Effect of soil wetness on air composition and nitrous oxide emission in a loam soil

Effects of cropping (bare fallow, grass), heavy irrigation and N fertilization (0, 100 kg ha ) on soil air (at depths of 15 and 30 cm) and N 2 O emission were studied in a factorial two-year field experiment in southern Finland. The responses of soil mineral N, dry-matter yield and uptake of N were also determined. Irrigation was performed during two periods in 1993 and one period in 1994. During sampling periods, the soil moisture ranged from 11% to 45% (v/v) and soil temperature from 0 °C to 21°C. Unirrigated bare fallow contained 14–21% O 2 , 0.1–2% CO 2 and 0.2–100 μl l -1 N 2 O (1993 maximum 27 μl l ) in the soil air. Cropping and irrigation lowered O 2 (minimum 3–7%) and raised CO 2 (maximum 9%) in soil air, but fertilization had no effect. Irrigation raised N 2 O in the soil air if nitrate was present abundantly. Consequently, fertilization increased N 2 O especially in the irrigated bare soil, which still contained plenty of nitrate in autumn 1993. Cropping decreased N 2 O. The variation in soil air composition was partly explained by that in soil air-space. The average daily N 2 O-N emission amounted to 0–40 g ha -1 (mean 7 g ha ) and correlated positively with N 2 O concentration in the soil air.


Introduction
Soil moisture affects the gas composition of soil air in different ways.Soil organisms affected by moisture consume and produce gases which alter the composition of soil air.Such changes are counteracted by the gas exchange between the soil and the atmosphere.Increasing moisture causes decreasing air content.Because gases are conducted almost entirely through air-filled pores, gas exchange between the atmosphere and soil pores will slow down with increasing moisture.
Excess soil moisture is known to be detrimental to the growth of various field crops.The change in soil air composition may play an important role (Glínski and Stépniewski 1985).Low O 2 concentration in soil air has been shown to retard plant growth independently of soil wetness in experiments where soil air composition is artificially regulated (e.g.Jaakkola et al. 1990).
Soil air composition has been measured in a few field studies in the Nordic countries (Lind-Fertilization and sowing was performed on 24 May 1993, with a combine drill being used to place the fertilizer (calcium ammonium nitrate) in rows 25 cm apart and 8 cm deep in the middle of every second sowing-row interval.The seed consisted of a mixture of winter rye (Secale cereale), Italian rye grass (Lolium multiflorum), Persian clover (Trifolium resupinatum), timothy (Phleum pratense) and meadow fescue (Festuca pratense).Plants from the bare fallow plots were removed by hand as they emerged.
Using a tractor-mounted sprayer producing approximately 10 mm water per hour, the ploughed layer was saturated with water during three periods at different stages of the growing season.In 1993 the field was irrigated with 120 mm of water between 15 June and 2 July, and with 110 mm of water between 27 July and 10 August.In 1994 84 mm of water was given during 18-22 August.
Porous cups made of sintered polyethylene (pore size Ø 100 µm), one for each depth (15 and 30 cm), were inserted into holes made in each plot with an auger (Ø 3 cm) immediately after sowing and fertilization.The air-filled space around and inside the cup was about 20 ml.Sampling of soil air was performed about once a week during the growing period of 1993 and about once a fortnight during the growing period of 1994, mostly between 6 and 8 p.m.A 4 ml sample was taken with a glass syringe through a silicon rubber septum connected to the cup with a narrow Teflon tube (volume approximately 1 ml).After discarding the first sample, 5 ml was taken for analysis in the same way.The air samples were stored for no more than two days in the glass syringes and then analyzed for N 2 , O 2 , CO 2 , CH 4 , C 2 H 4 and N 2 O. Two interconnected gas chromatographs (Hewlett Packard 5890) were used.One of them was equipped with a Molecular Sieve 5A packed column (1.8 m) for N 2 , O 2 +Ar, CH 4 and C 2 H 4 and a Porapak Q packed column (1.8 m) for CO 2 .Helium was the carrier gas (35 ml min -1 ).The oven temperature was 80°C.The detectors (200°C) were TC for N 2 , O 2 +Ar and CO 2 , and FI for CH 4 and C 2 H 4 .The other GC had a Porapak Q packed column (1.8 m) and an EC detector (300°C) for N 2 O.The carrier (95% Ar, 5% CH 4 ) flow was 35 ml min -1 and the oven temperature 40°C.The Ar concentration in air was assumed to be 0.9% for calculating the O 2 concentration.When calculating the results the sum of determined gas concentrations was adjusted to 100%.
Steel cylinders, 16 cm in diameter and 25 cm in height, were inserted 10 cm deep into the soil in nine plots (Fig. 1) at the beginning of the experiment in order to monitor the emission of N 2 O from the soil.One cylinder was placed on each unfertilized plot, but two cylinders on each N treated plot in order to cover the fertilizer rows and the space between them representatively.At each sampling of the soil air each cylinder was covered with an air-tight rubber sheet for 40-60 min.The daily emission of N 2 O was calculated assuming a linear increase of gas concentration in the closed chamber from the measured mean ambient level (0.322 µl l -1 ) to the concentration measured at the end of sampling period.
Soil moisture in the 0-20 cm layer was monitored by TDR (Tektronix 1502B) plotwise in blocks I-III (Fig. 1) as often as the soil air was sampled.The soil temperature at depths of 15 and 30 cm was monitored with Pt100 probes in three plots (Fig. 1) in connection with air sampling.
Soil samples were taken at depths of 0-15 cm and 15-30 cm from the area between unirrigated and irrigated plots on 15 June 1993, just before the first irrigation.The same soil depths were sampled on 4 July plotwise in the blocks I, II and III.All plots were sampled at the abovementioned depths on 2 September 1993.For determination of mineral nitrogen the samples were extracted with 2 M KCl.Ammonium and nitrate in the extract were determined colorimetrically.
The plant stand was cut from the cropped plots on 1 September 1993 and 14 June 1994, taking plotwise a sample from an area of 0.45 and 0.25 m 2 , respectively.The plant samples were dried at 70°C and weighed.Total nitrogen was determined using the common Kjeldahl digestion procedure.

Statistical analysis
The treatments were partly arranged systematically in the blocks (Fig. 1).However, no systematic change in soil properties was apparent.Therefore, in comparing the treatments, an analysis of variance for a blockwise randomized design was made.In cases where the interactions were significant, individual treatment means were compared by Tukey's test.Correlation analysis was performed between the plotwise N 2 O emission data and corresponding N 2 O concentrations in the soil air.
In order to reduce the random variation of gas concentrations in the soil air samples, averages over three subsequent samplings were statistically analysed.Soil moisture data were analysed similarly.A logarithm transformation was used for the N 2 O concentrations to approach a normal distribution.

Results
Nitrogen application increased the crop yield and the N uptake in the first year (Table 1).Irrigation increased the first-year yield significantly only when nitrogen was applied.The nitrogen uptake did not respond to irrigation.
Mineral nitrogen in the top 30 cm of soil did not significantly respond to nitrogen application or cropping in the middle of June three weeks after fertilization and sowing (Table 2) although the mean concentration was generally higher in the fertilized plots.About three weeks later (4 July) nitrogen application resulted in a significant increase, while cropping had a decreasing effect.Only nitrate in the topmost layer (0-15 cm) was affected.Irrigation did not have any effect.The crop reduced the nitrate concentrations in late summer (2 September), as did irrigation, but to a lesser extent.Nitrogen application still had a small increasing effect.Concentrations in the cropped soil were rather low.
Nitrogen application did not significantly affect the soil moisture or the response of soil air composition to other treatments.Therefore, averages over both N rates representing cropping and irrigation treatments are given in Figures 2  and 3, as well as in Tables 3, 4, 5 and 6.Variations of soil temperature during both years were rather similar, considering the dissimilar observation periods (Fig. 2).The soil moisture varied during the first year between 16% and 44% in the non-irrigated soil.The soil was dry when the experiment started (beginning of June), gained moisture for a couple of weeks Table 1.Crop (C 1 ) yield and uptake of N in unirrigated (I 0 ) and irrigated (I 1 ), as well as in unfertilized (N 0 ) and fertilized (100 kg ha -1 N, N 1 ) soil.being largest in the spring.Thus, even the crop had its biggest effect at the beginning of growing season.Irrigation performed in late summer also had only a small but significant effect on soil air O 2 and CO 2 (Tables 4 and 5).N application did not have any significant effects on O 2 concentration (Table 4) or CO 2 concentration (Table 5) in soil air at either depth.
The O 2 concentration in the soil air decreased in the ploughed layer to below 15% only when the volumetric soil moisture exceeded 30% (Fig. 4).
The concentration of CH 4 in the soil air varied between 0 and 43 µl l -1 independent of treat-ment or sampling date (data not shown).The concentration of C 2 H 4 did not exceed the detection limit of 0.5 µl l -1 in any sample (data not shown).
The concentration of N 2 O in bare, unirrigated soil (control treatment, C 0 I 0 , Fig. 5) varied at various depths and N rates between 0.4 and 27 µl l -1 in the first year.In the second year, the range was between 0.4 and 100 µl l -1 , but the concentrations did not exceed 7 µl l -1 after May.The peak concentration in 1993 took place by the end of August; higher values were found deeper in the soil.N application raised the peak.The irrigation in June 1993 raised the concentrations for CO 2 caused by compaction in the trial of Simojoki et al. ( 1991) was 5% at depths of 25 and 50 cm, but it lasted longer deeper in the soil.In the experiment of Hansen and Bakken (1993) a maximum CO 2 concentration of almost 5% at depths of 7-12 cm was caused by soil compaction.Soil respiration consuming O 2 and producing CO 2 was, no doubt, the most important phenomenon altering soil air composition in the present experiment.Respiration in cropped soil was probably 2-3 times higher than in uncropped soil (Currie 1975).In the first year, irrigation and N fertilization improved plant growth, which in turn probably also increased soil respiration.On the other hand, if plant water uptake had increased air-filled porosity, the enhanced gas exchange would have counteracted the effects of respiration.But since cropping and N fertilization had only minor effects on soil moisture in this experiment, the significant effects of cropping on O 2 and CO 2 concentrations in soil air were mainly due to differences in respiration.Hansen and Bakken (1993) also found no effect of N fertilization in sandy loam under ley.In contrast, Stépniewski (1977), working with several plant species, found that doubling the mineral fertilizer dose improved soil aeration status in a cropped loamy sand soil.
Soil air composition deviated from atmospheric air composition most during a period of very high moisture content in soil simultaneously with high temperature, occurring in July-August 1993.In wet soil, under a vigorously growing, oxygen-consuming plant, the O 2 concentration in soil air at the bottom of plough layer dropped below 4%.The concentration increased again during the second half of August and thereafter, although no marked increase in air-filled porosity occurred.Obviously the decreased consumption of oxygen allowed this increase.A similarly decreasing deviation from atmospheric air towards the end of growing season was observed e.g. by Simojoki et al. (1991) in a pot experiment with barley.In their study, decreasing respiration could have been related mainly to the developmental stage of the plant.However, in the present study the influences of oxygen deficiency, due to its low content in soil air, and of simultaneously decreasing temperature were most obvious, because similar changes were observed in both bare and cropped soils.In addition, the respiration rate in grass does not change remarkably with development stage as in cereals.Soil respiration is generally regarded as an exponential function of temperature (Glinski and Stépniewski 1985).Yearly variations in soil res-Table 6.Average (geometric mean) concentration of N 2 O in the soil air at two depths during various periods, µl l -1 , and ratio (effect of treatment) between treated and untreated soils.27.6.-11.7.1993 1.8.-15.8.1993 22.8.-5.9.1993 22.8.-19.9.1994 Treatment  (Currie 1975).The decrease of O 2 was connected with an increase in CO 2 .However, they were not equivalent, the former being bigger than the latter.This difference has been observed in many other studies (e.g.Russell and Appleyard 1915, Glinski and Stépniewski 1973, 1985) and is explained by the rather high solubility in water of CO 2 as compared with that of O 2 .If there had been strong anaerobic production of CO 2 in the soil, the sum of O 2 and CO 2 would have exceeded 21% (Glinski and Stépniewski 1973).Probably, CH 4 concentration would have also increased.In the present study, increases in the CH 4 concentration and in the sum of O 2 and CO 2 concentration were never observed.
The difference of O 2 concentration between depths was many times larger than the corresponding difference of CO 2 concentration.This is partly explained by the better solubility of CO 2 in water, but the more rapid diffusion of CO 2 in soil water may also play a role (Greenwood 1970).
A plant hormone C 2 H 4 is involved in plant response to hypoxia (see Jackson 1991).In relatively wet soils concentrations of several µl l -1 have been observed both in field (Dowdell et al. 1972, Smith andDowdell 1974) and pot experiments (Simojoki et al. 1991), although variation has been great.In contrast, no C 2 H 4 was found in soil air in the present study.In other investigations low concentrations (0.5 µl l -1 or less) have been measured in aerobic soils (Otani & Ae 1993), but sometimes also in wet soils (Meek et al. 1986).Taken together the results suggest that C 2 H 4 is not a sensitive indicator of hypoxia in soil.
The N 2 O concentrations (0.2-100 µl l -1 ) were in the range reported by Hansen and Bakken (1993) in the topsoil of Norwegian field experiment on a sandy loam with different soil compaction and fertilization treatments.In the present experiment the highest concentrations (May 1994) were probably caused by the spring thaw (see Nyborg et al. 1997).
At the beginning of the experiment a lot of nitrate was present in both non-fertilized and fertilized plots.By the end of the first irrigation period no effect of cropping on N 2 O in soil air was found; however irrigation had increased the concentration markedly and N application had done so to some extent.Granli and Bøckman 1994).Nitrification probably also contributed to N 2 O production, since the concentration of N 2 O in soil air was generally higher than the ambient level even when the soil was not wet (Bremner and Blackmer 1978).When the soil was very wet due to irrigation and soil nitrate was depleted by the crop, lower than ambient concentrations were found, suggesting that N 2 O was reduced to N 2 more rapidly than N 2 O was produced by denitrification or diffused from the atmosphere to the soil.N application increased N 2 O in soil air during several periods throughout the first growing season.The increases were largest in bare soil, when it was wet either due to irrigation or rain.The biggest increase during a fortnight was on average 2.5-fold.This compares well with the results of Hansen and Bakken (1993) in uncompacted soil.They reported a much higher (100fold) increase due to mineral N application in compacted soil only.
The emissions of N 2 O were in the range reported in other studies made in comparable conditions with similar methods (e.g.Kaiser et al. 1996, MacKenzie et al. 1997).Substantially higher emissions were observed only occasionally in their studies.The correlation between emission and N 2 O in soil air was expectedly better at a depth of 15 cm than deeper in the soil.

Vol. 7 (1998): 491-505.
An air-filled porosity of 10% v/v is commonly regarded as critical for the satisfaction of the oxygen demand of the crop (Wesseling 1974).However, the actual critical value depends on oxygen consumption rate, pore size and pore continuity in the soil, and will therefore change with microbial activity, temperature, vegetation, soil type and soil structure.Values from 8% to 15% have been reported (Wesseling 1974, Hodgson and MacLeod 1989, Chan and Hodgson 1995).There are also cases where no critical value could be determined (Chan and Hodgson 1995).
Assuming that the soil was water-saturated during the rather long period of wetness in the first year, an estimate (probably an underestimate) of total porosity is 43-45%.It is almost certain that in the present study plant growth was not affected by the limited gas exchange when the soil moisture was below 30% v/v.This corresponded to air-filled porosities of at least 13-15% v/v.Periods of limited gas exchange according to the adopted criterion were in the unirrigated soil from the middle of July onwards in the first year and until the middle of June in the second year.In the irrigated soil there was shortage of air-filled porosity already in the latter part of June.The irrigation in August of the second year caused a period of restricted gas exchange after the middle of August.The periods referred to above agree quite well with the periods of decreased O 2 and increased CO 2 concentrations.The agreement with elevated N 2 O concentrations is also reasonably good, if the existence of nitrate in the soil is also considered.
In conclusion, it can be stated that in the conditions prevailing in southern Finland the O 2 in soil air might be markedly decreased in wet periods.Especially under crop stands the O 2 concentration may drop substantially to a level where the plants, if the low concentration persists, may suffer from O 2 deficiency (see Glínski andStépniewski 1985, Jaakkola et al. 1990).Increases of CO 2 , although occasionally very large, probably do not reach detrimental levels.In wet soil, denitrification causing losses of nitrate N and increasing N 2 O emission is obvious.Marked increases of CH 4 or C 2 H 4 in soil air do not seem to be probable in the conditions of the present study.

Table 4 .
Average concentration of O 2 in the soil air at two depths during various periods, and concentration increase due to treatments, %.

Table 5 .
Average concentration of CO 2 in the soil air at two depths during various periods, and concentration increase due to treatments, %.