Chapter 1: Procedure for Determining Soil-Bound Organic Carbon and Nitrogen Rebecca J. Mahoney1, Alan L. Kafka2, and Kevin G. Harrison2*

Keywords: soil carbon storage, CO2 fertilization.

1

5 Mallard Cove, East Hampton, CT 06424, USA.

2

Department of Geology and Geophysics, Devlin Hall, 213, Boston College, Chestnut Hill, MA

02467, USA.

*Corresponding author ([email protected]).

ABSTRACT We present a method for determining soil-bound organic carbon, which we define as soil carbon having no carbonate, charcoal, litter fragments, or root fragments. Carbonates are removed by exposing the soil sample to acid vapor until the δ13C values reach a constant value. Charcoal, litter, and root fragments are removed by flotation. Rocks and other large debris are removed by inspection and sieves. Organic carbon and nitrogen concentrations of the processed soil are determined quantitatively using high-temperature flash combustion gas chromatography. At the Duke Free-Air CO2 Enrichment (FACE) site, our technique showed that the variability in litter-free soil was less than the variability in soil that contained litter and root fragments. Isolating the soil-bound fraction reduced the uncertainty of soil carbon uptake due to CO2 enrichment by a factor of about 20. This analytical technique will help researchers quantify the

1

flux of carbon between soil and the atmosphere and discern how changing climate, CO2 levels, nutrient availability, and land use alter this flux.

INTRODUCTION This paper presents a method and provides a rationale for determining soil-bound carbon and nitrogen. Measuring soil-bound organic material is an important step toward understanding how perturbations alter soil carbon storage. Perturbations such as climate change, CO2 fertilization, anthropogenic nitrogen deposition, and changing land use may be changing the soil carbon inventory. As Post et al. (1999) have stated, it is important that methods are developed which can accurately determine the rates of change of soil organic matter. We present our findings in the context of the Duke FACE experiment.

Rationale for Looking at Soil-bound Carbon There are two reasons for isolating soil-bound organic carbon, which we define as soil material that is free of carbonates, charcoal particles, litter fragments, and root fragments. First, the signal-to-noise ratio is much higher for the soil-bound fraction than for the soil organic matter fraction (Table 1). Second, the turnover time of soil-bound carbon differs from the turnover time of litter and root fragments, carbonates, and charcoal. The turnover time gives an indication of how quickly a pool of material will respond to a perturbation (Harrison et al., 1993). For example, if carbon dioxide levels double, it will take 25-years for soil-bound active soil carbon (τ = 25 years), to reach 70% of its new steady-state value. The turnover time and the change in inventory observed for a CO2 enrichment experiment can then be used to determine the additional increase in soil carbon storage due to observed increases in atmospheric carbon dioxide levels.

2

A review of the literature suggests starting points for modeling and measuring turnover times of various carbon pools found in soil. Litter and root fragments have turnover times of three years or less (Warneck, 1988; Schlesinger and Lichter, 2001). Charcoal is inert. For example, scientists use radiocarbon values of charcoal to date ecological events, such as forest fires (Pressenda et al., 1998; Gomes et al., 2000). Charcoal must be removed to get accurate turnover times for soil-bound organic material (Trumbore, 1988). Carbonates can have turnover times of about 85,000 years (Schlesinger, 1985). Hence, the soil carbon pools that might influence atmospheric carbon dioxide levels on a relevant time scale include litter and root fragments and active soil-bound carbon. Charcoal, carbonates, and passive soil carbon will not dramatically influence atmospheric carbon dioxide levels over the next several decades. Figure 1 shows the relationship between carbon turnover times and carbon uptake due to CO2 fertilization. In this case, the perturbation has been modeled as an increase in the flux of carbon to a pool. Carbon accumulates because there is a lag between the increased input and the increase in carbon respiration. Carbon pools having faster turnover times will attain the new steady-state value faster than carbon pools having slower turnover times. The new steady-state carbon inventory is higher than the initial steady-state carbon inventory. The carbon pools depicted in this figure are litter (~3 year turnover time), active soil-bound carbon (~25 year turnover time), and passive soil-bound carbon (~4,700 year turnover time). The passive carbon showed almost no response, and carbon pools having turnover times longer than 4,700 years will also show little response. Figure 2 shows how time series of radiocarbon measurements can be used to determine carbon turnover times. It is essential to isolate the soil carbon components so that accurate turnover times and inventory changes can be determined. Radiocarbon levels almost doubled in the atmosphere in 1964 because of nuclear bomb testing. This radiocarbon pulse can be used to estimate soil carbon turnover times, because the rate of radiocarbon penetration depends on the 3

carbon turnover time. Using soil radiocarbon measurements to determine soil carbon turnover times is analogous to “pulse-chase” experiments. The pulse of radiocarbon is global. CO2 enrichment experiments that use enrichment gas that is depleted in radiocarbon produce a "reverse-bomb" spike that can be used to derive soil carbon turnover times.

Further Importance of Isolating the Soil-bound Fraction Measuring the amount of carbon and radiocarbon in soil organic material that contains litter, charcoal, or carbonates produces results that are difficult to interpret. Consider the following hypothetical result, which illustrates the importance of isolating the soil-bound fraction. If soil organic carbon accumulates 15% faster in elevated rings compared to ambient rings after 4 years in an (ambient +200 ppm CO2) enrichment experiment, how much carbon would be sequestered for observed increases in atmospheric carbon dioxide levels? Figure 3 illustrates two possibilities for extrapolating these results globally. In the first case, the increase occurs in the litter fragments present in the soil. On average, litter has a turnover time of 1.2 years and a global inventory of about 60 Gt. C (Warneck, 1988). Extrapolating the 15% increase in soil litter results in global carbon sequestration of about 0.1 Gt. C/year at present. In the second case, the increase occurs in active soil-bound carbon. This carbon pool has a turnover time of about 25 years and a global inventory of 625 Gt. C (Harrison, 1996). Extrapolating a 15% increase in the active soil-bound carbon results in a global carbon sequestration of about 2.4 Gt. C/year at present. There is more than a 20-fold difference between global soil carbon accumulation estimates due to CO2 fertilization, depending on whether the observed carbon increase is in the litter pool or the active, soil-bound carbon pool. Thus, it is essential to measure how perturbations alter the inventory of active, soil-bound carbon.

MATERIALS AND METHODS 4

The FACE Site FACE stands for Free-Air CO2 Enrichment. This type of experiment was designed to study the effects of carbon dioxide enrichment on an ecosystem level, without altering environmental variables other than the CO2 level in the atmosphere.

The FACE facility

used in this study is located in the Blackwood Division of the Duke forest in North Carolina (35°97’N 79°09’W). It was built in 1996 in a loblolly pine (Pinus taeda L.) plantation planted in 1983 on former agricultural land. Loblolly pine is found in an area ranging from Virginia to Texas and it is an important tree species on the Piedmont and coastal plain. Sweetgum (Liquidambar styraciflua L.), yellow poplar (Liriodendron tulipifera L.), winged elm (Ulmus alata Michx.), and red maple (Acer rubrum L.) are other tree species that are found at the site. The soils in the FACE site are clay-rich ultic Alfisols of the Enon series (DeLucia et al., 1999; Schlesinger and Lichter, 2001) which are typical of many areas in the Southeast. The soil has a pH of 5.8. The average temperature is about 15oC and the average rainfall is about 1,140 mm. The FACE facility consists of six 30-meter diameter plots, or rings, within the pine forest. Each ring has 16 towers along its circumference, and each tower supports 2 pipes, which extend through the top of the canopy and contain apertures that allow the injection of carbon dioxide into the rings. The pipes can be extended once the canopy grows higher than the top of the towers. Carbon dioxide is delivered to three of the six rings to maintain an atmospheric concentration 200-ppm higher than ambient levels, or about 560 ppm. The ambient carbon dioxide levels in the control rings are continuously monitored and average about 360 ppm. The control rings receive the same volume of air as the elevated rings, in order to balance any microclimatic effects the FACE facility has on the ecosystem. Carbon dioxide was injected into the elevated rings beginning on August 27, 1996.

Soil Collection 5

We collected 72 soil samples in 1998, after 1.5 years of CO2 enrichment. Cores of the upper 35 centimeters of soil were collected beneath trees in all six rings using a volumetric coring device so that bulk density could be calculated. An attempt was made to select soil cores away from trees, roots, trampled areas, the edge of the rings, and away from the boardwalks. The high clay content of the deep soil made recovery of the lowest depth interval the most difficult. The cores were carefully inspected when collected. Incomplete or suspect cores were discarded in an attempt to minimize error associated with the collection process. We modified a conventional soil probe to facilitate sample collection. We replaced the probe handle with a stainless steel anvil (5 cm diameter by 7 cm height). The anvil allowed us to hammer the probe into the ground, which greatly facilitated sampling. Sample Processing Soil contains various forms of organic and inorganic carbon. We define soil organic matter as soil material that contains particles of litter, roots and charcoal, in addition to soilbound carbon. We define delittering as the process that separates litter, roots and charcoal particles by flotation with water, leaving only soil-bound carbon. Below, we describe a technique for removing carbonates without leaching organics from soil. After collection, the samples were stored in plastic ziploc bags until they could be dried and prepared for analysis. Once collected, it took less than three days to start drying the soil. The collected soil was dried to a constant weight at 60 degrees C.

This process removed

sufficient moisture to inhibit microbial oxidation of soil organic carbon and stabilized the soil for storage. The samples collected from this site contained no carbonates. If carbonates had been present, they would have been removed by exposing the soil to concentrated hydrochloric acid vapor in a desiccator until the δ13C measurements of the soil organic material reached a constant value. This technique uses acid vapor, rather than liquid acid, to eliminate the possibility of 6

organic material being leached from the soil. Liquid hydrochloric acid has previously been used by researchers to extract organic material from soil (Martel and Paul, 1974). The typical δ13C value for carbonates is 0‰, while the δ13C value for organic material ranges from -15 to -30‰. For this study, each sample was split into two subsamples. The first subsample was analyzed for soil organic carbon and nitrogen. The second sample was processed to isolate the soil-bound organic carbon fraction. The dried second subsample soils were sifted through 2-mm and 30-micron sieves. Visible rocks, charcoal, litter, and root fragments were removed by inspection. Any remaining litter and charcoal were removed by flotation. To delitter the samples by flotation, each sample was placed in a glass funnel lined with coarse, fast-flowing filter paper (Fisherbrand, P8 creped filter paper, 15-cm diameter). The samples were flushed with water so that the roots, charcoal, and leaves that were in the soil floated to the top and could be poured off. This process was repeated until no organic debris was left in the samples. After delittering, the soils were dried at 100˚C to a constant dry weight. The samples were ground up once more to homogenize the mixture, and stored in glass sample bottles for subsequent analysis. Following this process, small amounts of each sample were weighed out into tin cups (30 to 40 mg for carbon, and 40 to 200 mg for nitrogen, due to its lower concentration in soil). The cups were carefully compacted for gas chromatograph analysis.

Carlo Erba Calibration and Analysis A Carlo Erba NC-2100 Analyzer was used to determine the concentration of carbon and nitrogen in the soil samples. Soil samples underwent flash combustion and then nitrogen and carbon concentrations were determined by gas chromatography, as described below. How the Carlo Erba Works An autosampler injects the sample, in its tin cup, into the combustion chamber. The combustion chamber is maintained at 900°C, and filled with an oxidation catalyst (chromic 7

oxide, Cr2O3) which overlies silvered cobaltous-cobaltic oxide (Co3O4 + Ag) granules. Once the sample enters the chamber, 5 ml of oxygen are added to the chamber to create a highly oxidizing atmosphere, which completely combusts the sample and tin cup container. All of the carbon and nitrogen present in the samples is converted to CO2 and NOx, respectively. Helium is the carrier gas used to sweep the combustion products--CO2, NOx, and H2O--through the combustion column into the second column, the reduction chamber. The reduction agent is metallic copper and is maintained at 750°C. Excess oxygen is removed in the reduction chamber and NOx is reduced to N2. The helium carrier gas carries the remaining CO2, N2, and H2O through a desiccant (magnesium perchlorate, Mg(ClO4)2) to remove the water. The CO2 and N2 then enter the chromatographic column (a Porapak QS, 4 mm ID, 2-m long column), where the gases are separated. The detection of carbon and nitrogen is achieved with a thermal-conductivity detector. The Eager 200 software program converts the signal from the thermal conductivity detector into a chromatogram that displays the carbon and nitrogen peaks visually and provides the numerical data used to calculate the percentage of carbon and nitrogen present in the sample (Figure 4). The run lasts for 400 seconds.

Calibrating the Carlo Erba In order to maximize the performance of the Carlo Erba, we calibrate the machine using four samples of a standard, acetanilide, at the start of each analysis day and after every 50 samples. Acetanilide (C8H9NO) is 71.09% carbon, 10.36% nitrogen, 11.84% oxygen, and 6.71% hydrogen. The standard is weighed out to six significant figures on a Mettler Toledo AT21 Comparator balance into four tin cups in the following amounts: ~0.5 mg, ~1.0 mg, ~1.5 mg, and ~2.0 mg. The results of the standard runs are plotted as the area of carbon or nitrogen on the xaxis, and milligrams of carbon or nitrogen on the y-axis (Figure 5). The carbon and nitrogen amounts are calculated by multiplying the known percentage of C or N in acetanilide by the 8

amount of sample. The equation of the regression line that fits these points is used to calculate the amount of carbon and nitrogen in the soil samples. An acceptable regression line for carbon has an R2 value of at least 0.999. For nitrogen, the acceptable R2 value is at least 0.99. For R2 values less than these, another set of standards are run and a new calibration curve is derived. To monitor the accuracy of the Carlo Erba, one standard and one blank are analyzed between every six soil samples. The standards are treated as unknowns. Samples of ~1.0 mg of acetanilide are used. The measured values for these standards should produce results that are within 1% of the theoretical values. If they do not agree within 1%, maintenance or repairs are performed on the instrument and the samples are re-analyzed. If the blank values are not zero throughout the analysis, the samples are analyzed again.

RESULTS AND DISCUSSION Carbon Results Table 1 lists the results of the Carlo Erba analysis. The delittered samples had less carbon than their non-delittered counterparts. For example, the 0-7.5 cm average value for carbon in ring one for the soil with litter was 2.47% carbon, while the delittered sample had 0.673% carbon. The standard error was also significantly less for the delittered samples. For the 0 to 7.5 cm segment, the average standard error for the samples that had litter was 0.789, and 0.194 for the samples that had been delittered. The soil-bound values were more consistent and had a higher signal-to-noise ratio. This trend has continued for the samples collected from deeper depths. Because of the 10- to 30-year turnover time of active soil carbon (Harrison, 1996), 1.5 years of treatment is not long enough to discern other differences between the ambient and the elevated chambers, such as differences with regard to soil-bound carbon accumulation rates.

9

Nitrogen Results Table 1 also lists the nitrogen values, which were made on soil composites. The composites were made using sub-samples of soil that passed the bulk density box-plot test (below).

Quantifying Bulk Density The bulk density (g/cm3) of the soil samples must be calculated in order to determine the carbon and nitrogen inventory of the soil. However, bulk density may be affected by the manner in which the soil cores have been collected, and consequently may be miscalculated. For example, a stone or a pebble may have taken up a significant portion of volume in the core, or the soil may have collapsed in the core before it could be divided into depth intervals. In order to identify and eliminate outliers in the bulk density data, a histogram and box-plot analysis were used. Histograms were constructed for the bulk densities of all rings to observe the distribution of the data and to determine if any of the bulk densities measured in a ring were outliers. Since the distribution of a histogram is influenced by the subjective nature of choosing bin sizes, box plots were also constructed for each set of bulk densities. Box plots were expressly designed to isolate outliers in a data set (Mendenhall et al., 1999). Table 2 lists all of the sample bulk densities. The samples that failed the bulk density box plot test because of outliers are indicated in the table. Figure 6 shows how box plots were used to identify outliers.

CONCLUSION We have presented a technique for determining soil-bound organic carbon and have discussed the carbon and nitrogen results for one experiment at the Duke FACE site. Our results led us to propose that, by isolating the soil-bound fraction, researchers can begin to make 10

meaningful interpretations of changes in carbon inventories due to perturbations. These improved interpretations may help global change scientists balance the global carbon budget and predict future atmospheric carbon dioxide levels.

ACKNOWLEDGMENTS This research was supported by the US Department of Agriculture. We thank Emily Chapp, George Hendrey, Theresa Hensel, Matt Hoskins, Marla Knebl, Lindsey Kurnath, Kevin Mahoney, Adria Reimer, Josh Rollins, William Schlesinger, Shannon Smith, Lori Weeden, and Sara Wierzbicki.

REFERENCES DeLucia, E.H., J. G. Hamilton, S. L. Naidu, R. B. Thomas, J. A. Andrews, A. Finzi, M. Lavine, R. Matamala, J. E. Mohan, G. R. Hendrey, W. H. Schlesinger. 1999. Net primary production of a forest ecosystem with experimental CO2 enrichment. Science, 284, 11771179. Gomes, P. R. S., R. M. Anjos, J. C. Acquadro, G. M. Santos, K. C. D. Macario, R. L. Neto, N. Added, R. C. Cordeiro, B. Tureq, A. Sifeddine, M. M. Coimbra, C. R. Appoloni, M. di Tada, R. G. Cresswell and L. K. Fifield. 2000. Implementation of the Accelerator Mass Spectrometry Technique in Brazil and Environmental Studies in Central Amazon Forest. Heavy Ion Physics, 11, 485-496. Harrison, K.G. 1996. Using bulk soil radiocarbon measurements to estimate soil carbon turnover times. Radiocarbon, 38, 181-190. Harrison, K.G., W.S. Broecker, and G. Bonani. 1993. A strategy for estimating the impact of CO2 fertilization on soil carbon storage. Global Biogeochemical Cycles,7, 69-80. Martel, Y. A. and E. A. Paul. 1974. The use of radiocarbon dating of organic matter in the study of soil genesis. Soil Sci. Soc. Amer. Proc., 38, 501-506. Post, W.M., R.C. Izaurralde, L.K. Mann, and N. Bliss. 1999. Monitoring and Verifying Soil Organic Carbon Sequestration, p. 41-66, In Carbon Sequestration in Soils: Science, Monitoring, and Beyond. ed. N.J. Rosenberg, R. C. Izaurralde, E. L. Malone. Columbus, OH: Battelle Press. Pressenda, L. C. R., S. E. M. Gouveia, R. Aravena, B. M. Gomes, R. Boulet and A. S. Ribeiro. 1998. 14C dating and stable carbon isotopes of soil organic matter in forest-savanna boundary areas in the southern Brazilian Amazon region. Radiocarbon, 40, 1013-1022. Schlesinger, W. H. 1985. The formation of caliche in soils of the Mojave Desert, California. Geochimica et Cosmochimica Acta, 49, 57-66. Schlesinger, W.H. and J. Lichter. 2001. Limited carbon storage in soils and litter of experimental forest plots under increased atmospheric CO2. Nature, 411, 466-469. 11

Trumbore, S.E. 1988. Carbon cycling and gas exchange in soils, Ph.D. thesis, Columbia University, New York. Warneck, P. 1988. Chemistry of the Natural Atmosphere. New York: Academic Press.

12

Table 1. Carbon values with and without litter. The nitrogen values are for delittered composites. 1998 samples 0-7.5 cm Ring

1 5 6 2 3 4

Treatment

Ambient Ambient Ambient Elevated Elevated Elevated

%C (s.e.) with litter 2.47 (0.944) 2.67 (0.474) 2.59 (0.266) 2.07 (0.499) 2.06 (0.335) 2.48 (0.417)

%C (s.e.) delittered 0.673 (0.0956) 1.00 (0.107) 1.12 (0.250) 1.26 (0.369) 0.719 (0.238) 0.922 (0.104)

%N delittered 0.0354 0.0678 0.0566 0.0589 0.0607 0.0418

1998 samples 7.5-15 cm Ring

1 5 6 2 3 4

Treatment

Ambient Ambient Ambient Elevated Elevated Elevated

%C (s.e.) with litter

%C (s.e.) delittered

0.901 (0.300) 0.937 (0.140) 0.917 (0.115) 0.755 (0.150) 0.874 (0.0815) 0.989 (0.226)

0.387 (0.0484) 0.452 (0.115) 0.503 (0.0973) 0.442 (0.0856) 0.474 (0.0440) 0.396 (0.173)

%N delittered 0.0192 0.0249 0.0235 0.0229 0.0262 0.0309

1998 samples 15-35 cm Ring

1 5 6 2 3 4

Treatment

Ambient Ambient Ambient Elevated Elevated Elevated

%C (s.e.) with litter

%C (s.e.) delittered

0.504 (0.0760) 0.370 (0.0508) 0.469 (0.0184) 0.373 (0.0497) 0.454 (0.0613) 0.679 (0.183)

0.339 (0.0327) 0.207 (0.0443) 0.243 (0.0478) 0.304 (0.0580) 0.278 (0.0299) 0.387 (0.108)

%N delittered 0.0186 0.0131 0.0123 0.0177 0.0149 0.0240

13

Table 2. Bulk density of samples collected in 1998. *Indicates the sample failed the box-plot bulk density test. If one depth interval of a core failed the box-plot bulk density test, the entire core was not used.

Ring

Core

Depth (cm)

Bulk Density (g/cm3)

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 4

a a a b b b c c c d d d e e e a a a b b b c c c d d d a a a b b b c c c d d d a

0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5

1.159 1.823* 1.519 1.084 1.705 1.024* 1.216 1.597 1.354 0.793* 1.672 1.385 1.338 1.724 1.759 1.446 1.409 1.225 0.732 1.498 0.486 0.873 1.131 1.184 1.009 1.681 1.736 1.489 1.630 1.653 0.995 1.569 1.341 1.037 1.413 1.288 0.760 1.258 1.271 1.202 14

Ring

Core

Depth (cm)

Bulk Density (g/cm3)

4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6 6 6 6 6 6

a a b b b c c c d d d a a a b b b c c c d d d a a a b b b c c c d d d

7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35 0-7.5 7.5-15 15-35

1.653 1.500 1.103 1.531 1.200 1.315 1.606 1.639 1.324 1.700 1.255 1.108 1.578 1.380 1.023 1.587 0.784 0.764 1.446 0.840 0.943 1.479 1.646 1.112 1.818 1.666 1.004 1.258 1.410 1.301 1.672 1.694 0.619 0.826 1.343

15

30

% C Increase

25 atm CO

20

2

15 litter

10 5 passive

0 1900

1920

1940

1960

1980

active

2000

Year

Figure 1. Model of percent carbon increase vs. time for 3 carbon pools having different turnover times. The litter pool has a turnover time of 3 years, the active soil-bound carbon pool has a turnover time of 25 years, and the passive soil-bound carbon pool has a turnover time of 4,700 years. The turnover time equals the inventory of carbon in a pool divided by the flux into or out of the pool. Carbon pools having the shortest turnover times will show the fastest response to CO2 fertilization and will be the most likely candidates for the location of the “missing sink,” if they have a large enough inventory. These model results have been derived using an existing CO2 enrichment model (Harrison et al., 1993).

16

2

( 1 4C/C

sample

)/(

14

C/C 1850

wood

)

atmosphere

1.8 1.6

litter

1.4 1.2 active

1 0.8 1950

passive

1960

1970

1980

1990

2000

Year

Figure 2. Radiocarbon change for the atmosphere and three different carbon pools having different turnover times. These model results were derived using an existing radiocarbon model (Harrison et al., 1993). The litter pool has a 3-year turnover time, the active soil-bound carbon has a 25-year turnover time, and the passive soil-bound carbon has a 4,700 year turnover time. Atmospheric values of radiocarbon almost doubled in 1964 because of atmospheric nuclear bomb testing. This radiocarbon spike can be used to elucidate carbon turnover times. Using carbon dioxide that is depleted in radiocarbon can also be used to trace carbon dynamics in CO2 enrichment experiments.

17

3.5

3

Gt C/year

2.5

active mineral-bound carbon

2

1.5

1

0.5

litter

0 0

5

10

15

20

25

30

% C increase

Figure 3. Rationale for isolating the soil-bound fraction. Global carbon sequestration has been extrapolated from a hypothetical 15% increase in soil organic carbon after four years of CO2 enrichment. Carbon sequestration is much greater if the increase occurs in the soil-bound fraction than if the increase occurs in the litter fragment fraction. For example, the active soilbound carbon would have sequestered about 2.4 Gt. C/year. In contrast, the litter pool would have sequestered only about 0.1 Gt. C/year. These model results have been derived using an existing CO2 fertilization model (Harrison et al., 1993).

18

Figure. 4. Sample chromatogram. The nitrogen peak (about 74 seconds) occurs before the carbon dioxide peak (about 116 seconds).

19

1.5 y = -0.0014036 + 2.0779e-08x R2= 0.99999

mg C

1

0.5

0 0

1 107

2 107

3 107

4 107

5 107

6 107

7 107

8 107

Area

Figure 5. Carbon calibration curve. We obtain the areas of the carbon and nitrogen peaks using Eager 200 software. A nitrogen calibration curve would follow the same format.

20

1.4

Bulk Density (g/cm3 )

1.3 1.2 1.1 1 0.9 0.8 0.7

0-7.5 cm

Figure 6. Box plot of bulk densities from ring 1 (1-7.5 cm). One outlier was identified. Box plots were expressly designed to identify outliers in a data set. The distribution of a data set is illustrated in a box plot by depicting the median and the interquartile range of the data. The lower end of the box represents the lower quartile, the value that exceeds 25% of the measurements and is less than the remaining 75%. The upper end of the box represents the upper quartile, the value that exceeds 75% of the measurements and is less than the remaining 25%. The area between these two values is called the interquartile range and represents the middle 50% of the distribution. A horizontal line drawn within the interquartile range depicts the median. The values adjacent to the quartile measurements are connected to the box by vertical lines called whiskers. In this box plot, the adjacent value is equal to the lower quartile value, so there is no whisker on the bottom side of the box. Outliers are any measurements that lie outside of imaginary “fences” constructed using the interquartile range. The open circle indicates the outlier in this data set.

21