Stable Isotope Constraints on Net Ecosystem Production Under Elevated CO2

Elise Pendall, Jennifer Y. King, Arvin R. Moser, Jack Morgan,

Daniel Milchunas

Introduction: Approaches for Estimating Net Ecosystem Production

Net ecosystem production (NEP) is the total carbon accumulated in an ecosystem over a time interval, usually a year. The C is stored above- and belowground in plants, animals, and soils, and has residence times ranging from less than one to several hundred years or more. NEP varies in response to disturbance and during plant succession, and human-induced perturbations to climate and atmospheric constituents also affect NEP at local to global scales. Because NEP determines how much carbon dioxide is taken up from the atmosphere by terrestrial ecosystems, it is critical to evaluate what forcing factors might be regulating variations in NEP at various temporal and spatial scales. In particular, the effects of elevated atmospheric CO2 may be altering NEP in many ecosystems, and techniques such as stable isotope analysis are improving our ability to evaluate the magnitude of these effects (Pendall, 2002).

NEP has often been defined operationally, depending on the approach used to conduct measurements and the temporal and spatial scales of interest, but this has led to ambiguity and a lack of comparability among studies (Randerson etai, 2002). NEP was first estimated from changes in ecosystem pool sizes, after accounting for lateral fluxes (Olsen, 1963; Bormann et al, 1974; Aber and Melillo, 1991; Chapin et al, 2002):

NEP = (Aplant + Aanimal + ASOM)/At ± F]aterai (11.1)

It has been recognized that generally the changes in these inventories are extremely small relative to the entire pool size (particularly for SOM), making accurate measurement difficult unless long time intervals

(10 years or more) are allowed between sampling events. This has led to the development of flux-based methods, which allow NEP estimates to be made at more frequent intervals. A complete accounting can be formulated as the sum of fluxes into and out of the ecosystem (Olsen, 1963; Randerson et al., 2002):

% Fcpp + FRe + Ffire + Fi eaching "I" Ferosion + Fhydrocarbons

Fherbivory ^harvest (11-2)

where GPP is gross primary production and Re is ecosystem respiration. Net ecosystem exchange (NEE) has been defined as the balance of GPP and Re, and in the absence of fire or other disturbance, NEE provides an approximation of NEP at limited temporal and spatial scales (Chapin et al., 2002). NEE is measured at the ecosystem scale by eddy covariance techniques (e.g., Wofsy et al, 1993), or at the plot scale in flow-through chambers by monitoring the change in CO2 concentration going into and coming out of the chamber (Drake et al., 1996). NEP has also been formulated as the difference between net primary production (NPP) and heterotrophic respiration (Rh):

GPP and Ra (autotrophic respiration) are not generally measured, but NPP can be measured directly at the plot scale, or estimated from models based on leaf area index at regional to global scales. Empirical algorithms have been developed to estimate Rh; these are generally based on a relationship between microbial respiration and soil temperature (and sometimes moisture; Chapin et al., 2002). This approach also assumes that the smaller fluxes, such as leaching and erosional losses, can be neglected.

All of the methods used for estimating NEP (or NEE) have limitations. It appears that uptake measured by NEE in open-top chambers is often higher than can be accounted for by measurements of above- and belowground biomass (Drake et al., 1996; Niklaus et al., 2000). This has been termed a 'locally missing' carbon sink (Cheng et al., 2000; Niklaus et al., 2000), and may in part be caused by over-pressurization of chambers suppressing respiration, or by leaching or volatilization of C compounds. There are known issues with eddy covariance technology, particularly with respect to estimation of the respiration flux. Nighttime fluxes can be underrepresented if wind conditions are stable, or if air drainage patterns cause CO2 to be lost laterally from the system. Inventories miss a portion of the biomass to herbivory, senescence during the growing season, and incomplete sampling, especially belowground. Improvements in NEE methodology and in belowground inventory methods are narrowing the gap between inventory and exchange techniques.

It is inherently difficult to track small changes in NEP if C is being stored belowground, in part because many experiments are often not long enough (2-5 years) to produce changes in soil organic carbon (SOC) detectable using standard techniques. In certain situations, stable C isotopes improve the resolution with which C storage belowground can be detected. Recently-fixed C entering the soil pool has an isotopic signature distinct from preexisting SOC in many experiments on elevated CO2, as well as in cropping situations where a c4 crop has been planted into a C3 soil, or vice-versa (Balesdent et al, 1988; Leavitt et al, 1994). In these cases, 'new' inputs from rhizodeposition (root exudates and turnover of fine roots and mycorrhizal hyphae) can be quantified (Hungate et al, 1997; Van Kessel et al, 2000; Leavitt et al, 2001). Rhizodeposition is often too small to measure without isotopic techniques, and may thus constitute some portion of the 'missing' C accumulating in terrestrial ecosystems (Pendall et al, 2004 in review).

Carbon dioxide derived from decomposition of SOC will have a 5I3C value different from that of rhizosphere respiration in the same situations where there is an isotopic disequilibrium between new inputs and preexisting SOC (Rochette and Flanagan, 1997; Pendall et al, 2001b). This has allowed partitioning of the total soil respiration flux into its components of root/rhizosphere respiration and decomposition. No method exists to fully separate heterotrophic and autotrophic respiration fluxes in the field; decomposition by rhizosphere microbes is generally included with the root respiration component (Hanson, 2000). We suggest that in the case of estimating NEP, C loss by rapid turnover of recent (current year) photosynthate in the rhizosphere is functionally equivalent to autotrophic respiration, and the main C decomposition loss of interest is that which has a residence time of 1 year or longer.

Here we demonstrate a method that relies on biomass measurements and stable C isotopes for evaluating NEP in an experiment on elevated CO2. This approach is a modification of Eq. 11.5 in that we explicitly include measurement of rhizodeposition to the SOM pool, which is otherwise not included in the inventory:

where AG is annual aboveground C increment, or NPP, BG is annual belowground NPP, NSC is annual 'new' soil C input or rhizodeposition, and Rh is annual C loss by decomposition. We maintain that inventory estimates of NEP should include rigorous estimates of both decomposition and rhizodeposition. This approach would be suitable in any situation where an isotopic disequilibrium exists between recently fixed C and pre-existing soil organic matter, such as c3-c4 land-use or land-cover change. Stable C isotopes allow a more complete accounting of small increases in the soil carbon pool that are contributed by rhizodeposition, and also allow us to estimate decomposition losses by partitioning rhizosphere respiration and SOM decomposition.

Experimental Approach

An elevated CO2 experiment was conducted in the shortgrass steppe region of northeastern Colorado, at the USDA-ARS Central Plains Experimental Range (CPER; latitude 40°40' N, longitude 104°45' W), -55 km northeast of Fort Collins. The most abundant species at the study site were the c4 grass, Bouteloua gracilis (H.B.K) Lag. (blue grama), and the c3 grasses Stipa comata Trin and Rupr. (needle-and-thread grass) and Pascopyrum smithii (Rydb.) A. Love (western wheatgrass). Root biomass (including crowns) is responsible for ~70% of net primary production (NPP) in this ecosystem (Milchunas and Lauenroth, 2001). Beginning in 1997, open-top chambers (OTC; 4.5 m diameter) were used to evaluate the effects of C02 on the shortgrass steppe (SGS) ecosystem, with three replicate chambers at ambient (360 ± 20pmolmor1) and elevated (720 ± 20pmolmor1) C02. Three unchambered plots of the same area allowed evaluation of any chamber effects. Chambers were placed on the plots before growth started in late March or early April, and removed at the end of the growing season in late October. Blowers with ambient or elevated CO2 ran continuously. The experimental and chamber design has been described in detail by Morgan et al. (2001).

Figure 11.1 shows the <513C values of ecosystem components in elevated and ambient CO2 chambers at the SGS OTC experiment in 1999. Non-chambered plots had values that were not significantly different from those with the AC treatment. The tank gas used to double the atmospheric CO2 concentration had an approximate <513C value of —40%o during 1999, which produced air in the elevated chambers of —24.7 ± 1.4%o, which compared with background air <513C values of-8.1 ± 0.2%o (Pendall et al., 2003). The 13C-depleted air in the EC treatment produced depleted plant material, and by 1999 (after 3 years of elevated C02), the tracer was evident in soil organic matter. In the AC and NC plots, an isotopic disequilibrium between currently growing plants and SOM of ~4%o was attributed to a reduction in grazing c. 20 years prior to the start of our experiment (grazing was eliminated completely 2 years prior). Reduction of grazing pressure in the SGS tends to favor c3 over c4 grasses (Lauenroth and Milchunas, 1988).

NC AC C4's J Weigh,ed average new C |

Figure 11.1 Stable C isotope composition of ecosystem components in the shortgrass steppe open-top chamber experiment for 1999.

Figure 11.1 Stable C isotope composition of ecosystem components in the shortgrass steppe open-top chamber experiment for 1999.

Biomass Inventory Methods

In grassland ecosystems, aboveground biomass harvests are relatively straightforward. At the end of the growing season in October, aboveground biomass was clipped at the root crown from 1.5 m2 subplots, dried, and weighed. Because plots were clipped to the ground the prior fall, this harvest represents the annual aboveground NPP. We estimate that herbivory losses were less than 5% in all treatments; very occasional pesticide use was needed to reduce aphids. Different subplots were harvested in July and separated by species but this was not practical for plants collected in the fall. Tissue C content was determined by elemental analysis.

Belowground biomass inventories are typically labor-intensive and associated with a wide margin of error. The standing crop of roots in perennial grasslands may have a turnover time of 4—6 years (Milchunas and Lauenroth, 2001 ), and thus root ingrowth cylinders were used to assess the annual increment of root growth. In the fall prior to the growing season, two 20 cm diameter soil cores were removed from each plot, and replaced with 15 cm diameter PVC pipes to a depth of 40 cm. The outer diameter of each core was lined with plastic mesh (1mm square openings) and the 5 cm band gap was filled with sieved soil from near the experimental site. At the end of the following growing season, the PVC liners were carefully removed, roots were cut from the mesh liner, and all soil and roots were returned to the laboratory. Roots were separated from soil by flotation and washing over a 1 mm sieve, dried, ground, and analyzed for C content on an elemental analyzer. Subsamples of roots were ashed at 450°C to determine the ash-free C content.

Both above- and belowground NPP inventory methods tend to underestimate NPP, owing to losses such as herbivory and tissue senescence (Chapin et al., 2002). In the case of belowground NPP, our method of accounting for rhizodeposition should at least partly remedy this situation.

Rhizodeposition or 'New' Soil Carbon

Rhizodeposition is soil C derived from turnover of fine roots, secretion of soluble root exudates, and turnover of rhizosphere-associated microbial biomass. This C increment is generally too small to measure on an annual basis using total C analysis because it is within the measurement error, but by using an isotopic tracer, small changes may be detected in 1-2 years (Leavitt et al., 2001). The fraction of C that has been contributed by rhizodeposition (Fnew) is determined using the two-part mixing model described by Balesdent et al. (1988):

¿13CSOC = Fnew(513Cnew) + (1 " Fnew)(^13Cold) (H-7)

which can be rearranged to:

Fnew = (<51SCsOC - <513C0Id)/(<513Cnew - <51SCold) (11-8)

In order to use stable isotopes to estimate 'new' C inputs from rhizodeposition to the soil system, it is necessary to define the isotopic 'end-members,' i.e., the S'3C value of pre-existing SOM and that of the new C being added. Determination of the end-member isotopic values is not a trivial task in many ecosystems, because ¿13C of SOM often increases with depth; therefore, depth-dependent 'old' end-member values must be used. In some ecosystems, the <513C of 'new' input C is also challenging to determine. Mixed c3-c4 systems require assumptions of the proportion of C allocated belowground as well as the different isotopic values of the different functional groups.

In the case of the SGS OTC experiment, we assumed that the soils on the non-chambered (NC) plots remained at steady-state over the course of the experiment, and that the NC soils represented pre-existing C across the site. (This was confirmed by non-significant changes in <513C values in NC soils over time.) Our NC soils therefore provided the old C end-member values. An alternative approach is to use samples collected across the site prior to establishing an experiment (Leavitt et al., 1994). Collecting and analyzing steady-state soils together with treated soils provides analytical control over multi-year experiments. We collected two soil cores (40 cm deep, 15 cm diameter) from each plot at the end of each growing season. Cores were collected volumetrically and divided into five depth increments (0-5, 5-10, 10-20, 20-30, and 30-40cm). Gravel >2mm and most roots were removed immediately after sampling, with additional root picking done by hand on dried samples. Soils were then ground to a fine powder for analysis by elemental analyzer-mass spectrometry (EA-MS) for C content and <513C value, with precision of 0.1% and 0.2%o, respectively. Soils were carbonate-free to a depth of 40 cm, and we therefore did not acidify the samples.

In the mixed c3-c4 shortgrass steppe, the isotopic composition of new inputs can potentially vary spatially and temporally. Aboveground biomass clipping of 1.5 m2 subplots was done in July 1999, at peak green biomass. Leaves and stems were separated by species, dried, ground, and analyzed for C content and 513C value by elemental analyzer-IRMS. Roots were harvested by species at the end of the experiment in 2001, and averaged 1.4%o heavier than leaves (Pendall et al, 2004). We used the 1999 leaf <$13C values, weighted them by the 1999 aboveground biomass amounts, and then added 1.4%o for the new C end-member signatures * CnewS Table 11.1). July harvest samples were assumed to represent the proportion of c3 and c4 species for the entire growing season; samples harvested at the end of the season (peak total biomass) were not easily separated by species. The three dominant species maintain similar shoot: root ratios at ambient and elevated CO2, and therefore we assumed that the c3 : c4 ratio of rhizodeposition was proportional to aboveground biomass (Morgan et al, 1994).

Fnew was multiplied by the total mass of C in each horizon or depth increment to calculate rhizodeposition going into each horizon on a mass basis. This required measurement of bulk density, which can be a particularly challenging measurement in gravelly or organic-rich soils, but was straightforward in our sandy loam soils. Whole-profile rhizodeposition was calculated by summing the depth increments.

Uncertainties in estimating rhizodeposition were determined by accounting for variability in all of the components, including 513Csoc, <513C0id, and <513Cnew, following a first-order Taylor series approach (Phillips and Gregg,

Table 11.1 ¿>13C Values of C3 and C4 Plants, Proportion of C3 Biomass, and <513C Values of New C Inputs (After Adding 1.4%» to Account for Average Root 13C Enrichment) During the Growing Seasons of 1999 at the Shortgrass Steppe OTC Experiment

Table 11.1 ¿>13C Values of C3 and C4 Plants, Proportion of C3 Biomass, and <513C Values of New C Inputs (After Adding 1.4%» to Account for Average Root 13C Enrichment) During the Growing Seasons of 1999 at the Shortgrass Steppe OTC Experiment

<SI3C,C4 (%O)

<513C,CS (%0)

C3 biomass

<51SC new (%o)


-15.4 (0.21)

-25.2 (0.99)

0.72 (0.03)

-21.1 (1.01)


-33.2 (0.81)

-42.4 (1.92)

0.81 (0.04)

-39.3 (2.09)


-15.4 (0.27)

-26.5 (0.49)

0.67 (0.07)

-21.4 (0.56)

Standard errors in parentheses include error on the concentration and isotopic composition of C in end-members.

Standard errors in parentheses include error on the concentration and isotopic composition of C in end-members.

2001 ). Uncertainty estimates for <513Cnew included standard deviations of % biomass and <$13C of the dominant grass species. We further corrected for covariance of C content and <513C values among soil depth intervals when calculating the standard error for the whole profile (Donald L. Phillips, personal communication, 2002).

Decomposition (Rh)

Determination of the in-situ decomposition flux requires first assessment of the total soil respiration flux. We measured soil respiration in this grassland ecosystem by measuring the CO2 gradient with depth in the soil, and multiplying that by soil diffusivity to obtain the flux (Pendall et al, 2003). This allowed us to obtain a measurement of belowground respiration without disturbing the aboveground vegetation by clipping. In the fall before the growing season, stainless steel tubes (1/8" OD) were inserted horizontally 15 cm into a pit face at five depths (3, 5, 10,15, and 25 cm). The pit was then backfilled. Beginning in May, glass syringes, greased with Apiezon M and fitted with gas-tight valves, were used to collect and store the soil gas samples. Volumes of 10 mL were collected at all depths, except for the 3 cm depth, where 6 mL were collected, to ensure that no atmospheric air was pulled into the syringe. A small amount of magnesium perchlorate was used in-line to absorb moisture; this allowed us to analyze the oxygen isotopes as well as the carbon isotopes (Ferretti et al, 2003). Atmosphere samples were also collected on each sampling date, from each chamber or plot into 0.5 L flasks at ~1.5m above the ground, after flushing ~10 flask volumes. Diffusivity was calculated from soil temperature and moisture using an approach that accounts for inter- and intra-aggregate differences in diffusion rates (Potter etal., 1996).

Soil gas samples were analyzed within 24 hours of collection for CO2 concentration using an infra-red gas analyzer (Model LI-6251, LICOR, Inc., Lincoln, NE), with a precision of ±3pmolmol_1 over the concentration range 360-8000 pmol mol-1; most soil gas samples were about lOOOpmolmol^1. Soil gas was analyzed for stable isotopes of CO2 using gas chromatography-isotope ratio mass spectrometry (GC-IRMS; Isoprime model, Micromass, UK) (Miller et al, 1999). Tests indicated that standards stored in these greased syringes kept for up to one week without significant leakage or isotopic exchange. To ensure a linear response of the mass spectrometer, sample sizes varying from ~7 to ~ 250 pL of soil gas were injected into a carrier gas stream, which was further split before being introduced into the mass spectrometer. This allowed the peak height of the sample to be within ~10% of the peak height of the standard, and precision was better than ±0.1 %o. Flasks of atmospheric air were analyzed for CO2 mixing ratio by infra-red absorption (Conway et al., 1994) and for ¿>13C by dual-inlet mass spectrometry (Trolier et al, 1996).

We used the 'Keeling' plot approach to estimate the ¿13C value of soil respiration (<513Csr) (Keeling, 1958, 1961; Pendall et al, 2001b, 2003), in order to eliminate the influence of variable amounts of atmospheric CO2 that diffuse down into the soil profile. When <S13C is plotted against the inverse of the CO2 concentration, the y-intercept reflects the flux-weighted average <513C value of the biological source of CO2, resulting from root/rhizosphere respiration and microbial decomposition. The intercepts were calculated using geometric means to account for variability in both independent (CO2) and dependent (13C) variables (SokalandRohlf, 1995). We subtracted 4.4%o from the y-intercept values to account for kinetic fractionation during diffusion because the soil CO2 was sampled from within the soil rather than from the soil surface (Cerling, 1984; Amundson et al, 1998). This approach assumes that the biological end-member 513C value is constant with depth, reflecting a constant proportion of rhizosphere ('new') and microbial ('old') components.

In order to estimate Rh and ultimately NEP from Eq. 11.6, the soil respiration flux (SR) was partitioned into new and old components using a two-component mixing model similar to Eq. 11.8 (Pendall et al, 2001b, 2003):

where Frh is the proportion of CO2 generated by microbial decomposition of soil C. The main difference from Eq. 11.8 is that old end-member values (513CRh) were not taken from bulk soil C analyses, because a portion of bulk C is 'passive,' or unavailable to microbial decomposition. We instead used a biophysical fractionation approach to reflect decomposition of pre-existing SOC that was available to microbes. Soil samples collected at the end of the 1999 growing season were subjected to long-term (300 days) laboratory incubations (Paul et al, 2001). During the 'slow-pool' phase, after day 100, an approximate steady-state ¿13C value was achieved. Average <513C values (n = 4 for each of two depths) of CO2 evolved during decomposition of slow-pool C were used as the old C end-member signatures (<513Cj>h). For AC, <513Crh was -17.0 ± 0.47; for EC it was -21.9 ± 1.52; for NC it was -17.8 ± 0.48 (Pendall et al, 2001a).

The new C end-member signatures (<513Cnew) were determined as described above for rhizodeposition (Table 11.1). An offset of at least 4%o between old and new end-members allowed partitioning on all treatments. In pure C3 ecosystems, it is possible that photosynthate delivered to roots varies at daily to weekly timescales, depending on atmospheric vapor pressure deficit, leading to variability in the isotopic composition of soil-respired C02 (Ekblad and Hogberg, 2001).

The decomposition flux was estimated by multiplying Frh by the total soil respiration flux. Uncertainties in estimating the proportion of soil respiration from decomposition were determined by accounting for variability in all of the components, including 513Csr, <513CRh, and 513Cnew, following a first-order Taylor series approach (Phillips and Gregg, 2001). Uncertainty in 513Csr was based on least squares standard errors of the Keeling intercepts (which were determined by geometric means; Pataki et al., 2003). Uncertainty for <513CRh was estimated as the standard error of the average of 4 incubation subsamples over 7 sampling dates between 100 and 250 days after the start of the incubation experiment.

Was this article helpful?

0 0

Post a comment