| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Geological Sciences, Univ. of Saskatchewan, Saskatoon, SK, Canada, S7N 5E2
b National Water Research Inst., Environment Canada, 11 Innovation Blvd., Saskatoon, SK, Canada, S7N 3H5
c Dep. of Earth and Atmospheric Sciences, 126 Earth Sciences Bldg., Univ. of Alberta, Edmonton, AB, Canada, T6G 2E3
* Corresponding author (tkb130{at}mail.usask.ca)
Received 11 January 2006.
 |
ABSTRACT |
|---|
 TOPEXECUTIVE SUMMARY ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
18O of O2 as a constraint to create a dynamic link between these three parameters under transient conditions. The gas transport modeling was used to quantify the controls of biogeochemical processes and parameters (i.e., temperature and moisture content) on vadose zone distributions of CO2 and O2 gas concentrations. The investigation was conducted on a 3-m-thick, unvegetated, fine-sand vadose zone located in northern Alberta, Canada (56°40'N, 111°07'W). Using the modeled molar ratio of surface fluxes for O2 and CO2, the change in reaction rate for a temperature change of 10°C (Q10), moisture content at maximum reaction rates, and biogeochemical discrimination against consumption of 18O16O (
k), we determined that organic C oxidation by microbial respiration was the predominant mechanism consuming O2 and producing CO2. The mean k was determined to be 0.973, suggesting that subsurface respiration was via the alternative oxidase pathway, which may be common in cold climates. Modeling revealed that the moisture content of a moist, surficial clayey sand layer (0.1–0.3 m thick) had a dramatic effect on pore-gas CO2 and O2 concentrations and on 18OO2. The vadose zone in this study was at an unvegetated site to simplify the model application; however, it can be modified to include root respiration and applied to natural vadose zones to help quantify the role of subsurface respiration in global O2 and C budgets.
Abbreviations: CT, total carbon • TIC, total inorganic carbon • TOC, total organic carbon • VSMOW, Vienna Standard Mean Ocean Water
|
INTRODUCTION |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Globally, O2 gas is the most important oxidizing agent and biogenic gas (Stumm and Morgan, 1981). Terrestrial soil and subsurface respiration is believed to be the major O sink in the global O2 cycle, accounting for approximately 75% of global O2 consumption (Lasaga and Ohmoto, 2002). Far less scientific attention has been given, however, to the flux and controls on O2 in vadose zones compared with CO2 (c.f., Birkham et al., 2003; Jaynes et al., 1983; Wood and Petraitis, 1984).
Further, our ongoing and yet-unpublished studies of dissolved O2 in oxic groundwaters indicate that there are consistent but anomalous shifts in its 18O. These shifts appear to be controlled by yet-undefined vadose zone processes. As a result, more detailed investigations into the O2 in the vadose zone are warranted. The stable isotopic composition of atmospheric O2 (18OO2), while fairly constant throughout geologic timeframes, is controlled by the global balance between primary productivity and respiration processes in the soil and vadose zone (Angert et al., 2001; Bender et al., 1994). Soil and subsurface biological (e.g., microbial and root respiration) and mineral oxidation processes using O2 preferentially consume lighter isotologues (i.e., 16O16O), resulting in 18O enrichment of residual O2 (Hoefs, 1973; Schleser, 1979; Taylor et al., 1984).
Despite the fact that O2 concentrations in the subsurface may vary seasonally and with drought and rainfall events, to our knowledge only Angert et al. (2001) and Davis et al. (1986) considered transient models of O2 concentrations ([O2]) in unsaturated soils. We are unaware of any published studies that have combined transient [O2], [CO2], and 18O models for quantifying the biogeochemical reactions that govern the flux and production of subsurface biogenic gases. The objective of this research was to use a transient model for [O2], [CO2], and 18OO2 profiles in a simplified vadose zone as a first attempt to determine the rates and controls of biogeochemical reactions, as well as to quantify the parameters governing soil 18OO2 (including diffusive and consumptive fractionation). We measured in situ [O2], [CO2], 18OO2, temperature, and moisture content during a 990-d period. Surface flux of CO2 was measured on one occasion to supplement and verify the vadose zone [CO2] model. While we acknowledge that our use of an unvegetated site that has been impacted by mining activities is not representative of natural soils and vadose zones, the approach here can be applied to both natural and industrial settings to assess the impact of landscape disturbances and the efficacy of remediation and reclamation strategies.
Study Site
The study site was a simple 3-m-thick vadose zone constructed of fine tailings sand located at the Syncrude Canada Ltd. mine (56°40'N, 111°07'W, continental midlatitude climate) in northern Alberta, Canada, 480 km northeast of Edmonton and 40 km north of Fort McMurray, and is currently the site of a major industrial land restoration project. The mean annual precipitation and mean average daily temperature for 2001, 2002, and 2003 were 358.6, 422.3, and 496.2 mm and 2.1, –0.3, and 0.6 °C, respectively (Environment Canada, 2006). The Syncrude mine produces synthetic crude oil from excavated oil sands. After processing the oil sands, a slurry of water, sand, and residual hydrocarbon fines were transported by pipeline to the Mildred Lake Settling Basin (MLSB), where the solid particles (typically 
95% fine sand, 4% silt, and 1% fines) were separated by gravitational settling (MacKinnon, 1989). Sand deposition at the study area was completed in the early 1990s and a thin surficial layer of fine-grained soil was placed on top of the sand in 1995 to control the deposition of wind-blown sand. This surficial layer was typically 10 to 30 cm thick, but was absent in some isolated areas. No soil and vegetation reclamation had been undertaken on this site, so this data provides baseline data for future reclamation efforts.
|
MATERIALS AND METHODS |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Concentrations of O2 and CO2 were measured in the field using a GasTech CO2 Series portable gas analyzer (Gas Tech, Newark, CA) on 21 Feb., 13 Apr., 11 May, 29 May, 25 June, 18 July, and 18 Sept. 2001, and 20 Feb. and 8 May 2002. An internal pump was used to purge the tubing. The accuracy of the gas concentration measurements was ±0.5% by volume for [O2] and ±10% of the display reading for [CO2].
Pore-gas [O2] and [CO2] were measured with a field-portable Agilent MicroGas Chromatograph (M200 Series) gas chromatograph (Agilent Technologies, Santa Clara, CA) on 16 June, 13 July, 26 Aug., and 29 Oct. 2002, and on 19 Feb., 13 June, 8 July, 2 Oct., and 5 Nov. 2003. Concentrations of CH4 and N2 were also measured for completeness but are not reported here; [CH4] concentrations were below the detection limit (0.0005% v/v) and [N2], as expected, composed the majority of the remaining gas. Gas samples were collected in 140-mL syringes and injected manually into the gas chromatograph. A 4-m-long, 5-Å molecular sieve column and Ar carrier gas were used to separate O2 and N2 (column temperature of 70°C, carrier gas pressure of 206.8 kPa) and an 8-m Poraplot column and He carrier gas were used to separate CO2 and CH4 (column temperature of 65°C, carrier gas pressure of 103.4 kPa). Gas concentrations were measured with a thermal conductivity detector at a precision and accuracy of ±2 and ±5%. Both the gas analyzer and portable gas chromatograph were calibrated with atmospheric air (20.9% v/v O2 and 78.1% N2) and reference gas mixtures of 3% (v/v) O2 and 11.6% CO2, and 10% O2, 1.58% CO2, and 0.52% CH4 (balance of reference gas was N2).
Carbon Dioxide Surface Fluxes
Surface fluxes of CO2 were measured at 10 locations (within 5 m of the gas probes), at approximately 1500 h on 13 June 2003 by placing 160-mm-diameter closed chambers on the ground surface, at 0.5-m intervals, and measuring [CO2] in the chambers after 5 and 20 min. A more detailed description of the installation, sampling, and gas chromatography analyses of this surface flux technique is described by Kabwe et al. (2005). Although chamber-based surface flux techniques do interfere with normal surface flux transfer processes, they have been proven to be reasonably reliable and accurate for measuring CO2 surface flux for discrete areas (Kabwe et al., 2005). Knowing the volume and ground surface area covered by the collection chamber, the measured change in [CO2] in the sealed chamber with time was used to calculate the surface flux of CO2.
Pore Gas Oxygen Isotope Ratios
Gas samples for 18OO2 analyses were collected from the gas probes by suction into pre-evacuated 125-mL glass serum bottles that were crimp sealed with butyl blue stoppers (Bellco Biological Glassware, Vineland, NJ). Samples were collected on 18 July 2001, 8 May, 16 June, 13 July, 26 Aug., and 29 Oct. 2002, and 19 Feb., 13 June, 2 Oct., and 5 Nov. 2003. Birkham et al. (2003) provide a detailed description of the sampling procedure. The samples were analyzed for 18OO2 using a Micromass Optima isotope-ratio mass spectrometer (Micromass, Manchester, UK) using the method described by Wassenaar and Koehler (1999). Analytical error (SD, n = 10) for 18O was ±0.18
based on repeated injections of air. All results are reported relative to air which has a value of 23.5 relative to Vienna Standard Mean Ocean Water (VSMOW).
Drilling and Solids Sampling
Solid sediment samples were collected at the site on two occasions. In February 2001, five samples (200 g each) were collected using a 150-mm-diameter, hollow-stem auger and split-spoon sampler at depth intervals of 1.5 to 2.0, 3.0 to 3.5, 3.5 to 4.0, 4.5 to 5.0, and 6.0 to 6.5 m. Samples were stored cold and transported to Norwest Labs in Edmonton, AB, Canada, for grain-size distribution analyses. On 7 Oct. 2003, 24 samples (500 g each) were collected using a hammer-type drill rig. The samples were air-lifted up a 150 mm, double-walled casing, which was pounded into the sand with a one-cylinder diesel hammer. The samples were collected in plastic bags and placed in a freezer within 1 h of sampling. Samples were collected from depths of 0.15, 0.9, 1.5, 2.1, 2.7, 3.4, 4.0, 4.6, 5.2, and 5.8 m. These samples were used for C form content assays (see below).
Solids Chemistry
Sand samples collected in October 2003 were freeze-dried, homogenized in a rotary grinder, and analyzed for total carbon (CT) and total organic carbon (TOC). The CT concentrations were analyzed with a LECO CNS (LECO Corp., St. Joseph, MI) and TOC concentrations were measured with a LECO CR-12 Carbonator analyzer. The CT concentrations were determined using a soil standard with 3.84% (w/w) C, and a furnace temperature of 1100°C. The TOC concentrations were measured using a sucrose standard (42.10% organic C) and a furnace temperature of 840°C. The lower detection limit was 0.01% with an accuracy of ±1%. Total inorganic carbon (TIC) concentrations were calculated by subtracting TOC concentrations from CT concentrations.
Moisture Content
In situ moisture content profiles were measured using both a 503 DR Hydroprobe (neutron probe. Campbell Pacific Nuclear, Concord, CA) and a Diviner 2000 Series II system (Sentek Sensor Technologies, Stepney, SA, Australia). Moisture contents were measured with the Diviner system on 10, 18, and 26 June, 5, 12, and 14 July, 28 Aug., and 29 Oct. 2002, and 19 and 25 June, 2 and 12 July, and 2 Oct. 2003. To use the Diviner 2000 system, thin-walled 51-mm-diameter polyvinyl chloride (PVC) access tubes were installed to a depth of 2 m below ground surface at two locations (Diviner-1 and -2) within 4 m of the gas probes on 10 June 2002. To ensure close contact with the tailings sand, the access tubes were pushed into hand-augered holes of approximately 60-mm diameter. To measure moisture content depth profiles, a capacitance sensor attached to a 1.6-m rod was pushed down and then pulled up the installed tubes. The volumetric moisture contents were recorded by the Diviner 2000 computer at 10-cm intervals using a factory calibration for a sandy loam. For calibration purposes, 37 samples of sand were collected at 10- to 15-cm intervals for moisture content analysis during the installation of the Diviner access tubes.
A neutron access tube was installed on 10 June 2002 (within 2 m of the gas probes) and neutron counts were measured on 3 Oct. 2002 and 20 Feb., 10 June, 2 Oct., and 4 Nov. 2003. A linear calibration equation was used to calculate volumetric moisture contents from the measured neutron count ratio. The neutron probe was calibrated using moisture contents measured with the Diviner 2000 system.
In Situ Bulk Density and Porosity
In situ bulk densities were measured at three locations within 4 m of the gas probes at depths of 0.1 m (clayey sand) and 0.4 m (sand) on 2 Oct. 2003. Shallow holes were dug to create an even soil surface. A 44.5-mm length of 81.54-mm i.d. PVC pipe was then pushed into the soil until the top of the pipe was level with the ground surface. The PVC pipe was dug out of the sand and the sand on the bottom end of the pipe was carefully scraped until it was level with the pipe. The sand in the PVC pipe was collected in a plastic bag, weighed, and then dried overnight at 110°C. In situ dry bulk density was calculated by dividing the mass of dry sand collected by the volume of the PVC pipe mold. Total porosity was calculated assuming a specific gravity of 2.65 for the sand and 2.70 for the clayey sand.
Depth to Water Table
A piezometer with a 30-cm screen, constructed from 25.4-mm-diameter PVC pipe, was installed in a hand-augured 51-mm-diameter hole on 16 June 2002 to a total depth of 4.5 m below ground surface. Sand collapsed around the piezometer from 4.5 to 4.0 m and the annular space from 4.0 to 3.0 m was backfilled with sand (collected from the auger). The remainder of the annular space was backfilled with bentonite. The depth to the water table was measured on 16 June, 27 Aug., and 29 Oct. 2002, and 10 June, 2 Oct., 4 Nov., and 20 Feb. 2003.
Temperature
Temperatures were monitored by measuring the resistance across thermistors installed at 1, 2, and 3 m below the ground surface on 8 May 2002. The thermistors were placed in a 25.4-mm-diameter PVC pipe installed in the tailings sand. The pipe was filled with vegetable oil to improve the thermal contact with the tailings sand. The thermistors were monitored on 15 May, 18 June, 28 Aug., and 29 Oct. 2002, and 20 Feb., 10 April, 10 and 25 June, 2, 12, 21, and 31 July, 2 Oct., and 4 Nov. 2003. Daily average surface temperatures were measured at the site with a thermistor and recorded with a Campbell Scientific CR10X data logger (Campbell Scientific, Logan, UT).
Modeling the Oxygen and Carbon Dioxide Concentrations and Oxygen Isotope Ratio Profiles
The evolution of [O2], [CO2], and 18OO2 depth profiles was simulated using a one-dimensional numerical model developed and described by Hendry et al. (1999), and used by Birkham et al. (2003), Hendry et al. (2002), and Lee et al. (2003). The model has also been validated by comparison with the analytical solution of one-dimensional diffusion for simple steady-state cases. The model assumes that gas transport is by molecular diffusion, which is consistent with previous subsurface gas studies (c.f., Wood et al., 1993; Hendry et al., 1993; Keller and Bacon, 1998). The model uses the finite-element method to solve the one-dimensional diffusion equation:
 | [1] |
 | [2] |
 | [3] |
b = dry soil bulk density, G* = gas consumption or production rate (per unit mass of dry soil), z = depth below ground surface, t = time, nw = water-filled porosity, H = dimensionless Henry's law coefficient (c/cw, where cw is the equilibrium water concentration in consistent units), co = atmospheric concentration (20.9% v/v for O2 and 0.036% for CO2), and L = thickness of the unsaturated zone. Numerical modeling was performed using a 3.0-m grid corresponding to the unsaturated zone thickness. A uniform element size of 0.01 m and time steps of 0.5 d were used after a sensitivity analysis (data not presented) indicated that these values were suitable discretizations. Measured moisture contents and calculated air-filled porosities were linearly interpolated in time and space onto the grid. Effective diffusion coefficients (D*) were estimated using the Millington (1959) tortuosity formulation and the Fuller et al. (1966) temperature correction. This method of calculating effective diffusion coefficients has been demonstrated to accurately estimate diffusive gas transport in several previous vadose zone studies (c.f., Hendry et al., 1999; Elberling, 2003; Tang et al., 2003).
The O2 consumption and CO2 production rates in Eq. [1] were described by the function
 | [4] |
 | [5a] |
 | [5b] |
For modeling purposes, curves were fit to the measured temperature data using:
 | [6] |
= radial frequency (2
times actual frequency), 
o = –to (where to = a time shift), and d = damping depth (Hillel, 1980).
The Levenberg–Marquardt nonlinear fitting method (c.f., Press et al., 1989) was used to obtain Go and b values in Eq. [4] and k values in Eq. [5]
that resulted in the best match between modeled and measured [O2], [CO2], [16O16O], and [18O16O].
After simulating transient [O2] and [CO2] depth profiles, the consumption and redistribution of 16O16O and 18O16O isotopologues were modeled individually, whereby c in Eq. [1] represented either [16O16O] or [18O16O]. Measured [16O16O] and [18O16O] profiles were determined from
 | [7a] |
 | [7b] |
18OO2 of the atmosphere is 23.5 (Kroopnick and Craig, 1972).
For simulating the transport of 16O16O and 18O16O, the values of Do were related by the diffusive isotope fractionation factor:
 | [8] |
D in air is 0.986, as determined by the gas kinetic theory (Reid et al., 1977). Because 16O is the dominant isotope in O2 (i.e., 99.76%), we assumed that 16Do was the same as that for O2. The 18Do was then determined using Eq. [8], knowing 16Do and D.
For [16O16O] and [18O16O] modeling, best-fit G* values were determined by the model and can be defined as
 | [9] |
 | [10] |
k) at each node in the model grid were determined using 16k and 18k from Eq. [9] and [10]:
 | [11] |
18OO2 (VSMOW) profile. Thus, 18O assays provided an additional and important constraint on O2 sources (air) and subsurface consumption and transport processes (respiration and diffusion). |
RESULTS AND DISCUSSION |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Mean bulk densities of the sand (n = 3) and clayey sand cover (n = 3) were 1.57 g cm–3 (SD = 0.04 g/cm3) and 1.74 g cm–3 (SD = 0.06 g cm–3). Assuming a specific gravity of 2.65 for the sand and 2.70 for the clayey sand, a mean total porosity of 0.41 was calculated for the sand and 0.36 for the clayey sand cover. For these same samples, mean water-filled and air-filled porosities were 11.5% (SD = 1.5%) and 29.1% (SD = 3.0%) for the sand, and 27.6% (SD = 4.4%) and 8.1% (SD = 5.8%) for the clay cover.
The mean TIC concentration was 0.25% w/w (n = 9, SD = 0.17%) and mean TOC concentration was 0.17% w/w (n = 9, SD = 0.06%). Concentrations of TIC and TOC (data not presented) were relatively uniform with depth.
Moisture Content Profiles and Depth to the Water Table
Measured depth profiles of mean volumetric moisture contents are presented in Fig. 1
. Elevated moisture contents (25–40% v/v) were measured in the surficial clayey sand layer (0–0.3 m) using both the neutron probe and Diviner 2000 and were attributed to the higher matric potential expected for finer grained soils (Hillel, 1980). Elevated moisture contents with depth from 1.6 to 3.0 m were attributed to the capillary rise from the water table. The depth to the water table ranged from 3.12 to 3.70 m during the study period (Fig. 1). We assume that the variation in the depth to the water table was the result of seasonal variation in moisture recharge. This is supported by the fact that the depth to the water table was greatest in midwinter of 2003, when infiltration would be expected to be low.
|
|
|
Modeled Carbon Dioxide Concentration Depth Profiles and Surface Fluxes
Best-fit modeled [CO2] are presented in Fig. 3A. As an example of the shape and change in [CO2] depth profiles with time, three depth profiles of measured and modeled [CO2] are presented in Fig. 4A
. Field [CO2] measured on 8 May 2002 were used as initial concentrations in the model, corresponding to the commencement of field temperature measurements. The range in moisture contents used for the best-fit modeled [CO2] is presented in Fig. 1. Modeled [CO2] were very sensitive to changes in higher soil moisture contents from 0 to 0.3 m and from 1.9 to 3.0 m. This sensitivity was expected because D* is much more sensitive to changes in soil moisture for high degrees of saturation (moisture contents >25% v/v). The modeled [CO2] were also sensitive to the thickness of the moist clayey sand layer near the ground surface. For example, modeled [CO2] were overestimated by up to 5% (v/v) if the moist clayey sand layer was from 0 to 0.3 m, compared to 0 to 0.1 m for the best-fit scenario. Davis et al. (2005) also reported that a near-surface moisture retentive layer had a dramatic impact on vadose zone gas concentrations.
|
|
The best fit for parameter b in Eq. [4] was 3.3. This corresponded to a maximum G* at a degree of saturation of approximately 60%, which is typical for systems dominated by microbial respiration (Hendry et al., 1999). Best-fit [CO2] were simulated for a parameter k (Eq. [5]) value of 0.037, or a Q10 value of 1.4, and Tref and Tmin values of 7 and 0°C. This Q10 was only slightly less than the typical range of 1.5 to 3 reported in the literature (Hendry et al., 1999) for soil microbial respiration. The low Q10 in this study may have been the result of relatively dry soil moisture conditions. Given that Q10 is negatively correlated to moisture content (Xu et al., 2001; Liu et al., 2006) and the moisture contents in this study were relatively low (typically <15%), the sensitivity of reaction rates to temperature (i.e., Q10) may have also been low.
Modeled CO2 production rates ranged from 0 to 0.6 mg C kg–1 soil d–1 in the moist surficial clay layer (0–0.1 m), 0 to 0.4 mg C kg–1 soil d–1 from 0.2 to 2.0 m, and 0.4 to 0.7 mg C kg–1 soil d–1 from 2 to 3 m. The higher production rates from 0 to 0.1 and 2 to 3 m corresponded to higher moisture content (Fig. 1). Carbon dioxide production rates were similar to the lower production rates reported by Hendry et al. (1999), but less than CO2 production rates reported by de Jong and Schappert (1972) and Hendry et al. (1999) during the summer at vegetated sites, where high rates of root respiration are expected. These high production rates corresponded to high summer temperatures, while production rates of 0 mg C kg–1 soil d–1 corresponded to soil temperatures below 0°C (Tmin in Eq. [5]). Carbon dioxide production in the capillary rise had the greatest influence on [CO2] depth profiles, as modeled [CO2] depth profiles changed very little when CO2 production rates from 0 to 0.5 m were arbitrarily set to 0 mg C kg–1 soil d–1. Wood et al. (1993) and Affek et al. (1998) also identified the capillary rise as a zone of increased CO2 production. The [CO2] was less sensitive to CO2 production at shallow depths because CO2 produced near the ground surface diffused to the atmosphere with relatively little impact on [CO2] at depth.
Carbon dioxide surface fluxes (n = 10) measured on 13 June 2003 ranged from 7 to 82 mmol C m–2 d–1 (mean = 30, SD = 21.5 mmol C m–2 d–1). This range in surface flux is probably the result of varying thickness and moisture content of the near-surface clayey sand. Spatial variation in the surface flux is expected, given the sensitivity of the effective diffusion coefficient to moisture content, particularly approaching saturated conditions (Fig. 5). For example, the calculated effective diffusion coefficient for the moist near-surface clayey sand varies by more than an order of magnitude for a change in volumetric moisture content <5% at moisture contents >25%. Furthermore, thin, yet localized, layers of accumulated fine-grained materials (e.g., from local ponding of water following precipitation events) may interfere with gas losses to the atmosphere, causing shallow lateral migration to open areas. Thus, higher flux values are likely to be more representative of actual bulk fluxes to the atmosphere.
Modeled CO2 surface flux for 13 June 2003 was 86 mmol C m–2 d–1. Given the variation in CO2 surface flux for relatively small changes in moisture content, and ground surface conditions, the modeled flux of 86 mmol C m–2 d–1 was considered to be reasonably close to the measured values. The overestimation of the modeled surface flux corresponded to a slight overestimation (0.7% v/v) in the modeled [CO2] at 0.5 m for this day. The good agreement between both measured and modeled CO2 surface flux and [CO2] suggested that the transient, diffusive model used in this study was a good approximation of field conditions.
Modeled Oxygen Concentration Depth Profiles
Best-fit modeled [O2] are presented in Fig. 3B. Modeled [O2] in Fig. 3B were simulated using the same moisture content and temperature profiles, and T parameters that were used for best-fit [CO2] in Fig. 3A. The good agreement of both measured and modeled [CO2] and [O2] using the same moisture contents, temperature, and T parameters indicated CO2 production and O2 consumption were directly related (i.e., microbial respiration using organic C). As an example of the shape and change in [O2] depth profiles with time, three depth profiles of measured and modeled [O2] are presented in Fig. 4B.
Oxygen consumption rates ranged from 0 to 1.33 mg O2 kg–1 soil d–1 in the moist surficial clay layer, 0 to 0.9 mg O2 kg–1 soil d–1 from 0.2 to 2.0 m, and 1.4 to 3 mg O2 kg–1 soil d–1 from 2 to 3 m. These consumption rates were simulated using parameters b (Eq. [7]) and k (Eq. [8]) of 3.1 and 0.041 (Q10 = 1.5), which were very similar to values used for best-fit [CO2] modeling. Modeled O2 consumption rates in this study were similar to values reported by Lee et al. (2003) for a sandy soil and lake sediments in northern Saskatchewan (similar climatic conditions as this study).
Modeled O2 surface flux (into the soil) for 13 June 2003 was 125 mmol O2 m–2 d–1, compared with the modeled CO2 flux of 86 mmol C m–2 d–1. Furthermore, during the entire simulation period, the modeled average daily O2 surface flux was 133 mmol O2 m–2 d–1, compared with 92 mmol C m–2 d–1 for CO2. The annual average daily ratio of the molar CO2 surface flux to molar O2 surface flux was 0.69, which is within the range 0.5 to 0.7 determined by Lee et al. (2003) for microbial respiration in sandy lake sediments and soils, and similar to the molar ratio of 0.77 reported by Drever (1997) for a typical complex carbohydrate. The molar ratio is, however, much greater than the ratio of 0.2 reported by Lee et al. (2003) for a system dominated by pyrite oxidation (O2 consumption) and subsequent carbonate mineral dissolution (CO2 production).
Measured and Modeled Oxygen Isotope Ratios
As with the temporal trends in [CO2] and [O2], field-measured 18OO2 (Fig. 3C) also varied seasonally. The most positive soil 18OO2 values were measured in the late summer and early autumn (up to 39 at 3 m on 26 Aug. 2002) when [O2] was low. In contrast, the least positive 18OO2 values were measured in late winter and early spring (as low as 24.1 at 0.5 m on 8 May 2002; Fig. 3C). The range in 18OO2 measured in this study was greater than most other soil gas values reported in the literature. This greater range in 18OO2 was a result of comparatively lower [O2] measured in this study. Angert et al. (2003, 2001), Aggarwal and Dillon (1998), Lee et al. (2003), and Severinghaus et al. (1996) reported in situ 18OO2 ranging from 20.5 to 31.2 at much higher O2 concentrations.
Best-fit modeled 18OO2 profiles are presented in Fig. 3C. With the exception of the 26 Aug. 2002 values for depths of 2.5 and 3.0 m (most positive 18OO2), the modeled 18OO2 fit the measured 18OO2 well. This is illustrated in Fig. 4C with three depth profiles of measured and modeled 18OO2 values. In comparison to Fig. 4A and 4B, a winter sampling date has been added to Fig. 4C to demonstrate that simulated winter 18OO2 did match measured values well, despite a poor match in the late summer. The goodness of fit of modeled 18OO2 with measured 18OO2 is presented in Fig. 6
. To better show the goodness of fit, average differences (absolute values) between the model and measured 18OO2, normalized to the measured 18OO2, were calculated. Including all data, the average absolute difference between modeled and measured 18OO2 was 6.25%. Excluding the 26 Aug. 2002 values for depths of 2.5 and 3.0 m, the average absolute difference between modeled and measured 18OO2 was 4.86%.
|
18OO2 for low [O2] concentrations (2.5–4% v/v) at depths of 2.5 and 3 m on 26 Aug. 2002 (Fig. 4C) is noteworthy. For this date, simulated 18OO2 values were 7 to 15 more positive than the field-measured 18OO2. The reason for this discrepancy is currently not clear, but we suggest two possibilities. First, at low [O2] (<5% v/v) there is significantly greater analytical error in 18OO2. This may have resulted in measured 18OO2 being less positive than the theoretical modeled values. Second, it is possible that at increasingly lower O2 concentrations there is progressively less microbial discrimination against 18O (Angert et al., 2003) and, therefore, field-measured 18OO2 values are less positive than modeled values.
The mean k value for best-fit modeled 18OO2 was 0.973 with a standard deviation of 0.005. This mean value was calculated by averaging the modeled k value for each model-grid element every 50 d during the 550-d simulation (n = 3300). Patterns relating modeled k values to temperature, moisture content, consumption rates, and depth were also examined; however, no trends could be discerned from these data.
Our mean modeled k value was consistent with the lower k values reported in the literature. Lee et al. (2003) reported k values from 0.973 to 0.976 for a sandy, unsaturated zone in a subarctic climate (57°11'N, 105°34'W). Angert et al. (2003) also reported low k values (mean of 0.977, SD of 0.004) for O2 consumption in a vadose zone in the interior of Alaska (63°49'N, 144°59'W). They attributed the low k values to O2 consumption by the alternative cyanide-resistant oxidase pathway, which is more active in respiration at lower temperatures (Reyes and Jennings, 1997; Jarmuszkiewcz et al., 2001; Angert et al., 2003).
|
SUMMARY AND CONCLUSIONS |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
18OO2 profiles. The modeling of each of these three biogenic gas and isotope profiles provided independent and complementary information regarding subsurface biogeochemical processes occurring in the vadose zone. For example, modeling only the transient O2 concentrations did not clearly indicate if the O2 consumption was by mineral or organic C oxidation via respiration. Modeling of both the CO2 and 18OO2, however, provided strong evidence that the predominant geochemical reaction was oxidation of organic C via microbial respiration. These types of simulations can be used to greatly improve our understanding and quantification of the terrestrial O2 and CO2 budgets, controls of ground surface O2 and CO2 fluxes, and the dynamics of subsurface O2 isotope fractionation, particularly in large areas of continents where large seasonal temperature variation and variable soil moisture regimes occur. |
ACKNOWLEDGMENTS |
|---|
18OO2 analyses. |
REFERENCES |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
18O and 17O of gaseous and dissolved oxygen. Anal. Chem. 71:4965–4968.[CrossRef]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
