The study was conducted in a summer maize field in Zhangfatai village of Hebei Province (38° 02′ 02′ ′ N, 115° 49′ 12′ ′ E) from 28 June to 14 October 2014. The site is relatively flat with an elevation of 19 ma.s.l. Winter wheat and summer maize rotation is the most commonly practiced cropping system. The region is characterized by a temperate monsoon climate with the annual mean temperature of approximately 12.8° C and the mean annual precipitation of approximately 554 mm. The top soil (0-20 cm) has a pH of 8.49, electric conductivity of 360.75 µ s/cm, bulk density of 1.42 g/cm³, and organic matter content of22.1 g/kg.

In this experiment, two irrigation practices were established: flood irrigation (FI) and drip irrigation (DI). FI field was fertilized twice with urea containing N (46%): 187.5 kg/hm² on 19 July and 150.0 kg/hm² on 11 August 2014, and each followed byFI. A total of 155.2 kg N/hm²was applied via fertigation before each irrigation event. DIfield used different irrigation frequencies and water amountsduring different growing stages (Table 1). Again, compared with FI, DI was performed at more regular intervals in much smaller amounts with less intense water supply and slower water movement. As shown in Table 1, different water-soluble fertilizers were applied during different growing stages and the fertilizer types were differentiated by the ratio of nitrogen/phosphorus/potassium. Specifically, fertilizer I (the ratio, 33/6/11) was applied during the elongation and bell stages and fertilizer II (the ratio, 27/12/14) was applied during the tasseling and filling stages.

Table 1

Table 1

Table 1 Amounts of irrigation water and fertilizers under DI (drip irrigation) used in different growing stages Growing stage Irrigation date Irrigationamount (m3/hm2) Fertilization amount (kg N/hm2) Fertilizer type Elongation stage 9 Jul 2014 180 24.7 I Bell stage 30 Jul 2014 420 99.0 I Tasseling stage 20 Aug 2014 225 60.7 II Filling stage 10 Sep 2014 225 20.2 II Note: Fertilizer types are differentiated by the ratio of nitrogen/phosphorus/potassium. The ratio is 33/6/11 in fertilizer I and 27/12/14 in fertilizer II.

Table 1 Amounts of irrigation water and fertilizers under DI (drip irrigation) used in different growing stages

The fields were cultivated with wheat crop before maize crop. The planting spacing arrangement of maize was 60 cm× 25 cm(row spacing× plant spacing). The DI pipes with 30-cm-spaced emitters were placed near maize rows along the row direction. The distances between pipes were more or less the same with row spacing.

Gases were collected usinga static opaque chamber with a dimension of 30 cm (length)× 30 cm (width)× 40 cm (height). The chamber was made of 8-mm-thick black acrylic material with a tinfoil reflecting film attached to the external surface. The opaque chamber could eliminate the influence of plant photosynthesis and also prevent the temperature inside the chamber from rising during measurement. In FI plot(Fig.1), chambers were placed on plant rows (OR) and between plant rows (BR). In DI plot, chambers were placed on irrigation pipes (OP) and between irrigation pipes (BP). Bare areas (BAs) were purposely reserved under both FI and DI treatments. Plants in the bare area (BA) were removed after sprouting and BAsunder FI and DI situations were irrigated and fertilized in the same waysas in FI treatment and in DI treatment, respectively. Each of the two treatments (FI and DI) had threereplicates or experimental plots and the size of each irrigation plot was 6 m× 50 m.

Fig.1

	Figure Option ViewDownloadNew Window
	Fig.1 Locations of the chambers under FI (a) and DI(b)treatments.FI, flood irrigation; DI, drip irrigation; OR, on plant rows; BR, between plant rows; OP, on irrigation pipes; BP, between irrigation pipes.

The aboveground standing parts of maize within the experimentalplots (40 cm× 40 cm) were cut to ground level, and plant litter was removed before the measurements. Then, stainless steelframes sharpened at the bottom were inserted into the soil to a depth of 5 cm 24 h prior to the measurements to reduce perturbation of the soil structureresulted from steel frame insertion. During the course of measurements, the chambers were temporarily mounted on frames and maintained gas-tight by filling the groove with water. The lid of each chamber was installed with a mini fan inside the chamber driven by a 12 V lead-acid battery to ensure a uniform gas concentration within the chamber and also with a thermometer to measure temperature inside the chamber as well as silica gel guide tube connectingwith a 100-mLsyringe and a three-way stopcock for gathering gas.

Gas sampling lasted for 30 min each time. On each sampling day, gas samples were taken immediately after chamber closure and collected into polyethylene-coated aluminum gas bags at 10-min intervals (at 0, 10, 20, and 30 min) for CO₂ gas concentration analysis. Meanwhile, when gas sampling is doneeach time, air temperature (T-air) was simultaneouslymeasured by a DHM2 mechanical ventilated thermometer. The soil temperatures at the depths of 0 (T-0), 5 (T-5) and 10 cm (T-10) around the chamber were measured using a SN2202 digital thermos detector immediately before the closure of chamber and after the removal ofchamber. Temperature in the interiorof the chamber was measured immediately after each gas samplingwitha thermometer installed on the chamber lid.

Following the advices by Xu and Qi (2001), samplings were conducted between 09:00 and 11:00 (Beijing time) when the measured efflux is advocated to be representative for the daily mean efflux. Flux measurements were made in 10-day intervals during the growing season. The CO₂ concentrations were analyzed using a LI-6252 infrared CO₂ analyzer (LICOR Inc., Lincoln, NE, USA) in the laboratory within 7 days after gas sampling. The CO₂ flux rates were calculated from the rate of change in the CO₂concentration in the air inside the chambers with time.

The Q₁₀value, namely temperature sensitivity of soil respiration, indicated the temperature dependence of soil respiration. The Q₁₀ values were calculated from the b coefficient (Q₁₀=e^10b) of exponential regression functions between the soil CO₂ effluxes and temperature factors.

Aftergas sampling, soil samples were also collected using a soil drill at the depths of 0-10 and 10-20 cm from each one of chamber locations. The soil microbial biomass carbon (MBC) was analyzed with the chloroform fumigation-extraction method using 0.5 M K₂SO₄ as an extractant (Wu et al., 2006). The soil dissolved organic carbon (DOC) was extracted by deionized distilled water. For MBC analyses, the sub-samples were stored at 4° C; and for DOC analyses, the sub-samples were stored at -20° C. The air-dried samples were hand-ground through a 100-mesh sieve for total organic carbon (TOC) analysis. The concentrations of MBC, DOC and TOC were all analyzed by a TOC analyzer (Vario TOC Cube, Elementar, Germany). The soil water contents(SWCs) at the depths of 0-10 and 10-20 cm were determined at 105° C for 24 husing the oven-drying method.

According to the relative area of each location per irrigation treatment, the total emission was the averagedtotal emissionsof OP and BP under DI and of OR and BR under FI.The cumulative emissions for each irrigation type and each location were calculated by successivelinearinterpolationof all measurement times per day usingMatlab7.0. Graphs were prepared using Origin 8.5. Pearson correlation coefficients between the soil environmental factors and the CO₂ effluxes were calculated, and one-way ANOVA andpaired sample tests were used for statistical comparisons across locations and between irrigation practices (FI and DI) using SPSS17.0.

The soil CO₂ effluxes under FI and DI showed similar temporal patterns(Fig.2). Thefluxes decreased before the elongation stage, rapidly increased and reached the peak at the bell stage, and then declined until harvest. The daily CO₂ effluxes changed in the range of 233.86 to 1026.85 and 257.72 to 992.73 mg/(m²• h) under DI and FI, respectively. Over the entire growing season, the mean CO₂ effluxes were 694.66 mg/(m²• h) under DI and 609.54 mg/(m²• h) under FI, and no significant diurnal difference was detected between the two irrigation practices. In all, the cumulative CO₂ emissions at the field scalewere 1959.10 and 1759.12 g/m²under DI and FI, respectively.

Fig.2

	Figure Option ViewDownloadNew Window
	Fig.2 Temporal dynamics of soil CO2 effluxes under drip irrigation (DI) and flood irrigation (FI) during the maize growing seasonin 2014. Vertical bars indicate standard errors (n=6).

In the FI treatment, the CO₂ effluxes in the reserved bare area (BA) were larger than those in other locations (i.e., maize plots) before the elongation stage, and then the CO₂ effluxes on plant rows (OR) were larger than those in the BA and also than those between plant rows (BR) (Fig.3). The average CO₂ effluxes under FI for OR, BR and BA were 699.72, 549.89 and 601.86 mg/(m²• h), respectively. Thedifference between OR and BR is significant according tothe paired-samplet-test (P< 0.05). In the DI treatment, the CO₂ effluxes on irrigation pipes (OP) were significantly larger than those between irrigation pipes (BP)and in the BA (P< 0.01). The average CO₂ effluxes for OP, BA and BPunder DI were 774.33, 622.52 and 617.86 mg/(m²• h), respectively. The proportion of root respiration contributing to the soil CO₂ emissionson OR under FIranged from 17.49% to 46.66%, and the proportionon irrigation pipe (OP) under DI ranged from 3.45% to 46.55%.

Fig.3

	Figure Option ViewDownloadNew Window
	Fig.3 Spatial variations in soil CO2 effluxes in the maize field underflood irrigation (a) and drip irrigation (b) during the maize growing season in 2014. BR, between plant rows; OR, on plant rows; BA, bare area; BP, between irrigation pipes; OP, on irrigation pipes.Vertical bars indicate standard errors (n=3).

The cumulative CO₂emissions for OR, BR and BA under FI were 2034.17, 1575.92 and 1712.42 g/m², respectively. The cumulative emissions for OP, BP and BA under DI were 2205.56, 1720.55 and 1704.41 g/m², respectively. It means that the difference between DI and FI is not significant. The soil CO₂ emissions in the BA were larger than those between BR under FI, whereasthe soil CO₂ emissions in the BA weresmaller than those between BP under DI.

3.3.1 Soil water content (SWC)

SWC at the depth of 0-10 cmvaried between 14.14%and 29.51% under FI and between 14.54%and 27.92% under DI(Fig.4). SWC at the depth of 10-20 cm varied between 17.89%and 32.76% under FI and between 13.73%and 28.22%under DI. SWCexhibited a decreasing order of BA> BP> OP under DI and a decreasing order of BA> BR> OR under FI at each one of the measured soil depths. Generally speaking, SWC under FI was larger than that under DI at thesame depth and SWC at shallow depth (0-10 cm) was larger than that at deeper depth (10-20 cm). Correlation analysis indicated that the soil CO₂ effluxes were highly and positively correlated with SWC during the growing season. The stepwise regression analysis suggested that the main factor affecting CO₂ emissions was SWC at the depth of 0-10cmunder FI, whereas the main factor was SWC at the depth of 10-20cm under DI.

Fig.4

	Figure Option ViewDownloadNew Window
	Fig.4 Soil water content at the depths of 0-10 and 10-20 cm under flood irrigation (FI) and drip irrigation (DI)in 2014.Vertical bars indicate standard errors (n=3).

3.3.2Air temperature and soil temperature

The mean values of T-air (air temperature), T-0 (soil temperature at depth of 0 cm), T-5 (soil temperature at depth of 5 cm) and T-10(soil temperature at depth of 10 cm) during the growing seasonexhibiteda decreasing order of T-air> T-0> T-5> T-10 under both DI and FI (Fig.5). Correlation analysis indicated that the soil CO₂ effluxes were highly and positively correlated with air temperature (T-air)and also with soil temperaturesof different depths. The Q₁₀values (i.e., temperature sensitivity of soil respiration) increased with soil depths under the same irrigation treatment (Table 2), indicating that the soil respiration in the deeper soil horizons was more sensitive to the temperature change than that in the shallowsoil horizons. T-0 under FI and T-5 under DI were found to be the best-fitting factors for explaining the change in CO₂emissions, and they could explain approximately 41.3% (FI) and 45.9% (DI) of the variations in soil CO₂ emissions, respectively.

Fig.5

	Figure Option ViewDownloadNew Window
	Fig.5 Air temperature (T-air) and soil temperatures at the depths of 0 cm (T-0), 5 cm (T-5) and 10 cm (T-10)under DI and FI in 2014. Vertical bars indicate standard errors (n=3).

Table 2

Table 2

Table 2 Exponential regression functions between soil CO2 emissions and temperature under FI and DI Irrigation treatment Parameter Equation R2 Q10 FI T-air y=207.4e0.036x 0.382 1.43 T-0 y=173.2e0.050x 0.413 1.65 T-5 y=157.7e0.058x 0.400 1.79 T-10 y=128.7e0.068x 0.379 1.97 DI T-air y=164.6e0.043x 0.396 1.54 T-0 y=149.9e0.057x 0.450 1.77 T-5 y=123.6e0.071x 0.459 2.03 T-10 y=100.6e0.083x 0.442 2.29 Note:T-air, air temperature; T-0, soil temperature at depth of 0 cm; T-5, soil temperature at depth of 5 cm; T-10, soil temperature at depth of 10 cm.

Table 2 Exponential regression functions between soil CO2 emissions and temperature under FI and DI

The concentrations of soil dissolved organic carbon (DOC), microbial biomass carbon (MBC) and total organic carbon (TOC) of BR (i.e., between plant rows) and OR (i.e., on plant rows)situations under FI (i.e., flooding irrigation), OP(i.e., on irrigation pipes) and BP (i.e., between irrigation pipes) situations under DI (i.e., drip irrigation) during the growing season were averaged at each of the two measured soil depths (Table 3). The DOC, MBC and TOC concentrations decreased with increasing soil depths, and the average concentrations under FIwere larger than those under DI at all the measured soil depths. No significant differences were found between FI and DI in the soil TOC, DOC, and MBC concentrations for the same soil depth (either 0-10 or 10-20 cm). However, significant differences were found between shallow soil (0-10 cm) and deeper soil (10-20 cm) under the same irrigation treatment (either FI or DI).

Table 3

Table 3

Table 3 Concentrations of soil DOC, MBC and TOC at the depths of 0-10 and 10-20 cm under DI and FI DI FI 0-10 cm 10-20 cm 0-10 cm 10-20 cm DOC(mg/kg) 58.51± 4.00Aa 43.86± 3.65Ab 62.12± 3.67Aa 47.21± 5.68Ab MBC(mg/kg) 168.52± 25.30Aa 126.50± 15.78Ab 210.61± 25.38Aa 121.91± 21.95Ab TOC(g/kg) 22.7± 0.62Aa 21.8± 0.40Aa 25.1± 1.07Aa 23.9± 1.61Aa Note: DOC, dissolved organic carbon; MBC, microbial biomass carbon; TOC, total organic carbon. Different capital letters indicate significant differences in the soil DOC, MBC and TOC concentrations at the same soil depth under different irrigation treatments (i.e., FI and DI), and different lowercase letters represent significant differences in the soil DOC, MBC and TOC concentrations at different soil depths under the same irrigation treatment.

Table 3 Concentrations of soil DOC, MBC and TOC at the depths of 0-10 and 10-20 cm under DI and FI

Over the growing season, the DOC concentrations varied between 24.71and94.18 mg/kg under FI and between 30.20 and 98.20 mg/kg under DI at the depth of 0-10 cm. The DOC concentrationsvaried between 8.61 and 125.06 mg/kg under FI and between 19.21 and 88.84 mg/kg under DI at the depth of 10-20 cm. The soil MBC concentrations varied between 90.94 and 369.80 mg/kg under FI and between 48.12 and 326.15 mg/kg under DI at the depth of 0-10 cm. The MBC concentrations varied between 6.85 and 252.29 mg/kg under FI and between 42.94 and 202.33 under DI at the depth of 10-20 cm. The TOC concentrations varied between 22.4 and 32.8 g/kg under FI and between 20.9 and 26.0 g/kg under DI at the depth of 0-10cm. The concentrations varied between 20.6 and 37.2 g/kg and between 19.2 and 23.3 g/kg under DI at the depth of 10-20cm.

Correlation analysis showed that soil CO₂effluxes were significantly and positivelycorrelated withsoil MBC concentration at the depth of 0-10 cm under DI(r=0.505). And, soil CO₂ effluxes were significantly and negativelycorrelated with soil DOC concentration at the depth of 10-20 cmunder DI (r=-0.672).

Irrigation practicemay affect the production and emission of CO₂ (Kallenbach et al., 2010). Again, compared with FI, DI was performed at more regular intervals in much smaller amounts with less intense water supply and slower water movement(Chai et al., 2008). Therefore, the less-disturbed soil with more air movement under DI treatment could release more CO₂ from the soil into the atmosphere.Additionally, more frequent irrigation atregular intervals in small amounts under DI treatment could enhance soil wetting-drying cycles, and thus increase the CO₂ fluxes by promoting microbial activities, carbonmineralization, and respiration (Sparling and Ross, 1988; Van Gestel et al., 1993; Calderó n and Jackson, 2002). Our resultsindicated that soil CO₂ emission during the growing season was higher under DI than under FI, being supportive to the results obtained by Niu et al. (2014) and Zhang et al. (2014). Correlationanalysis suggested that the soil CO₂ emission was highly correlated withSWC and soil temperatures, lending further support to Wiseman and Seiler (2004) who concluded that SWC and soil temperaturescould affect soil CO₂ emission through modulating the decomposition rate of the soil organic matter. However, no significant differences in soil CO₂emissions were observed between FI and DI. This may be resulted from lower SWC under both DI and FI treatments and the lowered SWC could mitigate the effect of temperature on the production and emissions of CO₂ during the maize growing season (Qi et al., 2010).

Under DI treatment, SWC and soil nutrients distributions are highly heterogeneous due to vertical and lateral infiltrations(Allen et al., 1998). The spatial heterogeneities of the SWC and soil nutrients certainly have a significant impact on the spatial distribution of greenhouse gas emission from agricultural soil (Xu and Qi, 2001; Tang and Baldocchi, 2005; Kallenbach et al., 2010; Kennedy et al., 2013).Our results showed that the proportions of root respiration contributing to the soil CO₂ emissions of OP(on irrigation pipes)ranged from 3.45% to 46.55% under DI and from 17.49% to 46.66% to the soil CO₂ emissions of OR (on plant rows)under FI. Moreover, the average contribution of root respiration to the soil CO₂ emission of OP under DI was 22.72%, being significantly larger than that of OR under FI (15.82%), probably becauseirrigation water and fertilizers were applied along plant rows, providing more water and more nutrientsunder DI. The CO₂emissions from the plant rows (OR under FI and OP under DI) were higher than those from other locations (e.g., BR under FI and BP under DI), being consistent with the measurements by Lv et al. (2014). The soil CO₂ emissions of BA under DI had no significant difference with those of BP, probably because higherSWC in the BA thanin other locations (e.g., OP and BP) offsetthe contributionof root respiration to soil CO₂ emission of BP and the high SWC may have been resulted from lessened evapotranspiration due to shelter from maize plants. Correlation analysis showed that the soil CO₂ emissions were highly and positively correlated with the SWC and temperatures. The relatively high soil CO₂ emissionsfrom the plant rows (OR under FI and OP under DI) mean that higher SWC and root respiration might have released more CO₂. For the same spatial locations, DI treatment with less-disturbed soil and with more air movement might have released more CO₂and the SWC and soil temperatures might have affected soil CO₂ emission through modulating the decomposition rate of the soil organic matter.

Our results showed that the concentrations of soil DOC and MBC under FI were larger than those under DI, being consistent with the results obtained by Hanet al. (2010) who reported that high irrigation amounts under FI could result in temporary water saturation and could consequently inhibit the microbial activity, leading to higher DOC and MBC. Our results show that DOChad a negative relationship with SWConly at the 10-20cm depth (not at the depth of 0-10 cm), being somewhat supportive to Wang and Bettany (1993)and Han et al. (2010) who reported that FI could increase the leaching loss of soil DOC in shallow soil layers, leading to an increase in the DOC concentration in deep soil layers.

Either under FI or under DI, SWCcould be tightly associated with microbial activities, thus increasingor decreasing soil organic matter accumulation and decomposition (Gillabel et al., 2007; Butenschoen et al., 2011; McDowell and Smith, 2012; Arroita et al., 2013). In other words, either too less or too much of irrigated water in crop fields could result in anincreasein soil microbial activities, leading to a reduction in decomposition rate of the soil organic matter. Our results showed that the SWC and temperature conditions under DI were more favorable for organic matter decomposition. The microbial quotient (the ratio of MBC to TOC) was calculatedto further assess the quality of soil TOC, and it was regarded by some authors as a good index of the changes in soil organic matter quality (Insam and Merschak, 1997). Larger ratios implied an increase in the availability of fresh substrates (Anderson and Domsch, 1986). In the present study, the ratios of MBC/TOC under FI were larger than those under DI at the depth of 0-10 cm, whereas they were smaller than thoseunder DI at the depth of 10-20 cm.It suggested that DI was better in improving the soil quality at the depth of10-20 cm.