Published in Soil Sci. Soc. Am. J. 68:1242-1248 (2004).
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
DIVISION S-3—SOIL BIOLOGY & BIOCHEMISTRY
A System for Studying the Dynamics of Gaseous Emissions in Response to Changes in Soil Matric Potential
Helene De Wevera,*,
David T. Strongb and
Roel Merckxb
a Vito, Boeretang 200, 2400 Mol, Belgium
b Lab. for Soil and Water Management, Katholieke Universiteit Leuven, Kasteelpark Arenberg 20, 3001 Heverlee, Belgium
* Corresponding author (heleen.dewever{at}vito.be).
 |
ABSTRACT
|
|---|
An incubation system was developed to study the dynamics of greenhouse gas emissions from soil cores at different matric potentials. It combines the facility to establish and easily alter matric potential in soil columns, contained in nine parallel incubation chambers, with the feature of highly time resolved measurements of NO, N2O, N2, O2, and CO2. Studies on chamber performance demonstrated that the repeated application of specified suctions or pressures gave highly reproducible water contents and that the water release curves were similar to those obtained with traditional methods. The system was used to examine how soils preincubated at –5 or –75 kPa for 90 d responded to matric potential increases during a subsequent anaerobic incubation. The data showed that as large pores were filled with water rates of both CO2 evolution and total denitrification increased up to –2 kPa. This is in contrast to aerobic studies that show a decline in CO2 evolution at such high matric potentials. The data suggest that C that was not decomposed in large air-filled pores during the aerobic preincubation became colonized only when these pores became water-filled in the subsequent anaerobic incubation and provided substrate for denitrifying organisms. These studies demonstrate the capacity of the incubation system to explore the complex interaction between soil structure and the microbially mediated soil processes responsible for gaseous N and C emissions. This interaction is crucial to understanding factors controlling gaseous emission rates but has historically been difficult to be examined directly.
Abbreviations: GC, gas chromatograph • TCD, thermal conductivity detector
|
INTRODUCTION
|
|---|
THE PRONOUNCED dynamics of soil gaseous N emissions is one of several characteristics that make their study difficult. It arises from the multiple interacting factors that regulate these processes, each of which can vary over time. Perhaps the most temporally variable of these factors is soil water content, and it is this factor that most often explains the extremely episodic peaks of gaseous N emission in the field (Rolston et al., 1984; De Klein and Van Logtestijn, 1996). Attempts to simulate gaseous N emissions related to denitrification suggest that soil water is the factor that is in most urgent need of improved definition (Marchetti et al., 1997).
Water content affects the supply of O2 to the soil in a complex way because complex soil structural features determine the distribution of water through the soil volume and therefore the diffusion of O2 to the microsite of microbial activity. These soil structural features, especially pore-size distribution, can vary widely even between soils with similar textures. The ratio of the production of N2O to N2 that is so crucial for greenhouse gas studies is sensitive to soil water status and is greatly modified by soil structural characteristics (Parton et al., 1996). Moreover, the C substrates that fuel denitrification may reside in different pore classes in different concentrations (Strong et al., 1998; Thomsen et al., 1999), which would further complicate the response of denitrification to matric potential. Examinations of how pore-size distribution affects gaseous N emissions are few. To be able to make such a study one needs to be able not only to easily vary the water content of a soil sample, but also to fill specified pore classes with water. Both the characterization of pore-size distribution of a soil sample, and the specification of which pore classes are to be water-filled in that soil sample, requires a facility by which soil matric potential can be accurately established and altered.
Soil incubation systems that have the capacity to regulate soil matric potential include those of Wagenet and Starr (1977), Monaghan and Barraclough (1993), Shelton et al. (1997), and Strong and Fillery (2002). Both the systems of Swerts et al. (1995) and of Scholefield et al. (1997) have the clear advantage of being able to monitor N2O and N2 emissions separately—a feature essential for studies relating to global atmospheric change. We have combined the system of Swerts et al. (1995) with its capacity to make highly time resolved measurements of several greenhouse gas species, with the system described by Strong and Fillery (2002) with its capacity to establish and easily alter matric potential in a soil column. The aim was to provide a system capable of studying the dynamism of greenhouse gaseous emissions in response to matric potential changes and with an understanding of the pore-size distribution of the soil.
We describe the system and some preliminary experiments that demonstrate its capacities in the study of soil gaseous emissions in response to changes in matric potential with an attempt to take better account of the complex soil chemical, structural, and biological interactions.
|
MATERIALS AND METHODS
|
|---|
Experimental Apparatus
The experimental system described in this report represents the union between the automated gas monitoring system described by Swerts et al. (1995) and the soil incubation chambers described by Strong and Fillery (2002) involving modifications to each. Figure 1 illustrates how nine incubation chambers were coupled with the gas monitoring system. Strong and Fillery (2002) have described the incubation chambers in detail. In summary these chambers consisted of a polyvinyl chloride (PVC) shell within which was placed a stainless steel tube (50.8 mm o.d., 100 mm length) that contained soil. By means of neoprene seals and brass plates at the top and the bottom of the chambers the unit was made gas tight. The critical feature that provided the facility to alter matric potential was a membrane filter with 0.45-µm pores, which acted as a surrogate for a ceramic plate. Each incubation chamber had a 250-mL effluent bottle constructed from plexiglass that contained solution with which the soil could be saturated. The base of the effluent bottle was connected to the base of the chamber and by varying its height the soil could initially be saturated and then drained to specified matric potentials by establishing a hanging head (maximum of 60 cm). For lower matric potentials, each chamber was connected to a gas line by means of a three-way valve. By this means a headspace pressure could be applied to individual soil cores via a precision gas regulator (±0.1 kPa). The top of each effluent bottle was connected with Tygon tubing (Metrohm, Berchem, Belgium) (impervious to gas) to the top of the incubation chamber so that water could flow into the soil without opening the system to the atmosphere to make room for the displaced gas. Before measurement of gaseous concentrations, the pressure in the incubation chambers was equilibrated with the atmosphere using a routine that precluded atmospheric gaseous entry. Soil temperature in the chamber varied ±0.1°C.
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 1. Incubation system showing the union of incubation chambers with the automated gas sampling system. Incubation Chamber 1 is selected by the 10-port valve, which allows gaseous recirculation by the membrane pump before sampling by the sample loops.
|
|
The gas monitoring system is described in detail by Swerts et al. (1995) and here we only summarize the system and report where changes have been made. The system consisted of two gas chromatographs and several automatically activated valves, which allowed automatic sampling of nine soil columns. Figure 1 illustrates how each of the nine chambers were connected by 1/8'' stainless steel tubing to the 10-port selection valve in the gas monitoring system. With the selection valve in the home position, the nine circuits containing the chambers were closed static systems. Port 1 of the 10-port selection valve was connected to a stream of high purity He. When the gaseous phase of a specific soil column was to be sampled, the selection valve was switched to the appropriate position connecting the selected chamber (e.g., Chamber 1 in Fig. 1) via the outer loop of the valve with the valve outlet (inner loop) and the injection pathway (see Fig. 1). The circuit thus obtained consisted of stainless steel tubing and included a membrane pump (ASF 7005, Puchheim, Germany, Viton membrane) and the injection valve in the load position fitted with 100-µL sampling loops, both of which were operated by a programmed integrator. The pump was normally off. When the soil chamber was selected, it was switched on for 15 min to recirculate the gas through the system while the injection valve was in the load position. The pump then switched off again, the injection valve was switched to the inject position and a 100-µL gas sample was injected into the carrier stream of each gas chromatograph (GC). The gaseous samples were first directed through glass tubing filled with magnesium perchlorate (H2O trap) before they were swept into the chromatographic columns (Fig. 1). One GC was fitted with porapak Q columns (Alltech, Deerfield, IL) with a 95 to 5% argon-methane carrier and an electron capture detector (ECD) to detect O2, NO, and N2O (<1000 ppm). The second GC was fitted with a thermal conductivity detector (TCD) having a He carrier, and was used to detect N2, O2, CH4, CO2, and N2O (>1000 ppm). In this GC, after N2 and O2 from the gaseous sample had passed through the porapak Q column, an automatically operated switching valve diverted these species to a molecular sieve for further separation while the remaining gases were directed to the TCD. Elution of all gases was complete after 15 min. Purchased premixed gaseous standards were run and GCs calibrated with an entire standard curve for each gas before the commencement of the experiment.
The volume of the sampling circuit was approximately 20 mL. Between each sampling the 10-port valve was switched to Port 1 and thereby flushed the sampling circuit with He for 2 min, preventing gas from one chamber contaminating an adjacent chamber. This practice also meant a loss of approximately 20 mL of gas from each chamber at each sampling, which was compensated by the injection of 20 mL of He. This loss was accurately determined and accounted for in calculations for gaseous production rates.
In the present experiments, anaerobic conditions were chosen for the incubations. These were obtained by flushing the soil chambers and associated tubing with He and were maintained by compensating the gas losses due to sampling with He. However, aerobic conditions are also possible in this incubation system.
Studies of Chamber Performance
Experiment 1 was undertaken to examine whether the incubation chambers could be used to repeatedly establish matric potential over a range of soil types. Four soils (further denominated as Soils 2, 6, 9, and 13) were collected from within a 300-m landform transect ranging across a hill slope to a creek flat (Lovenjoel, Belgium). These soils were collected from 3- to 15-cm depth and had the range of textures shown in Table 1. Soils were passed through a 6.3-mm sieve and retained in the field moist condition. The equivalent of 100 g of dry soil was placed into incubation chambers with three replicates for each soil amounting to 12 incubation chambers. Eighty milliliters of water was placed in each of the 12 associated effluent bottles. The soils were then saturated and drained three times over a period of 2 h. Soils were saturated by raising the effluent bottle 20 cm above the height of the core and were drained by lowering the bottle 50 cm below the top of the core to produce a 50-cm hanging head. After the last saturation, the hanging head was imposed for 3 h and then the soil was allowed to stand at room temperature for 2 d. This procedure was to allow the soil to settle so that a relatively stable pore-size distribution would develop from the disturbed soil. The soils were then resaturated after which a range of increasing suctions and headspace pressures were applied to the soil corresponding to –0.5, –1, –2, –5, –10, and –30 kPa. Since our objective was to use changed matric potentials to study biological processes, it was not possible to wait for complete equilibrium to occur at the higher matric potentials before a suction or pressure was removed and water content was determined as would be necessary in pure soil physical studies. Instead the equilibration times were chosen so that >95% of equilibrium water loss (see below) had occurred before the suction or pressure was removed. To apply potentials of –0.5 to –2 kPa, we used a hanging head applied for 2 h and in the case of –5 kPa the hanging head was applied for 3 h. Potentials of –10 and –30 kPa were applied by pressure in the headspace for 3 h. Between each matric potential the weight of each incubation chamber was recorded. The soils were then resaturated and this cycle of matric potentials was applied two more times. By monitoring the weight of the effluent bottles during the period that potentials were applied (equilibration period) we were able to determine that equilibration was achieved for the four highest values (–0.5, –1, –2, –5 kPa) but for the other potentials equilibration was not complete. The water content of the soils was determined from the weight of the incubation chambers after the application of each potential and the weight of the dry components.
View this table:
[in this window]
[in a new window]
|
Table 1. Selected characteristics of four soils from the Lovenjoel transect used in the incubation chamber. Analyses of all parameters except total N were performed in duplicate. Data are averages ± standard deviations.
|
|
Experiment 2 compared water release curves determined using either the incubation chambers described here or the traditional ceramic plate and pressure chambers. Fourteen soils were sampled from the Lovenjoel transect in the same way as described above, and air-dried. Partially ground wheat straw (0.6 g) was added to moist soil (100 g dry soil equivalent). The wheat straw consisted of wheat husks which had been dried and ground and had a particle-size distribution >2000 µm: 19%, 1000 to 2000 µm: 14%, 500 to 1000 µm: 14%, 250 to 500 µm: 40%, and <250 µm: 13%. One sample from each soil was watered to achieve an approximate matric potential of –5 kPa or –75 kPa and was incubated aerobically for 90 d at 25°C. The target matric potentials were chosen to span field capacity (–10 kPa) and were included to examine if incubation matric potential during C decomposition affects pore structure and therefore the final water release curve. Furthermore, they correspond to water filled pore neck diameters of 60 and 4 µm, respectively, and these pore classes were found to be important for activity of soil biota and C decomposition (Strong et al., 2004). Water release curves were then determined using incubation chambers adopting procedures outlined in Exp. 1. The target potentials were 0 kPa (saturation), –2, –5, –6, –10, –20, –30, and –75 kPa. The –75 kPa potential was maintained for 12 h. Water release curves also were determined using the traditional pressure plate and pressure chamber techniques (Marshall and Holmes, 1979) on a replicate set of soils.
Studies of Carbon Substrate Location for Denitrification
Two studies were conducted using the whole operational system. The aim of these studies was to examine if the C used as substrate for denitrifying organisms was more concentrated in pore classes that were air-filled during a preincubation compared with those that were water-filled during the preincubation. It was hypothesized that if C decomposition was slower in air-filled pores (due to lack of water), then the addition of water, that is, increased matric potential should result in elevated CO2 production from the decomposition of previously undecomposed C, as well as in stronger denitrification rates. In both studies, the preincubation was aerobic, but after installation into the gas monitoring system anaerobic conditions were chosen. This was so that increasing the matric potential did not alter O2 availability to the microorganisms, which would decrease CO2 production, but the only effect was that the larger pores had progressively more free water as matric potential increased.
In Exp. 3, Soil 2 and 9 (Table 1) were collected from the Lovenjoel transect. After sampling, the soils were maintained at field moisture content and sieved to 6.3 mm so as to cause minimal destruction of the soil structure. To 100 g of soil (dry soil equivalent) was added 0.6 g of wheat straw. The soil was mixed thoroughly and transferred into stainless steel tubes described above. According to previously determined water release curves, the soils were brought to either –5 or –75 kPa by adding water. They were then preincubated aerobically at 25°C for 90 d at these two matric potentials as a factorial with three replicates. The preincubation treatments are referred to as P–75 and P–5, respectively. After 90 d, the columns of soil were transferred into the individual incubation chambers and connected to the automatic gas monitoring system. Soils were saturated with a 500 mg L–1 NO3–N solution and all were brought to a matric potential of –75 kPa. Each soil column was flushed for 15 min with high purity He to bring the O2 concentration to levels <0.03% (Parkin and Tiedje, 1984). Gaseous emissions were monitored for 110 h after which matric potential was sequentially increased to –10, –5, –2, and –1 kPa and held at each of these values for 55 h. Between each matric potential the columns were flushed with He for 15 min to produce an anaerobic environment and to remove any remaining N gases and CO2, produced at lower matric potential. During gas monitoring, concentrations of NO, N2O, N2, CO2, and O2 were automatically determined.
Changes in gaseous concentrations were monitored in the absence of soil to determine the importance of gaseous entry and leakage in the incubation system. However, changes were very small relative to those observed in the presence of soil. More indications on the gas-tightness of the system are the following. Potential leaks of the system were evaluated by pressure testing and were found to be negligible. In anaerobic conditions, no O2 was ever measured in the gaseous atmosphere of the chambers. Due to sampling losses, N2 concentration levels decreased toward the end of the tests, rather than increase due to gaseous entry. Finally, regular checks of the system with a He detector never indicated the loss of He.
Experiment 4 also had a 90-d preincubation period, which was the same as for Exp. 3 except that nine replicate soil samples of the two soils were incubated at –75 kPa, after being amended with wheat straw. After a 90-d preincubation, the soil columns were installed in the gas monitoring system and three replicates were brought either to –75, –5, or –2 kPa and were held at this matric potential for 7 d during which time gas monitoring was conducted under a He atmosphere in the same manner as for Exp. 3. Analysis of variance and regression analysis was conducted using Minitab 10.
|
RESULTS AND DISCUSSION
|
|---|
Accuracy and Repeatability of Establishing Matric Potentials
The soil water contents after matric potentials were repeatedly applied to replicates of different soil types are presented in Fig. 2. Each point on the plot is the mean of three replicate soils and three determinations for a given matric potential. In general errors were small. The data indicated that differences between replicate soils were larger than differences between repeated measurements on the same soil sample. It has been previously shown that the repeated application of specified suctions or pressures to sandy soils results in highly reproducible water content (Strong and Fillery, 2002). The present data demonstrate that over a range of soil textures these chambers give similarly reproducible results.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 2. Relationship between soil water contents and matric potential, established for four soils of the Lovenjoel transect. Each point represents three replicate soils and three determinations for a given matric potential.
|
|
The accuracy of the matric potentials achieved by means of these chambers as compared with the tradition ceramic plate and pressure chamber is illustrated in Fig. 3. Whereas in all our studies kPa is the unit of choice for matric potential, here we chose pF to accommodate the wide range of matric potentials examined and to clearly illustrate the result. Although it was expected that the incorporated wheat residue would affect the water release curves after 90 d of incubation at different matric potentials, this turned out not to be the case. The incubation treatments of –5 or –75 kPa behaved as duplicates for this analysis and were further treated as such. The 14 soils were divided into well-aggregated and less well-aggregated soils. Each point in Fig. 3, hence, represents the mean of either six duplicated well-aggregated soils or the mean of eight duplicated less well-aggregated soils. In general, across the range of potentials examined the water content achieved in the incubation chambers was higher than achieved using the ceramic plates. The variation between samples was similar for both techniques. In the less well-aggregated soils where variation was less, the difference between the two techniques was more pronounced. These results are generally consistent with expectations. Because the incubation chambers are designed to examine how soil structural features affect biological processes, we imposed potentials on soils for relatively short periods so as to have minimal effect on these processes. Because of these relatively short equilibration periods, especially at the potentials of –10 kPa and lower, it is not expected that equilibration will be complete. The results in Fig. 3 illustrate that this is in fact the case with the water contents in soils incubated with the ceramic plate/pressure chamber apparatus being consistently a little lower than those measured using our incubation chambers. Nevertheless, since water release curves determined in the different ways were quite similar and often not significantly different, we considered that the performance of the incubation chambers was sufficient to enable meaningful investigations of the influence of soil pore-size distribution, and the soil water matric potentials on biologically mediated production of gasses.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 3. Comparison of water release curves determined with traditional ceramic plate and pressure chambers or with the incubation chambers. For the less structured soils each point represents the mean of eight duplicate soils, for the well-structured soils the mean of six duplicate soils.
|
|
Location of Substrate Carbon for Denitrification within the Pore System
Using the system described above we examined the influence of preincubation and incubation matric potential on C decomposition and denitrification in an attempt to examine if C substrate for denitrification was concentrated in pores that had been air-filled during the 90-d preincubation. The total cumulative production of the measured gases at the end of the incubation is presented in Table 2. Rates were generally higher for Soil 9 than for Soil 2. Probably, the higher organic C content of Soil 9 (Table 1) led to higher CO2 production and overall denitrification rates. In addition to that, textural differences probably resulted in a higher volume of water-filled pores for Soil 9 than for Soil 2 at the same matric potential, and the larger volume for denitrification may have led to higher rates.
View this table:
[in this window]
[in a new window]
|
Table 2. Cumulative gas production (µmol kg–1) during anoxic incubation (average ± standard error) of Soil 2 (sandy loam) and Soil 9 (silty clay) in the incubation chamber.
|
|
For Soil 2 (sandy loam) the production of NO and N2O gases was higher for the soils preincubated at –75 kPa (P–75) compared with those preincubated at –5 kPa (P–5), but in Soil 9 variability between replicates was higher and in general differences could not be identified. An example from Soil 2 of how NO, N2O, N2, and CO2 evolved over time as matric potential increased is presented in Fig. 4. In general, the response of gaseous production to changing the matric potential during the incubation was not pronounced. An exception to this was the slowing of NO production and concomitant increasing N2O production when the matric potential was increased from –75 to –10 kPa in the P–75 treatment. Moreover, in Soil 2, as matric potential increased the proportion of total denitrification emitted as NO + N2O decreased (Fig. 5). This result was more pronounced for the P–75 treatment than for the P–5 treatment. A similar trend was observed for Soil 9 but in this case, it was generally less significant. Statistical analysis showed that the proportion of gaseous N losses emitted as NO + N2O was significantly greater for the P–75 treatment (p < 0.005) and that the slope of the decline of this proportion as matric potential increased (Fig. 5) was significantly steeper for the P–75 soils than for the P–5 soils.
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 4. Cumulative production of N gases and CO2 (average of three replicates) during anaerobic incubation of Soil 2 preincubated at –75 kPa. The arrows indicate the increases in matric potential.
|
|
View larger version (50K):
[in this window]
[in a new window]
|
Fig. 5. Partially oxidized species of gaseous N (NO + N2O) as a proportion of total gaseous N emissions. Legend refers to matric potential during anaerobic incubation and x-axis refers to aerobic preincubation matric potential.
|
|
In the anaerobic environment imposed it is not expected that redox potential would change in response to increasing matric potentials during the incubation. The declining proportion of NO + N2O as matric potentials increased (Fig. 5) may therefore be due to an increase in the diffusion path for NO and N2O through the soil water phase before it enters the gas phase. This would provide a longer time for these species to be reduced to N2. Indeed, as matric potential increased, the increase in N2 production was 59 and 36% for Soils 2 and 9, respectively. Another explanation for the faster decline in the proportion of NO + N2O as matric potential increased in the P–75 treatment (Fig. 5) might relate to C availability. During the preincubation the P–75 treatment had a greater volume of large air-filled pores (all pores >4 µm) compared with the P–5 soil (all pores >60 µm). The P–75 soil may therefore have had more undecomposed C substrate in the large pores (Strong et al., 1998; Thomsen et al., 1999). When these soils were saturated with nitrate solution and then brought to –75 kPa, microbial activity and denitrification in the water-films of the gas-filled pores was probably proceeding and since O2 concentrations were very low, denitrifying organisms became active in these thin films. The much higher rates of NO production in the P–75 treatment compared with the P–5 treatment may then be due to the higher availability of C in the large pores of the P–75 treatment. However, no corresponding increase in CO2 production could be observed. Nor did total gaseous N production provide supporting evidence that as larger pores were filled with water under anaerobic conditions, C substrate residing in the pores would decompose more quickly and fuel denitrification (data not presented). An alternative reason for the declining proportion of NO + N2O during matric potential increase may therefore be that time elapsed. That is, as matric potential was increasing, time was proceeding and enzymes were being synthesized or derepressed. In an attempt to eliminate the confounding effect of increasing matric potential and time, Experiment 4 was undertaken in which matric potentials varied between soil columns but remained static within soil columns over a 7-d incubation period.
Figure 6 presents the mean rates of NO, N2O, N2, and CO2 emissions over the 7-d period for Exp. 4. Figure 7 shows mean rates of total N emission (Ntot), the proportion of NO + N2O of total gaseous N losses (NO + N2O/Ntot), the sum of electrons consumed which was calculated by assuming that the reduction of NO–3 to the three measured gaseous N species accounts for all electron consumption, and the electrons consumed per unit of CO2.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 6. Average production rates of (a) NO, (b) N2O (c) N2, and (d) CO2 in soils anaerobically incubated during 7 d at three different matric potentials, after a 90-d aerobic preincubation at –75 kPa.
|
|
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 7. Average effect of incubation matric potential on (a) total gaseous N emission, (b) electrons consumed (c) NO + N2O as a proportion of total N gases, and (d) electrons consumed per unit CO2 produced. Both soils were preincubated aerobically at –75 kPa and then incubated anaerobically for 7 d at three different matric potentials.
|
|
Production rates of both CO2 and total N gases increased as matric potential increased up to –2 kPa. Furthermore, the total number of electrons consumed, which theoretically reflects the amount of substrate C oxidized, also increased with increasing matric potential. This is in contrast to aerobic studies where increasing matric potentials beyond –5 kPa result in a decrease in CO2 production (Linn and Doran, 1984). Moreover, we observed in an associated aerobic study using Soil 2 that the production rate of CO2 at –5 kPa was 21.7 µmol kg–1 h–1 but when raised to –2 kPa it fell 30% to 14.8 µmol kg–1 h–1. In the present study the anaerobic environment ensures that the confounding effect of increasing O2 deficiencies as matric potential increased were excluded. The effect of increasing matric potential therefore relates to other effects derived from making free water available in previously air-filled pores. The data suggest that during the aerobic preincubation organic matter residing in larger gas-filled pores did not decompose as fast as that residing in water-filled pores and therefore undecomposed C may have been somewhat concentrated in the large soil pores. As matric potential increased however, and the water-films became thicker, decomposition processes increased in the fabric of the large soil pores—presumably due to increased diffusion rates of enzymes and greater motility of microorganisms. The implication of this is that it represents an additional possible mechanism by which soil may be able to sequester C, and which may vary between soils of different structure.
These data may also provide information about the chemical quality of the organic matter residing in large pores. Paul et al. (1989) demonstrated that denitrification activity is positively related to the amount of electrons available per mole of C in the supplemented C source. In Exp. 4, we observed that the electron supply per unit of CO2 emitted increased with matric potential. Although this was not generally statistically significant, it suggests that as the pores were filled with water, more and more efficient C sources became available to the decomposing microorganisms. This must have been the young relatively undecomposed and chemically labile organic matter residing in the larger pores.
The rates of NO and N2O emissions varied very little with matric potential in Exp. 4 (Fig. 6a,b) but the rate of N2 production greatly increased (Fig. 6c). The proportion of partially reduced species therefore decreased with increasing matric potential, which is consistent with observations in Exp. 3. The higher rates of total denitrification with increased matric potential suggest that an increasing proportion of the pore system was becoming active in denitrification as pores were filled and water films became thicker. It is possible that NO and N2O were being predominantly produced in water films of gas-filled pores. As matric potential increased, the pores that were responsible for this production were larger and larger air-filled (predominantly He-filled and therefore anaerobic) even though rates of production remained relatively unchanged. It is acknowledged that because NO can be produced abiotically from NO2–, the contribution of NO to denitrification may have been overestimated. However, because we never detected NO2– in similar experiments, this route is not expected to be of major importance.
In conclusion, these studies show that this incubation–gas monitoring system effectively combines the capacity to alter matric potential with the capacity to intensively monitor a wide range of soil gaseous emissions. It provides the opportunity to explore in detail the complex interactions between soil biological processes and the physical soil structure. The examples provided here show that crucial aspects of the physical and chemical organization of soil, which have implications for understanding gaseous N emissions can be explored in a reproducible manner, allowing parameterization of the controlling factors.
|
ACKNOWLEDGMENTS
|
|---|
This research was conducted in the frame of the ‘Global Change and Sustainable Development Programme’, supported by the Belgian State, Prime Minister's Service, Federal Office for Scientific, Technical and Cultural Affairs. D.T. Strong acknowledges the National Foundation for Scientific Research (FWO) and the Research Council of the Katholieke Universiteit Leuven for a visiting postdoctoral fellowship and a junior fellowship, respectively.
Received for publication August 11, 2003.
|
REFERENCES
|
|---|
- De Klein, C.A.M., and R.S.P. Van Logtestijn. 1996. Denitrification in grassland soils in the Netherlands in relation to irrigation, N-application, soil water content and soil temperature. Soil Biol. Biochem. 28:231–237.
- Linn, D.M., and J.W. Doran. 1984. Aerobic and anaerobic microbial populations in no-till and plowed soils. Soil Sci. Soc. Am. J. 48:794–799.[Abstract/Free Full Text]
- Marchetti, R., M. Donatelli, and P. Spallacci. 1997. Testing denitrification functions of dynamic crop models. J. Environ. Qual. 26:394–401.[Abstract/Free Full Text]
- Marshall, T.J., and J.W. Holmes. 1979. Soil physics. Cambridge University Press, Cambridge.
- Monaghan, R.M., and D. Barraclough. 1993. Nitrous oxide and dinitrogen emissions from urine-affected soil under controlled conditions. Plant Soil 151:127–138.
- Parkin, T.B., and J.M. Tiedje. 1984. Application of a soil core method to investigate the effect of oxygen concentration on denitrification. Soil Biol. Biochem. 16:331–334.
- Parton, W.J., A.R. Mosier, D.S. Ojima, D.W. Valentine, D.S. Schimel, K. Weier, and A.E. Kulmala. 1996. Generalized model for N2 and N2O production from nitrification and denitrification. Global Biogeochem. Cycles 10:401–412.
- Paul, J.W., E.G. Beauchamp, and J.T. Trevors. 1989. Acetate, propionate, butyrate, glucose, and sucrose as carbon sources for denitrifying bacteria in soil. Can. J. Microbiol. 35:754–759.
- Rolston, D.E., P.S.C. Rao, J.M. Davidson, and R.E. Jessup. 1984. Simulation of denitrification losses of nitrate fertilizer applied to uncropped, cropped, and manure amended field plots. Soil Sci. 137:270–279.
- Scholefield, D., J.M.B. Hawkins, and S.M. Jackson. 1997. Development of a helium atmosphere soil incubation technique for direct measurement of nitrous oxide and dinitrogen fluxes during denitrification. Soil Biol. Biochem. 29:1345–1352.
- Shelton, D.R., A.M. Sadeghi, G.W. McCarty, and A.R. Isensee. 1997. A soil core method for estimating N-mineralization and denitrification during cover crop decomposition. Soil Sci. 162:510–517.
- Strong, D.T., P.W.G. Sale, and K.R. Helyar. 1998. The influence of the soil matrix on nitrogen mineralisation and nitrification. II. The pore system as a framework for mapping the organisation of the soil matrix. Aust. J. Soil Res. 36:855–872.
- Strong, D.T., and I.R.P. Fillery. 2002. Denitrification response to nitrate concentrations in sandy soils. Soil Biol. Biochem. 34:945–954.
- Strong, D.T., H. De Wever, R. Merckx, and S. Recous. 2004. Spatial location of carbon decomposition in the soil pore system. Eur. J. Soil Sci. (in press).
- Swerts, M., R. Uytterhoeven, R. Merckx, and K. Vlassak. 1995. Semicontinuous measurement of soil atmosphere gases with a gas-flow soil core method. Soil Sci. Soc. Am. J. 59:1336–1342.[Abstract/Free Full Text]
- Thomsen, I.K., P. Schjonning, B. Jensen, K. Kristensen, and B.T. Christensen. 1999. Turnover of organic matter in differently textured soils–II. Microbial activity as influenced by soil water regimes. Geoderma 89:199–218.
- Wagenet, R.J., and J.L. Starr. 1977. A method for the simultaneous control of the water regime and gaseous atmosphere in soil columns. Soil Sci. Soc. Am. J. 41:658–659.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
S. Peth, R. Horn, F. Beckmann, T. Donath, J. Fischer, and A. J. M. Smucker
Three-Dimensional Quantification of Intra-Aggregate Pore-Space Features using Synchrotron-Radiation-Based Microtomography
Soil Sci. Soc. Am. J.,
May 29, 2008;
72(4):
897 - 907.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
