Science Society of America ter and leaf litter that are respired, translocated as DOC, or 677 S. Segoe Rd. Madison WI 53711 USA All rights reserved. No part of this periodical may be reproduced or lost as DOC in soil leachate. To do so, we added 14C-labeled transmitted in any form or by any means, electronic or mechanical, root and leaf litter to 50-cm soil columns, we analyzed solution including photocopying, recording, or any information storage and leached from the columns during simulated spring snowmelt, retrieval system, without permission in writing from the publisher. we trapped respired 14CO2, and we extracted soil columns in Permission for printing and for reprinting the material contained layers for translocated 14C material. When we use the term herein has been obtained by the publisher. translocated within the soil, we are referring to DOC that was SSSAJ: Volume 71: Number 5 • September–October 2007 1555  solubilized from litter, moved some distance through soil, but The mudflows are all derived from andesitic bedrock and volcanic remained in the soil. In contrast, we use the term deep leach- ash from glacial outwash, originating from Mt. Shasta. Topography ing of DOC to refer to DOC that was solubilized from litter, and climate are uniform across the chronosequence. The flows were moved some distance through soil, and was then leached from first covered by a ponderosa pine forest, which later developed into the soil columns at a depth of 50 cm. a mixed-conifer forest during primary succession (Dickson and One of our specific objectives was to demonstrate the Crocker, 1953; Uselman et al., 2007). The 77-yr-old ecosystem site leaching of root-derived DOC, illustrating its importance can be described as a ponderosa pine forest, with soil that is a well- in soil formation processes. Therefore, we designed the drained loamy sand classified as a Typic Haploxerept, or a borderline experiment to simulate the water flux and soil temperature Vitrandic Haploxerept. The A horizon is approximately 10 cm in that occur under natural conditions during snowmelt. We depth. Climate can be generalized as winter precipitation– summer conducted the study using soil from the Mt. Shasta eco- drought temperate. Precipitation averages 1300 mm annually, mainly system chronosequence, composed of mudflows ranging in as snow between November and March. The average annual air tem- age from 77 to >850 yr old in 2001 (Dickson and Crocker, perature is 9.9°C, varying between 1.4°C in January and 20°C in July 1953; Lilienfein et al., 2003; Uselman et al., 2007). We used (Lilienfein et al., 2003). soil from the youngest ecosystem, the A flow, because it had The depth distribution of fine root biomass in the 77-yr-old the highest concentration of DOC leached from the root- ecosystem was 3% (forest floor), 37% (0–10 cm), 33% (10–20 cm), ing zone in the field (Lilienfein et al., 2004). Also, fertility 8% (20–30 cm), 14% (30–40 cm), 2% (40–50 cm), 1% (50–60 was lowest in this ecosystem, so loss of nutrients through cm), 1% (60–70 cm), and <1% (70–80 cm), measured as a per- leaching was believed to be most critical. We chose a depth centage of the roots encountered from 0 to 80 cm (Uselman et al., of 50 cm for the soil columns because 98% of the fine root 2007). Thus, 74% of the fine roots were in the top half and 20% biomass existed between 0 and 50 cm in the mineral soil were in the bottom half of the 0- to 50-cm depth interval. Fine root of the youngest ecosystem (Uselman et al., 2007). In this standing stock biomass was not measured in the 77-yr-old ecosys- study, we used three litter addition treatments: (i) leaf litter tem, but in the adjacent >850-yr-old ecosystem, the average stand- added to the surface of the soil, (ii) root litter added at 10 ing stock of live fine roots was 426 g m−2 in the A horizon (0–20 cm, where fine root density is greatest in the field (Uselman cm) and 131 g m−2 in the B horizon (20–30 cm). Fine root turnover et al., 2007), and (iii) root litter added at 40 cm. Although rate was 0.62 yr−1, based on measurements of sequential intact cores fine root density is diminished at 40 cm, fine roots do exist (Uselman et al., 2007). at this depth (Uselman et al., 2007), and this treatment Each of 10 50-cm soil microcosm columns were constructed of provides an important experimental comparison, as DOC a 60-cm length of clear Plexiglas pipe of 3.175-cm (1¼-inch) inside originating from this litter would be more likely to escape diameter, with sealed couplings fitted to the bottom and top. The A the soil columns in leachate solution. flow soil, sieved free of roots with a 2-mm sieve and homogenized by 0- to10-, 10- to 20-, and 20- to 50-cm depth increments, was gradu- MATERIALS AND METHODS ally packed into each column by depth increment to achieve a bulk Preparation of Carbon-14-Labeled Litter density as measured in the field (Lilienfein et al., 2003). At the bottom To produce 14C-labeled plant litter, we grew tree seedlings of of the column, a wicking system was used to maintain unsaturated Fremont cottonwood (Populus fremontii S. Watson) and California flow (Gee et al., 2002). The tension exerted (2.5 kPa) by the wetted black oak (Quercus kelloggii Newb.) in environmentally con- nylon wick, which was enclosed in Tygon tubing, was proportional trolled, sealed Plexiglas chambers for one and a half growing sea- to its length (25 cm), similar to a hanging water column. To achieve sons. Although ponderosa pine (Pinus ponderosa C. Lawson) is a full contact between the wick and soil, the wick was splayed radially, dominant species on the site, it was unsuitable for uniform labeling below a thin layer (∼1 g) of diatomaceous earth, which was placed in the column before adding the soil. The diatomaceous earth was also (2005). Briefly, the growth chambers were maintained at 370 μL L−1 because leaf longevity is >1 yr. For details, see Qualls and Bridgham used to prevent movement of soil out of the column. To collect 14CO2 CO2 and labeled with 14CO2 twice a week by injections of liquid from respiration, an air vent with a soda lime trap was added below NaH14CO3 through a septum into a mason jar filled with H3PO4 the soil column because a connection to the atmosphere is needed located inside each chamber. After senescence, leaf litter and fine for the wicking system to provide tension necessary for drainage. The leachate was collected into a sealed polypropylene bottle. A second in the experiment was ≤3 mm in diameter, with most (92%) being root material were collected and air dried. Fine root material used removable coupling was sealed to the top of the Plexiglas pipe, with a ≤2-mm diameter, and only fine roots that had grown during the small port for the addition of artificial snowmelt solution (see below). labeling period were used. Leaf litter was close to being uniformly To collect 14CO2 from respiration, one or two glass vials containing 1 labeled with 14C (Qualls and Bridgham, 2005). mL each of 1 M NaOH solution were placed in the headspace. Thus 14C from respiration was collected from two points along the column Construction of Soil Columns system. Soda lime and NaOH traps were replaced periodically well Soil was collected in the depth increments 0 to 10, 10 to 20, and before saturation occurred. 20 to 50 cm from the youngest 77-yr-old ecosystem, the A flow, of the Mt. Shasta Mudflows Research Natural Area, in the Shasta–Trinity Experimental Treatments National Forest, about 6 km northeast of McCloud, CA. The 77- There were three treatments with three replicate columns for yr-old ecosystem is part of an ecosystem chronosequence, composed each treatment (n = 3): (i) a leaf treatment with leaf material placed of mudflows of 77, 255, 616, and >850 yr old in 2001 (see Dickson on the surface of the soil, (ii) a shallow root treatment with root mate- and Crocker, 1953; Lilienfein et al., 2003; Uselman et al., 2007). rial placed at a depth of 10 cm within the soil, and (iii) a deep root 1556 SSSAJ: Volume 71: Number 5 • September–October 2007 treatment with root material placed at a depth of 40 cm within the snowmelt treatment was continued for 47 d, after which the col- soil. There was also one soil column with no litter added, which was umns were harvested. used to correct for background levels of 14C. The air-dried litter was cut into approximately 5- by 5-mm Analysis for Carbon-14 squares (leaf ) or 1-cm lengths (roots), and then added as uniform Collected leachate was analyzed for 14C activity using liquid layers to soil columns. Because of the small amount of total fine scintillation both before and after acidification. Acidification and root material, we used 0.25 g of Q. kelloggii roots and 0.15 g of P. purging was used to separate dissolved inorganic C (DIC) from fremontii roots, for a total of 0.4 g of fine root material for each root DOC. Thus, before acidification, leachate included total dissolved treatment column. Thus, the fine roots used in the 10- and 40-cm C (i.e., DIC + DOC); after acidification and purging, leachate treatments were the same material and the same mass, so that these included only DOC. To measure the 14C activity of the leachate treatments could be quantitatively compared. The total amount of solution, 5 mL was added to 15 mL of Ecolite scintillation fluid 0.4 g of fine root material added to each column corresponded to approximately 500 g m−2, which was similar to the standing stock (ICN Radiochemicals Corp., Irvine, CA), which was counted in a Beckman LS60001C liquid scintillation counter (Beckman Corp., of live fine roots measured in the >850-yr-old ecosystem of the Mt. Fullerton, CA) for 10 min. Shasta site (Uselman et al., 2007). We used 0.6 g of each species’ leaf Particulate 14C leached from the column was measured by litter, for a total of 1.2 g for each leaf treatment column, which cor- Pall Corp., Ann Arbor, MI), dissolving the filter in 100 μL of 2 filtering the leachate through a 0.45-μm membrane filter (GN-6, responded to approximately 3 yr of leaf litter accumulation on the forest floor (500 g m−2 yr−1 × 3). M HCl, adding 10 mL of scintillation fluid, and analyzing on the liquid scintillation counter. Because there was so little particu- Simulated Snowmelt late matter (and 14C), we filtered all three replicates from each An artificial snowmelt solution was made to approximate the pH treatment through one filter (per day) to get enough particulate and ionic strength of natural snowmelt based on snowmelt chemistry matter for analysis. measured under the forest canopy in Little Valley, near Reno, NV, in Trapped 14CO2 in the NaOH solution was analyzed for 14C the eastern Sierra Nevada Mountains by Johnson et al. (1997, 2001). activity, as in Qualls and Bridgham (2005). For the soda lime traps, To simulate spring snowmelt conditions, it was added to the micro- we composited the samples collected during the experiment into cosms at a rate of 19 mm d−1 using peristaltic pumps (with silicone one sample per column. The 14CO2 absorbed in the soda lime was tubing), and the microcosms were kept at 2 to 4°C by housing the removed by addition of excess HCl acid through a septum in a sealed experiment inside a dark refrigerator. The temperature was based on jar and retrapped into NaOH solution, which was then analyzed for field measurements (Uselman et al., 2007) during snowmelt. To facili- 14C as described above. tate even addition of the snowmelt solution to the soil, a glass fiber After 47 d of simulated snowmelt, the soil columns were sliced filter was placed on the surface of the soil (or leaves). into 2- or 10-cm depth increments, depending on the location of the To estimate maximum water flux due to snowmelt in March litter, and analyzed for translocated 14C content, as described in Qualls (maximum flux during 2001–2002, data not shown) at the Mt. and Bridgham (2005). Briefly, 200- to 800-mg samples weighed into Shasta site, we calculated winter precipitation in January through tin capsules were combusted in the combustion tube of a Shimadzu March as an upper estimate, and February through March as a lower 5050 TOC Analyzer (Shimadzu Corp., Columbia, MD) at 670°C, estimate, averaged for the period of 1991 to 2000. These estimates the effluent gas was trapped in two to four tubes connected in series, are based on the assumption that all precipitation falling during this each containing 10 mL of 1 M NaOH, and the 14C activity was ana- period will remain on the ground as snow, with minimal melting lyzed using liquid scintillation. until air temperatures warm in March. This estimate yielded a flux The 14C activity of the residual soil pore water was measured by of 424 to 705 mm per monthly period, so we used the average, extracting freshly collected moist soil (equivalent 5 g dry weight) with equivalent to 19 mm d−1. 10 mL of deionized water, centrifuging, filtering the supernatant, and This application rate resulted in unsaturated flow in this soil analyzing 5 mL of the filtrate with 15 mL of scintillation fluid for 14C type. Soil water retention curves and measurements of soil moisture activity using liquid scintillation. By using a 2:1 ratio of soil to water, for each depth increment showed that soils were 55 ± 1.2% of satu- we may have caused some minor desorption of DOC, thereby slightly ration during the experiment. The volumetric soil moisture content underestimating adsorbed DOC, but because hysteresis occurs during in the columns was 21 ± 0.4%. Field measurements of volumetric adsorption and desorption (Kaiser and Zech, 1999), we assumed this soil moisture content (0–30 cm) in the youngest ecosystem during error to be negligible. Soil solution remaining in the wick or glass- snowmelt in March 2002 were 15% on average, ranging from 12 to fiber filter was also extracted in deionized water, and analyzed for 14C 18% (Uselman et al., 2007). It should be noted that the absolute activity. All measurements of soil solution 14C activity were added to volumetric soil water content measured in the field was probably the total DOC or DIC leached from each column. lowered by the presence of rocks, which occupied an approximate The 14C content of the original leaf and root material was mea- volume of 8% within the 0- to 50-cm depth increment (Lilienfein sured by combusting three 30- to 40-mg replicate samples in the TOC et al., 2003). The resulting average residence time was 5.8 ± 0.2 d analyzer combustion tube as described above, with four tubes of NaOH, for snowmelt solution added to the columns. which resulted in a specific activity of 3.27 × 1011 Bq kg−1 dry weight The columns were first wetted with 100 mL deionized water for the fine roots and 3.74 × 1011 Bq kg−1 dry weight for the leaves. during a 4-d period, after which the application of the snowmelt All measurements were corrected for background radiation (in solution was started. The initial volume of water needed to wet units of becquerels) by subtracting the soil control column samples, the columns was determined on an additional test column dur- which were run at the same time as the treatment samples on the ing a preliminary wetting and equilibration test. The simulated liquid scintillation counter. SSSAJ: Volume 71: Number 5 • September–October 2007 1557  Analysis for Total Carbon column during the experimental period as well as residual DIC in pore Total C refers to all isotopic forms of C. Collected leachate was water remaining at the end of the experiment. The total CO2 in the soda analyzed for DOC concentration using a Shimadzu 5050 TOC ana- lime was distributed across the experimental time period in proportion to the CO2 trapped in the NaOH in the headspace. Adsorbed 14C was fication with HCl to pH ≤2. Leachate samples were also analyzed for lyzer (Shimadzu Corp.). Dissolved inorganic C was purged after acidi- calculated as total translocated 14C minus soil solution 14C, excluding the ultraviolet absorbance at 360 nm using a Shimadzu UV-1201 spec- 2-cm layer of soil in which root litter was originally placed. Cumulative 14C lost to DOC leaching, respiration, translocation, or particulates were trophotometer (Shimadzu Corp.), and specific ultraviolet absorbance at 360 nm (SUVA360) was calculated as absorbance divided by DOC calculated as percentages of the 14C added to the column (in 14C-labeled concentration (in milligrams per liter) and then multiplied by 100. The leaf or root plant material). Losses of 14C to DOC leaching, respiration, absorption spectrum of humic substances is broad, and wavelengths translocation, or particulates were also calculated as relative percentages of ranging from 254 to 400 nm have been used in studies of aquatic humic total 14C lost from the source in each treatment. substances (Thurman, 1985). The SUVA increases with increasing aro- For total DOC (all isotopic forms), cumulative C lost to DOC leach- maticity of organic compounds, and thus may be used as an index of ing was calculated as a percentage of the plant C added to the column. humic substance concentration (Thurman, 1985). The leaf material was assumed to be 50% C, and the fine root material Measurements of DOC leached from the treatment columns were was assumed to be 48% C, based on a database showing 48 ± 0.78% C corrected by subtracting DOC leached from the soil control column. content of fine roots (≤2 mm) compiled by Gordon and Jackson (2000). Thus net total DOC leached refers to DOC originating from the lit- ter, having subtracted DOC originating from SOC. Measurements of Statistical Analysis total DOC indicated the presence of additional DOC originating from Separate one-way ANOVAs and Tukey’s honestly significant dif- the wick material. Therefore, we performed the leaching experiment ference (HSD) tests were used to test for significant differences among again on a new control column with no soil for 30 d under identical treatments in the percentage of 14C lost from the source as (i) DOC in experimental conditions, until the values approached an asymptote. For deep leaching, (ii) CO2 in respiration, (iii) C translocated in soil, and (iv) SUVA360, we removed the DOC originating from the wick by subtract- total C lost. Similarly, separate one-way ANOVAs and Tukey’s HSD tests ing the wick DOC concentration from the treatment or soil control col- were used to test for significant differences among treatments in relative umn concentration. For total DOC, the effect of the wick was already percentage of 14C lost from the source as (i) DOC in deep leaching, (ii) removed by subtracting DOC from the soil control column. CO2 in respiration, and (iii) C translocated in the soil. For total DOC (all isotopic forms), we tested the percentage of C lost as DOC deep leaching Calculations during 47 d using a one-way ANOVA and Tukey’s HSD test. Cumulative dissolved organic 14C leached was calculated from To assure a normal distribution (Zar, 1996), proportion data (per- total DOC leached out of the column during the experimental period centages as proportions) were transormed by the arcsine of the square plus the residual DOC in pore water remaining at the end of the experi- root of the proportion; then all data met the assumptions of normal- ment. Cumulative 14CO2 from respiration was calculated from total ity (Shapiro–Wilk test) and equal variances (Levene test). Data were CO2 trapped in the NaOH solution in the headspace and in the soda analyzed using SPSS (SPSS, 2003). lime during the experimental period plus the DIC leached out of the RESULTS Carbon Losses from the Carbon-14- Labeled Litter Total losses of 14C to leaching, translocation, res- piration, and particulates differed significantly by treat- ment, with the leaf treatment highest (36.6%), shallow root treatment intermediate (11.5%), and deep root treat- ment lowest (7.3%) (one-way ANOVA, P < 0.001, Fig. 1). In general, the total C losses were highest for the leaf treat- ment (8.2% as DOC leach- ing, 13.4% as translocation, Fig. 1. Total losses of 14C to deep leaching of dissolved organic C (DOC), translocation within the soil and 14.8% as respiration), (but remaining in the soil), respiration, and particulates, as a percentage of the 14C added to the intermediate for the shallow columns in 14C-labeled leaf and root litter (mean ± SE). Separate one-way ANOVAs (P < 0.001, all) root treatment (2.3, 5.2, and on each fraction of total losses indicate significant differences among the treatments. Means with different lowercase letters are significantly different (Tukey’s HSD tests, P < 0.05). In addition, one- 3.9%, respectively), and low- way ANOVA (P < 0.001) and Tukey’s HSD tests on overall total loss indicate significant differences est for the deep root treatment among all treatments. (2.4, 1.9, and 2.9%, respec- 1558 SSSAJ: Volume 71: Number 5 • September–October 2007 tively) (separate one-way ANOVAs, P < 0.001 for all). However, the ferences among treatments: leaves (a), shallow roots (b), and percentage of C lost to DOC leaching was not significantly different deep roots (b). The relative differences among treatments and the shapes of the curves were similar to 14C-labeled DOC. ulates was ≤0.18% for all treatments. among the two root treatments. The percentage of C lost as partic- During the simulated snowmelt, the SUVA360 of the The 14C losses were also analyzed as relative percentages, DOC leached from the columns showed similar trends for i.e., as a percentage of total 14C loss in each treatment (100%; the root treatments and the soil control (Fig. 5). In contrast, data from Fig. 1). Separate one-way ANOVAs indicated sig- the leaf treatment had a much lower SUVA360 up to Day nificant differences among treatments for DOC leaching (P 18, compared with the other treatments and the soil control. = 0.01) and translocation (P = 0.007), but not for respiration After Day 18, the SUVA360 of the leaf treatment converged (P = 0.14). The relative percentage of 14C lost as DOC leach- with the other columns. ing was significantly higher for the deep root treatment (33%) than for the other treatments (21% for the shallow root treat- DISCUSSION ment and 22% for the leaf treatment). The relative percent- Using a 14C tracer, we showed that fine root litter contrib- age of 14C lost to translocation was significantly highest in the utes to DOC leaching as well as retention of C in soil, as has shallow root treatment (45%), significantly lowest in the deep been shown for DOC originating from leaf litter. Presumably, root treatment (26%), and intermediate in the leaf treatment DOC that is translocated within the soil is adsorbed to soil (37%). The relative percentages of 14C lost to respiration, particles or is incorporated into microbial biomass. Leaching which did not differ significantly among treatments, were 34% of DOM produced by leaf litter has been extensively studied, in the shallow root treatment, 40% in the deep root treatment, as documented in two reviews (Kalbitz et al., 2000; Michalzik and 41% in the leaf treatment. The relative percentages of 14C et al., 2001). In contrast, very little is known about DOM lost to particulates were <1%. Leaching Losses of Dissolved Organic Carbon-14 and Carbon-14 Dioxide during Simulated Snowmelt By the end of the simulated snowmelt treat- ment, cumulative losses of DOC and CO2 from the root treatments approached asymptotic val- ues (Fig. 2a and b). For the leaf treatment, the rates of increase in losses were much lower in the latter portion of the experiment than the early portion. In addition, there appeared to be an initial lag in the DOC leached from the col- umns compared with the root treatments. Distributions of Carbon-14 Remaining in the Soil The distribution of 14C that was lost from the leaves, translocated through the soil column, and remained in the soil, declined exponentially with depth (Fig. 3a). About half of the 14C was adsorbed or otherwise immobilized below the A horizon, i.e., below 10 cm. In contrast to the leaf treatment, the distribution in the shallow root treatment was fairly uniform below the “root layer” (Fig. 3b). The distribution in the deep root treatment was close to uniform, but appeared to decrease somewhat below the “root layer” (Fig. 3c). Total Dissolved Organic Carbon Loss from Litter After subtracting DOC losses (all isotopic forms) from the soil control with no litter added, we found that the DOC losses from litter were 4.7% for the leaf treatment, 1.8% for the shallow Fig. 2. Losses of 14C to (a) dissolved organic C leached from the columns and (b) root treatment, and 1.9% for the deep root treat- CO2 respired, during 47 d of simulated snowmelt, as a percentage of the 14C ment (Fig. 4). One-way ANOVA (P = 0.006) added to the columns in 14C-labeled leaf and root litter (mean ± SE). The first and Tukey’s HSD tests indicated significant dif- day of respiration includes respiration during the 4-d wetting period. SSSAJ: Volume 71: Number 5 • September–October 2007 1559  likely to be lost as DOC deep leaching in comparison to 14C originating from either shallow roots or leaf litter is presum- ably because this DOC produced at 40-cm depth bypassed a significant volume of soil, where it could have been adsorbed, respired, or incorporated into microbial biomass. The com- parison of the fate of DOC originating from shallow vs. deep roots demonstrates the importance of the location of origin: while a smaller fraction of the DOC originating from shal- low roots was lost to DOC deep leaching, a larger fraction was translocated and thus remained in the soil. In the field, the B horizon soils of the Mt. Shasta ecosystems are deeper than 50 cm and have a stronger adsorption capacity for DOC than the developing A horizon (Lilienfein et al., 2004). In addition, field measurements of soil solution DOC concen- trations showed that most of the reduction in concentration, from 58 to 20 mg/L, occurred between the forest floor and 40 cm (Lilienfein et al., 2004; unpublished data, 2006). Between 40 and 150 cm, however, there was an additional reduction in concentration from 20 to 8 mg/L, indicating that some of the DOC leached from deep roots at 50 cm may have been adsorbed to soil at depths >50 cm. By the end of the experi- ment, DOC concentrations leaching from the columns (data not shown) were comparable to concentrations of soil solu- tion at 150 cm from field measurements during snowmelt in the A flow (Lilienfein et al., 2004). As noted above, the fine roots used in both the shallow and deep root treatments were the same material and the same mass, so that they could be quantitatively compared. As described above, however, the distribution of fine root biomass declined with depth at the Mt. Shasta site (Uselman et al., 2007), as is typical of most ecosystems (Jackson et al., 1996). Thus the absolute contributions of DOC leached and translocated would be lower from roots in deeper soil layers, proportional to the fine root biomass found under field conditions and the loss of dissolved organic 14C as a percentage of total 14C-labeled root material. In addition, it is possible that the qualitative composi- tion of deeper roots may differ from that of shallower roots. The large percentage of 14C from leaf litter that was trans- Fig. 3. Distributions of 14C lost from the added litter that was located and remained in the soil (13.4%, Fig. 1), as well as the translocated through the soil and remained in the soil (at the pattern of exponential decline with depth, was evidence for its end of the experiment), shown as a percentage of the 14C role in the formation of an A horizon. The fact that about half added to the columns per centimeter of soil column (mean of that 14C was deposited below the A horizon was supporting ± SE). For depth, data is plotted at the midpoint of the depth increment sampled. The spikes in 14C activity in the root evidence that DOC originating from leaf litter can also con- treatments are the layers of soil that included the root litter. tribute to the accumulation of SOC in the B horizon during soil development. The substantial amount of translocated 14C originating from root litter, and we know of no studies, other that remained in the developing A horizon did so despite the than a companion study (Uselman, 2006), that have measured higher native organic C content and lower adsorption capacity production of DOM from root litter. Studies have shown the of this soil relative to the B horizon soil (Lilienfein et al., 2004). importance of DOC originating from live roots as root exu- Adsorption experiments done on this same soil demonstrated dates in feeding rhizosphere microbial activity and promoting a null-point DOC concentration of about 40 mg/L at 0 to higher rates of nutrient cycling (e.g., Swinnen et al., 1995; 10 cm, meaning that soil solution DOC concentrations >40 Verburg et al., 1998). Here we show that DOC originating mg/L would result in net adsorption. Although it was not pos- from fine root litter potentially contributes to both ecosystem sible to monitor concentration of DOC entering the A hori- C loss and SOC accumulation at depth. zon, our data suggest that concentrations were >40 mg/L and The considerable amount of deep leaching of dissolved that net adsorption did occur. It may also be that the observed organic 14C that we found in this experiment (2.3–2.4% from distribution of 14C is partly due to incorporation into micro- fine roots and 8.2% from leaves, Fig. 1) represents C that may bial biomass, which would be concentrated near the source of have the potential to be lost from ecosystems. Furthermore, the DOC, in this case the leaf litter at the surface of the soil. our finding that 14C originating from deep roots was more Finally, it should be noted that the soil used in this study has 1560 SSSAJ: Volume 71: Number 5 • September–October 2007 andic properties, and therefore it may have higher adsorption respired from DOC vs. the litter directly. In addition, some of capacity relative to nonandic soils of similar age. the 14C that was translocated in the soil could have been com- Additional evidence for microbial activity comes from the posed of particles of organic C that may have moved into the SUVA360 and respiration measurements. The DOC originat- soil. The percentage of 14C that was leached from the columns ing from leaf litter appears to have been less aromatic than the as particulate organic C was very low (0.05–0.18%, Fig. 1) in DOC originating from fine roots during the first portion (∼18 comparison to DOC, however, so we believe that most of the d) of the experiment (Fig. 5). This difference is probably due 14C that we measured as “translocated in the soil” must have to the higher proportion of monomeric sugars and hydrophilic originated from the DOC. In a companion study, we found that neutrals in DOC produced by leaf litter compared with that the percentage of total C in litter that was extractable as water- of root litter (Uselman, 2006). These compounds have been soluble organic C was about 6 to 15% for fine roots (recently shown to be the most biodegradable fraction of DOC (Qualls senesced and live fine roots, respectively) and about 20 to 30% and Haines, 1992; Jandl and Sollins, 1997; Qualls, 2005). The for recently senesced leaf litter (Uselman, 2006), which is simi- percentage of 14C lost as respiration was also highest in the leaf lar to the numbers observed in this column study. treatment, indicating the higher microbial activity relative to Carbon lost as DOC from fine root litter can also be the other treatments (Fig. 1). compared with DOC lost as exudation from live roots. In a Our data suggest that, in addition to DOC originating review article, Jones et al. (2004) stated that “it is likely that from leaf litter, DOC originating from fine roots can con- a true estimate of root exudation [is] 2–4% of net fixed C,” tribute to accumulation of SOC at depth. It should be stated, and other studies have found a similar percentage of total however, that many other sources may contribute to SOC net primary production in trees (Smith, 1976; Uselman et accumulation at depth, such as insoluble organic matter from al., 2000). Using this estimate, we calculated that DOC roots, refractory microbial biomass, and live root exudates. production from root litter is comparable in magnitude to Furthermore, the mass of refractory C originating from insol- DOC production from live root exudation (Uselman, 2006). uble sources, such as lignin, should be greater in magnitude Uselman et al. (2000) found that a relatively large fraction than the mass originating from the soluble fraction. Although (∼60%) of DOC originating from root exudation was rapidly DOC originating from leaf litter is translocated deeply in the respired, and thus very labile. In comparison, we found that columns (Fig. 3a), the contribution of DOC originating from only approximately 30% of the DOC produced by recently fine roots relative to the contribution from leaf litter increases senesced fine roots was composed of the most biodegradable with depth (Fig. 3a–3c). In other words, with increasing depth, fractions of DOC (Uselman, 2006). Because DOC from root the relative contribution of DOC originating from fine roots exudation appears to be more labile than DOC produced by incorporated into SOC is likely to increase, especially com- freshly senesced root litter, we suggest that DOC produced bined with the potentially lower biodegradability. The relative by root litter is more important in SOC accumulation. contribution of root-derived DOC would vary across different Using DOC lost to leaching and translocation as a conser- ecosystems based on the ratio of below- to aboveground litter vative estimate of DOC produced by fine root and leaf litter, production. For example, at the Mt. Shasta site, where the ratio of below- to aboveg- round litter production increased with eco- system age, we would expect an increasing relative contribution of fine roots to DOC that is adsorbed or incorporated into SOC at depth in soils. The soil on the youngest mudflow in the chronosequence had the least adsorption capacity for DOC. Soil fertility was also lowest in this youngest ecosystem so the losses due to leaching are most critical to the recycling of nutrients in the ecosystem. Had we used the oldest soil, having a higher adsorption capacity, we would expect to find a greater percent- age of 14C adsorbed to the soil. By adding 14C lost to DOC leaching and translocation, we conservatively esti- mate that 4.3 to 7.5% of the 14C that was added as fine roots and 21.6% of the 14C added as leaf litter was solubilized as DOC Fig. 4. Net total dissolved organic C leached during 47 d of simulated snowmelt, shown during the experiment. These values are as a percentage of the C added to the columns as leaf and root litter (mean ± SE). 14 likely to be underestimates because CO2 “Net” represents DOC that originated from the litter, because the DOC leaching from the soil (soil control with no litter added) was subtracted from the DOC losses may have originated from respiration of from the treatment columns. “Total” refers to all isotopic forms of C. One-way DOC during the experiment; however, ANOVA (P = 0.006) and Tukey’s HSD tests indicated significant differences between we could not discriminate between 14CO2 leaves and roots, but shallow and deep roots were similar. SSSAJ: Volume 71: Number 5 • September–October 2007 1561  we found that a greater fraction of the 14C-labeled plant litter was decomposed faster than fine roots (Fig. 1 and 2). In comparison, lost as DOC than as CO2 during the initial stages of decomposition. the rate of increase in cumulative DOC and CO2 lost from leaves The conditions present during spring snowmelt, with high water slowed by the end of the experiment, but continued to increase at flux and low temperature, may maximize the ratio of leaching of a rate that was noticeably higher than that of roots. In general, DOC to respiration. These conditions are typical in many temper- litter and older organic matter may continue to leach substantial ate and boreal forest ecosystems after autumn leaf fall. Cleveland et amounts of DOC for long periods of time, as suggested by the al. (2006) also found that a large percentage of the C lost during leaf studies of Hagedorn et al. (2004) and Fröberg et al. (2007), in litter decomposition was DOC relative to CO2 in a tropical ecosys- which 13C-depleted litter from a CO2 enrichment study was used tem under conditions of high water flux, indicating that hydrologic to distinguish newer from older organic matter. Although in our conditions are more important in determining DOC leaching than study leaves decomposed faster than roots, it is difficult to general- temperature. Still, given the strong effect of temperature on micro- ize about the difference in rates of decomposition between fine bial respiration in soils (Kirschbaum, 1995), we would expect that a roots and leaves given the relatively short duration of the study higher temperature in our study would have favored the production compared with decomposition studies conducted in the field using of CO2 rather than DOC during initial decomposition. In later litterbags, which are usually at least 1 yr in duration. As Ostertag stages of decomposition, we might expect the ratio of DOC to CO2 and Hobbie (1999) pointed out, there are few studies that have to shift in favor of CO2 as microbes slowly degrade the insoluble examined decomposition of fine roots and leaf litter at the same fraction of litter, as suggested by the trends in the latter days of this time, and the results of these studies were inconsistent. 47-d simulated snowmelt experiment. Finally, a large percentage of Although our measurements of total DOC (all isotopic forms) the soluble fraction was refractory, because even after 1 yr, about leaching from the columns during the experiment show the same 45% of the soluble organic C extracted from radio-labeled leaf lit- relative results as 14C-labeled DOC (i.e., that the percentage lost ter and added to cores of the same A flow soil remained in the soil from leaves is greater than from fine roots, with similarly shaped when incubated at 25°C (Qualls and Bridgham, 2005), suggesting curves, Fig. 4), these measurements also show the decreased sensitiv- that some DOC has the potential to contribute to long-term SOC ity achieved without the use of the 14C tracer. Without the label, it accumulation. is necessary to subtract the amount of DOC that originates from During this experiment representing the initial stages of native SOC, which introduces greater uncertainty in the results, decomposition, large percentages of the total 14C added to the col- hence the lower C loss percentages as total DOC leaching compared umns as fine roots (7.3–11.5%) and leaves (36.6%) were lost (Fig. with 14C-labeled DOC (Fig. 4 vs. 2a). In addition, without the use 1). As a comparison, we calculated that there would be approx- of a tracer it is not possible to separate respiration from root litter value (exponential decay coefficient, in units of yr−1) for broadleaf imately 6 to 8% mass loss during the period of our study using a k vs. respiration of native SOC, nor is it possible to separate DOC adsorbed or incorporated into SOC that originated from the added fine roots (≤2 mm) from fine root litterbag studies in a global data litter from native SOC without the use of labeling (e.g., Cheng and set (Silver and Miya, 2001). In our study, we found that leaf litter Johnson, 1998). In addition, the method we used in this study could be used to examine the effect of differing rates of hydrologic flux on the partitioning of C loss as CO2 vs. DOC. CONCLUSIONS In this experiment, we found that substantial per- centages of the total 14C added to the columns as fine root and leaf litter were lost during the first 47 d of decomposition. In general, losses from leaf litter were greater than losses from root litter. Considerable loss of 14C was attributable to DOC translocation in soil and deep leaching, both from root and leaf litter. The 14C lost as DOC leaching as a percentage of total losses Fig. 5. The SUVA360 (ultraviolet absorbance at 360 nm divided by DOC concentration and then mul- was significantly higher in tiplied by 100) during 47 d of simulated snowmelt (mean ± SE). The SUVA360 increases with aro- maticity of DOC, and is thus used as an index of humic substance concentration. Note that the the deep root treatment than SUVA360 of the DOC leached from the columns showed similar trends for the root treatments and the other treatments. In con- the soil control, whereas the leaf treatment had a much lower SUVA360 up to Day 18, after which it trast, relatively more DOC converged with the other columns. from shallow root litter was 1562 SSSAJ: Volume 71: Number 5 • September–October 2007  translocated and retained within the soil column than DOC Jackson, R.B., J. Canadell, J.R. Ehleringer, H.A. Mooney, O.E. Sala, and from deep roots. The distribution of 14C remaining in the soil in E.D. Schulze. 1996. A global analysis of root distributions for terrestrial biomes. Oecologia 108:389–411. the leaf litter treatment declined exponentially with depth, and Jandl, R., and P. Sollins. 1997. Water-extractable soil carbon in relation to the the SUVA360 measurements as well as chemical analyses from belowground carbon cycle. Biol. Fertil. Soils 25:196–201. Uselman (2006) indicated that there may be a greater portion of Johnson, D.W., R.B. Susfalk, and R.A. Dahlgren. 1997. Nutrient fluxes in labile DOC originating from leaf litter during the initial phase forests of the eastern Sierra Nevada mountains, United States of America. of leaching in comparison to DOC originating from root litter. Global Biogeochem. Cycles 11:673–681. Johnson, D.W., R.B. Susfalk, R.A. Dahlgren, T.G. Caldwell, and W.W. Miller. Our results suggest that DOC originating from leaf litter may 2001. Nutrient fluxes in a snow-dominated, semi-arid forest: Spatial and contribute to the formation of an A horizon and even to accu- temporal patterns. Biogeochemistry 55:219–245. mulation of SOC in the B horizon during soil development, Jones, D.L., A. Hodge, and Y. Kuzyakov. 2004. Tansley review: Plant and either by adsorption to soil or incorporation into microbial mycorrhizal regulation of rhizodeposition. New Phytol. 163:459–480. biomass. The 14C data further showed that DOC originating Kaiser, K., and W. Zech. 1999. Release of natural organic matter sorbed to oxides and a subsoil. Soil Sci. Soc. Am. J. 63:1157–1166. from root litter, especially root litter at greater depths, may help Kalbitz, K., S. Solinger, J.-H. Park, B. Michalzik, and E. Matzner. 2000. to explain the presence of SOC at depth in soil. Finally, our Controls on the dynamics of dissolved organic matter in soils: A review. results also suggest that a significant portion of DOC draining Soil Sci. 165:277–304. from soil profiles in forested ecosystems may originate from Kirschbaum, M.U.F. 1995. The temperature dependence of soil organic matter root litter as well as from leaf litter. decomposition, and the effect of global warming on soil C organic storage. Soil Biol. Biochem. 27:753–760. Across the Mt. Shasta ecosystem chronosequence, soil Lilienfein, J., R.G. Qualls, S.M. Uselman, and S.D. Bridgham. 2003. adsorption capacity for DOC and dissolved organic N Soil formation and organic matter accretion in a young andesitic (DON) increased (Lilienfein et al., 2004). Soil C and N chronosequence at Mt. Shasta, California. Geoderma 116:249–264. concentrations also increased, in total and with depth, dur- Lilienfein, J., R.G. Qualls, S.M. Uselman, and S.D. Bridgham. 2004. ing soil development (Lilienfein et al., 2003). In addition, Adsorption of dissolved organic carbon and nitrogen in soils of a weathering chronosequence. Soil Sci. Soc. Am. J. 68:292–305. fine roots became more deeply distributed, and the relative Michalzik, B., K. Kalbitz, J.-H. Park, S. Solinger, and E. Matzner. 2001. contribution of DOC and DON from fine roots to aboveg- Fluxes and concentrations of dissolved organic carbon and nitrogen—a round litter increased with ecosystem age (Uselman, 2006; synthesis for temperate forests. Biogeochemistry 52:173–205. Uselman et al., 2007). This study shows that DOC originat- Ostertag, R., and S.E. Hobbie. 1999. Early stages of root and leaf decomposition in ing from root litter, especially at greater depths, may be one Hawaiian forests: Effects of nutrient availability. Oecologia 121:564–573. Qualls, R.G. 2000. Comparison of the behavior of soluble organic and of a number of sources that help to explain the presence of inorganic nutrients in forest soils. For. Ecol. Manage. 138:29–50. SOC at depth in soil. As a result, we suggest that fine roots Qualls, R.G. 2005. Biodegradability of fractions of dissolved organic carbon leached may play an increasingly fundamental role in the accumula- from decomposing leaf litter. Environ. Sci. Technol. 39:1616–1622. tion of SOC during primary succession. Qualls, R.G., and S.D. Bridgham. 2005. Mineralization rate of 14C-labelled dissolved organic matter from leaf litter in soils of a weathering chronosequence. Soil Biol. Biochem. 37:905–916. ACKNOWLEDGMENTS Qualls, R.G., and B.L. Haines. 1992. Biodegradability of dissolved organic We would like to thank Peter Van Susteren and the U.S. Forest matter in forest throughfall, soil solution, and stream water. Soil Sci. Soc. Service McCloud Ranger Station, without whose support this Am. J. 56:578–586. project would not have been possible. Thanks to Scott Tyler for Silver, W.L., and R.K. Miya. 2001. Global patterns in root decomposition: the data on soil water retention. This research was funded by a Comparisons of climate and litter quality effects. Oecologia 129:407–419. National Science Foundation Ecosystem Studies Grant (DEB Smith, W.H. 1976. Character and significance of forest tree root exudates. 9974062), Graduate Student Association of the University of Ecology 57:324–331. Nevada-Reno Merit Research Grant, and in part by the Nevada Sollins, P., G. Spycher, and C. Topik. 1983. Processes of soil organic matter Agricultural Experiment Station. Additional support was provided accretion at a mudflow chronosequence, Mt. Shasta, California. Ecology 64:1273–1282. by the Soil Science Society of America, the Nevada, Pacific Region, SPSS. 2003. SPSS Version 12.0 for Windows. Release 12.0.0. SPSS, Chicago. and National Garden Clubs, Inc., the Nevada Women’s Fund, and Swinnen, J., J.A. van Veen, and R. Merckx. 1995. Root decay and turnover of USA Funds to S.M.U. rhizodeposits in field-grown winter wheat and spring barley estimated by 14C pulse-labelling. Soil Biol. Biochem. 27:211–217. REFERENCES Thurman, E.M. 1985. Organic geochemistry of natural waters. Martinus- Cheng, W., and D.W. Johnson. 1998. Elevated CO2, rhizosphere processes, Nijhoff, Dordrecht, the Netherlands. and soil organic matter decomposition. Plant Soil 202:167–174. Uselman, S.M. 2006. Production and fate of soluble organic carbon, nitrogen, Cleveland, C.C., S.C. Reed, and A.R. Townsend. 2006. Nutrient regulation of and phosphorus during forest ecosystem development: Root versus leaf organic matter decomposition in a tropical rain forest. Ecology 87:492–503. litter. Ph.D. diss. Univ. of Nevada, Reno (Diss. Abstr. 3222985). Dickson, B.A., and R.L. Crocker. 1953. A chronosequence of soils and Uselman, S.M., R.G. Qualls, and J. Lilienfein. 2007. Fine root production vegetation near Mt. Shasta, California: I. Definition of the ecosystem across a primary successional ecosystem chronosequence at Mt. investigated and features of the plant succession. J. Soil Sci. 4:123–141. Shasta, California. Ecosystems (in press) doi:10.1007/s10021–007– Fröberg, M., D. Berggren Kleja, and F. Hagedorn. 2007. The contribution of 9045–8. fresh litter to dissolved organic carbon leached from a coniferous forest Uselman, S.M., R.G. Qualls, and R.B. Thomas. 2000. Effects of increased floor. Eur. J. Soil Sci. 58:108–114. atmospheric CO2, temperature, and soil N availability on root exudation Gee, G.W., A.L. Ward, T.G. Caldwell, and J.C. Ritter. 2002. A vadose zone water of dissolved organic carbon by a N-fixing tree (Robinia pseudoacacia L.). fluxmeter with divergence control. Water Resour. Res. 38:16.1–16.7. Plant Soil 222:191–202. Gordon, W.S., and R.B. Jackson. 2000. Nutrient concentrations in fine roots. Verburg, P.S.J., A. Gorissen, and W.J. Arp. 1998. Carbon allocation and Ecology 81:275–280. decomposition of root-derived organic matter in a plant–soil system Hagedorn, F., M. Saurer, and P. Blaser. 2004. A C-13 tracer study to identify of Calluna vulgaris as affected by elevated CO2. Soil Biol. Biochem. the origin of dissolved organic carbon in forested mineral soils. Eur. J. 30:1251–1258. Soil Sci. 55:91–100. Zar, J.H. 1996. Biostatistical analysis. 3rd ed. Prentice Hall, Englewood Cliffs, NJ. SSSAJ: Volume 71: Number 5 • September–October 2007 1563 View publication stats